WO1998040008A1

WO1998040008A1 - Validating and processing fluorescence spectral data for detecting the rejection of transplanted tissue

Info

Publication number: WO1998040008A1
Application number: PCT/CA1998/000229
Authority: WO
Inventors: Peter D. Whitehead; Nicholas B. Mackinnon; Calum E. Macaulay; Haishan Zeng
Original assignee: Biomax Technologies, Inc.
Priority date: 1997-03-13
Filing date: 1998-03-12
Publication date: 1998-09-17
Also published as: AU6491198A; EP0973436A1; CA2280880A1

Abstract

The present invention validates, conditions, and analyzes fluorescence spectral data for detecting the rejection of transplanted tissue. A spectral analysis dataset intended to represent the fluorescent response of animale tissue is obtained which indicates the luminous intensity of light collected at each of a range of wavelengths. In one embodiment of the invention, integrated intensities of the dataset are compared to maximum and for minimum intensity thresholds to determine validity. The validated dataset is then conditioned to render it suitable for analysis. The conditioned dataset is then analyzed by comparing extracted features to features of datasets from healthy tissue to detect possible tissue rejection.

Description

VALIDATING AND PROCESSING FLUORESCENCE SPECTRAL DATA FOR DETECTING THE REJECTION OF TRANSPLANTED TISSUE

Technical Field The invention relates generally to the field of data acquisition, validation, and processing, in particular with respect to physiological data.

Background of the Invention

The transplanting of tissues such as organs into a host is a well recognized technique in surgery. Unfortunately, a major, long-standing difficulty is the rejection of the transplanted tissue by the host. Briefly, the immune system of the host recognizes a foreign body, (i.e., the transplanted tissue) and then rejects that foreign body. A variety of techniques exist for the suppression of rejection, and improved rates of success are now being achieved. A popular technique is to suppress the recipient's immune system, for example with cyclosporin. However, such immunosuppression techniques carry risks for the patient, and are therefore minimized, when possible, by attempting to determine prior to immunosuppression if the tissue exhibits characteristics of rejection.

A standard means of determining whether an organ is being rejected is the conduction of physical biopsies (such as an endomyocardial biopsy (EMB) for the heart). In the case of heart transplants, accurate diagnosis is vital for the effective care of the heart transplant, and percutaneous transvenous EMB is a standard method for such assessment of rejection. Crudely described, this means inserting a catheter comprising a device known as a bioptome, which comprises a wire with tiny jaws at the distal end, into a blood vessel. Many varieties of catheters and bioptomes are known in the art. See, e.g., U.S. Patent No. 3,964,468; U.S. Patent No. 4,953,559; U.S. Patent No.

4,884,567; U.S. Patent No. 5,287,857; U.S. Patent No. 5,406,959; WO 96/35374; WO 96/35382; WO 96/29936; WO 96/35374. The distal end of the catheter is fed into an entry point, typically on the leg or neck, and then on to the heart chamber where a tiny piece is clamped in the jaws of the bioptome and removed for analysis. This biopsy permits accurate detection of the presence and the severity of histologic changes in the transplanted tissue once the site of rejection is found.

A patient may require an average of 5 and as many as 10 biopsies per biopsy structure. Thus, over the first year of a heart transplant recipient, as many as 180 EMBs are taken. A typical schedule for EMBs is as follows: Table 1 Right Ventricular Biopsy Protocol for Heart Transplant

Period Time Frequency Procedures

Immediate post0-4 weeks from day five, 6 operative twice weekly

4-6 weeks weekly 3

Late post-operative 2-3 months bimonthly 4

4-6 months monthly 3

6-12 months quarterly 2

Total First Year 18

After one year yearly

(in the absence of rejection)

After rejection 14-21 days therapy

EMBs, and other biopsies, are problematic. However, because during each biopsy a number of potential complications may occur. These complications include the following:

right ventricular perforation cardiac tamponade ventricular and supraventricular arrhythmia embolus (thrombus or air) pneumothorax air in the pleural cavity infection bleeding

EMBs are the principle method for monitoring cardiac allograft rejections. The heart material obtained from the biopsy is then graded for the level or severity of the rejection. The International Society for Heart and Lung Transplantation (ISHLT) Koblect et al., Transplant Pathology, p. 200, (Am. Soc. Clin. Path., 1994), rate cardiac rejection as follows:

Table 2 International Society for Heart and Lung Transplantation

Grade 0 No evidence of acute rejection

Grade 1 Mild

A. Focal/Perivascular

B. Diffuse/Interstitial

Grade 2 Moderate, Uni-focal

Grade 3 Moderate, Multi-focal

A. Several foci

B. Diffuse

Grade 4 Severe

Ongoing

Mild, moderate, severe

Resolving

In an alternative formulation, Billingham's Kolbeck et al., Transplant Pathology, p. 199, (Am. Soc. Clin. Path., 1994), Histopathologic Classification of Rejection establishes the features of tissue rejection as follows:

Table 3

Billingham's Histopathologic Classification of Acute

Rejection in Human Heart Allograft

Severity of Acute Features Prognostic and Rejection Implications Therapeutic

Mild Rare (usually 1 -2) localized Reversible, typically perivascular collections of without augmentation of mononuclear cells with limited immunosuppressive therapy. extension into the interstitium. No definite myocardial injury. Moderate Collection of "activated" Reversible, typically with perivascular and interstitial augmentation of therapy mononuclear cells with associated and rebiopsy. myocyte injury.

Severe Widespread inflammatory Reversible, but with infiltrates including mononuclear difficulty. Requires cells and often polymorphonuclear augmentation of therapy. leukocytes and eosinophils. Multifocal tissue and small vessel necrosis is associated with fresh hemorrhage.

Resolving Granulation tissue at various Reversed rejection, stages if collagenization. Includes spontaneously or numerous fibroblasts with therapeutically induced. scattered mononuclear cells, plasma cells and phagocytosed lipochrome pigment.

Thus, the EMB, which is a physical biopsy and diagnostic aid, is hazardous for the patient. Attempts have been made to reduce the number of biopsies per patient, but these attempts have not been successful, due in part to the difficulty in pinpointing the sites where rejection starts and to the difficulty in assessing tissue without performing the actual biopsy.

One approach to reducing the number of biopsies is described in U.S. Patent Application No. 60/046,368 "METHODS AND APPARATUS FOR DETECTING THE REJECTION OF TRANSPLANTED TISSUE" and U.S. Patent Application No. , filed March 12, 1998, and entitled "METHODS AND APPARATUS FOR

DETECTING THE REJECTION OF TRANSPLANTED TISSUE" (these applications, as with all references cited herein, are explicitly incorporated by reference herein in their entirety). This application describes a method and apparatus that reduces the number of EMBs needed for a patient and that assists in pinpointing sites where rejection starts. The method described therein entails collecting information from transplanted animale tissue for the purpose of determining whether that tissue may be undergoing rejection. (The term "animale tissue" as used herein describes the tissue of both humans and non- human animals.) As to this method, a number of points should be made. First, the type of information being collected from the environment differs from that conventionally obtained. Second, the means by which the information is collected is unconventional, and third, the purpose for which the information is being collected differs from the purposes for conventional data acquisition processes. Thus, it would be helpful to have available techniques for processing this information, including validation, conditioning, and analysis of this information. To satisfy this need, the present invention provides techniques for validating, conditioning, and analyzing information from transplanted animale tissue for the purpose of determining whether that tissue may be undergoing rejection.

Summary of the Invention The present invention validates and processes fluorescence spectral data for detecting the rejection of transplanted tissue. In accordance with a preferred embodiment of the invention, a software facility for validating and processing fluorescence wavelength data ("the facility") validates, conditions, and analyzes fluorescence spectral datasets ("datasets") collected from transplanted tissue in order to determine whether the tissue may be undergoing rejection. The datasets indicate, for each of a number of wavelengths within a range of wavelengths, the intensity of the tissue's fluorescence response at that wavelength. In various embodiments, validation may encompass testing the intensity of the dataset at a particular testing wavelength, and/or measuring the integrated, or "aggregate," intensity across a subrange of the wavelengths of the dataset. Conditioning may encompass performing baseline correction on the dataset to correct for idiosyncrasies of the collection device used to collect the dataset, performing spectral smoothing on the dataset to minimize noise in the dataset, and/or scaling the dataset to a uniform scale. Analysis may encompass determining whether the dataset indicates that the tissue may be undergoing rejection on the basis of the relative integrated intensities in two or more different wavelength subranges of the dataset and/or on the basis of the spectral width or wavelength of maximum intensity of the largest peak in the dataset or in a subrange of the dataset. By applying these techniques, the facility is able to ensure the validity of the dataset, correct for incidental artifacts common to such datasets, and effectively analyze the dataset to ultimately determine whether the dataset indicates that the tissue whose fluorescence response it represents may be undergoing rejection. Brief Description of the Drawings

Figure 1 is a high-level block diagram of the computer system cup on which the facility preferably executes.

Figure 2 is an overview flow diagram showing the steps preferably performed by the facility to process each dataset received via the data input device 12.

Figure 3 is a detailed flow diagram showing the steps preferably performed by the facility to validate the dataset as part of step 201.

Figure 4 performed by the facility is a dataset graph showing a sample dataset that satisfies the "intensity at sampling wavelength" test. Figure 5 is a dataset graph showing a sample dataset that fails to satisfy the

"intensity at sampling wavelength" test.

Figure 6 is a dataset graph showing a dataset that fails to satisfy the "integrated intensity over sampling wavelength range" test.

Figure 7 is a dataset graph showing the sample dataset that fails to satisfy the "integrated intensity over sampling wavelength range" test.

Figure 8 is a detailed flow diagram showing the steps preferably performed by the facility as part of step 204 to condition the dataset.

Figure 9 is a dataset graph showing a dataset affected by the idiosyncrasies of a particular collection device used to collect it, as well as the calibration correction vector for this collection device.

Figure 10 is a dataset graph showing the application of a correction vector by the facility in accordance with step 801.

Figure 1 1 is a dataset graph showing a sample dataset before spectral smoothing.

Figure 12 is a dataset graph showing the result of performing spectral smoothing on the sample dataset shown in Figure 11.

Figure 13 is a dataset graph showing a sample dataset before scaling.

Figure 14 is a dataset graph showing the same dataset after scaling.

Figure 15 is a detailed flow diagram showing the steps performed by the facility in a first preferred embodiment of the invention to analyze the dataset as part of step 205.

Figure 16 is a dataset graph illustrating the performance of the steps shown in Figure 15.

Figure 17 is a detailed flow diagram showing the steps performed by the facility in a second preferred embodiment of the invention to analyze the dataset as part of step 205.

Figure 18 is a dataset graph illustrating the performance of the steps shown in Figure 17. DETAILED DESCRIPTION OF THE INVENTION

The present invention validates and processes fluorescence spectral data for detecting the rejection of transplanted tissue. In a preferred embodiment, a software facility for validating and processing fluorescence wavelength data ("the facility") validates, conditions, and analyzes fluorescence spectral datasets ("datasets") collected from transplanted tissue in order to determine whether the tissue may be undergoing rejection. The datasets indicate, for each of a number of wavelengths within a range of wavelengths, the intensity of the tissue's fluorescence response at that wavelength. In a preferred embodiment, a dataset is an array or vector of values representing the intensities of wavelengths at approximately 2 nm intervals between approximately 480 nm and 800 nm. The values of a dataset may alternatively represent other similar or equivalent measurements of photonic or illuminative magnitude, such as energy. A dataset is preferably collected by illuminating animale tissue at a range of wavelengths and an intensity that causes the tissue to fluoresce. This fluorescence response is collected by optical sensors and transformed into an electrical signal, and ultimately a dataset as described above. The basic procedure for gathering data may be summarized as a) illuminating the transplanted tissue under conditions suitable to cause the transplanted tissue to fluoresce; b) collecting the fluorescence to provide a transplant fluorescence signature; and c) comparing the transplant fluorescence signature with a healthy fluorescence signature representative of the same type of tissue as the transplanted tissue, and therefrom determining whether the transplanted tissue exhibits one or more characteristics indicative of rejection. The known fluorescence signature is obtained from a sample tissue having a known rejection status, which is preferably healthy but can be grade I, II, III or IV rejection.

It should be appreciated that healthy tissue exhibits a characteristic fluorescence response in reply to excitation with ultraviolet to visible light. The present inventors have discovered that the fluorescence response of transplanted tissue changes as the transplanted tissue is rejected by the host organism. The present invention measures changes in the fluorescence properties of transplanted tissue, both in vitro and in vivo. The changes in fluorescence properties identify characteristics of rejection of the transplanted tissue. The detection of such characteristics of rejection assist in determining whether a tissue biopsy is needed in a transplanted organ, and thus permits the elimination of needless biopsies to the benefit of the patient. Such detection also assists in selecting sites within an organ for tissue biopsies for pathological analysis.

A transplanted tissue is a tissue for an organ such as the heart, liver, kidney, skin, or lungs that has been transferred from a first, donor organism or a synthetic source such as a tissue culture (e.g., for blood or skin), to a second, donor organism (also referred to as a host or recipient). The transplant can be from any combination of donor and donee organisms or sources, including homogeneic, syngeneic, allogeneic or heterogeneic organisms. The transplanted tissue exhibits or comprises one more characteristics indicative of rejection by the host when the tissue appears to suffer at least Grade 1 or mild rejection as discussed relative to the above Tables. In a preferred embodiment, where the transplanted tissue exhibits characteristics indicative of rejection, the method further comprises determining the level of rejection, which can be correlated to the grades and/or levels discussed in the Tables above. In various embodiments of the present invention, validation of the dataset may encompass testing the intensity of the dataset at a particular testing wavelength, and/or measuring the integrated, or "aggregate," intensity across a sub-band of the wavelengths of the dataset. Conditioning may encompass performing baseline correction on the dataset to correct for idiosyncrasies of the collection device used to collect the dataset, performing spectral smoothing on the dataset to minimize noise in the dataset, and/or scaling the dataset to a uniform scale. Finally, analysis may encompass determining whether the dataset indicates that the tissue may be undergoing rejection on the basis of the relative integrated intensities in two or more different wavelength sub-bands of the dataset and/or on the basis of the spectral width or wavelength of maximum intensity of the largest peak in the dataset or in a subrange of the data set.

By applying these techniques, the facility is able to ensure the validity of the dataset, correct for incidental artifacts common to such datasets, and effectively analyze the dataset to ultimately determine whether the dataset indicates that the tissue whose fluorescence response it represents may be undergoing rejection. Figure 1 is a high-level block diagram of the computer system upon which the facility preferably executes. The computer system 100 may either be a general-purpose computer system or a dedicated special-purpose computer system. The computer system 100 contains a central processing unit (CPU) 110, input/output devices 120, and a computer memory (memory) 130. The input/output devices include a storage device 121 , such as a hard disk drive and a computer-readable media drive 122, which can be used to install software products, including the facility (which are provided on a computer-readable medium, such as a CD-ROM). The input/output devices also include a data input device 123 for receiving wavelength datasets. This data input device 123 may be realized as a network connection to another computer system or an interface device for connecting the computer system 100 to an apparatus for collecting fluorescence wavelength datasets A display device 124, such as a video monitor, may be provided for displaying the results of processing the datasets. The memory 130 preferably contains the facility 131 for processing datasets, as well as one or more datasets 132. The contents of the memory 130 may also be stored persistently on the storage device 121. While the facility is preferably implemented on a computer system configured as described above, those skilled in the art will recognize that it may also be implemented on computer systems having different configurations.

Figure 2 is an overview flow diagram showing the steps preferably performed by the facility to process each dataset received via the data input device 123. Each dataset includes intensity values for different wavelengths. Datasets processed by the facility may either be stored in the memory 130 until processing begins, or may be processed by the facility in real time as they are received via the data input device. In various additional embodiments, the dataset processed in accordance with these steps may be received by the computer system 100 in other ways, such as via the computer-readable media drive 122. It should be noted that, while the steps of Figure 2 are shown in a particular sequence, these steps may be advantageously performed in various other sequences. Indeed, the substeps discussed below of the steps shown in Figure 2 are interleaved in certain embodiments in order to coordinate interaction between the substeps.

In order to gain an appreciation for the nature of the data in the datasets, it is helpful to review the steps that are taken to determine whether tissue is being rejected. Initially, light having wavelengths ranging from ultraviolet light to visible light are transmitted to illuminate the transplanted tissue under conditions suitable to cause the transplanted tissue to fluoresce or otherwise respond. Transmitting the light to the transplanted tissue comprises delivering light from a light source (such as a lamp) to the tissue. The light is typically transmitted by a light guide, such as an optical fiber, fiber bundle, liquid light guide or hollow reflective light guide or lens system. Preferably, and particularly where the methods are implemented in vivo, the light does not comprise UV light because such light can be harmful to the tissue. Further preferably, the light consists essentially of blue light, and even further preferably light of a wavelength of about 430 nm - 450 nm. Preferred specific wavelengths include about 405 nm, 436 nm and/or 442 nm +/- about 5 nm. Conditions to induce fluorescence in tissue are well known in the art. See, e.g., U.S. Patent No. 4,836,203; U.S. Patent No. 5,042,494; U.S. Patent No. 5,062,428; U.S. Patent No. 5,071,416; U.S. Patent No. 5,421,337; U.S. Patent No. 5,467,767; U.S. Patent No. 5,507,287.

"Fluorescence" and "fluoresce" are used herein in their ordinary sense, which includes the emission of, or the property of emitting, electromagnetic radiation, typically in the visible wavelength range, resulting from and occurring following the absorption of the light that is transmitted to the transplanted tissue as a part of the method. Fluorescence includes fluorescent light produced by either endogenous fluorophores or exogenous fluorophores; exogenous fluorophores include those provided by drugs, chemical labels or other external sources. Autofluorescence is fluorescence from endogynous fluorophores. Preferably, the fluorescence is collected at a plurality of wavelengths to facilitate analysis of the transplant fluorescent signature. For example, the collection and analysis of a plurality of wavelengths permits observation of a change of intensity from one wavelength to another. The transplant fluorescent signature is then compared with a healthy fluorescent signature, which means a fluorescent signature that represents healthy tissue that is preferably the same type of tissue as the transplant tissue as the transplant tissue (e.g., the signature for a transplanted human heart is compared to the signature for a healthy human heart). If the transplant fluorescent signature is similar to the healthy fluorescent signature, then the transplanted tissue does not comprise, or exhibit, characteristics of rejection. Thus, a biopsy is typically not needed for the transplanted tissue, and therefore the scanning of the transplanted tissue using the methods of the present invention prevents the unnecessary extraction of tissue from the transplanted tissue, along with the attendant risks discussed above. If the transplant fluorescent signature shows one or more indicia of rejection, such as a red-shift relative to the healthy fluorescent signature, then the transplanted tissue comprises characteristics of rejection, and further action, typically including a biopsy, should be taken.

Fluorescence characteristics that contribute to the changes observable in transplanted tissue undergoing rejection are affected by the wavelength of excitation, the concentration, absorption coefficients, scattering coefficients, quantum efficiency, and the emission spectra of the fluorophores inside the tissue. For example, in vivo determination of the presence or absence of characteristics of rejection of a transplanted heart preferably includes measurement and analysis at the endocardium, epicardium, myocardium and/or arterial tissue of the fluorescence characteristics described above, as well as changes in fluorescence characteristics due to physiological changes associated with rejection such as thickening of the endothelium and increase in collagen content. Different wavelengths of illumination or excitation light can excite difference fluorophores inside the transplanted tissue, and therefore can lead to different quantum efficiencies for exciting tissue fluorescence. Thus, the user can select one or more desired excitation wavelengths in order to achieve better or more complete detection sensitivity. In one embodiment, a Laser/Spectrometer system is used for various excitation wavelengths because such a system conveniently facilitates utilizing excitation wavelengths from about 360 UV nm to about 700 IR nm. In addition to using different wavelengths of illumination light, multiple wavelengths of illumination light can be used simultaneously or sequentially, thereby providing at least two photons of different wavelengths for absorption by the transplanted tissue. For example, combining simultaneous excitation by one photon at 400 nm with excitation by a second photon at 500 nm can provide enhanced detection because the long wavelength light can penetrate deeper into the tissue to sample a large tissue volume. In addition, different fluorophores may be excited, and the absorption of the fluorescence spectra by interfering matter can be reduced.

In one preferred embodiment, the induction of fluorescence comprises the simultaneous excitation of the fluorophore by multiple photons, each having a certain fraction of the energy of a single photon at the desired excitation wavelength. In particular, when the multiple photons (which are of a longer wavelength) simultaneously contact the fluorophore, the energies of the photons combine to provide the same excitation that is achieved by the use of the wavelength. An advantage of this approach is that the longer wavelength, lower energy photons can penetrate deeper into the tissue, and therefore sampling can take place at different and/or deeper tissue depths. Typically, this multi-photon excitation is effected using two photons that each have one- half the energy of the desired photon, although it is possible to use three photons each having one-third the energy, etc. The resulting fluorescence is the same as the fluorescence induced using other excitation methods discussed herein, and therefore the analysis of the fluorescence is also the same. In a preferred embodiment, the illumination light guide(s) comprises a focusing device at its distal end, for example a gradient refractive index (GRIN) lens, a microlens, or a diffractive optic lens. The spectroscopic analysis can comprise comparing a full width at half maximum (FWHM) of the measured fluorescence spectrum that comprises the transplant fluorescent signature with a FWHM of the fluorescence spectrum that comprises a healthy fluorescent signature characteristic of healthy tissue when the healthy tissue is the same type of tissue as the transplant tissue. The FWHM is the full width of the measured fluorescence spectrum at a level that is one-half the maximum height of the spectrum.

The spectroscopic analysis can alternatively, or also, comprise comparing the ratio of the integral intensity of two or more wavelength ands of the spectrum that comprises the transplant fluorescent signature to the same ratio from healthy tissue. The wavelength bands for such an analysis an be selected, for example, by using numerical techniques to select sub-regions from the measured fluorescence spectrum acquired with a spectrometer or by using optical techniques, for example optical bandpass filter, to select specific spectral bands that are measured by broadband optical detectors. Thus, a wavelength band is a range of wavelengths of light defined by a selected shorter wavelength limit and a selected longer wavelength limit. In some embodiments, the wavelength band is measured by a broad band optical detector, which is characterized by a response to light across a broad spectral region, typically greater than several hundred nanometers. Examples of broad band detectors include silicon detectors, photomultiplier tubes (PMTs) and CCD assays. Additionally, the wavelength bands can be specific spectral bands, which can be selected using optical bandpass filters in conjunction with the broad band detector.

As another alternative, the step of comparing can comprise comparing the wavelength of maximum intensity of the fluorescence spectrum of the transplanted tissue with the wavelength of maximum intensity of the fluorescence spectrum from the healthy tissue. The wavelength of maximum intensity is the wavelength at which the fluorescent spectrum reaches its maximum intensity; a red-shift in the wavelength of maximum intensity indicates that the transplanted tissue comprises characteristics of rejection.

In one embodiment of the present invention, the illumination and collection are both performed during a single diastole of a single heart beat (or other selected motion of the target tissue). This embodiment is particularly preferred when the target tissue is the heart. Determination of the diastole of the heart eat can be effected by a variety means that will be apparent to one of ordinary skill in the art in view of the present specification. For example, the user can detect an electrocardiogram of the heart beat of the host, and then use one or more signals, such as the QRS wave or other identifiable event, of the electrocardiogram to initiate or trigger the steps of transmitting and collecting during a single diastole of the heart beat. Alternatively, the user can detect a pulse of the host using a blood pressure monitor, and then use the pulse to trigger the steps of transmitting and collecting. A pulse oximeter, which measures the oxygen content of the blood, may be used to provide the trigger than induces the scanning or date gathering.

In an alternative embodiment, a plurality of measurements are obtained throughout the duration of the heart beat (or other motion). When the tissue is then repeatedly induced to fluoresce and the corresponding fluorescence is synchronously collected, the information obtained provides a generally repetitive series of sequentially increasing and decreasing data points, the increases and decreases correspond to the movement of the heart during a bat, and therefore provide a measure of the heart beat. The data points can then be selected to provide optimal information about the target tissue, for example, by selecting only data points above a certain threshold, by selecting only peak data points and/or by selecting data points that only occur in a certain temporal locale within the beat. In addition, these data point selection criteria can be combined with physiological triggers such as an CG or pulse measurement.

In step 201 , the facility validates the dataset to ensure that it represents valid fluorescence response data for the type of tissue to which it corresponds. This validation is based on empirical analysis. In validating the dataset, the facility determines whether the dataset can form the basis for a reliable determination of whether can form the basis for a reliable determination of whether the tissue to which it corresponds may be undergoing rejection. Dataset samples may be invalid, for example, because of failure in the system used to illuminate the tissue and collect its fluorescence response, or because of a misorientation of the system with response to the tissue. Figure 3 is a detailed flow diagram showing the steps preferably performed by the facility to validate the dataset as part of step 201. In various preferred embodiments, the performance of steps 301 , 303 and 305 shown in Figure 3 is reordered, or one or more of these steps is omitted. In step 301, if the intensity value of the dataset at a particular sampling wavelength is less than a minimum intensity, or greater than a maximum intensity then the facility continues in step 302 to determine that the dataset is invalid, else the facility continues in step 303. Figures 4 and 5 illustrate the "intensity at sampling wavelength" test performed in step 301. Figure 4 is a dataset graph showing a sampling dataset that satisfies the intensity at sampling wavelength test. Figure 4 shows that, at the sample wavelength 650 nm, the dataset reflects an intensity of 700 illumination units ("a.u."). Based upon the examination of a number of datasets known to be valid, a minimum and maximum intensity for this wavelength, such as 400 a.u. and 1400 a.u., respectively, are chosen. With these intensity limits, the dataset shown in Figure 4 passes the intensity at sampling wavelength test, at 700 a.u. is greater than the minimum intensity of 400 nm, and less than the maximum intensity of 1400 a.u.

Figure 5 is a dataset graph showing a sampling dataset that fails to satisfy the "intensity at sampling wavelength" test. It can be seen that, at the sampling wavelength of 650 nm, the dataset shown at Figure 5 has an intensity of only 201 a.u. Because this intensity is less than the minimum intensity of 400 a.u., the facility determines in step 301 that the dataset shown in Figure 5 is invalid.

Returning to Figure 3, if the "intensity at sampling wavelength" test of step 301 does not determine that the dataset is invalid, then the facility continues at step 303. In step 303, if the integrated intensity over a sampling wavelength band is less than a minimum integrated intensity, or greater than a maximum integrated intensity, then the facility continues at step 304 to determine that the dataset is invalid, else the facility continues at step 305. The integrated intensity is preferably obtained by summing the intensity values for the wavelengths in the sampling wavelength band. The sampling wavelength band may either be the entire wavelength band of the dataset, or a sub-band thereof. As was discussed above, the spectroscopic analysis can involve comparing the ratio of the integral intensity of two or more wavelength bands. Those skilled in the art will recognize that the integrated intensity may also be obtained in other ways, including utilizing appropriate approximation processes.

Figures 6 and 7 show the application of the "integrated intensity over sampling wavelength band" test. Figure 6 is a dataset graph showing a dataset that satisfies the "integrated intensity over sampling wavelength band" test. It can be seen from Figure 6 that, within the sampling wavelength band of 550 nm - 700 nm, the dataset has an integrated intensity corresponding to the area of the region 601 under the dataset plot within this range. Step 303 involves comparing this integrated intensity to a minimum and maximum integrated intensity. Like the minimum and maximum sampling wavelength intensity, this maximum integrated intensity is identified by studying the integrated intensities of datasets known to be valid. Figure 7 is dataset graph showing a sample dataset that fails to satisfy the

"integrated intensity over sampling wavelength band" test. It can be seen that the integrated intensity corresponding to the area of region 701 under the plot of this dataset is substantially larger than the integrated intensity of the dataset shown in Figure 6. Because the integrated intensity of this dataset is larger and exceeds the maximum integrated intensity, the facility determines in step 303 that the dataset shown in Figure 7 is invalid.

Returning to Figure 3, if the "integrated intensity over sampling wavelength band" test of step 303 does not determine that the dataset is invalid, then the facility continues at step 305. In step 305, if the standard deviation of the dataset over a further sampling wavelength range that is a subrange of the wavelength band of the dataset exceeds a maximum standard deviation for that further sampling wavelength band, then the facility continues at step 306 to determine that the dataset is invalid, else the facility continues at step 307 to determine that the dataset is valid. In step 305, the facility may employ a number of sampling wavelength ranges over which the standard deviation of the dataset is determined and compared to maximum standard deviations corresponding to each of the multiple sampling wavelength ranges.

Returning to Figure 2, after the facility determines whether the dataset is invalid in step 201, in step 202, if the dataset is valid, then the facility continues in step 204 to condition the dataset, else the facility continues in step 203 to indicate that the dataset is invalid. After step 203, these steps preferably conclude without conditioning and analyzing the dataset. Figure 8 is a detailed flow diagram showing the steps preferably performed by the facility as part of step 204 to condition the dataset. It should be noted that, in various embodiments, one or more of the steps shown in Figure 8 is omitted, and/or these steps are performed in a different order. In step 801, the facility performs baseline correction on the dataset. Baseline correction corrects idiosyncrasies in the dataset that result from the use of a particular collection device to collect the dataset that are discernible in the absence of an optical input signal. Various conventional collection devices for collecting spectral data systematically introduce certain errors, or "shifts," in the intensity values at certain wavelengths when not exposed to any optical signal. Such errors are known as variation in the collection device's "null response." In accordance with the baseline correction procedure, a correction vector is first generated for the collection device. The correction vector indicates, for each effective wavelength, the amount that the intensity for that wavelength reported by the collection device should be adjusted, either up or down, to correct the systematic errors introduced by the collection device. The correction vector for baseline correction is preferably generated for a particular collection device by depriving the collection device of any optical signal, comparing the dataset generated in response to the absence of optical signal to a true null response of 0 a.u. at each wavelength of the wavelength range, and creating a correction vector for baseline correction identifying the differences between the dataset and the true null response. The generation of a correction vector for baseline correction is similar to the generation of a correction vector for calibration correction, which is illustrated in greater detail below in conjunction with step 802. As part of step 801, the facility modifies the dataset in accordance with the correction vector generated for baseline correction in order to correct for idiosyncrasies of the dataset that are discernible in the absence of an optical input signal. The steps outlined above are preferably performed in real time.

In step 802, the facility performs calibration correction on the dataset. Calibration correction corrects idiosyncrasies in the dataset that result from the use of a particular collection device to collect the dataset that are discernible in the presence of an optical input signal having a known spectral distribution. For calibration correction, the facility again generates a correction vector. The correction vector is preferably generated for a particular collection device by exposing the collection device to a light source having a known spectral distribution, comparing the dataset generated using this device to the known spectral distribution, and creating a correction vector identifying the differences between the dataset and the known spectral distribution. In step 802, the facility modifies the dataset in accordance with this correction vector in order to correct for the idiosyncrasies of the collection device that are discernible in the presence of an optical input signal.

Figure 9 is a dataset graph showing a dataset affected by the idiosyncrasies of a particular collection device used to collect it, as well as the calibration correction vector for this collection device. It can be seen in Figure 9 that intensity values 901, 902, and 903 appear to constitute significant deviations from the overall shape of the dataset curve. These deviant intensity values are the result of idiosyncrasies of the collection device used to collect the dataset. A correction vector developed for the dataset is shown in Figure 9 as correction values 911, 912 and 913, corresponding to the wavelengths of errant intensity values 901, 902 and 903, respectively.

Figure 10 is a dataset graph showing the application of a correction vector by the facility in accordance with step 802. It can be seen that, when the calibration correction process adds the correction vector into the dataset, once-errant intensity values 1001, 1002 and 1003 are corrected to fall within the general shape of the graph. Returning to Figure 8, in step 803, the facility performs stray light correction on the dataset. Stray light correction corrects for portions of the intensity values of the dataset attributable to light not transmitted directly from the sample animale tissue, and thus properly not part of the spectral fluorescence response of the animale tissue. In order to perform stray light correction, the facility determines the intensity at each of two wavelengths at either end of the wavelength range of the dataset at which a zero signal is expected. Where the intensities at both of these wavelengths are non-zero, the facility interpolates an intensity at each wavelength between these two wavelengths that is likely attributable to stray light, and subtracts the interpolated intensity from the intensity value of the dataset at each such wavelength. While a straightforward interpolation technique such as linear interpolation is preferably used in step 803, those skilled in the art will appreciated that other interpolation techniques are equally applicable.

In step 804, the facility performs spectral smoothing of the dataset. Spectral smoothing refers to a group of techniques used to eliminate unwanted noise from data. Techniques for performing such spectral smoothing, such as Fourier transformation and neighborhood averaging, are well known to those skilled in the art. Figures 1 1 and 12 illustrate a performance of step 803. Figure 11 is a dataset graph showing a sample dataset before spectral smoothing. It can be seen that, in Figure 11, the graph of the dataset, besides exhibiting significant gross features such as features 1101 , 1202 and 1103, also exhibits more minor oscillations, e.g., between intensity values 11 1 1, 1112 and 1113.

Figure 12 is a dataset graph showing the result of performing spectral smoothing on the sample dataset shown in Figure 11. It can be seen that, though the gross features 1 101, 1102 and 1 103 relied upon by some embodiments of the facility remain as features 1201, 1102 and 1203, the noisy oscillations between intensity values 11 11, 11 12 and 1 113 have been eliminated between intensity values 1211, 1212 and 1213.

Returning to Figure 8, in step 805, the facility scales the dataset to a uniform scale. This process, also called "normalization," is well known to those skilled in the art, and involves mapping intensity values of the dataset from absolute units, such as illumination units, to a relative scale. This involves identifying the largest absolute intensity value in the dataset or in a wavelength subrange of the dataset, and dividing each absolute intensity value in the dataset by this maximum absolute intensity of the dataset in order to transform the absolute intensity values of the dataset to relative intensity values. After this transformation, each intensity value in the dataset is a relative intensity value, indicating in each case a fraction of the selected normalization value constituted by each intensity value. This scaling process places the dataset in uniform format for further processing. Figure 13 is a dataset graph showing a sample dataset before scaling. It can be seen in Figure 13 that each intensity value is an absolute intensity value between 0 and 1950 a.u. Figure 14 is a dataset graph showing the same dataset after scaling. It can be seen in Figure 14 that the intensity values of the dataset are relative values ranging between 0 and 1 scaled units. After step 805, the steps in Figure 8 conclude. Returning to Figure 2, after conditioning the dataset in step 204, the facility analyzes the dataset in step 205 to determine whether it indicates that the tissue whose fluorescence response it represents may be undergoing rejection. Different preferred embodiments of the invention have different implementations of step 205. Two such implementations are shown in Figures 15 and 17, discussed below. In each implementation, the facility compares some characterization of the dataset to a similarly-obtained characterization of datasets collected from healthy tissue of the same type. For example, where the datasets collected for healthy human heart tissue. If the characterization of the dataset is within an acceptable tolerance of the characterization of the datasets collected from the healthy tissue, then the facility determines that the tissue to which the dataset corresponds is not undergoing rejection. Otherwise, the facility determines that the tissue corresponding to the dataset may be undergoing rejection. The facility preferably uses one or more different kinds of characterizations of the dataset for this comparison.

Figure 15 is a detailed flow diagram showing the steps performed by the facility in a first preferred embodiment of the invention to analyze the dataset as part of step 205. The steps shown in Figure 15 characterize the dataset using the ratio of the integrated intensities of two different wavelength sub-bands of the dataset. Figure 16 is a dataset graph illustrating the performance of the steps shown in Figure 15. In step 1501, the facility determines the integrated intensity across that first wavelength sub- band. While various embodiments of the invention utilize a first sub-band of 480 nm - 520 nm, those skilled in the art will recognize that other sub-bands could be used. The integrated intensity determined in step 1501 for the first wavelength sub-band as shown has the area region 1601 under the dataset curve. In step 1502, the facility determines the integrated intensity across a second wavelength sub-band. While a preferred embodiment of the invention utilizes a second wavelength sub-band of 540 nm - 580 nm, those skilled in the art will recognize that other sub-bands may be used in view of the present specification. The integrated intensity determined in step 1502 is shown as the area of region 1602 under the dataset curve. In step 1503, the facility divides the first integrated intensity by the second integrated intensity in order to obtain the ratio of the integrated intensities of the first and second wavelength sub-bands. The facility then compares the ratio obtained in step 1503 to a corresponding ratio obtained for healthy tissue of the same type. In step 1504, if the ratio obtained in step 1503 for the dataset is within a predetermined tolerance of the ratio determined for healthy tissue, then the facility determines in step 1505 that the tissue in not undergoing rejection, else the facility determines in step 1506 that the tissue may be undergoing rejection.

Figure 17 is a detailed flow diagram showing the steps performed by the facility in a second preferred embodiment of the invention to analyze the dataset as part of step 205. In accordance with the second embodiment, the facility characterizes the dataset using the spectral width of the highest peak in the dataset curve. Figure 18 is a dataset graph illustrating the performance of the steps shown in Figure 17. In step 1701, the facility identifies in the dataset the peak with the largest intensity, i. e. , the highest peak. In Figure 18, it can be seen that the facility identified intensity value 1801 as the highest peak in the dataset. In step 1702, the facility multiplies the intensity of the identified peak by a predetermined percentage in order to obtain a target intensity of 0.5 while the facility preferably uses a predetermined percentage of 50%; those skilled in the art will recognize that other predetermined percentages may be used in view of the present disclosure. In step 1803, the facility identifies the wavelengths at which the dataset intersects the target intensity at wavelengths above and below the wavelength of the identified peak. In Figure 18, it can be seen that, at the target intensity of 0.5, the facility identifies the wavelengths 485 nm and 610 nm at intensity values 1802 and 1803, respectively. In step 1704, the facility determines the difference between the intersecting wavelengths identified in step 1703 to obtain the spectral width of the peak identified at step 1701. Such distance is determined by determining the difference between the wavelengths at which the dataset intersects the target intensity at wavelengths above and below the wavelengths of the identified peak. It can be seen in Figure 18 that this identifies difference corresponds to distance 1804, 125 nm. The facility then compares the spectral width obtained in step 1704 with a spectral width similarly obtained for healthy tissue of the same type. In step 1705, if the spectral width determined for the in step 1704 is within a tolerance of the spectral width obtained for the healthy tissue, then the facility continues in step 1706 to determine that the tissue is not undergoing rejection, else the facility determines that the tissue may be undergoing rejection. Returning to Figure 2, after analyzing the dataset in step 205, these steps conclude. While this invention has been shown and described with reference to preferred embodiments, it will be understood by those skilled in the art that various changes or modifications in form and detail may be made without departing from the scope of the invention. For example, the facility may utilize feature wavelength subranges other than those discussed herein. Further, a number of feature wavelength subranges other than two may be used to characterize the significance of the dataset. Further, features other than the overall intensity of feature wavelength subranges may be used to characterize the significance of a dataset. Yet further, minimal datasets may be used that merely reflect the overall intensity for one or more feature wavelength subranges, thereby bypassing the structural and procedural complexity involved in collecting and processing the more detailed datasets discussed herein.

Claims

1. A method utilizing a computer system for collecting and assessing the validity of a spectral analysis dataset representing the fluorescence response of animale tissue, the method comprising the steps of: illuminating the tissue in a manner suitable to cause the tissue to fluoresce; collecting fluorescence produced by the illuminating step; producing from the collected fluorescence a dataset, the dataset indicating, at each of a range of wavelengths, a luminous intensity of the light collected at that wavelength; determining the aggregate intensity of the dataset over one or more subranges of the range of wavelengths to obtain an aggregate intensity value; comparing the obtained aggregate intensity value to a predetermined minimum aggregate intensity value; and if the obtained aggregate intensity value does not exceed the predetermined minimum aggregate intensity value, determining that the dataset is invalid, in that it does not properly represent the fluorescence response of animale tissue.

2. The method of claim 1, further comprising the step of, if the determined aggregate intensity exceeds a predetermined maximum aggregate intensity value, determining that the dataset is invalid.

3. The method of claim 1, further comprising the steps of: determining the intensity of the dataset of a predetermined sample wavelength; and if the determined intensity at the sample wavelength is less than a sample wavelength intensity threshold, determining that the dataset is invalid.

4. The method of claim 3, further comprising the step of, if the determined intensity at the sample wavelength is greater than a second sample wavelength intensity threshold, determining that the dataset is invalid.

5. A method in a data processing system to validate and process a spectral analysis dataset intended to represent the fluorescence response of animale tissue to assess the rejection of the tissue, the dataset indicating the luminous intensity of light collected at each of a range of wavelengths, the method comprising the steps of: determining the overall intensity of the dataset across the range of wavelengths; if the determined overall intensity falls below a predetermined minimum overall intensity or above a predetermined maximum overall intensity, determining that the dataset is invalid; and if it is not determined that the dataset is invalid: determining the overall intensity for each of two preselected subranges of the range of wavelengths; determining the ratio of the overall intensities of the preselected subranges of the dataset, comparing the determined ratio to a ratio determined for corresponding healthy tissue, and based on the ratio comparison, assessing the rejection of the tissue.

6. The method of claim 5, further comprising the steps of: determining the intensity of the dataset at a predetermined sample wavelength; and if the intensity determined at the sample wavelength falls below a predetermined minimum sample wavelength intensity or above a predetermined maximum sample wavelength, determining that the dataset is invalid.

7. The method of claim 5, further comprising the steps of, if it is not determined that the dataset is invalid: determining the overall intensity for one or more additional preselected subranges of the range of wavelengths; determining the ratio of one or more additional pairs of determined overall intensities of subranges of the range of wavelengths; and wherein the comparing step further comprises the determined ratios of one or more additional pairs of determined overall intensities to ratios determined for corresponding healthy tissue.

8. The method of claim 5 wherein the assessing step includes the steps of: if the ratio determined for the dataset is within a predetermined tolerance of the ratio determined for the corresponding healthy tissue, indicating that the dataset suggests that the animale tissue is not undergoing rejection; and if the ratio determined for the dataset is not within a predetermined tolerance of the ratio determined for the corresponding tissue, indicating that the dataset suggests that the animale tissue is undergoing rejection.

9. A method in a data processing system for assessing the validity of a spectral analysis dataset intended to represent the fluorescence response of animale tissue, the dataset indicating the illuminative magnitude of light collected at each of a plurality of different wavelengths within a range of wavelengths, the method comprising the steps of:

(a) determining the aggregate illuminative magnitude of the dataset over one or more wavelength subranges of the range of wavelengths represented by the dataset;

(b) comparing the determined aggregate illuminative magnitude to a predetermined minimum aggregate illuminative magnitude value; and (c) if the determined aggregate illuminative magnitude does not exceed the predetermined minimum aggregate illuminative magnitude value, determining that the dataset is invalid, in that it does not properly represent the fluorescence response of animale tissue.

10. The method of claim 9, further comprising the steps of: comparing the determined aggregate illuminative magnitude to a predetermined maximum aggregate illuminative magnitude value; and if the determined aggregate illuminative magnitude exceeds the predetermined maximum aggregate illuminative magnitude value, determining that the dataset is invalid.

11. The method of claim 9 wherein the wavelength subranges used in step (a) cover the entire range of wavelengths of the dataset.

12. The method of claim 9, further comprising the steps of : determining the illuminative magnitude of the dataset at a predetermined sample wavelength; and if the determined illuminative magnitude at the sample wavelength is less than a sample wavelength illuminative magnitude threshold, determining that the dataset is invalid.

13. The method of claim 12, further comprising the step of, if the determined illuminative magnitude at the sample wavelength is greater than a second sample wavelength illuminative magnitude threshold, determining that the dataset is invalid.

14. The method of claim 9, further comprising the steps of: determining a standard deviation of the illuminative magnitudes of a predetermined wavelength subrange of the dataset; and if the determined standard deviation of the illuminative magnitudes of the predetermined wavelength subrange of the dataset is larger than an illuminative magnitude standard deviation threshold for the predetermined wavelength subrange, determining that the dataset is invalid.

15. The method of claim 9, further comprising the step of determining a standard deviation of the illuminative magnitudes of each of one or more additional predetermined wavelength subranges of the dataset, and wherein the step of determining if the dataset is invalid determines that the dataset is invalid if the determined standard deviation for any predetermined wavelength subrange is larger than an illuminative magnitude standard deviation threshold for that predetermined wavelength subrange.

16. The method of claim 9 wherein the dataset is generated using a sensing device, further comprising the steps of, before determining the aggregate illuminative magnitude of the dataset: accessing a plurality of illuminative magnitude offsets for the sensing device each representing an adjustment to be made to the illuminative magnitude of the dataset at a particular wavelength in order to correct the illuminative magnitude at that wavelength; and applying the calibration offsets to the dataset in order to eliminate from the dataset inaccuracies relating to the operating of the sensing device.

17. The method of claim 16, further comprising the steps of generating the illuminative magnitude offsets based on a comparison of a proper output of the sensing device in the absence of an optical input signal and the actual output of the sensing device in the absence of an optical input signal.

18. The method of claim 16, further comprising the step of generating the illuminative magnitude offsets by comparing the proper output of the sensing device for an optical input signal having a known spectral distribution to the actual output of the sensing device when exposed to the optical input signal having a known spectral distribution.

19. The method of claim 9 wherein the generation of the dataset is susceptible to the inclusion of the intensity of light not forming part of the fluorescence response of the animale tissue, further comprising the steps of, before determining the aggregate illuminative magnitude of the dataset: determining the illuminative intensity at each of two selected wavelengths at opposite ends of the wavelength range of the dataset at which the expected intensity for the fluorescence response of animale tissue is zero; interpolating, at the wavelengths between the two selected wavelengths, an intensity value between the intensity values at the selected wavelength; and subtracting from the intensity value of the dataset at each wavelength the interpolated intensity value.

20. The method of claim 9, further comprising the step of, before determining the aggregate illuminative magnitude of the dataset, performing a spectral smoothing operation on the dataset.

21. The method of claim 9, further comprising the steps of identifying in the dataset a selected feature; comparing the selected feature identified in the dataset with a corresponding feature recorded for healthy animale tissue; and based upon the comparison, determining whether the animale tissue whose fluorescence response the dataset represents is healthy.

22. The method of claim 9, further comprising the steps of: determining the aggregate illuminative magnitude of the dataset across a first predefined wavelength subrange; determining the aggregate illuminative magnitude of the dataset across a second predefined wavelength subrange; dividing the determined integrated illuminative magnitudes to obtain a characteristic ratio; comparing the characteristic ratio obtained for the dataset with a corresponding characteristic ratio recorded for healthy animale tissue; and based upon the comparison, determining whether the animale tissue whose fluorescence response the dataset represents is healthy.

23. The method of claim 9, further comprising the steps of: identifying in the dataset a first wavelength at which the largest illuminative magnitude occurs; determining a targeted illuminative magnitude at a predetermined percentage of the largest illuminative magnitude in the dataset; identifying a second wavelength that is the largest wavelength smaller than the first wavelength at which the dataset intersects the targeted illuminative magnitude; identifying a third wavelength that is the smallest wavelength larger than the first wavelength at which the dataset intersects the targeted illuminative magnitude; determining the distance between the second and third wavelengths to obtain the spectral width of the dataset; comparing the spectral width of the dataset with a corresponding spectral width recorded for health animale tissue; and based upon the comparison, determining whether the animale tissue whose fluorescence response the dataset represents is healthy.

24. An apparatus for collecting and assessing the validity of a spectral analysis dataset representing the fluorescence response of animale tissue, comprising: a source of illumination for illuminating the tissue in a manner suitable to cause the tissue to fluoresce; an optical sensor for determining the luminous intensity of the fluorescence produced in the tissue by the source of illumination at each of a range of wavelengths; an accumulator for determining the overall intensity of the fluorescence over one or more subranges of the range of wavelengths; a comparator for comparing the determined overall intensity to a predetermined minimum overall intensity value; and a discriminator for determining that the dataset is invalid, in that it does not properly represent the fluorescence response of animale tissue, where the overall intensity determined by the accumulator does not exceed the predetermined minimum overall intensity value.

25. The apparatus of claim 24, further comprising an optical probe for conveying light from the source of illumination to the tissue, and for conveying the fluorescence produced by the tissue to the optical sensor.

26. The apparatus of claim 24 wherein the comparator further compares the determined overall intensity to a predetermined maximum overall intensity value, and wherein the discriminator further determines that the dataset is invalid, in that it does not properly represent the fluorescence response of animale tissue, where the overall intensity value determined by the accumulator exceeds the predetermined maximum overall intensity value.