WO2008122056A2

WO2008122056A2 - Medical apparatus and method associated therewith

Info

Publication number: WO2008122056A2
Application number: PCT/US2008/059190
Authority: WO
Inventors: Fei Baowei
Original assignee: Case Western Reserve University
Priority date: 2007-04-02
Filing date: 2008-04-02
Publication date: 2008-10-09
Also published as: WO2008122056A3

Abstract

A medical apparatus and associated method are provided in various embodiments. In one embodiment, the medical apparatus includes a storage device storing medical imaging data related to treatment of a tumor using photodynamic therapy (PDT) in conjunction with a photosensitizer, an image processor to process the medical imaging data, and an assessment logic to evaluate efficacy of the PDT and photosensitizer. Silicon phthalocyanine 4 (Pc 4) or another suitable photosensitizer may be used, In one embodiment, the method includes processing the medical imaging data to form comparative data and analyzing the comparative data to evaluate the PDT and photosensitizer. hi other embodiments, a photosensitizer may be administered to a subject having a prostate tumor and the prostate tumor may be treated using PDT. In further embodiments, medical imaging may be using to guide the treatment during PDT.

Description

MEDICAL APPARATUS AND METHOD ASSOCIATED THEREWITH by Baowei Fei

CROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application claims the benefit of U.S. Provisional Patent Application

Numbers 60/909,528 (Attorney Docket Number 27708.04104), filed April 2, 2007, and 61/040,509 (Attorney Docket Number 27708.04107), filed March 28, 2008. The contents of all above-identified patent application^) are fully incorporated herein by reference.

STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH OR

DEVELOPMENT

[0002] The work leading to the present invention was supported by one or more grants from the U.S. Government, including NIH/NCI Grants R21CA120536, R24CA110943, and R01CA083917 and DOD Grant DAMD 17-02- 1-0230. The U.S. Government may have certain rights in the invention.

BACKGROUND

[0003] MEDICAL IMAGING

[0004] Medical imaging refers to the techniques and processes used to create images of the areas of interest, such as a human body (or parts thereof), for clinical purposes (e.g., medical procedures seeking to reveal, diagnose or examine disease) or medical science (e.g., including the study of normal anatomy and function). As a discipline and in its widest sense, it is part of biological imaging and incorporates radiology (in the wider sense), radiological sciences, endoscopy, (medical) thermography, medical photography and microscopy (e.g. for human pathological investigations).

[0005] Common forms of medical imaging include magnetic resonance imaging

(MRI), positron emission tomography (PET) imaging, ultrasound imaging, fluoroscopy imaging, and photoacoustic imaging. There are many other forms medical imaging within nuclear medicine, radiography, and tomography. Additionally, measurement and recording techniques which are not primarily designed to produce images, such as electroencephalography (EEG) and magnetoencephalography (MEG) and others, but which produce data susceptible to be represented as maps (i.e. containing positional information), can be seen as forms of medical imaging. [0006] PROSTATE CANCER

[0007] Prostate cancer is the second leading cause of cancer mortality in American males, and it is estimated that in 2004 there were 230,110 cases of prostate cancer and 29,900 prostate cancer deaths. The current principal treatment options are: 1) radical prostatectomy; 2) external beam radiation therapy; 3) interstitial brachytherapy; and 4) hormone ablation therapy. These methods can have serious side effects such as incontinence and sexual dysfunction. External beam radiation therapy and hormone therapy also require repetitive treatments over weeks or months. If radiation therapy fails, there are currently only a limited number of salvage options available for treatment of recurrent prostate cancer. Therefore, new treatment methods are of great potential value. [0008] PHOTODYNAMIC THERAPY

[0009] Photodynamic therapy (PDT) is a therapeutic modality for cancer treatment.

With PDT, a tumor-localized photosensitizer is irradiated with red light to generate reactive oxygen that efficiently kills cells and ablates tumors. PDT can be administered deep into tumors using minimally invasive techniques as only the small laser fiber that delivers the light to the tumor needs to be inserted into the lesions. PDT with porfimer sodium (e.g., Photofπn^®) is US-FDA approved for treating early and advanced lung cancer, advanced esophageal cancer, and Barrett's esophagus. An important advantage of PDT is that both the photosensitizer and the light are inert by themselves, and the light can be precisely delivered to a selected region, allowing high specificity in the localization of the photodynamic effect. Consequently, systemic toxicities are minimized. PDT-treated tumors often have a rapid response; in many cases a single treatment session may be sufficient to ensure successful eradication.

SUMMARY

[0010] In one aspect, a medical apparatus is described herein. In one embodiment, the medical apparatus includes a storage device adapted to store first and second medical imaging data related to a tumor, the first and second medical imaging data being in time- shifted relation to treatment of said tumor using photodynamic therapy (PDT) in conjunction with a photosensitizer, the second medical imaging data being time-shifted from the first medical imaging data, an image processor to process the first and second medical imaging data to form comparative data for at least one characteristic of the tumor, and an assessment logic to analyze the comparative data to evaluate efficacy of at least one of the PDT and photosensitizer. [0011] In another aspect, a method associated with the medical apparatus is described herein. In one embodiment, the method includes: a) providing first and second medical imaging data related to a tumor, the first and second medical imaging data being in time- shifted relation to treatment of said tumor using photodynamic therapy (PDT) in conjunction with a photosensitizer, the second medical imaging data being time-shifted from the first medical imaging data, b) processing the first and second medical imaging data to form comparative data for at least one characteristic of the tumor, and c) analyzing the comparative data to evaluate efficacy of at least one of the PDT and photosensitizer. [0012] In another embodiment, the medical apparatus includes an administering device to administer a photosensitizer to a subject having a prostate tumor and a light emitting device to treat the prostate tumor using photodynamic therapy (PDT) by selectively positioning an optic component proximate to a target area encompassing the prostate tumor and selectively delivering light to activate the photosensitizer.

[0013] In another embodiment, the method includes: a) administering a photosensitizer to a subject having a prostate tumor, b) selectively positioning an optic component of a light emitting device proximate to a target area encompassing the prostate tumor, and c) selectively delivering light to activate the photosensitizer to treat the prostate tumor using photodynamic therapy (PDT).

DRAWINGS

[0014] These and other features, aspects, and advantages of the present invention will become better understood with regard to the accompanying drawings, following description, and appended claims.

[0015] FIG. 1 shows an exemplary series of multiscale images in conjunction with medical imaging data;

[0016] FIG. 2 shows exemplary results from classification of medical imaging data;

[0017] FIG. 3 shows additional exemplary results from classification of medical imaging data;

[0018] FIG. 4 shows other exemplary results from classification of medical imaging data;

[0019] FIG. 5 shows more additional exemplary results from classification of medical imaging data;

[0020] FIG. 6 shows even more additional exemplary results from classification of medical imaging data; [0021] FIG. 7 shows exemplary results from registration of multiple sets of medical imaging data;

[0022] FIG. 8 shows additional exemplary results from registration of multiple sets of medical imaging data;

[0023] FIG. 9 shows other exemplary results from registration of multiple sets of medical imaging data;

[0024] FIG. 10 shows an exemplary process for fiuorodeoxyglucose (FDG) uptake in tumor cells;

[0025] FIG. 11 shows exemplary results from registration of multiple sets of medical imaging data;

[0026] FIG. 12 shows additional exemplary results from registration of multiple sets of medical imaging data;

[0027] FIG. 13 shows other exemplary results from registration of multiple sets of medical imaging data;

[0028] FIG. 14 shows more exemplary results from registration of multiple sets of medical imaging data;

[0029] FIG. 15 shows exemplary results from processing FDG imaging data to determine FDG uptake;

[0030] FIG. 16 shows exemplary results from processing MRI data to assess photodynamic therapy (PDT) for tumor treatment;

[0031] FIG. 17 shows additional exemplary results from processing MRI data to assess PDT for tumor treatment;

[0032] FIG, 18 shows other exemplary results from processing MRI data to assess

PDT for tumor treatment;

[0033] FIG. 19 shows more exemplary results from processing MRI data to assess

PDT for tumor treatment;

[0034] FIG. 20 shows even more exemplary results from processing MRI data to assess PDT for tumor treatment;

[0035] FIG. 21 shows additional exemplary results from processing MRI data to assess PDT for tumor treatment;

[0036] FIG. 22 shows other exemplary results from processing MRI data to assess

PDT for tumor treatment;

[0037] FIG. 23 shows additional exemplary results from processing MRI data to assess PDT for tumor treatment; [0038] FIG. 24 shows more exemplary results from processing MRI data to assess

PDT for tumor treatment;

[0039] FIG. 25 shows even more exemplary results from processing MRI data to assess PDT for tumor treatment;

[0040] FIG. 26 shows an exemplary protocol for tumor treatment using PDT with photosensitizer Pc 4 and assessment of the treatment using MRI;

[0041] FIG. 27 shows exemplary results from processing MRI data to assess PDT for tumor treatment;

[0042] FIG. 28 shows exemplary results from processing MRI data to assess PDT for tumor treatment;

[0043] FIG. 29 shows additional exemplary results from processing MRI data to assess PDT for tumor treatment;

[0044] FIG. 30 shows an exemplary protocol for tumor treatment using PDT with photosensitizer Pc 4 and assessment of the treatment using positron emission tomography

(PET);

[0045] FIG. 31 shows exemplary results from processing MRI data to assess PDT for tumor treatment;

[0046] FIG. 32 shows exemplary results from processing choline-PET imaging data to determine choline uptake to assess PDT for tumor treatment;

[0047] FIG. 33 shows other exemplary results from processing choline-PET imaging data to determine choline uptake to assess PDT for tumor treatment;

[0048] FIG. 34 shows additional exemplary results from processing choline-PET imaging data to determine choline uptake to assess PDT for tumor treatment;

[0049] FIG. 35 shows more exemplary results from processing choline-PET imaging data to determine choline uptake to assess PDT for tumor treatment;

[0050] FIG. 36 shows even more exemplary results from processing choline-PET imaging data to determine choline uptake to assess PDT for tumor treatment;

[0051] FIG. 37 shows other exemplary results from processing choline-PET imaging data to determine choline uptake to assess PDT for tumor treatment;

[0052] FIG. 38 shows exemplary results from processing medical imaging data to assess PDT for tumor treatment;

[0053] FIG. 39 shows exemplary protocol for image-guided PDT for tumor treatment;

[0054] FIG. 40 is block diagram of an exemplary embodiment of a medical apparatus; [0055] FIG. 41 is a flow chart of an exemplary embodiment of a method associated with the medical apparatus of FIG. 40;

[0056] FIG, 42 is block diagram of an exemplary embodiment of a medical apparatus;

[0057] FIG. 43 is a flow chart of an exemplary embodiment of a method associated with the medical apparatus of FIG. 42;

[0058] FIG. 44 is block diagram of an exemplary embodiment of a medical apparatus;

[0059] FIG. 45 is a flow chart of an exemplary embodiment of a method associated with the medical apparatus of FIG. 44;

[0060] FIG. 46 is block diagram of an exemplary embodiment of a medical apparatus; and

[0061] FIG. 47 is a flow chart of an exemplary embodiment of a method associated with the medical apparatus of FIG. 46.

DESCRIPTION

[0062] In one aspect, a medical apparatus and an automatic image classification method are described herein. In one embodiment, the apparatus and method provide for differentiation of necrotic tumor tissue from live tumor tissue, comprising (a) acquiring a first set of diagnostic image data from a tumor, (b) applying an interventional therapy to at least a portion of the tumor in order to treat the tumor, (c) acquiring a second set of diagnostic image date from the tumor, (d) inputting the first set and the second set of diagnostic image data to a computer to be analyzed using a multiscale fuzzy C-means (FCM) algorithm to arrive at a classification that is capable of differentiating necrotic tumor tissue from live tumor tissue. [0063] In another aspect, a medical apparatus and a method for monitoring treatment of a tumor for evaluating efficacy of interventional therapy in a patient is described herein. In one embodiment, the apparatus and method comprising: a) subjecting the patient to a first set of diagnostic imaging to determine an initial signal intensity value in the tumor; b) applying an interventional therapy to at least a portion of the tumor in order to treat the tumor; and c) subjecting the patient to a second set of diagnostic imaging to determine a second signal intensity value in the tumor; d) comparing the second signal intensity value with the first signal intensity value, wherein a change in the initial signal intensity value as compared to the second signal intensity value is indicative of the efficacy of the interventional therapy.

[0064] In some examples, the diagnostic imaging comprises magnetic resonance imaging (MRI). In one example, the signal intensity value is the MRI T2 value and an increase in the T2 value after interventional therapy is indicative that the interventional therapy is effective. In another example, the signal intensity value is the apparent diffusion coefficient (ADC). hi other examples, the diagnostic imaging comprises Positron Emission Tomography (PET) imaging. In one example, the PET imaging comprises ¹⁸F- Fluorodeoxyglucose PET imaging. In another example, the PET imaging comprises ¹¹C- choline PET imaging. In yet other examples, the diagnostic imaging comprises choline MR spectroscopy imaging (MRSI). In yet another example, the diagnostic imaging comprises a combination of MRI and PET or a combination of MRI and MRSI. The second set of diagnostic imaging can be obtained at any time during or after the interventional therapy. In one example,, the second set of diagnostic imaging is obtained immediately after therapy. In other examples, further sets of diagnostic imaging may be obtained at various time intervals after the therapy, hi another example, a surrogate biomarker may be obtained from noninvasive diagnostic imaging of a tumor, wherein a change in the value of the biomarker after administration of cancer therapy as compared to the value of the marker prior to cancer therapy is predictive of the success of the cancer therapy, hi one example, the surrogate biomarker is the T2 value of MRI images. In other examples, the surrogate biomarker is FDG update, and choline as measured from MRSI and/or PET imaging.

[0065] In yet another aspect, a medical apparatus and a method for treating prostate cancer are described herein, hi one embodiment, the apparatus and method comprising: (a) administering an effective amount of a phthalocyanine compound Pc 4 having the formula HOSiPcOSi(CH₃)₂(CH₂)₃N-(CH₃)₂, and (b) applying light of sufficient wave length and intensity to the prostate cancer to activate the Pc 4, wherein the activated Pc 4 exerts a cytotoxic effect on the prostate cancer.

[0066] In one example, the tumor is a solid cancerous tumor. In another example, the tumor is selected from the group consisting of prostate cancer, colon cancer, breast cancer, ovarian cancer, and bladder cancer. The interventional therapy can he one or more of the following: thermal ablation, cryoablation, injection of a denaturing liquid, injection of a chemotherapeutic agent, radiation therapy, brachytherapy, hormone ablation therapy, and photodynamic therapy. In one example, the interventional therapy is Pc 4-based photodynamic therapy of solid tumors. In another example, the interventional therapy is Pc A- based photodynamic therapy of prostate cancer. In one example, the interventional therapy is PDT. hi a further example, the tumor is selected from the group consisting of prostate cancer, colon cancer, breast cancer, ovarian cancer, and bladder cancer. [0067] A method of treating prostate cancer using photodynamic therapy (PDT) with a second-generation photosensitizing drug Pc 4 is disclosed herein. Additionally, imaging methods to assess the efficacy of PDT in various cancers are disclosed herein. [0068] The class of dyes known as phthalocyanines (Pc's) are synthetic macrocycles related to poφhyrins but having very strong absorption in the red region of the spectrum at wavelengths that penetrate tissue well. One member of this group, the silicon phthalocyanine photosensitizer Pc 4 [HOSiPcOSi(CH₃)₂(CH₂)₃N-(CH₃)₂] is found to be very effective in photodynamic therapy (PDT) of cancer. It is believed that its mechanism of action is an oxidative stress associated with induction of apoptosis in various cell types. PDT with Pc 4 is a strong inducer of apoptosis, or programmed cell death.

[0069] As a photosensitizing drug, Pc 4 prepares cancerous tissue to be broken down by light and reactive oxygen species. A small amount of the drug is administered intravenously to the patient over a two-hour period. About twenty-four to forty-eight hours later, a red laser light is applied, which is absorbed by the tumor-localized Pc 4. The light- activated photosensitizing compound produces forms of reactive oxygen that kill cancer cells and break down the tumor while leaving surrounding normal cells virtually untouched. [0070] Without being bound by the following theory, it is believed that because the photosensitizer localizes in mitochondria, immediate photooxidative damage to these organelles causes the release of factors that trigger the late stages of apoptosis. PDT also activates a series of stress signals, initiated as a result of membrane damage, including phospholipase activation, ceramide release, and stress kinase activation. Tumors respond very rapidly to PDT, with a high incidence of apoptosis in the early hours after a single treatment and complete loss of visible tumor in a few days.

[0071] Provided are imaging methods to aid in the early assessment of the therapeutic efficacy of cancer treatment. Imaging techniques can provide tools for the assessment of cancer therapy efficacy. The ability to determine the spatial and metabolic distribution of cancer cells can be important in assessing the initial stage, prognosis, and treatment efficacy. Additionally, combined anatomic and metabolic imaging (e.g. MRI and PET and/or MRSI) allow structural and metabolic evaluation of cancer and can improve the diagnostic accuracy, as well as the assessment of treatment efficacy.

[0072] In addition, the combination of multiple imaging modalities can improve the monitoring of treatment efficacy. Each imaging modality has its own strengths and weaknesses. For example, in the case of PDT, PET can image the rapid biochemical and physiological responses of tumors to PDT; whereas MRI provides superior anatomical information, locations, and morphological changes within tumors. Combining PET and MRI can have several advantages. For example, MRI scans provide anatomical reference for the PET images. In addition, fusion of MRI and PET images can enhance our ability to visualize the distribution of a radiolabeled pharmaceutical. MRI provides tumor shape and size information that can be used to improve the accuracy of the PET data analysis, such as drawing regions of interests (ROIs) and performing quantitative analyses. Furthermore, MRI can be used to correct PET data for partial volume effects to clarify that the PET-measured changes induced by PDT are due to metabolic and hemodynamic changes and not to artifacts of changes in tumor size. Other imaging techniques that can be used in the present invention include: diffusion-weighted MRI, perfusion MRI, Proton MRSI, choline MRSI, and P31 MR spectroscopy

[0073] PET imaging with choline is useful for the diagnosis of prostate cancer.

Choline is a substance that is present in cellular membranes. When it is marked with carbon- 11 or fluorine-18, this radiotracer has an affinity for prostate tissues and allows the differentiation of malignant from benign processes. Compared to ¹¹C, ¹⁸F has a longer half- life and a shorter positron range. Choline-PET is particularly useful for re-staging patients who experience increasing serum prostate-specific antigen (PSA) levels following prostatectomy. We have found that choline-PET can provide an in vivo tool to monitor tumor response to PDT.

[0074] In the case of Pc 4-based PDT, it has been found that Pc 4-PDT works predominantly through the direct killing of malignant cells. PDT-induced apoptosis is a common mechanism of cell death both in vitro and in vivo. The rapid tumor response to Pc 4- PDT provides a mechanism for in vivo imaging. Positron emission tomography (PET) can provide functional and biochemical information about cancers. PET has relatively high sensitivity, full quantitative capability, and can provide dynamic pictures of the distribution of a radiopharmaceutical over time. Magnetic resonance imaging (MRI) can provide high- resolution anatomic details regarding lesions. MR spectroscopic imaging (MRSI) is another tool for detecting metabolites containing protons, phosphorus, fluorine or other nuclei. Pc 4- PDT of prostate tumors can cause changes in cell membranes that alter their ability to incorporate choline into their constituent phospholipids. The resulting changes can be detected by ^I8F-fluorocholine_s ^πC-choline PET imaging, and/or choline MR spectroscopy imaging. Therefore, the profound and rapid effects of Pc 4-PDT can be detected by one or more in vivo molecular imaging techniques. [0075] The use of non-invasive imaging methods can improve the prediction of PDT efficacy and result in optimizing each patient's therapy. Imaging can be applied before, during and after treatment. In vivo images may provide an early alert to the physician that additional treatment is needed or that treatment has already reached a sufficient threshold for response. This strategy has great potential to improve the treatment outcome by tailoring the therapy to each individual.

[0076] The noninvasive imaging techniques described herein can be applied to other cancers. They can be used to detect very early cancers, optimize treatment planning for individualized therapy, utilize image-guided minimally invasive therapy, and monitor therapeutic efficacy of various treatment regimes.

[0077] With reference to FIG. 40, an exemplary embodiment of a medical apparatus includes a storage device, an image processor, and an assessment logic to analyze the comparative data to evaluate efficacy of at least one of the PDT and photosensitizes The storage device is adapted to store first and second medical imaging data related to a tumor, the first and second medical imaging data being related to treatment of said tumor using photodynaraic therapy (PDT) in conjunction with a photosensitizer. The image processor may process the first and second medical imaging data to form comparative data for at least one characteristic of the tumor. The assessment logic may analyze the comparative data to evaluate efficacy of at least one of the PDT and photosensitizer.

[0078] The first and second medical imaging data may include at least one of magnetic resonance imaging (MRI) data, positron emission tomography (PET) imaging data, and ultrasound imaging data. In one embodiment, at least one of the first and second medical imaging data is from a time during the treatment. In another embodiment, the first and second medical imaging data is in time-shifted relation to the treatment. The second medical imaging data may be time-shifted from the first medical imaging data. [0079] In one embodiment, the assessment logic also analyzes the comparative data in conjunction with at least one of further diagnosis of the tumor and further treatment of the tumor. The tumor may include at least one of a malignant tumor, prostate cancer, breast cancer, ovarian cancer, colon cancer, glioma, and a benign tumor. The photosensitizer may include at least one of silicon phthalocyanine and porfimer sodium.

[0080] In one embodiment, the first and second medical imaging data may include magnetic resonance imaging (MRI) data. In this embodiment, the image processor may include a mapping logic, a statistical analyzer, and a comparator. The mapping logic may produce T2 maps of the tumor from the first and second medical imaging data. The statistical analyzer may produce T2 values for at least one of a histogram, a mean, and a standard deviation from the T2 maps. The comparator may identify changes in T2 values between the first and second medical imaging data. In this embodiment, the assessment logic may use changes in T2 values as a surrogate biomarker to evaluate the efficacy of at least one of the PDT and the photosensitizer.

[0081] In another embodiment, the first and second medical imaging data may also include magnetic resonance imaging (MRI) data. In this embodiment, the image processor may include a mapping logic, a statistical analyzer, and a comparator. The mapping logic may produce apparent diffusion coefficient (ADC) maps of the tumor from the first and second medical imaging data. The statistical analyzer may produce ADC values for at least one of a histogram, a mean, and a standard deviation from the ADC maps. The comparator may identify changes in ADC values between the first and second medical imaging data. In this embodiment, the assessment logic may use changes in ADC values as a surrogate biomarker to evaluate the efficacy of at least one of the PDT and the photosensitizer. [0082] In still another embodiment, the first and second medical imaging data may include choline imaging data. In this embodiment, the image processor may include an image reconstruction logic, a statistical analyzer, and a comparator. The image reconstruction logic may produce images of the tumor from the first and second medical imaging data. The statistical analyzer may produce a time activity curve showing choline uptake from the images. The comparator may identify changes in choline uptake between the first and second medical imaging data. In this embodiment, the assessment logic may use changes in choline uptake as a surrogate biomarker to evaluate the efficacy of at least one of the PDT and the photosensitizer.

[0083] In yet another embodiment, the first and second medical imaging data may include fluorodeoxyglucose (FDG) imaging data. In this embodiment, the image processor may include an image reconstruction logic, a statistical analyzer, and a comparator. The image reconstruction logic may produce images of the tumor from the first and second medical imaging data. The statistical analyzer may produce a time activity curve showing FDG uptake from the images. The comparator may identify changes in FDG uptake between the first and second medical imaging data. In this embodiment, the assessment logic may use changes in FDG uptake as a surrogate biomarker to evaluate the efficacy of at least one of the PDT and the photosensitizer.

[0084] With reference to FIG. 41, an exemplary embodiment of a method associated with the medical apparatus of FIG. 40 includes: a) providing first and second medical imaging data related to a tumor, the first and second medical imaging data being related to treatment of said tumor using photodynamic therapy (PDT) in conjunction with a photosensitizer, b) processing the first and second medical imaging data to form comparative data for at least one characteristic of the tumor, and c) analyzing the comparative data to evaluate efficacy of at least one of the PDT and photosensitizer. In one embodiment, at least one of the first and second medical imaging data is from a time during the treatment. In another embodiment, the first and second medical imaging data is in time-shifted relation to the treatment. The second medical imaging data may be time-shifted from the first medical imaging data. In another embodiment, the method may also include: d) analyzing the comparative data in conjunction with at least one of further diagnosis of the tumor and further treatment of the tumor.

[0085] In still another embodiment, the first and second medical imaging data may include magnetic resonance imaging (MRI) data. In this embodiment, the method may also include: d) mapping the first and second medical imaging data to produce T2 maps of the tumor, e) analyzing the T2 maps to produce T2 values for at least one of a histogram, a mean, and a standard deviation, f) comparing the T2 values to identify changes between the first and second medical imaging data, and g) using changes in T2 values as a surrogate biomarker to evaluate the efficacy of at least one of the PDT and the photosensitizer. [0086] In yet another embodiment, the first and second medical imaging data may include magnetic resonance imaging (MRI) data. In this embodiment, the method may also include: d) mapping the first and second medical imaging data to produce apparent diffusion coefficient (ADC) maps of the tumor, e) analyzing the ADC maps to produce ADC values for at least one of a histogram, a mean, and a standard deviation, f) comparing the ADC values to identify changes between the first and second medical imaging data, and g) using changes in ADC values as a surrogate biomarker to evaluate the efficacy of at least one of the PDT and the photosensitizer.

[0087] In still yet another embodiment, the first and second medical imaging data may include choline imaging data. In this embodiment, the method may also include: d) reconstructing the first and second medical imaging data to produce images of the tumor, e) analyzing the images to produce a time activity curve showing choline uptake, f) comparing choline uptake to identify changes between the first and second medical imaging data, and g) using changes in choline uptake as a surrogate biomarker to evaluate the efficacy of at least one of the PDT and the photosensitizer. [0088] In another embodiment, the first and second medical imaging data may include fluorodeoxyglucose (FDG) imaging data. In this embodiment, the method may also include: d) reconstructing the first and second medical imaging data to produce images of the tumor, e) analyzing the images to produce a time activity curve showing FDG uptake, f) comparator to identify changes in FDG uptake between the first and second medical imaging data, and g) using changes in FDG uptake as a surrogate biomarker to evaluate the efficacy of at least one of the PDT and the photosensitizer.

[0089] With reference to FIG. 42, an exemplary embodiment of a medical apparatus includes an administering device and a light emitting device. The administering device may administer a photosensitizer to a subject having a prostate tumor. The light emitting device may treat the prostate tumor using photodynamic therapy (PDT) by selectively positioning an optic component proximate to a target area encompassing the prostate tumor and selectively delivering light to activate the photosensitizer. The prostate tumor may include at least one of a malignant tumor, prostate cancer, and a benign tumor. The photosensitizer may include at least one of silicon phthalocyanine and porfimer sodium.

[0090] In another embodiment, the medical apparatus may also include a medical imaging device. The medical imaging device may provide an image of the prostate tumor and surrounding area in relation to at least one of the positioning of the optic component and the delivering of the light for image-guided PDT. The medical imaging device may include at least one of a magnetic resonance imaging (MRI) device, a positron emission tomography (PET) device, and an ultrasound imaging device.

[0091] With reference to FIG. 43, an exemplary embodiment of a method associated with the medical apparatus of FIG. 42 includes a) administering a photosensitizer to a subject having a prostate tumor, b) selectively positioning an optic component of a light emitting device proximate to a target area encompassing the prostate tumor, and c) selectively delivering light to activate the photosensitizer to treat the prostate tumor using photodynamic therapy (PDT). The prostate tumor may include at least one of a malignant tumor, prostate cancer, and a benign tumor. The photosensitizer may include at least one of silicon phthalocyanine and porfimer sodium.

[0092] In another embodiment, the method may also include: d) imaging the prostate tumor and surrounding area in relation to at least one of the positioning in b) and the delivering in c) for image-guided PDT. The imaging in d) may be provided by at least one of a magnetic resonance imaging (MRI) device, a positron emission tomography (PET) device, and an ultrasound imaging device. [0093] With reference to FIG. 44, an exemplary embodiment of a medical apparatus includes a storage device, a rigid-body registration logic, a deformable registration logic, and an image fusion logic. The storage device may store first and second medical imaging data. The rigid-body registration logic may identify normalized mutual information common to the first and second medical imaging data and may align the first and second medical imaging data based at least in part on the normalized mutual information. The deformable registration logic may identify at least one deformable volumetric characteristic of the aligned first and second medical imaging data based at least in part on a finite element model. The image fusion logic may deform at least one of the first and second medical imaging data to form hybrid medical imaging data based at least in part on the at least one deformable volumetric characteristic.

[0094] The first and second medical imaging data may include at least one of magnetic resonance imaging (MRI) data, positron emission tomography (PET) imaging data, and ultrasound imaging data. The hybrid medical imaging data may be used in conjunction with at least one of detection of a tumor, diagnosis of a tumor, treatment of a tumor, and assessment of tumor treatment. The tumor may include at least one of a malignant tumor, prostate cancer, breast cancer, ovarian cancer, colon cancer, glioma, and a benign tumor. The treatment may include the use of photodynamic therapy (PDT) in conjunction with a photosensitizer. The photosensitizer may include at least one of silicon phthalocyamne and porfimer sodium.

[0095] With reference to FIG. 45, an exemplary embodiment of a method associated with the medical apparatus of FIG. 44 includes: a) identifying normalized mutual information common to first and second medical imaging data based at least in part on a rigid-body registration model, b) aligning the first and second medical imaging data based at least in part on the normalized mutual information, c) identifying at least one deformable volumetric characteristic of the aligned first and second medical imaging data based at least in part on a finite element model, and d) deforming at least one of the first and second medical imaging data to form hybrid medical imaging data based at least in part on the at least one deformable volumetric characteristic. In one embodiment, the first medical imaging data may include magnetic resonance imaging (MRI) data and the second medical imaging data may include positron emission tomography (PET) imaging data. In another embodiment, the first medical imaging data may include magnetic resonance imaging (MRT) data from a first time and the second medical imaging data may include MRI data from a second time. In still another embodiment, the first medical imaging data may include magnetic resonance imaging (MRI) data and the second medical imaging data may include ultrasound imaging data.

[0096] With reference to FIG. 46, an exemplary embodiment of a medical apparatus includes a storage device, an anisotropic diffusion filter, a clustering logic, and a fuzzy classifier. The storage device may store medical imaging data. The anisotropic diffusion filter may process the medical imaging data to form a plurality of multiscale images ranging in resolution from a first multiscale image at an original resolution to a last multiscale image at a coarser resolution. The clustering logic may process the last multiscale image to form an initial estimate of class prototypes for a plurality of tissue types associated with the medical imaging data based at least in part on a k-means clustering algorithm. The fuzzy classifier may classify components of the first multiscale image into at least one tissue type of the plurality of tissue types based at least in part on processing the plurality of multiscale images using a multiscale fuzzy C-mean algorithm.

[0097] The medical imaging data may include at least one of magnetic resonance imaging (MRI) data, positron emission tomography (PET) imaging data, and ultrasound imaging data. The plurality of tissue types may include at least one of live tumor tissue, necrotic tumor tissue, and intermediate tumor tissue. The classifying of the first multiscale image may be used in conjunction with at least one of detection of a tumor, diagnosis of a tumor, treatment of a tumor, and assessment of tumor treatment. The tumor may include at least one of a malignant tumor, prostate cancer, breast cancer, ovarian cancer, colon cancer, glioma, and a benign tumor. The treatment may include the use of photodynamic therapy (PDT) in conjunction with a photosensitizer. The photosensitizer may include at least one of silicon phthalocyanine and porfimer sodium.

[0098] With reference to FIG. 47, an exemplary embodiment of a method associated with the medical apparatus of FIG. 46 includes: a) filtering medical imaging data to form a plurality of multiscale images ranging in resolution from a first multiscale image at an original resolution to a last multiscale image at a coarser resolution based at least in part on an anisotropic diffusion filter, b) determining an initial estimate of class prototypes for a plurality of tissue types associated with the medical imaging data based at least in part on processing the last multiscale image using a k-means clustering algorithm, and c) classifying components of the first multiscale image into at least one tissue type of the plurality of tissue types based at least in part on processing the plurality of multiscale images using a multiscale fuzzy C-means algorithm. In one embodiment, b) may be repeated until a current initial estimate reaches a predetermined convergence in relation to a last initial estimate. In another embodiment, c) may be performed for multiple multiscale images in a coarser to finer resolution sequence toward the original resolution and results from processing at coarser resolutions are used to initialize the processing of the multiscale image at the next finer resolution. In still another embodiment, the method may also include: d) assigning a high membership value to a voxel whose intensity is close to a center of a class, e) allowing membership in neighborhood pixels to regulate classification toward piecewise-homogeneous labeling, and f) incorporating supervision information from the processing of the multiscale image at the previous coarser resolution.

[0099] TISSUB-BASED CLASSIFICATION FOR MEDICAL IMAGING

[00100] In one embodiment, a medical apparatus provides tissue-based classification technique for medical imaging. The image classification may be used for therapeutic assessment. In one embodiment, the classification technique may include a multiscale image classification method. In another embodiment, the classification technique may be used to classify magnetic resonance (MR) images.

[00101] Multiple medical images may be acquired. The medical images may include multiple weighted medical images, such as MR images. The medical images may include images of tumors, such as those associated with prostate cancer. The medical images may be acquired in conjunction with photodynamic therapy (PDT), a therapeutic modality for cancer treatment, including pre-PDT, post-PDT, or twenty-four hours after PDT. The tissue-based classification may be used to differentiate live, necrotic and intermediate tissues within the treated tumor on the medical images. This may result in medical images with more clearly defined live, necrotic and intermediate tissue regions.

[00102] A multiscale diffusion filter may be used to process MR images before classification. A multiscale fuzzy C-means (MFCM) classification method may be applied along the scales of the diffusion filter. The object function of the standard fuzzy C-means (FCM) may be modified to allow multiscale classification processing where results from a coarser scale may be used to supervise classification in the next finer scale. The MFCM method may take noise levels and partial volume effects into account during the classification processing.

[00103] In certain studies, the classification method was validated by simulated MR images with various noise levels. For simulated data, the classification method has achieved a 96.0 ± 1.1% overlap ratio. For real mouse MR images, the classification results of treated tumors were validated by histologic images. The overlap ratios were 85.6 ± 5.1%, 82.4 ± 7.8% and 80.5 ± 10.2% for the live, necrotic, and intermediate tissues, respectively. These MR imaging and classification methods may provide a useful tool for in vivo assessment of tumor response to PDT. Other medical imaging techniques with similar capabilities may also provide useful tools for assessment of PDT. Moreover, these medical imaging technique can be applied to PDT for various types of cancer, including human prostate cancer. [00104] PDT is a modality for treatment of cancer. The therapy uses a tumor-localized drug called a photosensitizer excited by irradiation with a laser light of a particular wavelength, which generates reactive singlet oxygen that efficiently kills cells and ablates tumors. Both the photosensitizer and the light are inert by themselves. Therefore, systemic toxicities in PDT are minimized. PDT is minimally invasive as a small laser fiber may be mounted externally to deliver the light to tumors.

[00105] Medical imaging techniques may provide a tool for assessment of PDT efficacy. More specifically, as discussed herein, in vivo medical imaging techniques may be used for assessment of tumor response to PDT using, for example, Pc 4 as the a photosensitizer. For example, high-resolution magnetic resonance imaging (MRI) can show anatomical and morphological changes of lesions. In one embodiment, an MR imaging system may be used to acquire multiple weighted MR images before, after and twenty-four hours after PDT. The tumor response to the treatment may be defined by the degree of tumor necrosis or apoptosis. To quantitatively evaluate PDT, an image classification method may be used to differentiate live, necrotic and intermediate tissues within the treated tumor on the MR images.

[00106] MR images may be affected by multiple factors, such as noise, intensity inhomogeneity and partial volume effects. Partial volume effects, for example, occur where pixels contain a mixture of multiple tissue types. These effects may make assignment the pixel to a single class more difficult and may create boundary regions. A Gaussian mixture based classification model may be used to estimate the mixture of each pixel by modeling the image histogram. This imaging processing method may assume the intensity of a single tissue type is a Gaussian distribution. In actuality, due to partial volume effects and image smoothing from post processing, the intensity distribution may deviate from a Gaussian model.

[00107] An FCM algorithm may be used to employ fuzzy partitioning that allows one voxel to belong to tissue types with different membership grades between 0 and 1. Various modified FCM may be used to compensate for intensity inhomogeneity and spatial information. However, FCM may be sensitive to the initial guess and noise with regard to both speed and stability. Therefore, an anisotropic diffusion filter may be applied to smooth noise, while preserving edge boundaries. Accordingly, the result of a k-means classification on a coarse level may be sufficient for the initial guess of the FCM method. An MFCM classification method may be applied along the scales of the anisotropic diffusion filter to provide more accurate classification in a step by step fashion with faster convergence at fine scales. The object function of the standard FCM may be modified to allow multiscale classification where the result from a coarse scale is used to supervise the classification in the next scale.

[00108] Multiscale Space from Anisotropic Diffusion Filtering

[00109] Multiscale space permits representation of images by using a series of images at varying spatial resolution in which an image contains less local information as the scale increases. An anisotropic diffusion filter is a partial differential diffusion equation model in which the image processing can achieve more smoothing while preserving inter-region edges through discrete time step increasing. For additional information on anisotropic diffusion filters, see Perona et at, Scale-space and edge detections using anistropic diffusion, IEEE Trans. Pattern Anal. Machine Intell., 1990, 12(7), p. 629-639, the contents of which are rally incorporated herein by reference. The multiscale description of images may be generated by anisotropic diffusion filter with the time step as scale.

[00110] The anisotropic diffusion equation may be described as shown in the following equation:

where I(x_st) is the intensity of MR volumes at time step or scale t; ^ and div are spatial gradient and divergence operator.

[00111] The g(x, t) component is the diffusion coefficient and chosen as a function of the magnitude of the gradient of intensity images as shown in the following equation; g (Il W(V)I) _ -(|w(*,/)|/*)²

(2), where the constant k is chosen to be the gradient magnitudes produced by noise, and can be fixed manually or estimated using a noise estimator. For additional information on noise estimators, see Canny, A computational approach to edge detection, IEEE Trans. Pattern Anal. Machine Intel!., 19S6, Vol. 8, p. 679-698, the contents of which are fully incorporated herein by reference. By applying an anisotropic diffusion filter to the original MR images, a series of images may be generated and the scale spaces may be formed. For reference, the scale level of original images is typically zero (0). When the scale increases, the images become blurred and contain more general information. FIG. 1 illustrates an exemplary scale space constructed from anisotropic diffusion filtering. As shown, the scale space is composed by the stack of the original image filtered at different time steps. t=0 is the original image. At higher scale levels, less local information appears in the image. [00112] Multiscale Fuzzy C-means (MFCM)

[00113] The classification may be begin at the coarsest scale and proceed to the original images. The classification result at a coarser level (t+1) may be used to initialize the classification at a higher scale level (t). The final classification may be the result at the scale level zero (0). During the classification processing at level (t+1), the pixels with the highest membership above a threshold may be identified and assigned to the corresponding class. These pixels maybe labeled as training data for the next level (t).

[00114] The objective function of the FCM at level (t) may be expressed as shown in the equation below:

(3), where u,_k stands for the membership of the pixel i belonging to the class k, and Vk is the vector of the class k center, X₁ is the feature vectors from multi-weighted MR images, N, stands for the eight (8) neighboring pixels of x, for 2D images, and the parameter p is a weighting exponent and is selected as 2. The objective function is the sum of three terms, where a and β are scaling factors to maintain balance between them. The first term is the standard fuzzy c-means object function that assigns a high membership to the voxel whose intensity is close to the center of the class. If only this term would be used, this is standard FCM. The second term allows the membership in neighborhood pixels to regulate the classification toward piecewise-homogeneous labeling. If both the first and second terms are used, this is modified FCM. The third term incorporates the supervision information from the

classification of the previous scale, ^UΛ is the membership obtained from the classification in the previous scale. If all three terms are used, this is multiscale FCM (MFCM).

[00115] The ^u* component may be determined using the equation below:

, rw,';',ifmax(ι4⁺¹)>κ

{ 0, otherwise / Λ\ where K is the threshold to determine the pixels with known class in the next scale classification, and is set as 0,85 in our implementation. The classification is implemented by minimizing the object function J. The minimization of J happens when the first derivative of J with respect to u_& and V_k are zero. From the equation below:

the class center may be updated using the equation below:

where

for every pixel, according to the Lagrangian approach as defined by the equation below:

(7). [00116] For optimization with respect to u*_k, the equation below is used: i l2 -A 1 = O n

(8).

[0Ol 17] The membership of every pixel i belonging to class k is updated according to equation below:

[00118] MFCM Algorithm

[00119] Typically, MFCM is used as an iterative algorithm based on an initial estimation of the class prototypes. Generally, proper selection of the initial classification improves clustering accuracy and reduces the number of iterations. The k-means method may be used on the coarsest image to estimate the initial class prototypes because noise and inhomogeneities have been effectively attenuated by anisotropic filtering at the coarsest image.

[00120] In one embodiment, the MFCM algorithm for classifying MR images includes; 1) anisotropic diffusion filtering the images to the scale level t = n, 2) obtaining the initial class prototypes using k-means clustering method in the coarsest level image by setting

{^u'k } = 0, 3) running clustering in the filtered image at level t =n-l, and 4) thresholding

membership functions according to equation (4) and getting matrix { ^M« } . Steps 3) and 4) may be repeated on the next scale image until the classification at original image (t=0) is complete.

[00121] In one embodiment, step 3) may include: a) updating the membership using equation (9), and b) computing class centroids {V_k} using equation (6). Steps a) and b) may be repeated until convergence, where convergence may be defined using the equation below: new old < ε

(10), where the iteration is terminated when Euclidean distance of class centers between iterations is less than a small number e (e - 0.01). [00122] PDT Experiments and Image Acquisition

[00123] Human prostate tumor PC-3 cells may be grown in as monolayers in E-MEM supplemented with 15% fetal bovine serum. Two tumors may be initiated in athymic nude mice by injection of PC-3 cells subcutaneously on the back. Tumors may be treated and imaged when they reach 6-10 mm in diameter. The photosensitizer Pc 4 may be injected in the tumor-bearing mice via tail vein by 0.6 mg/kg of body weight. After forty-eight hours, one of the tumors may be exposed to red laser light (e.g., 672 nm) from a diode laser with a dose, for example, of 150 J/cm² and a fluence rate of 100 mW/cm². MR images of the mice may be acquired before, immediately (i.e., as soon as possible, such as within five minutes or one hour) after, and twenty-four hours after the therapy to monitor the PDT treatment. During each imaging session, the mice may be mounted on a plastic holder and may be provided with a continuous supply of 2% isoflurane in oxygen to minimize motion artifacts in MR images.

[00124] Tl-, T2-, and FLASH weighted MR images may be acquired for the tumor- bearing mice. The MR images may be acquired using a 1.5 T scanner (e.g., Siemens Sonata 1.5 T scanner from Siemens Medical Systems, Erlangeα, Germany). A custom-designed whole-body mouse coil (e.g., 2-element phased-array, ID = 32 mm) may be used to minimize noise levels. The acquired coronal scan may provide images at, for example, 256 x 120 matrix, 80 x 36 mm FOV and 1 mm slice thickness. The number of signal averages may be set, for example, at six to obtain images with low noise.

[00125] With reference to FIG. 2, exemplary classification results for a 3D simulated image volume using the MFCM algorithm may be assessed by visual observation. More specifically, to evaluate the classification method, a 3D tumor model with four tissue classes was simulated. Each class was given a gray level between 0 and 255 and combined with 10% Gaussian noises. The images were filtered by three pixel Gaussian filter for partial volume effects. As shown, three image slices (Slice 6, 21, and 52) cover the whole tumor volume. The (a) and (b) frames are simulated MR image volumes with different weightings. The (c) frames are the classification result. The (d) frames are the ground truth. Pn the (c) and (d) frames, four gray levels show the different tissue classes.

[00126] Table 1 shows the results of a more detailed tissue-dependent quantitative analysis in which the sensitivity and specificity that is evaluated by the ground truth is computed. As shown, for the simulated images, the MFCM algorithm can correctly classify 96.0 ± 1.1% of the tissue in the image. The values in the confusion table on simulated data are the percentages computed over all voxels of each class (C1-C4) in the reference. False positive (FP) and false negative (FN) rates are computed in percentages using the reference.

Table 1

[00127] The MFCM method was also applied to the digital brain phantom data generated by the BrainWeb MR simulator described in Collins et al., Design and construction of a realistic digital brain phantom, IEEE Trans. Med. Imaging, 1998, 17(3), p. 463-468, the contents of which are fully incorporated herein by reference. The MFCM classification method was applied to Tl and T2 weighted MR images with different noise levels and 20% intensity non-uniformity. Prior to the classification, extracranial tissue, such as skull, meninges, and blood vessels were removed so that the brain MRI images consisted of three types of tissue: i) gray matter, ii) white matter and iii) cerebrospinal fluid (CSF). The classification is evaluated by the overlap ratio between the classification result and the realistic model for every class, which is defined as twice the number of corrected classified pixels divided by the total number of pixels in the ground truth and classified results for each class. FIG. 3 illustrates the classification results and corresponding ground of truth. FIG. 4 demonstrates the overlap ratio for each class between classified results and ground of truth, which decrease by less than 6.0% with added noises. The MFCM method was compared with the standard FCM method and the modified FCM method. This was implemented by setting the constant Gf= β ~0 for the standard FCM method, a - 0.85 and /3=0 for the modified FCM method, and «=0.85 and β -0.80 for the multiscale FCM (MFCM) method, respectively. FIG. 5 shows the overlap ratio change with respect to different methods applied on Tl and T2 MRI with 9% noise and 20% intensity non-uniformity.

[00128] For real MR images from treated mice, the tumors were first manually segmented on each slice and then classified into three tissue classes (e.g., live, necrotic, and intermediate tissues) using the MFCM method. The MFCM classification results were evaluated by histology images which were classified into three classes by a pathologist. FIG. 6 illustrates the Tl, T2, FLASH weighted MR images and the MFCM classification results of a real mouse tumor in MR images twenty-four hours after PDT. The (a), (b), and (c) frames in FIG. 6 are the original Tl, T2, FLASH tumor MR images, respectively. The (d) frame is the MFCM classification result showing three classes. The (e) frame of FIG. 6 is the corresponding histology. As shown, the necrotic (right) and intermediate (left) regions are marked on the image. The (f) frame shows the overlap of the histologic marking and the MFCM classified result.

[00129] Additionally, the MFCM classification method was performed on MR images of three mice. The overlap ratios were 85.6 ± 5.1%, 82.4 ± 7.8%, and 80.5 ± 10.2% for the live, necrotic, and intermediate tissues, respectively.

[00130] As described herein, a multiscale fuzzy c-means (MFCM) classification method may be used for assessment of PDT. An anisotropic filter may be used to attenuate the noise within regions while preserving edges between different tissue types. A scale space may be generated by the anisotropic filtering and the general structure information may be kept in the images at a coarser scale. Therefore, a k-means method on the coarsest images may be used for the initial guess. The classification may be advanced along the scale space to include local information in fine-level images and to compensate the partial volume effects due to smoothing. The result from a coarser scale may provide the initial parameter for the classification in the next scale. Meanwhile, the pixels with a high probability of belonging to one class in the coarse scale may belong to the same class in the next level. Therefore, these pixels in the coarser images may be considered as points with a known class and may be used as training data to constrain the classification in the next scale. In this way, an accurate, step by step classification may be provided that avoids being trapped in local minima. Furthermore, a term that constrains a pixel may be included that can be influenced by its immediate neighborhood so as to achieve a piecewise-homogeneous solution. The MFCM algorithm provides an accurate and robust method for both simulated and real MR images. [00131] Field inhomogeneity may be smooth compared to MR images and segmented tumors may be classified with a small volume on the whole MR images. Therefore, field inhomogeneity may be neglected in the MFCM classification. However,, heavy field inhomogeneity cannot be attenuated by anisotropic filtering and may corrupt the result severely. A gain field term maybe incorporated in the objective function of FCM methods to estimate both the tissue classification and the bias field.

[00132] The MFCM imaging technique may be used to study the therapeutic effects of cancer treatment. The imaging and classification method may provide a tool to differentiate necrosis from viable tumor cells on MR images. This could be used for early assessment of therapeutic effects in human cancer therapy, including PDT therapy of prostate cancer. [00133] IMAGE REGISTRATION QF MEDICAL IMAGES

[00134] In one embodiment, a medical apparatus provides an image registration technique for medical images. In one embodiment, the registration technique includes a finite element model (FEM)-based deformable volume matching (DVM) method. In another embodiment, the registration technique includes a hybrid intensity and model based deformable registration algorithm. In still another embodiment, the registration technique includes a thin-plate spline image registration method. In one embodiment, the registration technique may be used for registering position emission tomography (PET) and magnetic resonance (MR) images. In another embodiment, the registration technique may be used for registering MR and ultrasound images. In still another embodiment, the registration technique may be used for registering multiple MR images. In yet another embodiment, the registration technique may be used for registering multiple longitudinal MR images. [00135] Multiple medical images may be acquired. The medical images may include multiple medical images, such as positron emission tomography (PET), magnetic resonance imaging (MRI), including magnetic resonance spectroscopy imaging (MRSI), or ultrasound images. The medical images may include images of tumors, such as those associated with prostate cancer. The medical images may be acquired in conjunction with photodynamic therapy (PDT), a therapeutic modality for cancer treatment, including pre-PDT, post-PDT, or twenty-four hours after PDT. The image registration technique may be used, for example, to study the tumor response to PDT. PET images can provide physiological and functional information. High-resolution MRI can provide anatomical and morphological changes. [00136] Image registration can be used to combine MRI and PET images for improved tumor monitoring. For example, high-resolution MRI and microPET

[¹⁸F]fluorodeoxyglucose (FDG) images may be acquired from C3H mice with RIF-I tumors that were treated with Pc 4-based PDT. For tumor registration, a finite element model (FEM)-based deformable registration scheme may be implemented. To assess the registration quality, slice by slice review of both image volumes may be performed, the volume overlap ratios may be computed, and both volumes may be visualized in color overlay. In one study, the mean volume overlap ratios for tumors were 94.7% after registration. Registration of high-resolution MRI and microPET images can combine anatomical and functional information of the tumors and can be used as a tool for evaluating PDT.

[00137] PDT is a therapeutic modality for cancer treatment. With PDT, a tumor- localized photosensitizer is irradiated with visible light to generate reactive oxygen that efficiently kills cells and ablates tumors. PDT can be administered deep into tumors using minimally invasive techniques as a small laser fiber is inserted into the lesions and delivers the light to the tumor. PDT with porfimer sodium (e.g., Photofrin^®) is US-FDA approved for treating early and advanced lung cancer, advanced esophageal cancer, and Barrett's esophagus. Both the photosensitizer and the light are inert by themselves. The light can be precisely focused onto a selected region, allowing specificity in the localization of the photodynamic effect. Consequently, systemic toxicities are minimized. [00138] Imaging techniques may provide tools for assessing PDT efficacy. For example, MRI may be used to evaluate PDT-induced vascular damage followed by hemorrhagic necrosis in murine Ml tumors in mice. Blood oxygenation level-dependent (BOLD) contrast MRI may show attenuation (e.g., 25-40%) of MR signal at the treated tumor site. Decreases in contrast agent uptake rates following PDT may be observed by gadolinium contrast MRI. Additionally, in vivo ³¹P nuclear magnetic resonance (NMR) may be used to monitor tumor metabolic status before and after the treatment of RIF-I tumors and mammary carcinoma. The NMR data may reveal significant differences in the time course of high energy phosphate levels in combined hyperthermia and photodynamic therapies. There may be a relationship between NMR measurements immediately (i.e., as soon as possible, such as within one minute or one hour) following PDT and the ultimate effect on the tumor. Moreover, diffusion-weighted MRI may show a biphasic change in the apparent diffusion coefficient (ADC) within the first twenty-four hours post-PDT, indicating the early response of PC-14 tumors to PDT. Additionally, use of PET with ⁾⁸F-fluorodeoxyglucose (FDG) to image mice after PDT may show that the tumor FDG uptake decreased immediately after PDT.

[00139] In one embodiment, multiple imaging modalities for monitoring PDT efficacy may be combined. For example, PET can image the rapid biochemical and physiological responses of tumors to PDT whereas MRI provides superior assessment of anatomical information, location, and morphological changes within tumors. Combining PET and MRI may provide several advantages. For example: 1) MRI scans may provide an anatomical reference for the PET images, 2) fusion of MRI and PET images can enhance one's ability to visualize the distribution of a radio-labeled pharmaceutical, 3) MRI can provide tumor shape and size information that may be used to improve the accuracy of the PET data analysis (e.g., drawing regions of interests (ROIs) and performing quantitative analyses), and 4) MRI can be used to correct PET data for partial volume effects, for example, to clarify whether PET- measured changes induced by PDT are due to metabolic and hemodynamic changes or artifacts of changes in tumor size.

[00140] In one embodiment, the method may provide registration for MRI and PET images. Rigid-body registration algorithms for MRI and PET images, for example, have been used for tumors and human brain. Deformable registration may be used when the subject is in different positions or the organ is deformed. Finite element models (FEMs) may be used, for example, for registration of images of the brain, lung, prostate, and coronary arteries. These methods may be applied to register images from the same modality. For example, thin-plate spline based registration techniques may be implemented for MRI. These methods may be used for human image registration.

[00141] In one embodiment, a deformable method may provide registration for tumor microPET and MR images. Imaging and PDT experiments, for example, may be performed on tumor-bearing mice. Registration results may be provided for visual inspection and quantitative measurements. [00142] Animal preparation

[00143 J RIF (Radiation-induced fibrosarcoma)- 1 cells may be grown as monolayers in

E-MEM supplemented with 15% fetal bovine serum. Prior to inoculation, C3H/HeN mice may be shaved and depilated. Two tumors may be initiated in each mouse by injection of 10⁵ - 10⁶ RiF-I cells intradermally on the shoulder flanks. Tumors may be treated and imaged when they reached 3-5 mm in diameter, which may be 7-10 days after implantation. Animals may be given the photosensitizer Pc 4 (e.g., 1 mg/kg) by tail vein injection. It is known that neither the light nor the photosensitizer alone produces any response. After forty-eight hours, one of the tumors in each animal may be exposed to red light (e.g., 670 ran) from a diode laser (e.g., 150 J/cm ; 150 mW/cm ). The other tumor in each animal may serve as a control (i.e., receiving photosensitizer, but no light). The animals may be studied by microPET and MR imaging.

[00144] Image Acquisitions

[00145] Two days after photosensitizer injection, MR images may be acquired using a

1.5 T scanner (e.g., Siemens Sonata 1.5T scanner from Siemens Medical Systems, Erlangen, Germany). A dedicated custom-designed whole-body mouse coil (e.g., 2-element phased- array, ID = 32 mm) may be used to minimize noise levels. A Tl -weighted spin echo pulse sequence (e.g., TR/TE=600/13ms) with a slice thickness of 1 mm may be used to generate high-resolution coronal images (e.g., Matrix: 256 x 120, FOV: 80 x 36-mm₅ Pixel size: 0.3 x 0.3-mm). The acquisition time for an image slice, for example, may take about 72 seconds, m these Tl -weighted images, the tumors may be delineated by the bright subcutaneous fat signal. Three to five MR image volumes may be acquired for each mouse. [00146] After MR image acquisition, a microPET scanner (e.g., a micropET R4 scanner from Concorde Microsystems, Inc., Knoxville, TN 37932) may be used to follow a single animal over, foe example, a 90-minute period of time to monitor the response to PDT and the outcome. A standard radiopharmaceutical (e.g., ¹⁸F-FDG) may be used in PET scanning for tumor diagnosis and assessment. Both transmission and emission images may be acquired from the same mouse. The animal may be anesthetized so that it remains in the same position during the imaging session and so that no movement can be assumed between the PET transmission and emission scans. One PET image volume may include 63 transverse slices covering the whole mouse with each slice including, for example, 128 x 128-pixel with an in-plane pixel size of 0.85 x 0.85-mni and a thickness of 1,2 mm. For example, 10-22 dynamic PET image volumes may be acquired from each mouse. The total FDG activity for the imaging period may also be computed to create another PET image volume. These volumes may be used for registration experiments.

[00147] Image Processing

[00148] Interpolation may be used to create isotropic MR volumes before registration.

The input MR volume may be a 2D MR acquisition with a pixel size of 0.3 x 0.3-mm and a slice thickness of 1.0-mm. For example, twenty nine coronal slices may cover the whole mouse. Using a sine-linear interpolation, for example, 0.3 mm isotropic voxels may be created on a side for both PET and MR image volumes. For example, IDL (i.e., Interactive

Data Language from Research System Inc., Boulder, CO) may be used as the programming language.

[00149] For the purposes of deformable registration, image slices that are not of interest may be optionally cropped. For example, if the tumors are on the mouse back, near the shoulder, images at the head and abdomen may be cropped. An exemplary image volume covering the whole mouse before cropping may be 350 x 250 x 250-voxels. After cropping, the volume near the region of interest may be a 148 x 80 x 90-voxels. Cropping out regions that are not of interest can increase image consistency for the mutual information registration.

For example, flexible areas of a mouse body, such as the abdomen, may have deformations that cause inconsistency for image registration, It can be beneficial to crop areas these areas.

The smaller number of voxels after cropping can also increase the speed of image registration.

[00150] Deformable Tumor Registration

[00151 ] In a first step, the rigid-body normalized mutual information-based registration algorithm may be applied to align the cropped MRI and microPET images. After registration, the tumor may be manually segmented, slice-by-slice, on both high-resolution

MRI and microPET image volumes.

[00152] In a second step, a finite element model (FEM)-based deformable registration algorithm may be applied. For a linear elastic continuum with no initial stresses and strains, the deformation energy E of an elastic body submitted to externally applied forces can be expressed using the equation, below:

E = - [ σ^τs d Ω+ [F u d Ω

2 ^ ^h (11), where u is the displacement vector, Ω is the elastic body, ^σ is the stress vector, ^ε is the strain vector, and F is the force applied to the elastic body. For a material with the maximum symmetry (i.e., an isotropic material), the material properties are the same in every direction. There are only two independent parameters for the stress and strain vectors ( ^σ and ^ε ): the Young's modulus that relates tension and stretch, and the Poisson ratio that is the ratio of the lateral contraction due to the longitudinal stretch.

[00153] The displacement field u within each element may be approximated as an assembly of discrete elements interconnected at the nodal points on the element boundaries. Tetrahedral elements may be used for the volumes and triangles for the surfaces. A commercially-available software application (e.g., AMIRA from Mercury Computer Systems, Inc, Chelmsford, MA) may be used to build the meshes for the tumor surfaces. The built tumor surfaces may be imported to a commercially-available finite element analysis software application (e.g., FEMLAB from COMSOL, Inc., Burlington, MA). The tumor, as defined by the surface, may be partitioned into a union of tetrahedral elements using an unstructured meshing method in the finite analysis software application (e.g., FEMLAB). For example, over 500,000 tetrahedral solid elements may be created to represent the solid tumor model. The boundary condition may be defined at the surface vertices (e.g., > 800). For each surface vertex on the MRI model, its distances to the surface vertices on the PET model may be computed. The closest vertex is the corresponding point. The displacement fields of the surface vertices may serve as the boundary motion of the tumor. Additional external forces do not need to be applied to the tumor model.

[00154] The registration approach may deform the tumor surface from the MRI volume toward that from the PET image. The displacement forces at the surface vertices may drive the elastic surface from the MRI image toward that from the PET image. The tumor may be modeled as a linear isotopic elastic material with Young's modulus of 60 kPa and Poisson' s ratio of 0.49. The FEM model may be used to infer volumetric deformation of the tumor from the surface. The force may be integrated over each element and may be distributed over the nodes belonging to the element using its shape functions. After obtaining the displacement field for all vertices, a linear interpolation may be used to obtain the deformed image volume of the tumor. [00155] Registration Evaluation

[00156] A variety of qualitative and quantitative methods may be used to evaluate the registration of microPET and high-resolution MRI. For example, visual inspection methods may be used to evaluate the registration quality. Color overlay displays may provide a useful tool to evaluate structure overlap. For example, rendering one image in gray and the other in red with a manually adjustable transparency scale may provide a way to visually determine registration accuracy. A checkerboard display, where the reference and registered images are divided into sectors, may be used to create an output image by alternating sectors from the two input images. Even small shifts of edges, may be clearly visible in the checkerboard display. 3D volume rendering and color overlap may also be used to visualize registration results.

[00157] Additionally, quantitative registration errors may be computed. For example, the tumor boundaries in image slices may be manually segmented and copied to corresponding slices from other registered volumes. From each segmented slice, the center of the lesion may be computed. From the segmented boundaries across all slices, the centroids of the lesion in 3-D space may be computed. This enables offline visual determination of the registration quality. By manually segmenting the lesion from multiple volumes, centroid distances and volume overlap ratios (VOR) may be derived to evaluate the registration quality. The VOR may be defined as the overlap volume and divided by the average of the volumes measured from MRI and PET images. A VOR value, for example, ranges from 0 (no overlap) to 1 (full overlap). The consistency errors for the deformable registration may be measured. For example, a voxel in Volume A may be transformed to Volume B and then transformed back to A. The distances between the corresponding voxel after the two deformable transformations may serve as a measure of the registration consistent errors.

[00158] With reference to FIG. 7, exemplary results of tumor registration after rigid- body transformation are shown. The color overlay of the MRI and microPET images demonstrates good registration of the tumor in both transverse and coronal slices indicating that the tumors are aligned in three dimensions. FIG. 7 shows registration and fusion of MRI and microPET images in transverse (left frames) and coronal (right frames) orientations. In the top frames, MR images that cover the tumor region are shown. Corresponding PET emission images are shown in the middle frames. Color overlays of the MRI (gray) and microPET (red) images are shown in the bottom frames. The fusion images show that the tumors were aligned.

[00159] To evaluate the rigid-body registration, the tumor may be manually segmented from both MRI and microPET images and 3D meshes may be used to represent the tumor surfaces. In FIG. 8, an exemplary 3D visualization shows that the tumor deformed between the two imaging sessions. In order to evaluate manual segmentation errors, two observers segmented each tumor three times. The volume overlap ratios of the six segmentations are 95.0% ± 1.0% and 92.0% ± 2.6% for MRI and PET images, respectively. This indicates excellent repeatability. FIG. 8a shows the tumor segmented from a high-resolution MR volume. FIG. 8b shows the same tumor segmented from corresponding microPET emission images. FIG. 8c shows a color overlay of the tumor from MRI (yellow) and microPET (red) images. As shown, the tumor deformed during the two imaging sessions. [00160] With reference to FIG. 9, the results of exemplary rigid and exemplary deformable registration are shown for comparison. The contour overlap shows that the deformable method may be better than the rigid-body registration. This is consistent with quantitative measures. The NMI values increased from 0.06 & 0.01 to 0,12 ± 0.02 after deformable registration. The volume overlap ratios were also improved from 86.3% ± 2.5% to 94.7% ± 1.5 % with deformable registration. The mean consistence error is less than 0.1 mm for the deformable registration. The corresponding MRI (a) and microPET (b) images after rigid-body registration are shown in the (a) and (b) frames of FIG. 9. The tumor on both images was manually segmented for registration evaluation in the (c) and (d) frames of FIG. 9. The tumor contour from the microPET image (d) is copied to the MR image (c). The contour mismatch is due to the tumor deformation. After deformable registration, the tumor on the MRI is warped and matched with that from the microPET image in the (e) frame. Other slices are also matched indicating excellent tumor registration in three dimensions. [00161] Moreover, the treated tumor shows less FDG uptake than the control indicating the effect of PDT. This is consistent with the microPET images (see FIG. 7). The fusion of PET with MRI images may aid in defining regions of interest on PET images for quantitative measurements. The tumor registration and fusion methods may be useful for this application.

[00162] The deformable registration method is accurate for tumor registration.

Because the tumors on both MRI and microPET images were already segmented, the registration quality was well controlled. The deformable registration performs better than the rigid-body method whenever there are deformations of the tumors. For example, using a Pentium IV computer (e.g., 3.4 MHz CPU and 3.0 GBytes memory) and commercially- available software (e.g., FEMLAB), the computation time for the deformable transformation was less than four minutes.

[00163] The MR image quality was excellent when using a dedicated mouse coil. A high in-plane pixel size of 100 x 100-μm for small animal imaging was achieved using a clinical 1.5 T MR scanner. Additional experiments may be performed on 7 T and 9.4 T superconducting MR imaging systems (e.g., Bruker Biospec superconducting MR imaging systems).

[00164] Tumors may respond rapidly to PDT. PET and MRI images either during the photo irradiation or within a short time thereafter may be used to assess the in vivo response of tumors to PDT. It will be important to ensure that changes in metabolic parameters, as measured by PET imaging, are properly assigned to the treated tumor or other tissue of interest. Deformable image registration should improve the ability to quantitatively evaluate the desired responses.

[00165] An exemplary deformable registration method for tumor MRI and microPET images is provided herein. The image registration and fusion of the method may provide both functional and anatomic information for evaluating PDT, The method could also serve as a tool for other applications of imaging in cancer biology, functional genomics, and drug development.

[00166] MAGE REGISTRATION OF MEDICAL IMAGES

[00167] hi another embodiment, a medical apparatus provides image registration of medical images from multimodality imaging to assess the efficacy of cancer therapy. For example, the efficacy of Pc 4-based PDT for the human prostate cancer model. For one imaging modality, FIG. 10 shows an exemplary embodiment of a process for F-

Fluorodeoxyglucose imaging of glucose uptake and metabolism with Positron Emission

Tomography (PET).

[00168] Animal Preparation

[00169] Human prostate tumor cells PC-3 and other cancer cells such as RIF-I cells, may be prepared and implanted on the back of athymic nude mice as described herein. The mice may be treated after the implanted tumors reached a size of 5-10 mm. Treatment may include injection of Pc 4. PDT may be provided approximately forty-eight hours after the Pc

4 injection as described herein.

[00170] Image Acquisitions

[00171 ] MR and PET image acquisition may be provided as described herein.

[00172] Image Processing

[00173] As described herein, interpolation may be used to create isotropic MR volumes before registration. The input MR volume may be a 2D MR acquisition with a pixel size of 0.3 x 0.3-mm and a slice thickness of 1.0-mm. For example, twenty nine coronal slices may cover the whole mouse. Using a sine-linear interpolation, for example, 0.3 mm isotropic voxels may be created on a side for both PET and MR image volumes. For example, IDL (i.e., Interactive Data Language from Research System Inc., Boulder, CO) may be used as the programming language.

[00174] The PET data may be discretized to 256 gray levels for image display and processing. The scaled data may be used for NMI registration. Registration performance may be examined using different intensity scaling such as 512, 256, 128, 64, or 32 bins for both volume data sets. Scaling between zero and the maximum value may be linear. Registration quality may be analyzed by NMI values and by visual inspection. [00175] As described herein, for the purposes of deformable registration, image slices that are not of interest may be optionally cropped. For example, if the tumors are on the mouse back, near the shoulder, images at the head and abdomen may be cropped. An exemplary image volume covering the whole mouse before cropping may be 350 x 250 x 250-voxels. After cropping, the volume near the region of interest may be a 148 x 80 x 90- voxels. Cropping out regions that are not of interest can increase image consistency for the mutual information registration. For example, flexible areas of a mouse body, such as the abdomen, may have deformations that cause inconsistency for image registration. It can be beneficial to crop areas these areas. The smaller number of voxels after cropping can also increase the speed of image registration. [00176] Automatic whole body registration

[00177] An automatic three-dimensional (3D) rigid-body registration algorithm may be used for alignment of the whole mouse body. Normalized mutual information (NMI) may be chosen as the similarity measure for the rigid-body registration because it does not require a linear relationship between the intensity values of the two images and it is suitable for multimodality image registration. One image R is the reference, and the other image F is floating. NMI for image R and image F is given by the following equation:

_{M(Λ|/0 =}J*WL

H(R) ₊ H[F) (₁₂)₍ where:

H(Λ)=-∑>_Λ«logp_Λ(r) r (13),

MI(R,F) =* ∑p^rJVog-^^-r

T_f p_R(ή -p_F(f) (15)_, [00178] The joint probability PXF^V >J ) , marginal probability of the reference image

P^R ^ ' , and marginal probability of the floating image PF ^J ' can be estimated from the normalized joint intensity histograms. When two images are geometrically aligned, NMI is maximal.

[00179] The PET transmission and emission images may be combined to form one data set by taking a weighted sum. The combined PET data and high-resolution MR image may be used for registration of whole mouse body. The transmission images may provide anatomic information to aid in the NMI registration. The MRI data may be used as the floating image because it is higher resolution than the PET images. Rigid-body transformation (e.g., three translations and three rotations) and trilinear interpolation may be used. A downhill simplex method may be used for optimization. For additional information on downhill simplex methods, see Nelder et al., A simplex method for function minimization, Comput. J., 1965; 7:308-313, the contents of which are fully incorporated herein by reference. Optimization of similarity ends either when the maximum number of calculations (e.g., 800) is reached or the fractional change in similarity function is smaller than a tolerance (e.g., 0.001). Typically, the latter is achieved within about 200 iterations. The first initial guess for the three displacements and three angles may be all zeros.

[00180] FIG. 11 shows a mutual information based method for image fusion and a corresponding joint histogram. FIG. 12 shows an example of a whole mouse body overlay of MRI and PET images in which the heart and bladder can be observed. FIG. 13 shows an example of a whole mouse body overlay of MRI and PET images in which the kidney can be observed. FIG. 14 shows an example of a whole mouse body overlay of MRI and PET images with the cancer model described herein and a corresponding 3D rendering of the whole mouse body.

[00181] FIG. 15 shows an exemplary FDG uptake after PDT therapy for treated and control tumors. As shown, within 20 minutes after PDT therapy, the uptake for the treated tumor was lower in comparison to the control tumor, FIG. 16 shows serial Tl and T2 MR imaging of the treated tumor pre-PDT, immediately post-PDT, and twenty-four hours after PDT. FIG. 17 shows an exemplary correlation of the Tl- and T2-weighted MR images, an exemplary image classification result showing the necrotic region on the MR images, and a corresponding histological image for evaluation of the classification. As demonstrated, in vivo imaging can provide a useful tool to monitor early tumor response to therapy. Moreover, combining MRI and PET images can provide both anatomic and functional information about tumor response to PDT.

[00182] IN VIVO MR IMAGING FOR ASSESSMENT OF PDT

[00183] In one embodiment, a medical apparatus provides an in vivo magnetic resonance (MR) imaging technique for assessment of photodynamic therapy (PDT). The in vivo MR imaging (MRI) technique may assess early effects of PDT for treatment of cancer, including human prostate cancer, In one embodiment, the in vivo MRI technique may use changes in T2 values as a surrogate biomarker for evaluating PDT efficacy. In another embodiment, the in vivo MRI technique may use changes in apparent diffusion coefficient (ADC) as a surrogate biomarker for evaluating PDT efficacy.

[00184] PDT is a therapeutic modality for treatment of various cancers. A second- generation photosensitizing drug, silicon phthalocyanine 4 (Pc 4), may be used in conjunction with PDT for cancer treatment. Various medical imaging techniques, including magnetic resonance imaging (MRI) techniques, may provide monitoring and early assessment of tumor response to PDT.

[00185] In certain studies, human prostate cancer xenografts in athymic nude mice were generated. For imaging experiments, a high-field 9.4-T small animal MR scanner (e.g., Bruker Biospec) was used. High-resolution MR images were acquired from the treated and control tumors pre- and post-PDT and twenty-four hours after PDT. Multi-slice multi-echo (MSME) MR sequences were utilized. During imaging acquisitions, the animals were anesthetized with a continuous supply of l%-2% isoflurane in oxygen and were continuously monitored for respiration and temperature. After imaging experiments, the tumors were manually segmented on each image slice for quantitative image analyses. Three-dimensional (3D) T2 maps were computed for the tumor regions from the MSME images. Histograms of the T2 maps were plotted for each tumor pre- and post-PDT and twenty-four hours after PDT. After the imaging and PDT experiments, the tumor tissues were dissected and histologic slides were used to validate the MR images.

[00186] In these studies, six mice with human prostate cancer tumors were imaged before and after treatment. The T2 values of treated tumors increased by 24 ± 14% twenty- four hours after the therapy. The control tumors did not demonstrate significant changes of the T2 values. Inflammation and necrosis were observed within the treated tumors twenty- four hours after the treatment. Thus, Pc 4-PDT may be effective for the treatment of human prostate cancer. The MR imaging technique may provide a useful tool to evaluate early tumor response to PDT. Other medical imaging techniques with similar capabilities may also provide useful tools for assessment of PDT. Moreover, these medical imaging technique can be applied to PDT for various types of cancer, including human prostate cancer. [00187] PDT is a therapeutic modality for cancer treatment. With PDT₅ a tumor- localized photosensitizer is irradiated with red light to generate reactive oxygen that efficiently kills cells and ablates tumors. PDT can be administered deep into tumors using minimally invasive techniques as a small laser fiber is inserted into the lesions and delivers the light to the tumor. PDT with porfimer sodium (e.g., Photofrin^®) is US-FDA approved for treating early and advanced lung cancer, advanced esophageal cancer, and Barrett's esophagus. Both the photosensitizer and the light are inert by themselves. The light can be precisely focused onto a selected region, allowing specificity in the localization of the photodynamic effect. Consequently, systemic toxicities are minimized. [00188] MRI techniques may provide tools for assessing PDT efficacy. For example,

MRI may be used to evaluate PDT-induced vascular damage followed by hemorrhagic necrosis in murine Ml tumors in mice. Blood oxygenation level-dependent (BOLD) contrast MRI may show attenuation (e.g., 25-40%) of MR signal at the treated tumor site. Decreases in contrast agent uptake rates following PDT may be observed by gadolinium-contrast MRI. Additionally, in vivo ³¹P nuclear magnetic resonance (NMR) spectroscopy may be used to monitor tumor metabolic status before and after the treatment of RIF-I tumors and mammary carcinoma. The NMR data may reveal significant differences in the time course of high energy phosphate levels in combined hyperthermia and photodynamic therapies. There may be a relationship between NMR measurements immediately following PDT and the ultimate effect on the tumor. Moreover, diffusion-weighted MRI may show a biphasic change in the apparent diffusion coefficient (ADC) within the first twenty-four hours post-PDT, indicating an early response of PC- 14 tumors to PDT.

[00189] MRI and positron emission tomography (PET) may be used to image C3H mice bearing RIF-I tumors after PDT. PET with ¹⁸F-fluorodeoxyglucose (FDG) may provide metabolic information of the tumors. High-resolution MRI may provide anatomical and morphological changes of the lesions. Both rigid and deformable image registration methods may be used to combine MRI and PET images for improved tumor monitoring. Fusion of MRI and PET images may provide both anatomical and functional information of the tumors for evaluating PDT effects. The tumor FDG uptake may be decreased immediately after PDT. In one embodiment, MR imaging and analysis methods for monitoring the efficacy of Pc 4-based prostate PDT in vivo may be provided. Quantitative image analysis techniques may identify subtle changes immediately after PDT for evaluating therapeutic efficacy. [00190] Pc 4 Formulation

[00191] The chemical synthesis of Pc 4 was described in Oleinick et al., "New phthalocyanine photosensitizers for photodynamic therapy," Photochem. Photobiol., vol. 57, pp. 242-247, Feb, 1993, the contents of which are fully incorporated herein by reference. A stock solution (e.g., 1 mg/ml) may be made by dissolving Pc 4 at 10 mg/ml in 50% Cremophor EL, 50% absolute ethanol, then adding nine volumes of normal saline with mixing. For injection, the Pc 4 stock solution may be mixed with an equal volume of 5% Cremophor EL, 5% ethanol. Each animal may be weighed at the time of injection, and the volume of injected solution may be adjusted to give the desired dose in mg/kg. [00192] PC-3 Human Prostate Cancer Model

[00193] Human prostate cancer PC-3 cells may be grown as monolayers in E-MEM supplemented with 15% fetal bovine serum. Male athymic nude mice of 4-8 weeks old may be obtained and housed under pathogen- free conditions. They may be maintained under controlled conditions (e.g., 12-hour dark-light cycles; temperature 20-24⁰C) with free access to sterilized mouse chow. Two tumors may be initiated in each mouse by injection of 105 - 106 PC-3 cells intradermally on the shoulder flanks. One of the tumors may be treated and the other may serve as the control. [00194] PDT Protocol

[00195] Tumors may be treated and imaged after reaching 5-8 mm in diameter, which may be 2-4 weeks after implantation. Animals may be given Pc 4 (e.g., 0.6 mg/kg) by tail- vein injection. It is known that neither the light nor the photosensitizer alone produces any response. After forty-eight hours, a 1-cm area encompassing the tumor may be irradiated with red light (e.g., 672 nm; 150 J/cm²; 150 mW/cm²) from a diode laser (e.g., Applied Optronics Corp., Newport) coupled to a fiber optic terminating in a microlens that distributes light uniformly throughout the treatment field. One of the tumors in each animal may be exposed to red light and the other tumor may serve as a control (i.e., receiving photosensitizer, but no light). To measure histologic responses following Pc 4-PDT, mice may be sacrificed at different time points after therapy. The tumors may be surgically removed and immediately stored in 10% formalin for 2-7 days before histologic processing. [00196] Image Acquisitions

[00197] Two days after photosensitizer injection, high-resolution MR images may be acquired from each mouse pre- and post-PDT and twenty-four hour after PDT. The mouse MR images may be acquired using a high-field (e.g., 9.4-T) small animal MR scanner (e.g., Broker BioSpin GmbH, Rheinstetten, Germany). A dedicated whole body mouse coil may be used for the image acquisitions. A multi-slice multi-echo (MSME) sequence (e.g., TR=6929 ms and TE-28, 56, 84, and 111 ms) with a slice thickness of 0.5 mm may be used to generate high-resolution coronal images (e.g., Matrix: 256 x 256, Pixel size: 0.27 x 0.13-mm). In these T2-weighted images, the tumors may be clearly delineated by the bright subcutaneous fat signals.

[00198] Image Processing And Analysis

[00199] A commercially-available software application (e.g., Analyze from

AnalyzeDirect, Inc., Overland Park, KS) may be used to segment the tumor on each image slice from the MR image volumes. The segmented images may be used for the calculation of T2 maps and tumor volumes. For example, four 0.5 mm MSME image slices may be used to generate the T2 maps over a 2.0 mm tumor region. A commercially-available software application (e.g., Paravision 3.1 from Bruker BioSpin GmbH, Rheinstetten, Germany) may be used for computation of T2 maps. Once the T2 values for each voxel within the tumor region are obtained, the histogram, mean, and standard deviation of the T2 values of the tumor may be calculated.

[00200] With reference to FIG. 18, an exemplary set of MSME time-phased images of a treated tumor are shown. The images were acquired pre- and post-PDT and twenty-four hours after PDT. As shown, changes within the treated tumor region can be observed twenty- four hours after PDT. The MSME MR images show the treated tumor pre-PDT (left), immediately post-PDT (middle), and twenty-four hours after PDT (right). As shown, the signal intensity values changed twenty-four hours after the treatment. [00201] With reference to FIG. 19, the corresponding T2 maps calculated from the

MSME images of FIG. 18 are shown. The T2 maps show the treated tumor pre-PDT (left), immediately post-PDT (middle) and twenty-four hours after PDT (right). The MSME time- phased images of FIG. 18 were used to calculate the T2 maps. As shown, the T2 values of the treated tumor increased twenty-four hours after PDT.

[00202] With reference to FIG. 20, the corresponding histograms calculated from the

T2 maps for pre- and post PDT and twenty-four hours after PDT of FIG. 19 are shown. As shown, the peak in the histogram shifted to the right twenty-four hours after PDT, indicating that the T2 values increased after the treatment. This is consistent with the observed changes on the MSME images of FIG. 18 and the T2 maps of FIG. 19. The calculated means for the T2 values were 56.1 ± 16.0 ms, 53.6 ± 15.7 ms, and 65.5 ± 13.1 ms for pre-PDT, immediately post-PDT, and twenty-four hours post-PDT, respectively. This shows that twenty-four hours after PDT, the mean T2 values increased by 9.5 ms, which is a 17% increase. [00203] With reference to FIG. 21, the T2 values of the treated and control tumors for six mice (M1-M6) as well as the mean T2 values are provided. For the treated tumors, the mean T2 values are 53.4 ± 7.8 ms, 52.6 ± 6.2 ms, and 65.9 ± 9.4 ms pre-PDT, post-PDT, and twenty- four hours after PDT, respectively. For the control tumors, the mean T2 values are 48.4 ± 6.9 ms, 49.8 ± 6.9 ms, and 52.0 ± 11.2 ms pre-PDT, post-PDT, and twenty-four hours after PDT, respectively. For the treated tumors, the mean T2 values increased by 24 ± 14% twenty-four hours after PDT. As shown, there is a significant difference between the T2 values pre-PDT and twenty-four hours after PDT (p=0.002). For the control tumors, the T2 values increased by 6 ± 9%, indicating a non-significant difference between the T2-values of pre-PDT and twenty-four hours after PDT (p=0.114). Not that the first mouse (Ml) had only one tumor, which was treated.

[00204] Histologic slides showed that an inflammatory response with edema was observed in the treated tumor, which was not seen within the control tumor. This is consistent with the increases of the T2 values of the treated tumor tissues. [00205] As described herein, various MR imaging and analysis methods may be used to assess the efficacy of PDT on, for example, human prostate cancer growing as a xenograft in athymic nude mice. High-resolution MR images can detect early tumor response to the therapy twenty-four hours after the treatment. For treated tumors, the T2 values may significantly increase one day after the treatment. For control tumors, no significant difference between the T2 values before and twenty-four hours after the therapy may be observed. MR parameters (e.g., T2 values) may provide a surrogate biomarker to predict the success of the therapy. This may provide a tool for other applications of medical imaging in cancer biology, functional genomics, and drug development. [00206] HIGH-FIELD MR IMAGING FOR ASSESSMENT QF PDT

[00207] In one embodiment, a medical apparatus provides a high-field magnetic resonance (MR) imaging technique for assessment of photodynamic therapy (PDT). The high-field MR imaging (MRI) technique may assess early effects of PDT for treatment of cancer, including human prostate cancer. In one embodiment, the high-field MRI technique may use changes in T2 values as a surrogate biomarker for evaluating PDT efficacy. In another embodiment, the high-field MRI technique may use changes in apparent diffusion coefficient (ADC) as a surrogate biomarker for evaluating PDT efficacy. [00208] High-field MRI is a technique that provides a non-invasive tool for in vivo studies of cancer therapy in animal models. PDT is a modality for treatment of cancer, including prostate cancer which is the second leading cause of cancer mortality in American males. As described herein, a high-field MRI technique may bed used to evaluate the response of human prostate tumor cells growing as xenografts in athymic nude mice to Pc 4- sensitized PDT.

[00209] In certain studies, PG-3, a cell line derived from a human prostate malignant tumor, was injected intradermally on the back flanks of athymic nude mice. Two tumors were initiated on each mouse. One was treated and the other served as the control. A photosensitizing drug, Pc 4 (e.g., 0.6 mg/kg body weight) was delivered to each animal by tail vein injection forty-eight hours before laser illumination (e.g., 672 nm, 100 niW/cm², 150 J/cm²). A high-field (e.g., 9,4 Tesla) small-animal MR scanner was used for image acquisitions. A multi-slice multi-echo (MSME) technique, permitting noninvasive in vivo assessment of potential therapeutic effects, were used to measure T2 values and tumor volumes. Each animal was scanned immediately before and after (i.e., as soon as possible, such as within five minutes or one hour) before and after PDT and twenty-four hours after PDT. T2 values were computed and analyzed for the tumor regions.

[00210] In these studies, for the treated tumors, the T2 values significantly increased

(e.g., p < 0.002) twenty-four hours after PDT (e.g., 68.2 ± 8.5 ms), compared to the pre-PDT values (e.g., 55.8 ± 6.6 ms). For the control tumors, there was no significant difference (e.g., p = 0.53) between the pre-PDT (e.g., 52.5 ± 6.1 ms) and twenty-four hours post-PDT (e.g., 54.3 ± 6.4 ms) values. Histologic analysis showed that PDT-treated tumors demonstrated necrosis and inflammation that was not seen in the control. As shown, changes in tumor T2 values measured by multi-slice multi-echo MR imaging may provide an assay that could be useful for clinical monitoring of PDT, including PDT for treatment of prostate tumors. [00211] PDT is a therapeutic modality for cancer treatment. With PDT, a tumor- localized photo sensitizer is irradiated with red light to generate reactive oxygen species that efficiently kills cells and ablates tumors. Both the photosensitizer and the light are inert by themselves. The light can be precisely delivered to a selected region, allowing specificity in the localization of the photodynamic effect. Consequently, side effects are minimized. PDT with porfimer sodium (e.g., Photofrin^®) is US-FDA approved for treating early and advanced lung cancer, advanced esophageal cancer, and Barrett's esophagus. PDT with second- generation photosensitizers may be used for treating a variety of cancers, including prostate cancer.

[00212] Prostate cancer is the second leading cause of cancer mortality in American males. The current therapy options for patients with clinically localized prostate cancer include: a) radical prostatectomy; b) external beam radiation therapy; and c) interstitial brachytherapy. These methods can have serious side effects, such as incontinence and sexual dysfunction. If radiation therapy fails, there may only be a limited number of salvage options available for treatment of recurrent prostate cancer. PDT may be a salvage treatment modality for recurrent localized prostate cancer. PDT can be administered deep into tumors using minimally invasive techniques as a small laser fiber is inserted into the lesions and that deliver the light to the tumor. Second-generation photosensitizing drugs such as Pc 4, motexafin lutetium (Lu-Tex), Pd-bacteriopheophorbide (TOOKAD), aminolevulinic acid (ALA), mTHPC, and SnET2 may be used for treating various forms of cancer, including prostate cancer.

[00213] MRI and MR spectroscopy (MRS) techniques may be useful tools for assessing PDT efficacy. For example, MRI may be used to evaluate PDT-induced vascular damage followed by hemorrhagic necrosis in murine Ml tumors in mice. Blood oxygenation level-dependent (BOLD) contrast MRI may show attenuation (e.g., 25-40%) of MR signal at the treated tumor site. Decreases in contrast agent uptake rates following PDT may be observed by gadolinium-contrast MRI. For example, gadolinium diethylenetriamene pentaacetate (DTPA) contrast-enhanced MRI may be used to assess the boundary of PDT- induced tissue necrosis in a canine model and in human patients. Additionally, in vivo ³¹P nuclear magnetic resonance (NMR) spectroscopy may be used to monitor tumor metabolic status before and after the treatment of RIF-I tumors and mammary carcinoma. The NMR data analysis may reveal significant differences in the time course of changes in high energy phosphate levels in response to combined hyperthermia and photodynamic therapies. There may be a relationship between NMR measurements immediately following PDT and the ultimate effect on the tumor. Moreover, diffusion-weighted MRI may show a biphasic change in the apparent diffusion coefficient (ADC) within the first twenty-four hours post- PDT, indicating the early response of PC-14 tumors to PDT.

[00214] MRI and positron emission tomography (PET) may be used to image C3H mice bearing RIF-I tumors after PDT. PET with ¹⁸F-fluorodeoxyglucose (FDG) may provide metabolic information of the tumors. High-resolution MRI may provide anatomical and morphological changes in the lesions. As described herein, registration methods may be used to combine MRI and PET images for improved tumor monitoring. Fusion of MRI and PET images may provide both anatomical and functional information about the tumors for evaluating PDT effects. The tumor FDG uptake may be decreased immediately after successful PDT. [00215] As described herein, high-field MRI may be used to monitor the early response of cancer, including prostate cancer, to PDT. For example, this is an in vivo imaging technique that may be used to assess Pc 4-based PDT of prostate cancer. The medical apparatus and associated method disclosed herein provides non-invasive imaging and quantitative analysis techniques to identify subtle changes that occur within twenty-four hours after PDT for evaluating its therapeutic efficacy. [00216] Pc 4 Formulation

[00217] A second-generation photosensitizing drug, silicon phthalocyanine 4 (Pc 4),

[HOSiPcOSi(CH₃)₂(CH₂)₃N(CH₃)₂], was developed for treating a variety of cancers. For additional information on the chemical synthesis of Pc 4, see Oleinick et al., New phthalocyanine photosensitizers for photodynamic therapy, Photochem Photobiol 1993; 57:242-247, the contents of which are fully incorporated herein by reference. A stock solution (e.g., 1 mg/mL) may be made by dissolving Pc 4 in 50% Cremophor EL, 50% absolute ethanol, then adding nine volumes of normal saline with mixing. For injection, the Pc 4 stock solution may be mixed with an equal volume of 5% Cremophor EL, 5% ethanol, 90% saline to give a final concentration of 0.05 mg/mL (i.e., 0.07 niM). [00218] Tumor Model

[00219] The PC-3 cell line is derived from a primary malignant human prostate tumor.

PC-3 cells may be grown as monolayers in E-MEM supplemented with 15% fetal bovine serum at 37⁰C. Cells may be harvested by trypsinization in ethylenediaminetetraacetic acid/trypsin, washed in Hank's balanced salt solution (HBSS) without Ca²⁺ and Mg²⁺, and centrifuged at 150g for five minutes. Cells may be counted in a hemacytometer using 0.4% trypan blue. The cell suspension may be brought to a final concentration of 1 x 10⁶ cells/mL and kept on ice for immediate injection.

[00220] Male athymic nude mice of 4-8 weeks old may be obtained and housed under pathogen-free conditions. They may be maintained under controlled conditions (e.g., 12-hour dark-light cycles; temperature 20-24°C) with free access to sterilized mouse chow. Two tumors may be initiated in each mouse by injection of 50 μh containing 5 x 10⁴ PC-3 cells intradermally on each flank at least 20 mm apart and as far from the lung and heart as possible to minimize motion effects in MRI. [00221] Experimental Protocol

[00222] Tumors may be treated and imaged when they reached 8-10 mm in diameter, which may be 2-4 weeks after implantation. Each animal may be weighed at the time of injection, and a volume of Pc 4 solution may be injected intravenously into the tail vein to give 0.6 mg/kg (e.g., 240 μL to a 20 g mouse). This dosage was found to be optimal in another xenograft model (i.e., OVCAR-3 ovarian epithelial carcinoma). Appropriate controls of photosensitizer without light and light without photosensitizer are known to produce no response. Forty- eight hours after photosensitizer injection, the animals may be taken for imaging and PDT. For PDT, a diode laser (e.g., Applied Optronics Corp., Newport, CT) may be used to deliver light (e.g., 672-nm light, the longest wavelength absorption maximum of Pc 4). The laser may be coupled to a fiber optic cable terminating in a microlens. The treatment light may cover the entire tumor and may be distributed uniformly throughout the treatment field. One of the two tumors on each animal may be irradiated, for example, with a fluence of 150 J/cm² and an irradiance of 100 mW/cm². This has been shown to produce a complete response and some cures in other tumor models. The low power of the laser light may preclude thermal effects. The other tumor in each animal may serve as a control (i.e., receiving photosensitizer, but no light). Mice may be euthanized twenty-four hours after PDT to measure early histologic responses to Pc 4-based PDT. The tumors may be harvested and immediately stored in 10% formalin before histologic processing. In one study, a total of thirteen tumor-bearing animals were treated and imaged. Each mouse had two tumors, but mice 1 and 7 each had a small control tumor. Therefore, data was only obtained from twenty- four tumors. [00223] MRI

[00224] Forty-eight hours after photosensitizer injection, high-resolution MR images may be acquired from each mouse pre- and post-PDT, and an additional MR image may be obtained twenty-four hours after PDT. The mice may be imaged immediately after light treatment and twenty-four hours later to focuses on detecting early tumor response to PDT. The mouse MR images may be acquired using a high-field (e.g., 9.4-T) small-animal MR scanner (e.g., Bruker BioSpin GmbH, Rheinstetten, Germany). A dedicated whole body mouse coil may be used for the image acquisitions. During each imaging session, the animals may be placed on a plastic holder and may be provided with a continuous supply of 2% isoflurane (EZAnesthesia, Palmer, PA) in air. To minimize motion artifacts, respiration- gated MR image acquisitions may be used. Animal respiration rates and core-body temperatures may be monitored throughout the experiments; temperature may be maintained via a feedback system that provided warm air to the bore of the magnet. For example, the respiration rate may be maintained at 40/min and the core-body temperature at 35 - 37 ⁰C. [00225] Images with varying echo times (TE's) may be obtained using a commercial multi-slice multi-echo (MSME) sequence to enable T2 calculation. Two sets of imaging parameters may be used. First, MR images may be obtained at four different echo-times (e.g., 27,85, 55.71, 83.56, and 111.42 ras) in one single acquisition (e.g., TR = 6929 ms, field of view 7.50 x 3.75 cm, matrix size 256 x 128, slice thickness 0.5 mm, receiver bandwidth 30.864 kHz, no average). Typically, 15-20 coronal slices may be acquired to cover the two tumors. The total scan time to simultaneously acquire the four T2-weighted images may take 14 minutes, 47 seconds. These MR parameters may be used for a first group of mice (e.g., mice 1-6). Then, another set of echo times (e.g., 10.25, 20.50, 30.75, and 41.00 ms) may be used for a shorter acquisition time. Other parameters may be modified accordingly (e.g., TR = 1250 ms, field of view 3.5 x 3.5 cm, matrix size 128 x 128, slice thickness 0.5 mm, receiver bandwidth 25 kHz, no average). The total scan time for the four T2-weighted images acquired simultaneously may take 2 minutes, 43 seconds. As can be seen, the second set of MR parameters results in a much shorter scan time than the first set. The second set of MR parameters may be used for a second group of mice (e.g., mice 7-13). Even if two sets of imaging parameters are used for the image acquisitions, they can be from the same MR imaging sequence, such as an MSME sequence that is a standard MR imaging sequence installed in the MR scanner by the manufacture, hi one study, no variation was detected in T2 calculations on the same tissues using the different MSME echo times (data not shown). [00226] Image and Data Analysis

[00227] Quantitative image analysis may be performed for the MSME images. First, the MSME images may be used to generate T2 maps by performing a linear least squares fit to the semilogarithm at each voxel using the equation below:

S= S₀ exp(-TE/T₂) (16), where So and S are the initial signal and signal at echo time TE, respectively. A commercially- available software application (e.g., Paravision 3.1 from Bruker BioSpin GmbH, Rheinstetten, Germany) may be used to compute the T2 maps. Second, the tumor on each slice of the MSME image volumes may be manually segmented. A commercially- available software application (e.g., Analyze from AnalyzeDirect, Inc., Overland Park, KS) may be used for the segmentation. On the T2-weighted MR images, the tumor may appear as a bright region, The boundary of the tumor may be drawn on the image and then the object map of the tumor may be saved. Additionally, the image may be examined and the object map may be loaded to verify the segmentation. If necessary, the object map may be edited. The final boundaries of the segmented tumor may be saved and copied to the corresponding T2 map. The T2 value for each voxel may be determined within the tumor region. Third, the histogram, mean, and standard deviation of the T2 maps may be calculated. The mean T2 values for the treated and control tumors may be compared. [00228] Histologic Analysis

[00229] Histologic analyses may be performed by dissecting the prostate tumors 1-7 days after PDT. In one study, sixteen tumors (nine PDT-treated, seven control) were harvested twenty-four hours after PDT and eight tumors (4 PDT-treated, 4 control) were dissected seven days after PDT. Excised tissues may be fixed in a large volume of 10% formalin for a minimum of three days to allow complete tissue fixation. Subsequently, the tissue may be sectioned along approximately the same plane as the coronal MR images to permit correlation of histologic and MR images. The tumors may be stained with hematoxylin and eosin (H&E) for histopathologic assessment of tumor features. Tissue sections of the specimen may be examined by a pathologist specially trained in genitourinary pathology with a microscope (e.g., Olympus BX40 microscope) at magnifications ranging from 4OX to 400X. [00230] Statistical Analysis

[00231] Statistical analyses may be performed to compare the T2 values obtained at three different time points (e.g., pre-PDT, post-PDT, and twenty-four hours after PDT). Microsoft Excel 2007 (Microsoft, Seattle, WA) may be used to compute a two-tailed two- sample Student's t-test for the T2 values. A p-value ≤0.05 may be assigned statistical significance.

[00232] With reference to FIGs. 22a and 22b, exemplary MR images of a tumor- bearing mouse pre-PDT and twenty-four hours after PDT are shown. The treated and control tumors are clearly delineated on the images. As shown, the signal intensity values changed twenty-four hours after the treatment. The MR images may be used to calculate corresponding T2 maps (see FIGs. 22c and 22d). As shown, the T2 values of the treated tumor increased twenty-four hours after PDT compared to the T2 map before PDT. After the treatment, inflammation at the tumor region and the surrounding tissues was observed. On both the MR images and the T2 maps, visible intensity variation was also observed within the treated tumor indicating possible heterogeneity of the tumor response to the therapy. On the twenty-four hour T2 map, the intensity also increased at other regions outside of the tumor. The intensity increases in the peritumoral area may be due to inflammation from PDT. [00233] The exemplary images in the (a) and (b) frames of FIG. 22 were acquired using a multi-slice multi-echo (MSME) MR sequence with the following imaging parameters: TE - 10.25, 20.50, 30.75, and 41.00 ms; TR = 1280 ms; FOV - 3.5 x 3.5 cm; Matrix size: 128 x 128. The MR images were acquired from the first echo (TE = 10.25 ms). As shown, the signal intensity values changed twenty-four hours after the treatment. MR images from four echoes were used to calculate T2 maps. Compared to the T2 map before PDT in the (c) frame, the T2 values increased twenty-four hours after the treatment as shown in the (d) frame, especially within the treated tumor (arrow).

[00234] With reference to FIG. 23, the T2 histograms of the treated and control tumors pre-PDT, post-PDT and twenty-four hours after PDT are shown. For the treated tumors, the T2 histogram shifted to the right twenty- four hours after the treatment, indicating increases in the T2 values within the treated tumor. The T2 histograms of the control tumor did not demonstrate significant changes immediately or twenty- four hours after PDT as compared to the pre-PDT values. It should also be noted that the T2 histogram of the treated tumor immediately after PDT shows increased numbers of voxels with low T2 values. One possibility is that the level of deoxyhemoglobin was changed immediately upon PDT, as has been observed by others.

[00235] As shown in the (a) frame of FIG. 23, for the treated tumor, the T2 histogram shifted to the right twenty-four hours after the treatment, indicating increases in the T2 values. For the control tumor, the (b) frame shows that the T2 histograms did not demonstrate significant change pre-PDT, post-PDT and twenty-four hours after PDT. The treated and control tumors were for the same mouse (M2).

[00236] With reference to FIG. 24, the mean T2 values for the thirteen treated mice are shown. For the treated tumors, the mean T2 values are 55.8 ± 6.6 ms and 68.2 ± 8.5 ms pre- PDT and twenty-four hours after PDT, respectively, and are significantly different (p < 0.0002). For the control tumors, the mean T2 values are 52.5 ± 6.1 ms and 54.3 ± 6.4 ms pre- PDT and twenty-four hours after PDT, respectively, which are not significantly different (p = 0.53). For both treated and control tumors, there was no significant difference between the T2 values obtained pre-PDT and immediately post-PDT (data not shown). [00237] As shown in the (a) frame of FIG. 24, for the treated tumors, the mean T2 values are 55.8 ± 6.6 ms and 68.2 ± 8.5 ms pre-PDT and twenty-four hours after PDT, respectively. An asterisk is placed at the mean T2 value of treated tumors and the T2 values are significantly different for these two time points (p <0.0002). For the control tumors, the (b) frame shows that the mean T2 values are 52.5 ± 6.1 ms and 54.3 ± 6.4 ms pre-PDT and twenty-four hours after PDT, respectively. Ml and M6 each had a small control tumor, which was not included in this study. There is no significant difference between the T2 values at the two time points studied (p = 0.53). [00238] With reference to FIG. 25, histologic images of treated and control tumors are shown. These images are typical of those obtained from the other tumors. An inflammatory response with edema was observed in the treated tumor, which was not seen within the control tumor. The treated tumor cells were massively damaged by the PDT and the tissues became necrotic. Substantial intra-tumor variation in response to the treatment was also observed. Factors that may contribute to the heterogeneity of the tumor response include variations in drug distribution within the tumor, oxygen supply from the microvasculature system and laser light distribution. As shown herein on the MR images and the T2 maps, variations in intensity within the tumor may also be observed. Thus, the MR images are consistent with the histologic findings. The most likely explanation is that the biological effects of the treatment result in altered water distribution within the treated tissue, including substantial edema, which contributes to changes in the T2 values.

[00239] Histologic images of treated and control tumors from a representative mouse

(M7) twenty- four hours after PDT are shown in FIG. 25. An inflammatory response with edema was observed in the treated tumor shown in the (a) frame, which was not seen within the control tumor of the (b) frame. The rectangular areas on the images in the (a) and (b) frames are magnified and shown in the (c) and (d) frames, respectively. On the image in the (c) frame, massive areas of tumor cells were damaged by PDT, and the tissues became necrotic. However, the control tumor cells were intact in the (d) frame. The laser light was focused approximately perpendicular to the plane of the tissue slice. The two tumors were from the same mouse (M7).

[00240] MR imaging and analysis methods for non-invasively assessing the efficacy of

PDT on cancer, such as prostate cancer, are provided herein. This includes demonstration of the effect of Pc 4-based PDT for human prostate cancer (PC-3) in an animal model which is detectable with imaging at an early stage. High-resolution MSME MR images may be able to reveal tumor response to the therapy twenty-four hours after the treatment. For treated tumors, the T2 values may significantly increased one day after treatment, whereas no significant difference in T2 values may be observed in untreated tumors over the same time. Histologic images verified the therapeutic effect on the treated tumors. The MR imaging parameter (T2 value) may provide a useful tool to monitor early tumor response and to determine the effectiveness of the treatment regimen.

[00241] The targets of PDT include tumor cells and cells of and within tumor microvasculature, and photodynamic damage to these targets leads to direct tumor cell death and to inflammatory and immune responses by the host. The photosensitizer Pc 4 localizes in and has a major influence on mitochondria, and Pc 4-based PDT produces cytotoxic reactive oxygen species which lead to cell apoptosis and necrosis. Rapid tumor responses to Pc A- based PDT include acute edema and inflammation a few hours after the treatment. PDT- induced lesions are characterized by marked necrosis a few days after therapy. Given the mechanism of action of PDT with photosensitizers, such as Pc 4, one might expect alterations in MR imaging parameters in the treated area based on increased water content from edema, vascular occlusion and necrosis. As shown on the histologic images (FIG. 25), there are massive areas of inflammation and necrosis within the treated tumor. T2-weighted MR imaging is sensitive to alterations in tissue water content. The change of T2 values one day after PDT may be related to the increased edema and the changes of water distribution in the treated tissues, consistent with necrosis and inflammation.

[00242] The imaging and analysis methods may provide a useful tool to monitor tumor response to PDT, to study therapeutic mechanisms, and to evaluate new PDT drugs. Potential clinical applications of the imaging technique include PDT efficacy assessment and prediction of long-term tumor cure or regrowth.

[00243] HIGH-FIELD MR IMAGING FOR ASSESSMENT OF PDT

[00244] In another embodiment, a medical apparatus provides a high-field magnetic resonance (MR) imaging technique for assessment of photodynamic therapy (PDT). The high-field MR imaging (MRI) technique may assess early effects of PDT for treatment of cancer, including human prostate cancer. In one embodiment, the high-field MRI technique may use changes in T2 values as a surrogate biomarker for evaluating PDT efficacy. In another embodiment, the high-field MRI technique may use changes in apparent diffusion coefficient (ADC) as a surrogate biomarker for evaluating PDT efficacy. [00245] Animal Preparation

[00246] Human prostate tumor cells CWR22 and PC-3 may be prepared and implanted on the back of athymic nude mice. The preparation and implantation of PC-3 cell tumors may be as described herein. The CWR22 cells may be derived from a primary human prostatic carcinoma and form androgen-dependent xenografts. Male athymic nude mice of 4-8 weeks old may be obtained and housed under pathogen-free conditions. They may be maintained under controlled conditions (e.g., 12-hour dark-light cycles; temperature 20-24°C) with free access to sterilized mouse chow. Two tumors may be initiated in each mouse by injection of the CWR22 cells intradermally on the shoulder flanks. One to three weeks after the injection of the cells, the tumor may be ready for treatment. The mice may be treated 3-4 weeks after tumor implantation using Pc 4-based PDT as described herein. An exemplary protocol for the PDT treatment is shown in FIG. 26. [00247] MRI Experiments

[00248] Two days after photosensitizer injection, high-resolution MR images may be acquired from each mouse pre-PDT, post-PDT, twenty-four hours after PDT, and seven days after PDT. The mouse MR images may be acquired using a high-field (e.g., 9.4-T) small animal MR scanner (e.g., Bruker BioSpin GmbH, Rheinstetten, Germany). A dedicated whole body mouse coil may be used for the image acquisitions. A multi-slice multi-echo (MSME) sequence (e.g., TR-6929 ms and TE=28, 56, 84, and 111 ms) with a slice thickness of 0.5 mm may be used to generate high-resolution coronal images (e.g., Matrix: 256 x 256, Pixel size: 0.27 x 0.13-mm). In these T2-weighted images, the tumors may be clearly delineated by the bright subcutaneous fat signals. Other imaging methods, such as diffusion- weighted MR imaging may also be implemented.

[00249] During each imaging session, the animals may be mounted on a plastic holder and provided with a continuous supply of l%-2% isoflurane (EZAnesthesia, Palmer, PA) in oxygen to minimize motion artifacts. The mice may be in a prone position in the plastic holder and may be placed in a similar posture to minimize the body deformation in different imaging sessions.

[00250] Quantitative Image Analysis

[00251] A commercially-available software application (e.g., Analyze from

AnalyzeDirect, hie, Overland Park, KS) may be used to manually segment the tumor on each image slice from the MR image volumes. The segmented images may be used for the calculation of T2 maps and tumor volumes. The four MSME images may be used to generate the T2 maps over the tumor regions. A commercially-available software application (e.g., Paravision 3.1 from Bruker BioSpin GmbH, Rheinstetten, Germany) may be used for computation of T2 maps. Once the T2 values are obtained for each voxel within the tumor region, the histogram, mean, and standard deviation of the T2 values of the tumor may be calculated.

[00252] FIG. 27 shows exemplary MR images and T2 Maps of tumors pre-PDT treatment, immediately post-PDT treatment and twenty- four hours after PDT treatment. The treated tumors had significant higher T2 values twenty-four hours after treatment as compared to those before treatment. The control tumors did not show significant changes in T2 values after treatment. For the control tumor, the cells were healthy and were not damaged. However, the cells within the treated tumor were damaged by the Pc 4-based PDT and were either dead or dying. Additionally, there were many areas of inflammations within the treated tumor tissues. This verified the effectiveness of the treatment. [00253] FIG. 28 shows that Pc 4-PDT effectively cured the human prostate tumor, as evidenced by the PSA chart showing a dramatic decrease and final disappearance of PSA after PDT. This was verified by the observation of the mouse where the tumor disappeared after the treatment. Fig. 29 shows the PSA levels in each experimental mouse. The data showed that the PSA values decreased for all mice seven days after the treatment. This further verified the efficacy of the treatment for treating this human prostate cancer model. [00254] This study evaluated a second- generation PDT drug (i.e., Pc 4) for treating human prostate cancer in vivo. Pc 4-PDT was shown to be effective for treating human prostate cancer in two mouse models (PC-3 and CWR22). MR imaging was shown to provide a non-invasive image-based parameter for assessing the efficacy of PDT at an early stage.

[00255] MOLECULAR IMAGING FOR ASSESSMENT OF PDT

[00256] In one embodiment, a medical apparatus provides a molecular imaging technique for assessment of photodynamic therapy (PDT). The molecular imaging technique may include a positron emission tomography (PET) technique to assess early effects of PDT for treatment of cancer, including human prostate cancer. In one embodiment, the PET technique may use changes in choline uptake as a surrogate biomarker for evaluating PDT efficacy. In another embodiment, the PET technique may use changes in fluorodeoxyglucose (FDG) as a surrogate biomarker for evaluating PDT efficacy.

[00257] PDT is a therapy for treating various cancers. Choline imaging with positron emission tomography (PET) may provide an early surrogate biomarker for monitoring tumor response to PDT.

[00258] For example, choline imaging has been used to monitor the response of human prostate cancer cell lines (e.g., androgen independent (PC-3) and androgen dependent (CWR22)) to PDT. Tumor cells were injected subcutaneously on the back flanks of athymic nude mice. A second-generation photosensitizer, Pc 4 (e.g., 0.6 mg/kg body weight), was delivered to each animal by tail vein injection forty-eight hours before laser illumination (e.g., 672 nm, 100 mW/cm², 150 J/cm²). Dynamic microPET images with ^πC-choline were acquired from each mouse before PDT and one hour, twenty- four hours, and forty-eight hours after PDT. Time activity curves and standard uptake values of ^l ^-choline were analyzed for each tumor at different time points. [00259] A total of fifteen mice were treated and imaged. For the treated tumors

(N = 15), the choline uptake decreased significantly twenty-four hours and forty-eight hours after PDT, compared to the same tumors pre-PDT, However, for the control tumors (N-8), choline uptake increased significantly twenty-four hours and forty- eight hours after PDT due to the growth of the tumors. For mice bearing CWR22 tumors, the level of prostate specific antigen decreased twenty-four hours and forty-eight hours after PDT. Histologic examination detected the presence of necrosis and inflammation in PDT-treated tumors, which were not seen in the controls.

[00260] Changes in tumor choline uptake detected by PET imaging are useful for monitoring Pc 4-mediated PDT of prostate cancer with these animal models. [00261] PDT is recognized as a treatment for various cancers. PDT requires a photosensitizing drug, light of a specific wavelength, and oxygen. Upon absorption of photons, the drug generates toxic singlet oxygen and other reactive oxygen species that react with nearby lipids, proteins, and nucleic acids. The primary role of PDT is to kill cancer cells by both direct and indirect mechanisms. Direct modes of cell death relate to nonspecific necrosis and the initiation of signaling pathways that elicit apoptosis, autophagy or both. Indirect effects of PDT include vascular damage and occlusion. Both the photosensitizer and the light are inert by themselves. The light can be precisely delivered to a selected region, allowing tumor specificity in the localization of the photodynamic effect. Consequently, side effects are minimized.

[00262] Pc 4 is a second-generation photosensitizing drug. Pc 4 is a silicon phthalocyam^'ne with a strong absorption maximum at a relatively long wavelength (672 nm). This permits deep tissue penetration of laser light. It has been demonstrated in animal models that Pc 4-PDT is effective for treating human prostate tumors, breast cancer, human ovarian epithelial carcinoma, human colon cancer, and human glioma. The National Cancer Institute's Drug Decision Network sponsored preclinical toxicity and pharmacokinetic evaluations of Pc 4 and developed a formulation appropriate for its use in humans. [00263] Extensive in vitro studies have shown that the drug Pc 4 exhibits mitochondrial localization and binds at or near cardiolipin. Cardiolipin is a phospholipid that comprises ~22% by weight of the inner membrane lipid of mitochondria and participates in membrane bilayers. Choline is a precursor for the synthesis of phospatidylcholine which is a major constituent of membrane phospholipids, including cardiolipin (Kennedy pathway). In cancer cells, membrane synthesis is activated during cell proliferation, and the phosphocholine level is elevated. This provides a basis for detecting cancer using positron emission tomography (PET) with radiolabeled choline. PET imaging methods may be used to monitor tumor response to Pc 4-PDT. Pc 4-PDT may decrease the ability of cell membranes to incorporate choline into their constituent phospholipids. This change may be detectable by PET imaging with radiolabeled choline.

[00264] PET with ¹⁸F-fluorodeoxyglucose (FDG) has been used to monitor tumor metabolic response to PDT. PET with FDG can image glucose metabolism after PDT in mice. FDG uptake may be reduced in treated tumors. Dynamic FDG-PET may be used for monitoring transient glucose metabolic processes during PDT, where the ¹⁸F-FDG time- activity curves during PDT may show distinct transient patterns characterized by a drop and subsequent recovery of the tumor FDG uptake rates. However, FDG uptake in tumors is not specific, because inflammation can increase FDG uptake and because muscle tissue usually has a high rate of glucose metabolism. FDG-PET may have less sensitivity and/or specificity for assessing some types of cancer, such as prostate cancer.

[00265] PET imaging with radiolabeled choline may be used for detecting early tumor response to Pc 4-PDT. Since cell membrane biosynthesis is a good indicator of cellular metabolic activity as well as cell proliferation and since choline is an important constituent of cell membrane, choline PET imaging could be a specific and sensitive method for PDT assessment. In various embodiments, Pc 4-PDT may be used to treat tumor-bearing mice, "C-choline PET may be used before and after the PDT, and quantitative image analysis of the PET data may provide an assessment of the treatment. The imaging may verify that Pc 4- PDT interferes with choline uptake into prostate cancer cells in vitro. [00266] Pc 4 formulation

[00267] The chemical synthesis of Pc 4 [HOSiPcOSi(CH₃)₂(CH₂)₃N(CH₃)₂] was described in Oleinick et al., "New phthalocyanine photosensitizers for photodynamic therapy," Photochem.PhotobioL, vol. 57, pp. 242-247, Feb, 1993, the contents of which are fully incorporated herein by reference. A stock solution (e.g., 1 mg/mL) may be made by dissolving Pc 4 in 50% Cremophor EL (e.g., Sigma- Aldrich Co., St. Louis, MO), 50% absolute ethanol, then adding nine volumes of normal saline with mixing. For injection, the Pc 4 stock solution may be mixed with an equal volume of 5% Cremophor EL, 5% ethanol, 90% saline to give a final concentration of 0.05 mg/mL (i.e., 0.07 mM). [00268] Tumor model

[00269] The CWR22 xenograft model of human androgen-dependent prostate cancer may be maintained as described in Pretlow et al., Transplantation of human prostatic carcinoma into nude mice in Matrigel, Cancer Res 1991 ;51:3814-7, the contents of which are fully incorporated herein by reference. A cell suspension containing approximately 1 x 10⁷ cells in 0.2 πϊL of Matrigel (e.g., Collaborative Research, Bedford, MA) may be injected through a 19-gauge needle subcutaneously into the rear flank of male athyraic nude mice, Mice with CWR22 may be given 12,5-mg sustained release testosterone pellets (e.g., Innovative Research of America, Sarasota, FL) s.c. before receiving tumors and at intervals of 3 months until death. One CWR22 tumor may be initiated on each mouse. These animals may be used to measure prostate-specific antigen (PSA) after treatment to monitor the therapeutic efficacy.

[00270] The PC-3 cell line may be derived from a primary malignant human prostate tumor. PC-3 cells may be grown as monolayers in E-MEM supplemented with 15% fetal bovine serum at 37⁰C. Cells may be harvested by trypsinization in ethyl enediaminetetraacetic acid/trypsin, washed in Hank's balanced salt solution (HBSS) without Ca²⁺ and Mg²⁺, and centrifuged at 15Og for five minutes. Cells may be counted in a hemacytometer with 0.4% trypan blue. The cell suspension may be brought to a final concentration of 1 x 10⁶ viable cells/mL and kept on ice for immediate injection. Two PC-3 tumors may be initiated in each mouse by injection of 50 μL containing 5 x 10⁴ PC-3 cells subcutaneously on each flank at least 20 mm apart and as far from the lung and heart as possible to minimize motion effects in PET imaging.

[00271] Male athymic nude mice 4-6 weeks old weighing between 25 and 30 g may be obtained and housed one mouse/cage under pathogen-free conditions. They may be maintained under controlled conditions (e.g., 12-hour dark-light cycles; temperature 20-24°C) with free access to sterilized mouse chow. [00272] PDT protocol

[00273] Tumors may be treated and imaged when the shortest dimension of the tumor size reached 4-5 mm, which may be 2-4 weeks after implantation. Each animal may be weighed at the time of injection, and a volume of Pc 4 solution may be injected intravenously into the tail vein to give 0.6 mg/kg (e.g., 240 μh to a 20 g mouse). Forty-eight hours after photosensitizer injection, the animals may be taken for PDT and imaging. For PDT, a diode laser (e.g., Applied Optronics Corp., Newport, CT) may be used to deliver light (e.g., 672-ran light, the longest wavelength absorption maximum of Pc 4). The laser may be coupled to a fiber optic cable terminating in a microlens. The treatment light may cover the entire tumor and may be distributed uniformly throughout the treatment field. The tumor on each animal may be irradiated with a fluence of 150 J/cm² and an irradiance of 100 mW/cm²'. This has been shown to produce a complete response and some cures in PC-3 tumors. The low power of the laser light may preclude thermal effects. For the PC-3 tumor model, the second tumor in each animal may serve as a control (i.e., receiving photosensitizer, but no light). [00274] For example, fifteen animals may be treated, among which eight mice with

PC-3 tumors and seven mice with CWR22 tumors. As described herein, each PC-3 mouse may have two tumors, one for treatment and the other serving as the control, and each CWR22 mouse may have one tumor. The PSA may be used to determine the treatment effect in the CWR22 mice. Ih this example, data is obtained from 23 tumors. [00275] Radiosvnthesis of ' ' C-Choline

[00276] The synthesis method for ⁿC-Choline is provided in Pascali et al,

[l lC]Methylation on a Cl 8 Sep-Pak cartridge: a convenient way to produce [N-methyl- HC]choline, Journal of Labeled Compounds and Radiopharmaceuticals 2000;43:195~203, the contents of which are incorporated herein by reference. ^πC-Carbon dioxide may be produced by a Scanditronix MC 17 cyclotron and bubbled into a reaction vial previously filled with LiAlH₄ in tetrahydrofiirane (THF) solution (0.1 mol/L, 1 niL) at room temperature. After THF was completely evaporated, hydriodic acid (HI, 57%, 1 mL) may be added, and the reaction vial may be heated to 12O⁰C. ¹¹C-CH₃I obtained by this "wet" chemistry may then be distilled, dried and trapped onto an Accell Plus CM Sep-Pak cartridge which was previously loaded with precursor N,N-dimethylaminoethanol (60 μL) at room temperature. The methylation reaction may take place immediately. The final product may be eluted from the cartridge by saline after being washed with ethanol and water and then passed through a 0.2-μm sterile filter. The radiolabeling yield may be about 80% (corrected to ¹¹C-CH₃I). The radiochemical purity may be greater than 99% as determined by high-performance liquid chromatography (HPLC) (e.g., Partisil SCX cation exchange column, 250 mM NaH₂PO₄/ CH₃CN (9:1, v/v), flow rate: 1.8 mL/min). [00277] PET imaging

[00278] With reference to FIG. 30, an exemplary PDT and imaging protocol is provided. Forty-eight hours after the injection of the photosensitizer Pc 4, PET images may be acquired from each mouse during therapy or immediately (i.e., as near as possible, such as within five minutes or one hour) before the laser irradiation for PDT. One group of mice (N = 4) had an additional PET scan one hour after prjT. The second group of mice (N = 5) was scanned twenty-four hours after PDT. The third group (N = 6) was imaged forty-eight hours after PDT. A total of fifteen mice were scanned by PET before and after PDT at different time points. The mice were imaged no more than two days after PDT, because our study focuses on detecting the early tumor response to PDT. The mouse PET images were acquired with a dedicated microPET imaging system (R4, Siemens Preclinical Solutions, Knoxville, TN). Approximately 18.5 MBq of ^πC-choline in 0.12 mL of physiological saline were injected into each animal via the tail vein. Mice were immediately scanned for 60 min with a list-mode acquisition that allowed retrospective determination of time-binning of dynamic data. During each imaging session, the animals were taped onto a plastic holder and were provided with a continuous supply of 2% isoflurane (EZAnesthesia, Palmer, PA) in air. Animal respiration rates were monitored throughout the entire experiment; typically, the respiration rate was maintained at 40/min,

[00279] FIG. 30. Protocol for PDT and PET imaging. Two days before PDT, the animal was injected with the photosensitizing drug Pc 4. Forty-eight hours later, the animal was taken to the PET imaging facility for a baseline image acquisition immediately (i.e., as soon as possible, such as within 30 minutes or two hours) before PDT, The tumor was then exposed to laser light for PDT. Another PET image dataset was acquired one hour (N ~ 4), twenty-four hours (N = 5), or forty-eight hours (N = 6) after PDT. [00280] Image Reconstruction and Data Analysis

[00281 ] Dynamic PET data were re-binned into a total of 7 frames (600, 600, 600, 600,

600, 300, 300 sec). The emission scans were reconstructed using an ordered subset expectation maximum reconstruction algorithm with an interpolated pixel size of 0.8 x 0.8 mm and a thickness of 1.2 mm. The dimension of reconstructed volume is 128 x 128 x 63 voxels. The unit of pixel value is MBq/cc. ASJPro (Acquisition Sinogram and Image Processing) software which was packaged with the microPET system was used to visualize and analyze the reconstructed PET volumes. Localization of ^l ^-choline accumulation in the PET images in relation to anatomical structures was assured by visually comparing PET images with transmission images. Regions of interest were placed on various tissue areas, in particular the tumors. A time-activity curve was generated from manually segmented regions of interest. Dividing the tissue uptake by the injected activity per gram of body weight yields the standardized uptake value (SUV). Time-activity curves were analyzed to determine the uptake difference between the treated and control tumors. [00282] Histopathology

[00283] Mice were euthanized twenty-four or forty-eight hours after PDT to measure early histologic responses to Pc 4-PDT. For the CWR22 mice, the tumors were harvested twenty-four hours (N = 4) and forty-eight hours (N = 3) after PDT. For the PC-3 mice, eight tumors (4 PDT-treated, 4 controls) were dissected forty-eight hours after PDT. Dissected tumors were sliced into 2-3 slices and excised tissues were fixed in a large volume of 10% formalin overnight. Histologic slides were prepared at the Case Comprehensive Cancer Center Histology Core Facility. All tumors were stained with hematoxylin and eosin (H&E) for histopathologic assessment of tumor features. Tissue sections of the entire specimen were then examined with an Olympus BX40 microscope at magnifications ranging from 4OX to 400X.

[00284] In vitro PDT in PC-3 cells

[00285] PC-3 cells were cultured in RPMI 1640 medium in three wells of six-well plates at a concentration of 2-3 x 10⁵ cells per well. Half of the plates served as control and the other half were treated with PDT (200 nM of Pc 4, 200 mW/cm², 200 mJ/cm², the LD₉₀ dose kills about 90 % of treated PC3 cells as determined by trapan blue exclusion assay). Immediately after PDT, culture medium was replaced with HBSS supplemented with 10 mM HEPES (pH 7.3) and 2 % glucose, because RPMI medium contains choline cholide. ¹¹C- choline (0.5 MBq) was loaded to each well and cultures were incubated at 37° C for 5, 30, and 45 minutes. After incubation, the medium was removed from the wells, and the cells were washed two times with 2 ml HBSS, lysed with 2 ml of 1% Sarkosyl NL-97 (ICN), and another wash with HBSS. The radioactivity in the incubation medium, lysates, and all the washes was then determined separately with a 1282 Compugamma gamma well counter (Wallac, Inc., Gaithersburg, MD). Each procedure was carefully planned and timed during the experiment. All measured radioactivities were corrected for decay. [00286] Statistical Analysis

[00287] Statistical analyses were performed to compare the standard uptake values obtained at three different time points (pre-PDT and twenty-four hours or forty-eight hours after PDT), We used Microsoft Excel 2007 (Microsoft, Seattle, WA) to compute a two-tailed two-sample Student's t-test for the T2 values. A p-value <0.05 was assigned statistical significance.

[00288] In vivo PET imaging showed that the choline uptake of PDT-treated tumors decreased at one, twenty- four, and forty-eight hours after PDT. FIG. 31 shows the microPET images of a tumor-bearing mouse before and forty-eight hours after PDT. The mouse was implanted with two PC-3 tumors. One was treated and the other served as the control, receiving Pc 4 but no light. Forty-eight hours after PDT, necrotic tissue was visible within the treated tumor. The PET images showed that the "C-choline uptake by the PDT-treated tumor decreased forty-eight hours after PDT. Within the tumor region, choline uptake demonstrated heterogeneity but the overall uptake of the entire tumor decreased after therapy. The control tumor did not show visible change on the PET images before and after therapy (not shown).

[00289] FIG. 31. MicroPET imaging with ^] ^-choline for Pc 4-PDT of human prostate cancer in athymic nude mice. Pictures on the left were taken (A) pre-PDT and (D) forty- eight hours after PDT. One of the two tumors (arrows) was treated and the other was the control. Forty-eight hours after PDT, necrotic tissue was visible in the treated tumor (D). ^πC-choline PET images of the treated tumor (B) before and (C) forty-eight hours after PDT. The tumor region was magnified (E and F) and is shown below the corresponding whole- body image. These images show that ^uC-choHne uptake decreased forty-eight hours after PDT. Quantitative analysis confirmed the visual results.

[00290] FIG. 32 shows the normalized time-activity curves of ^πC-choline uptake, which were calculated for four mice bearing PC-3 tumors before and forty-eight hours after PDT. A decrease in choline uptake was observed in all of the treated tumors forty-eight hours after therapy. For the four treated tumors from the four mice, the mean SUVs at the seven time points (5, 15, 25, 35, 45, 53, and 57 min) were 0.177 ± 0.002 and 0.109 ± 0.0005 (MBq/cc)/( MBq/g) immediately before therapy and at forty-eight hours after PDT, respectively. Therefore, the uptake by the treated tumors was decreased by 38.4% two days after PDT (p < 0.001). In contrast, ^] ^-choline uptake by the control tumors was increased at the 48-hour time point. For the four untreated tumors that were from the same four mice, the mean SUVs at the seven time points were 0.195 ± 0.020 and 0.263 ± 0.013 (MBq/cc)/(MBq/g) immediately before and forty-eight hours after PDT, respectively. Therefore, the uptake by the untreated tumors was increased by 35.0% in two days (p < 0.001). The increase in the choline uptake may have been caused by tumor growth, as verified by the tumor sizes.

[00291] FIG. 32. ^uC-cholme uptake into (A) PDT-treated (N = 4) and (B) untreated tumors (N = 4) pre-PDT and forty-eight hours after PDT. Normalized time-activity curves were measured from the dynamic ^πC-choline microPET images acquired. Each mouse had two human prostate cancer PC-3 tumors. One was treated and the other was the control. The error bars are the standard errors.

[00292] This PET study was repeated in mice bearing CWR22 tumors. Again, choline uptake was decreased twenty-four hours (N = 5) or forty-eight hours (N = 2) after PDT for the treated tumors (FIG. 33). For five PDT-treated tumors from five different mice, the mean SUVs at the seven time points were 0.174 ± 0.019 and 0.043 ± 0.004 (MBq/cc)/(MBq/g) immediately before therapy and twenty-four hours after PDT, respectively. Therefore, the uptake by the treated tumors was decreased by 75.5% twenty-four hours after PDT (p < 0.001), In a different group of two mice, two PDT-treated tumors had mean SUVs of 0.157 ± 0.017 and 0.086 ± 0.015 (MBq/cc)/(MBq/g) immediately before therapy and forty-eight hours after PDT, respectively. In this case, the uptake by the treated tumors was decreased by 45.3% forty-eight hours after PDT (p < 0.001). Note that the twenty-four hours group showed a greater decrease than the forty-eight hours group, which indicates that an optimal time may exist for monitoring the response to PDT. This encouraged us to image the animals at an even earlier time.

[00293] FIG. 33. ⁿC-choline uptake into human prostate cancer CWR22 xenografts.

(A) Twenty- four hours after PDT, choline uptake into PDT-treated tumors was less than 30% compared to pre-PDT. (B) Choline uptake into the PDT-treated tumors decreased forty- eight hours after PDT. The twenty-four hours group (N = 2) and the forty-eight hours group (N = 3) were from different mice. The error bars are standard errors.

[00294] FIG. 34 shows the PET imaging results one hour after PDT. The treated tumors had less choline uptake one hour after therapy compared to pre-PDT; however, the control tumors showed slightly increased choline uptake one hour after therapy. For the four PDT-treated tumors from four mice, the mean SUVs at the seven time points were 0.185 ± 0.012 and 0.130 ± 0.005 (MBq/cc)/(MBq/g) immediately before therapy and one hour after PDT, respectively. Therefore, uptake by the treated tumors was decreased by 29.8% just one hour after PDT (p < 0.001). For the four untreated tumors from the same four mice, the mean SUVs at the seven time points were 0.207 ± 0.018 and 0.227 ± 0.001 (MBq/cc)/(MBq/g) for the baseline and one hour later, respectively. The uptake by the untreated tumors was increased by 9.7% one hour after PDT (p = 0.005). As each mouse had one treated and one untreated tumor, the increased uptake of the untreated tumor may be due to a change of blood flow or other systemic response to the therapy immediately after PDT. [00295] FIG. 34. ^πC-choline uptake by (A) PDT-treated and (B) control tumors pre-

PDT and one hour after PDT. Each mouse bore two PC-3 tumors. One was treated and the other was the control. PDT-treated tumors (N = 4) had less choline uptake one hour after PDT compared to pre-PDT. The control tumors (N = 4) had slightly increased uptake at the one-hour point.

'[00296] Changes in choline uptake may indicate early therapeutic effect as verified by

PSA levels. In FIG. 35, we report our preliminary observation on two mice bearing human prostate CWR22 tumors. The two mice (A and B) were scanned and treated at the same time (FIG. 35A). Forty-eight hours after PDT, tumor necrosis was visible on both mice. Mouse B demonstrated more necrosis than A (FIG. 35B). PSA levels decreased in both mice (FIG. 35C), but the decrease in PSA was greater in Mouse B. Note that Mouse B had a higher PSA level (66.57 ng/niL) than Mouse A (44.55 ng/mL) before treatment. However, two days after PDT, both mice had similar PSA levels (28.08 ng/mL for Mouse A and 29.31 ng/mL for Mouse B). As shown in FIG. 35D and 35E, the choline uptake decreased forty-eight hours after PDT for both mice compared to their pre-PDT values, but the decrease was greater in Mouse B. For Mouse A, the mean SUVs at the seven time points were 0.125 and 0.095 (MBq/cc)/(MBq/g) before therapy and forty-eight hours after PDT, respectively. Therefore, the uptake was decreased by 24.1% forty-eight hours after PDT. For Mouse B, the mean SUVs were 0.190 and 0.078 (MBq/cc)/(MBq/g) before therapy and forty-eight hours after PDT, respectively. The uptake was decreased by 59.2% forty-eight hours after PDT. The choline uptake was consistent with both the visible necrosis and the PSA levels. This observation is encouraging because the change in choline uptake may be able to predict therapeutic efficacy. These results stimulated us to investigate the mechanism of the PDT- induced effect on tumor choline uptake, in order to determine if the change in choline uptake was due to cell metabolic activities or blood flow, we performed an experiment in cell culture.

[00297] FIG. 35. Tissue necrosis, PSA and ^πC-choline uptake of human prostate tumors (CWR22). Pictures on the top were taken from two mice A and B pre-PDT and forty- eight hours after PDT. Each mouse had only one tumor (arrow). The two mice were imaged and treated at the same time. Forty-eight hours after PDT, the tumors became necrotic. Mouse B showed more necrosis than Mouse A. Graph (C) shows the PSA levels pre-PDT and twenty-four and forty-eight hours after PDT. Graphs D and E show the normalized time activity curves of ^πC-choline forty-eight hours after PDT. The tumors from both mice had decreased choline uptake compared to pre-PDT. Mouse B had a greater decrease than Mouse A, which is consistent with the change of PSA levels.

[00298] In vitro Pc 4-PDT of prostate cancer cells caused a marked decrease in choline uptake compared to that of control cells. The suppression of choline uptake observed in tumors following PDT could be resulted from either a reduction of blood flow to tumors, as a consequence of photodamage to blood vessels, or suppression of choline uptake into tumor cells, or both. To differentiate these possibilities, the effect of PDT on the uptake of ¹¹C- choline in cultured human prostate cancer PC-3 cells was studied. As shown in FIG. 36, the choline uptake decreased at all three time points by more than 50% (56.2% ± 6.0%) in the treated cells compared to the control cells. Note that the standard deviation at each time point was small (5% ± 2%), The uptake rate of ^πC-choline in the treated cells is only 46% of that in the control cells. This in vitro study demonstrates that Pc 4-PDT caused metabolic changes in cells that resulted in decreased choline uptake independent of blood flow. [00299] FIG. 36. Pc 4-PDT-induced changes in ^πC-choline uptake as a function of post-PDT incubation time (5, 30 and 45 min). Human prostate cancer cells (PC-3) were prepared and treated with Pc 4-PDT as indicated in Methods. The activity of ^πC-choline was measured by a gamma counter and the unit is count per minute (CPM). (A) Activity of ¹¹C- choline in treated cells is 43% ± 6% compared to that in the control cells for the three time points. (B) The uptake rate of ⁿC-choline in the treated cells was 46% of that in the control cells. Each data point represents three wells of cells, and the standard deviation is less than the size of symbol.

[00300] Choline PET imaging and analysis methods for non-invasively assessing PDT effect on human prostate cancer xenografts in mice was accomplished. This demonstrates the utility of PET imaging with radiolabeled choline for detecting tumor response to PDT. Dynamic PET images may reveal early tumor response to Pc 4-PDT several hours after the treatment. For treated tumors, choline uptake may significantly decreased twenty-four hours and forty-eight hours after treatment, whereas increases in choline uptake may be observed in the untreated tumors over the same time. Histologic images may be used verify the therapeutic effect on the treated tumors. The PET imaging parameter (choline uptake) may provide a useful tool to determine the effectiveness of the treatment regimen. [00301] The PDT-induced decrease of choline uptake into prostate cancer cells in culture is consistent with observations from previous in vitro mechanism studies. It has been found that Pc 4 accumulates in cytoplasmic membranes of tumor cells, including mitochondria where the photosensitizer localizes near the phospholipid cardiolipin. Cardiolipin is an important constituent of the membrane bilayers. Pc 4-PDT has profound effects on cellular membranes. Mitochondrial reactive oxygen species were detected within minutes of exposure of cells to Pc 4 and photoactivating light. This was followed by mitochondrial inner membrane permeabilization, depolarization and swelling, cytochrome c release, and apoptotic cell death. As choline is the substrate for the synthesis of cellular membranes, the reduced choline uptake as measured both in cell cultures and in animals may represent an early molecular response to PDT that is detectable by dynamic PET imaging with radiolabeled choline.

[00302] The targets of PDT include tumor cells and cells of and within tumor microvasculature, and photodynamic damage to these targets leads to direct tumor cell death and to inflammatory and immune responses by the host. PDT effects on all these targets may influence each other, producing a plethora of responses; the relative importance of each for the overall tumor response has yet to be fully defined and may differ for different tumor types. Rapid tumor responses to Pc 4-PDT include acute edema and inflammation a few hours after the treatment. PDT-induced lesions are characterized by marked necrosis a few days after therapy. Given the mechanism of action of PDT with photosensitizers such as Pc 4, one might expect alterations in PET imaging parameters in the treated area based on membrane metabolism, mitochondrial damage, apoptosis, and necrosis. As shown on the histologic images (FIG. 37), there are massive areas of inflammation and necrosis within the treated tumor.

[00303] FIG. 37. Histologic images of treated and control tumors from a representative mouse forty-eight hours after PDT. An inflammatory response with edema was observed in the treated tumor (A), which was not seen within the control tumor (B). The rectangular areas on images (A) and (B) are magnified and shown in images (C) and (D), respectively. On image (C), massive areas of the treated tissue were damaged by PDT and the tissue became necrotic. However, the control tumor cells were intact (D).

[00304] The findings from this study could be translated to the clinic for monitoring therapy response at an early stage (several hours to two days after treatment). The microPET scanner has analogs in all modern radiology departments, thus ensuring the ability to translate our studies to human patients. Furthermore, we evaluated two human prostate cancer cell lines (PC-3 and CWR22).

[00305] In conclusion, the PET imaging and analysis methods may provide a useful tool to monitor tumor response to PDT, to study therapeutic mechanisms, and to evaluate new PDT drugs. Our study indicates that PET imaging parameters may be related to intratumor properties altered by PDT in an animal model. Potential clinical applications of the imaging technique include PDT efficacy assessment at an early stage and prediction of long-term tumor cure or regrowth.

[00306] MOLECULAR IMAGING FOR ASSESSMENT OF PDT

[00307] In another embodiment, a medical apparatus provides a molecular imaging technique for assessment of photodynamic therapy (PDT). The molecular imaging technique may include a positron emission tomography (PET) technique to assess early effects of PDT for treatment of cancer, including human prostate cancer. In one embodiment, the PET technique may use changes in choline uptake as a surrogate biomarker for evaluating PDT efficacy. In another embodiment, the PET technique may use changes in fluorodeoxyglucose (FDG) as a surrogate biomarker for evaluating PDT efficacy.

[00308] PDT is a therapy for treating various cancers. Choline imaging may provide an early biomarker for monitoring tumor response to PDT at cellular and molecular levels. [00309] For example, choline imaging has been used to monitor the response of PC-3 cancer cells, a cell line derived form a human prostate malignant tumor line. The PC-3 tumor cells may be injected intradermally on the back flanks of athymic nude mice. Two tumors may be initiated on each mouse. One tumor may be treated and the other may serve as the control. A second-generation photosensitizer drug, Pc 4 (e.g., 0.6 mg/kg body weight), may be delivered to each animal by tail vein injection forty-eight hours before laser illumination (e.g., 672 nm, 100 mW/cm², 150 J/cm²). A microPET scanner may be used to acquire dynamic images from each mouse before PDT and twenty- four hours and forty-eight hours after PDT. ^πC-cholme may be synthesized for the imaging. Time activity curves and standard uptake values may be computed for each tumor.

[00310] For the treated tumors (N = 12), the choline uptake decreased significantly twenty-four hours and forty-eight hours after PDT₅ compared to those pre-PDT. However, the uptakes of the control tumors (N=12), significantly increased twenty-four and forty-eight hours after PDT. Histologic analysis showed that PDT-treated tumors demonstrated apoptosis, necrosis and inflammation that were not seen in the control. [00311] Changes in tumor choline uptake detected by PET imaging may provide an assay that could be useful for clinical monitoring of PDT of prostate cancer at an early stage. [00312] PC 4-BASED PDT

[00313] In one embodiment, a medical apparatus provides a technique for treating cancer using photodynamic therapy (PDT), The PDT may be Pc 4-mediated. The cancer may include prostate cancer. In another embodiment, a medical apparatus provides an image-guided technique for treating cancer using PDT. The image-guided technique may implement a multimodality imaging technique. In one embodiment, the multimodality imaging technique may include a magnetic resonance imaging (MRI) technique. In another embodiment, the multimodality imaging technique may include an ultrasound imaging technique.

[00314] For treatment of prostate cancer in a patient, a photosensitizer, such Pc 4, may be administered to a patient at an optimum dose for treatment of the prostate cancer and the prostate tumor tissue may be irradiated with a source of red light. [00315] In one example, the Pc 4 is administered at about 0.1 mg/Kg to about 10 mg/kg body weight, or from about 0.3 to about 2.0 mg/Kg body weight. In another example, the Pc 4 is administered at about 0.6 mg/kg. The mode of administration can include: intravenous, intra-arterial, intra-peritoneal, subcutaneous, intramuscular administration, or direct injection to the tumor tissue.

[00316] In one example, the source of red light is an optic fiber that is inserted directly into the targeted prostate tumor. The red light can have a spectrum above 600 nm. In one aspect, the source of red light has a wavelength of 670 ± 10 nm. In another aspect, the source of red light has a wavelength of about 672 ± 3 nm.

[00317] In one embodiment, the optic fiber can be inserted directly into the prostate tumor under MRI or ultrasound guidance. This procedure can be similar to the well-known procedure for a transrectal ultrasound (TRUS)-guided prostate biopsy, which is an outpatient diagnostic procedure performed in urology clinics. For additional information on the TRUS- guided prostate biopsy procedure, see Clements R, et al., (1993) Clin. Radiol., Vol. 47, pp. 125-126 and Collins, et al., (1993), Br. J. Urol., Vol. 71, pp. 460-463.). The contents of both of these documents are fully incorporated herein by reference. For example, in image-guided PDT, a long needle-like catheter may be inserted into the target tissue. After needle placement, the stylet of the needle may be replaced by a laser fiber in a coaxial manner. Then, the needle may be retracted and the laser fiber may deliver the light to the target tissue. Needle insertion can occur either through the rectum (transrectal) or the perineum (trans- perineal).

[00318] PDT is a therapeutic modality for clinical treatment of cancer, including prostate cancer. With PDT, a tumor-localized photosensitizing drug is irradiated with red light to generate reactive oxygen that efficiently kills cells and ablates tumors. PDT has little or no systemic toxicity and thus may avoid systemic side effects. It permits treatment via minimally invasive techniques. In one embodiment, medical imaging can be used to aid the insertion of optical fibers for light delivery. For example, PDT may be a salvage therapeutic modality for recurrent localized prostate cancer. In addition, PDT treated tumors often have a rapid response, sometimes a single treatment session may be sufficient to ensure successful eradication. In comparison, external beam radiation therapy often requires repetitive treatments over weeks or months.

[00319] A second-generation photosensitizing drug, Pc 4, may be used for PDT of various cancers. For example, Pc 4-based PDT may be effective for treating two types of human prostate cancer. Additionally, an image-guided minimally invasive PDT technique may be used for the treatment of prostate cancer.

[00320] With reference to FIG. 38, in certain studies using PDT to treat human prostate cancer in athymic nude mice, the PSA value decreased by half one day after PDT and became less than 0.1 ng/ml seven days after PDT. After one month, PSA was less than 0.05 ng/ml.

[00321] With reference to FIG. 39, an image-guided minimally invasive photodynamic therapy (PDT) for prostate cancer is shown. With image guidance, a small needle may be inserted into the prostate and then a laser fiber may be used to deliver light and achieve interstitial PDT.

[00322] While the invention is described herein in conjunction with one or more exemplary embodiments, it is evident that many alternatives, modifications, and variations will be apparent to those skilled in the art. Accordingly, exemplary embodiments in the preceding description are intended to be illustrative, rather than limiting, of the spirit and scope of the invention. More specifically, it is intended that the invention embrace all alternatives, modifications, and variations of the exemplary embodiments described herein that fall within the spirit and scope of the appended claims or the equivalents thereof. Any element in a claim that does not explicitly state "means for" performing a specified function, or "step for" performing a specific function, is not to be interpreted as a "means" or "step" clause as specified in 35 U.S.C. § 112, | 6. In particular, the use of "step of in the claims herein is not intended to invoke the provisions of 35 U.S.C. § 112, *|f 6.

Claims

1 A medical apparatus, including: a storage device adapted to store first and second medical imaging data related to a tumor, the first and second medical imaging data being related to treatment of said tumor using photodynamic therapy (PDT) in conjunction with a photosensitizer; an image processor to process the first and second medical imaging data to form comparative data for at least one characteristic of the tumor; and an assessment logic to analyze the comparative data to evaluate efficacy of at least one of the PDT and photosensitizer.

2. The medical apparatus of claim 1 wherein the first and second medical imaging data includes at least one of magnetic resonance imaging (MRI) data, positron emission tomography (PET) imaging data, and ultrasound imaging data.

3. The medical apparatus of claim 1 wherein at least one of the first and second medical imaging data is from a time during the treatment.

4. The medical apparatus of claim 1 wherein the first and second medical imaging data is in time-shifted relation to the treatment.

5. The medical apparatus of claim 1 wherein the second medical imaging data is time- shifted from the first medical imaging data.

6. The medical apparatus of claim 1 wherein the assessment logic also analyzes the comparative data in conjunction with at least one of further diagnosis of the tumor and further treatment of the tumor.

7. The medical apparatus of claim 1 wherein the tumor includes at least one of a malignant tumor, prostate cancer, breast cancer, ovarian cancer, colon cancer, glioma, and a benign tumor.

8. The medical apparatus of claim 1 wherein the photosensitizer includes at least one of silicon phthalocyanine and porfimer sodium.

9. The medical apparatus of claim 1 wherein the first and second medical imaging data includes magnetic resonance imaging (MRI) data, the image processor including: a mapping logic to produce T2 maps of the tumor from the first and second medical imaging data; a statistical analyzer to produce T2 values for at least one of a histogram, a mean, and a standard deviation from the T2 maps; and a comparator to identify changes in T2 values between the first and second medical imaging data; wherein the assessment logic uses changes in T2 values as a surrogate biomarker to evaluate the efficacy of at least one of the PDT and the photosensitizer.

10. The medical apparatus of claim 1 wherein the first and second medical imaging data includes magnetic resonance imaging (MRI) data, the image processor including: a mapping logic to produce apparent diffusion coefficient (ADC) maps of the tumor from the first and second medical imaging data; a statistical analyzer to produce ADC values for at least one of a histogram, a mean, and a standard deviation from the ADC maps; and a comparator to identify changes in ADC values between the first and second medical imaging data; wherein the assessment logic uses changes in ADC values as a surrogate biomarker to evaluate the efficacy of at least one of the PDT and the photosensitizer.

11. The medical apparatus of claim 1 wherein the first and second medical imaging data includes choline imaging data, the image processor including: an image reconstruction logic to produce images of the tumor from the first and second medical imaging data; a statistical analyzer to produce a time activity curve showing choline uptake from the images; and a comparator to identify changes in choline uptake between the first and second medical imaging data; wherein the assessment logic uses changes in choline uptake as a surrogate biomarker to evaluate the efficacy of at least one of the PDT and the photosensitizer.

12. The medical apparatus of claim 1 wherein the first and second medical imaging data includes fluorodeoxyglucose (FDG) imaging data, the image processor including: an image reconstruction logic to produce images of the tumor from the first and second medical imaging data; a statistical analyzer to produce a time activity curve showing FDG uptake from the images; and a comparator to identify changes in FDG uptake between the first and second medical imaging data; wherein the assessment logic uses changes in FDG uptake as a surrogate biomarker to evaluate the efficacy of at least one of the PDT and the photosensitizer.

13. A method, including: a) providing first and second medical imaging data related to a tumor, the first and second medical imaging data being related to treatment of said tumor using photodynamic therapy (PDT) in conjunction with a photosensitizer; b) processing the first and second medical imaging data to form comparative data for at least one characteristic of the tumor; and c) analyzing the comparative data to evaluate efficacy of at least one of the PDT and photosensitizer.

14. The method of claim 13 wherein at least one of the first and second medical imaging data is from a time during the treatment.

15. The method of claim 13 wherein the first and second medical imaging data is in time- shifted relation to the treatment.

16. The method of claim 13 wherein the second medical imaging data is time-shifted from the first medical imaging data.

17. The method of claim 13, further including: d) analyzing the comparative data in conjunction with at least one of further diagnosis of the tumor and further treatment of the tumor.

18. The method of claim 13 wherein the first and second medical imaging data includes magnetic resonance imaging (MRI) data, the method further including: d) mapping the first and second medical imaging data to produce T2 maps of the tumor; e) analyzing the T2 maps to produce T2 values for at least one of a histogram, a mean, and a standard deviation; f) comparing the T2 values to identify changes between the first and second medical imaging data; and g) using changes in T2 values as a surrogate biomarker to evaluate the efficacy of at least one of the PDT and the photosensitizer.

19. The method of claim 13 wherein the first and second medical imaging data includes magnetic resonance imaging (MRI) data, the method further including: d) mapping the first and second medical imaging data to produce apparent diffusion coefficient (ADC) maps of the tumor; e) analyzing the ADC maps to produce ADC values for at least one of a histogram, a mean, and a standard deviation; f) comparing the ADC values to identify changes between the first and second medical imaging data; and g) using changes in ADC values as a surrogate biomarker to evaluate the efficacy of at least one of the PDT and the photosensitizer.

20. The method of claim 13 wherein the first and second medical imaging data includes choline imaging data, the method further including: d) reconstructing the first and second medical imaging data to produce images of the tumor; e) analyzing the images to produce a time activity curve showing choline uptake; f) comparing choline uptake to identify changes between the first and second medical imaging data; and g) using changes in choline uptake as a surrogate biomarker to evaluate the efficacy of at least one of the PDT and the photosensitizer.

21. The method of claim 13 wherein the first and second medical imaging data includes fluorodeoxyglucose (FDG) imaging data, the method further including: d) reconstructing the first and second medical imaging data to produce images of the tumor; e) analyzing the images to produce a time activity curve showing FDG uptake; f) comparator to identify changes in FDG uptake between the first and second medical imaging data; and g) using changes in FDG uptake as a surrogate biomarker to evaluate the efficacy of at least one of the PDT and the photosensitizer.

22 A medical apparatus, including: an administering device to administer a photosensitizer to a subject having a prostate tumor; and a light emitting device to treat the prostate tumor using photodynamic therapy (PDT) by selectively positioning an optic component proximate to a target area encompassing the prostate tumor and selectively delivering light to activate the photosensitizer.

23. The medical apparatus of claim 22 wherein the prostate tumor includes at least one of a malignant tumor, prostate cancer, and a benign tumor,

24. The medical apparatus of claim 22 wherein the photosensitizer includes at least one of silicon phthalocyanine and porfimer sodium.

25. The medical apparatus of claim 22, further including: a medical imaging device to provide an image of the prostate tumor and surrounding area in relation to at least one of the positioning of the optic component and the delivering of the light for image-guided PDT.

26. The medical apparatus of claim 23 wherein the medical imaging device includes at least one of a magnetic resonance imaging (MRI) device, a positron emission tomography (PET) device, and an ultrasound imaging device.

27 A method, including: a) administering a photosensitizer to a subject having a prostate tumor; b) selectively positioning an optic component of a light emitting device proximate to a target area encompassing the prostate tumor; and c) selectively delivering light to activate the photosensitizer to treat the prostate tumor using photodynamic therapy (PDT).

28 The method of claim 27 wherein the prostate tumor includes at least one of a malignant tumor, prostate cancer, and a benign tumor.

29 The method of claim 27 wherein the photosensitizer includes at least one of silicon phthalocyanine and porfimer sodium.

30. The method of claim 27, further including: d) imaging the prostate tumor and surrounding area in relation to at least one of the positioning in b) and the delivering in c) for image-guided PDT.

31. The method of claim 27 wherein the imaging in d) is provided by at least one of a magnetic resonance imaging (MRI) device, a positron emission tomography (PET) device, and an ultrasound imaging device.

32. A medical apparatus, including: a storage device adapted to store first and second medical imaging data; a rigid-body registration logic to identify normalized mutual information common to the first and second medical imaging data and to align the first and second medical imaging data based at least in part on the normalized mutual information; a deformable registration logic to identify at least one deformable volumetric characteristic of the aligned first and second medical imaging data based at least in part on a finite element model; and an image fusion logic to deform at least one of the first and second medical imaging data to form hybrid medical imaging data based at least in part on the at least one deformable volumetric characteristic.

33. The medical apparatus of claim 32 wherein the first and second medical imaging data includes at least one of magnetic resonance imaging (MRI) data, positron emission tomography (PET) imaging data, and ultrasound imaging data.

34. The medical apparatus of claim 32 wherein the hybrid medical imaging data is used in conjunction with at least one of detection of a tumor, diagnosis of a tumor, treatment of a tumor, and assessment of tumor treatment.

35. The medical apparatus of claim 34 wherein the tumor includes at least one of a malignant tumor, prostate cancer, breast cancer, ovarian cancer, colon cancer, glioma, and a benign tumor.

36. The medical apparatus of claim 34 wherein the treatment includes the use of photodynamic therapy (PDT) in conjunction with a photosensitizer.

37. The medical apparatus of claim 36 wherein the photosensitizer includes at least one of silicon phthalocyanine and porfϊraer sodium.

38. A method, including: a) identifying normalized mutual information common to first and second medical imaging data based at least in part on a rigid-body registration model; b) aligning the first and second medical imaging data based at least in part on the normalized mutual information; c) identifying at least one deformable volumetric characteristic of the aligned first and second medical imaging data based at least in part on a finite element model; and d) deforming at least one of the first and second medical imaging data to form hybrid medical imaging data based at least in part on the at least one deformable volumetric characteristic.

39. The method of claim 38 wherein the first medical imaging data includes magnetic resonance imaging (MRI) data and the second medical imaging data includes positron emission tomography (PET) imaging data.

40. The method of claim 38 wherein the first medical imaging data includes magnetic resonance imaging (MRI) data from a first time and the second medical imaging data includes MRI data from a second time.

41. The method of claim 38 wherein the first medical imaging data includes magnetic resonance imaging (MRI) data and the second medical imaging data includes ultrasound imaging data.

42. A medical apparatus, including: a storage device adapted to store medical imaging data; an anisotropic diffusion filter to process the medical imaging data to form a plurality of multiscale images ranging in resolution from a first multiscale image at an original resolution to a last multiscale image at a coarser resolution; a clustering logic to process the last multiscale image to form an initial estimate of class prototypes for a plurality of tissue types associated with the medical imaging data based at least in part on a k-means clustering algorithm; and a fuzzy classifier to classify components of the first multiscale image into at least one tissue type of the plurality of tissue types based at least in part on processing the plurality of multiscale images using a multiscale fuzzy C-mean algorithm.

43. The medical apparatus of claim 42 wherein the medical imaging data includes at least one of magnetic resonance imaging (MRI) data, positron emission tomography (PET) imaging data, and ultrasound imaging data.

44. The medical apparatus of claim 42 wherein the plurality of tissue types includes at least one of live tumor tissue, necrotic tumor tissue, and intermediate tumor tissue.

45. The medical apparatus of claim 42 wherein the classifying of the first multiscale image is used in conjunction with at least one of detection of a tumor, diagnosis of a tumor, treatment of a tumor, and assessment of tumor treatment.

46. The medical apparatus of claim 45 wherein the tumor includes at least one of a malignant tumor, prostate cancer, breast cancer, ovarian cancer, colon cancer, glioma, and a benign tumor.

47. The medical apparatus of claim 45 wherein the treatment includes the use of photodynamic therapy (PDT) in conjunction with a photosensitizer.

48. The medical apparatus of claim 47 wherein the photosensitizer includes at least one of silicon phthalocyanine and porfimer sodium.

49. A method, including: a) filtering medical imaging data to form a plurality of mulliscale images ranging in resolution from a first multiscale image at an original resolution to a last multiscale image at a coarser resolution based at least in part on an anisotropic diffusion filter; b) determining an initial estimate of class prototypes for a plurality of tissue types associated with the medical imaging data based at least in part on processing the last multiscale image using a k-means clustering algorithm; and c) classifying components of the first multiscale image into at least one tissue type of the plurality of tissue types based at least in part on processing the plurality of multiscale images using a multiscale fuzzy C-means algorithm.

50. The method of claim 49 wherein b) is repeated until a current initial estimate reaches a predetermined convergence in relation to a last initial estimate.

51. The method of claim 49 wherein c) is performed for multiple multiscale images in a coarser to finer resolution sequence toward the original resolution and results from processing at coarser resolutions are used to initialize the processing of the multiscale image at the next finer resolution.

52. The method of claim 49, further including: d) assigning a high membership value to a voxel whose intensity is close to a center of a class; e) allowing membership in neighborhood pixels to regulate classification toward piecewise-homogeneous labeling; and f) incorporating supervision information from the processing of the multiscale image at the previous coarser resolution.