CA2177784A1

CA2177784A1 - Registration of nuclear medicine images

Info

Publication number: CA2177784A1
Application number: CA002177784A
Authority: CA
Inventors: Alex Natanzon; Naor Wainer; Gideon Berlad; Shoulamit Cohen-Shwartz
Original assignee: Individual
Current assignee: GE Medical Systems Israel Ltd
Priority date: 1995-05-31
Filing date: 1996-05-30
Publication date: 1996-12-01
Also published as: DE19621540A1; IL118255A0; FR2734935B1; US5871013A; FR2734935A1; JP3022773B2; JPH09133771A

Abstract

A method of registering a plurality of functional images comprising providing a plurality of functional images, providing a plurality of structural images each one of which has a known positional relationship to at least one of said plurality of functional images, finding a first mapping transformation between pairs of functional images based on said first mapping transformation and said positional transformation.

Description

995~:~57~LZt~;OmlDC ~ 1 7 7 7 8 ~L
REGISTRaTION OF NUCLE~R MFnIrT?~F IMAGES.
FIELD OF THE INVENTION
The present invention relates to the art of diagnostic imaging. In particular, the invention relates to nuclear imaging systems incorporating simultaneous transmission and emission , L c~hy .
BACKGF~OUND OF THE INVENTION
SPECT (Single Photon Emission Computerized Tomography) is used to study the three dimensional di3tribution o~ a radionuclide in a patient. Typically one or more radiopharmaceuticals are ingested or are in~ected into the patient . When radiopharmaceuticals are in~ ected it is usually into the patient ' s blood stream, to image the,,~ardio-vascular system or to image specific organs which absorb the in~ected radiopharmaceuticals. One or more gamma or scintlllation detectors are positloned near the patlent to record emltted radlation .
SPECT lmages' are generally produced by:
(a) rotating the aetectol(s) around the patient in order to record emlssions from a plurallty of dlrectlons; and (b) transformlng the recorded f.m~qq~nq, uslng methods well known ln the art, lnto a I - ,L~hlcal multi-slice lmage, a three dlmenslonal image or some other representation of the dlstrlbutlon of the radlopharmaceutical in~ ected into the patient ' s body.
One problem with SPECT is that the tissues surroundlng the organs being imaged attenuate and scatter the radiation emltted by the radlopharmaceutical, dlstortlng the resultlng SPECT
lmages. To solve thls problem, a SPTCT ( Slngle Photon Transmlsslon Computerized Tomography ) image of the region being lmaged, ls ac~ulred, slmultaneously wlth the SPECT lmage. The SPTCT lmage provldes information regarding the attenuation and scatteri~g characteristics of the region being imaged, so that the multi-view ~m~q.q~n data can J~e corrected.
In order to acSIuire the simultaneous SPTCT image, a source of radiation is placed opposite the patient ' s body from the 1, .
/

995XI~ ~Q~ 2~777~
detectors(3) and rotated with the detector(s). Preferably, but not necessarily, the energy of the SPTCT source is different from that of the radiorh;~rm=~-eutical so that the detector ls able to easily differentiate the two radiations.
Since the ~m~ cy I nn image is ac~uired at the same time as the transmission image, and the relative geometry of the SPTCT and SPECT systems are known, the images are easily registered to one another .
The diagnostic method that uses SPECT and SPTCT
simultaneously is known as STET ( Simultaneous Transmission and n Tomography~. This method is ~ r;h~ in further detail in US Patent 5,210,421, the disclosure of which is in-,oly~JL~ d herein by reference.
One aspect of the present lnvention relates to the use of STET imaging technigues for functional imaging. In this use, the resultant STET image shows the metabolic activity of body tissue, since dead or damaged body tissue absorbs the radioph~rr-^~utical at a different rate (or not at all) from healthy tissue. When used in this manner, the STET image shows the functional activity of the body tissue, not its structural detail.
However, STET images have two drawbacks. First, as indicated above, the STET image does not show much structural detail;
therefore, it is difficult to pinpoint where the imaged function is occurring in the patient ' s body. Many diagnostic imaging methods, in modalities other than nuclear ~-l~r~n~, reveal almost exclusively structure and not function, therefore, it is hard to compare STET images with other types of diagnostic images.
Second, a common me~hr ~ gy, especially in cardiac examination, is to acguire a STET image shortly after in~ection of the radio pharmaceutical and to ac auire another STET image of the same region after a certain period of time. By comparing these two (or more ) images, it is possible to learn still more about the function of the tissue studied, such as the speed at which different portions of tissue absorb and metaboli~e the radiopharmaceutical. However, if the two STET images are too different, it is not possible to losely compare them because the 3:1S ~. ~O~IZC
~ ``'~ 2177~84 operator can not match the dif ferent parts of the images to each other .

3995 3:~5 11~55LIBX
.` ``~ ~ 21777~4 SUMMARY OF THE INVENTION
The present lnvention contemplates a method for registering STET images and other functional images to images of other modalities, and for matching two STET images taken at different times of the same body region, thereby solving the above mentioned problems.
In accordance with one preferred embodiment of the present invention, a method for matching two STET images acquired at different times uses the SPTCT data in order to identify structure in the patient ' s body . When two STET images are to be compared, the two respective SPTCT images are registered, preferably, using a correlation method or another known image matching method. Since the STET image is registered to its SPTCT
image, registering the two SPTCT images automatically registers the two STET images.
In accordance with another preferred embodiment of the present invention, a method for registering a STET image and a structural diagnostic image ( such as an MRI, ultrasound or X-ray CT image ) uses the SPTCT data in order to identify structure in the patient ' s body . When the STET image is to be registered to the structural diagnostic image, the structural SPTC~ image and the structural diagnostic image are registered. This registration is preferably ~ h.of~ through the choosing and comparing of pLI 'n~nt body structures, such as the skeleton, organs or body outlines. Once this matching is accomplished, a mapping between the images can be defined, based on the mapping between the prominent body structures chosen. This mapping is used to transform one image so that it can be superimposed over the other image .
Alternatively, prominent body markings on the SPTCT image are saved as ~ u~ ~y marks with the STET image. These marks are used to match the STET image to another ~u~.l,uLal image.
In accordance with yet another preferred embodiment of the present invention, a method or registering a irst SPECT image to a structural diagnostic image uses a second SPECT image to serve as a structural image. Two SPECT images are acquired o the "S~ 217778~
studied region, the first image is acquired using a first radiopharmaceutical, which is selected so that the resultant SPECT image shows the desired function, The second SPECT image is acquired using a second radiopharmaceutical, which is selected so that the resultant image shows some structure, such as outlines of organs which can be used to register the second SPECT image to another structural image. Alternatively, parameters other than the radiopharmaceutical are varied in order to generate the different SPECT images.
Matching between the second SPECT image and the structural diagnostic image is accomplishea through the choosing and eomparing of prominent body structure shown in both images.
Preferably, the two SPECT images are acquired simultaneously using a dual isotope gamma camera, so that they are automatically registered .
A mapping between the first SPECT image and the structural diagnostic image is then created based on the inherent registration between the two SPECT images and the matchlng between the second SPECT image and the structural diagnostic image. It should be noted that this preferred "o~l~mPnt does not require a STET device, a SPECT device is sufficient.
In a simple situation, the sl~e and shape of the images is not affected and only translation and/or rotation is required.
Where scaling is required, one of the images is scaled in aceordanee with the eorrelation of a plurality of ehosen struetural features or of the images as a whole. In one embodiment of the invention, warping and other complex eorreetions ean be applied to improve the match between the images .
The term " structural image " as used herein means an image that ls used to compare struGtures. The term "functional image"
as used herein means a functional image that is not used to determine registration. As can be appreciated, functional images may show structure and a substantial amount of structure in struetural images may be eaused by funetionality.
Preferably for many types of studies, the aequisition of SPECT, 6PTCT and STET lmages 18 synchronlzed to the cardlac rhythm, the resplratory rhythm or other body motlons by gatlng. In such gated lmages data acqulred durlng the lmaglng process 18 blnned (or wlndowed) accordlng to a gatlng slgnal derlved f rom the body rhythm.
Thus, ln a preferred ~ 1. of the lnventlon, lmage acqulsltlon 18 gated to body rhythms and motlons.
Preferably, the structural lmages are also synchronlzed ln the same manner. For example, gated CT lmages are used as structural lmages lnstead of res~ular CT lmages when the STET
images are ~ated. An advantage o~ comblnlng STET lmaglng wlth gatlng 18 the ablllty to correct blnned data for patlent motlon durlng data acqulsltlon by reallgnment based on the resJlstratlon of the lma51es. Thls corrects for smearlng otherwlse produced by patlent motlon. Addltlonally, data from separate blns 18 more easlly comblned.
Another advantage 18 the ablllty to correct organ motlon caused by the gated rhythm, by applylng a geometrlc ~transformatlon to data acqulred based on the phase of the gated~hythm. Yet another advantage 18 the ablllty to reglster transmlsslon lmages to emlsslon lmages even when they are not acqulred slmultaneously. A transmlsslon lmage of a patlent whlch 18 gated to body rhythms can be automatlcally reglstered to lts correspondlng gated emlsslon lmage, slnce most of the mlsallgnment between the two lmages 18 caused by body rhythms whlch are, ln general repet lt lve .

~199S3:1S~ 2 ~ 7 7 7 8 4 BRIEF DESCF~IPTION OF THE DE~AWINGS
Fig. 1 is a partial, s~mrl~fied schematic view of a slice of the human body in the chest region, showing the heart, ribs and a portion of functioning heart tissue;
Fig. 2A is a simplified schematic of a SPTCT scan of the body slice from Fig. l;
Fig. 2B is a simplified schematic of a STET image of the body slice shown in Fig. l;
Fig. 2C is a simplified schematic of a STET image of the body slice shown in Fig. 1, acquired at a different time from Fig. 2B;
Fig. 3 is a simplified schematic X-ray CT image of the body slice shown in Fig. 1. ~
Fig. 4A is a simplified correlated STET image created by al ;gn~n~ and superimposing the STET images from Fig. 2B and Fig.
2C;
Fig. 4B is a superposition image created from the functional STET image in Fig. 2B and the structural image from Fig. 3;
Fig. 5 is a simplified schematic STET image with fiduciary marks for aiding in correlation with structural images such as X-ray CT scans; and Fig. 6 is a simplified block diagram of a STET system lncluding equipment for cardiac and respiratory gating.

~ 5. ~Im 3:15 ~i. I~K
~ `~ 217778~
DETAILED DESCKIPTION OF THE J~Kk-~hKKkl~ EMBODIMENT
The present invention does not require the use of any specific STET device, and for most devices the invention can be practiced by changes and/or additions in image processing and registration. In addition, lt is possible to use the present invention with NON-STET devices, provided that the SPECT and SPTCT images can be registered to each other.
Fig. 1 in US patent 5,210,421 shows a typical STET camera assembly which is used for acquiring STET images.
The process for acquiring these images typically lnrl-l~lPq ~ a) placing a patient on a couch, so that the part to be studied will be in an examlnation area;
(b) injecting a radioph~ eutical into the patient;
( c ) acquiring pairs of SPTCT and SPECT images using one or more detectors;
( d ) rotating the detector( s ) around- the examination area, in order to acquire a plurality of image pairs;
(e) transforming the plurality of image pairs into a multi-slice tomographical STET image, a three ~ nc~oni~l STET
image or another representation of STET data, the SPTCT images being employed to correct the attenuation and scattering artifacts in the SPECT images to produce the STET images;
(f) optionally, after an attending physician Px~m~nP~
this image, the patient is sent to rest and/or exercise and/or rein~ ection;
( g ) after a period of rest or exercise, the image acguisition process is typically repeated, with the patient placed in as nearly as poss~ hl P the same position as during the previous study, so as to facilitate comparing the new images with the old ones.
Preferably for many types of studies, the acquisition of SPECT, SPTCT and STET images is synchronized to the cardiac rhythm, the respiratory rhythm or other body motions by gating.
In such gated images data acquired during the imaging process is binned ( or windowed ) according to a gating signal derived from the body rhythm.

jW5~:15~=L~ 21777~4 The following discussion refers to a section of the patient ' s body being imaged, shown in Fig . 1. Fig. 1 is simplified to include only a heart 1 including a functionally active area 2 of the heart, ribs 8 and a backbone 3. In order to simplify the discussion, only one slice is shown, even though the STET image is three 1~ q~nAl Application of the invention to three dimensions and choosing the correct slices is described below .
Fig. 2B shows a STET image 6 of the body slice shown in Fig.
1, such as would be acquired in a heart study. In such studies, most of the radiophArTn?c~utical is concentrated in the blood or in soft tissues and specific organs such as the heart and liver, so that the acquired STET image 6 shows mostly portions of target organs and a fuzzy outline 9 of the patient ' s body. Fig . 2C shows a later STET image 6 ' of the same region in the same patient.
With the passage of time, the radiorhAr~=~~tical is Ah5~rh~o~1 and metabolized by the body tlssues, and the STET image changes, as can be seen by oomparing image 6 with image 6 ' . In Fig. 2C a functionally active area 2 ' is imaged which is larger than area 2.
Fig. 2B and Fig. 2C are STET images 6 and 6 ' o the region shown in Fig. 1. The images 6 and 6 ' show functionally active areas 2 and 2 ' respectively but not bones such as the ribs 8 or even the non-active areas of heart 1. Fig. 2A shows a very simplified SPTCT image 7 which is a structural image, much like a standard X-ray CT, except for poorer resolution and lower organ definition ability. The SPTCT image 7, shows heart 1, ribs 8 and even baclcbone 3, but does not specifically differentiate the functionally active areas of the heart.
In the later STET image 6 ', of Fig. 2C, there are significant changes from the earlier STET image 6, of Fig. 2B, making it difficult, if not impossible, to match correctly functioning area 2 in image 6 with functioning area 2 ' in image 6 ' . In addition, it is difficult to identify correctly the structural areas ~hich are functioning as revealed by the radiorhAr~--Putical .
I

, Apdl 5~1~5 ~:15 rUI.. ZOri~
177784 v ~ second SPTCT image is acquirea simultaneously with image 6 ' . The SPTCT images acquired with images 6 and 6 ' are very similar, since the patient ' s body structure does not change much between the images, and the continuing diffusion of the radiorh~rm~ t~tical which plays a crucial part in images 6 and 6 ' doeg not play a part in SPTCT imaging. Two types of differences between the two SPTCT images are caused by:
(a) changes due to patient movement caused, for example, by breathing; and (~) changes due to different placement of the patient on the e2~amination table.
Since the respective emission and transmission images are acquired with the same known system geometry, the mapping o the emission image to its respective tr~n -m~ ssl on image is also known, so the two respective images can be r~r~ncr~r~red registered to each other. The following discussion assumes that any necessary registration between the two respective images has been performed .
A preferred embodiment of the invention uses the following process in order to transform a SPTCT structural image, which has an associated registered STET image, so that it is registered to a structural image:
(a) marking L~L~ 'n~nt body structures in the two structural images;
( b ) correlating the prominent structures between the structural images;
(c) det~rm1n~nJ a tranaLulll,Lion between the two structural images, based on the correlation between the structures; and ( d ) transforming the SPTCT image in accordance with the transformation found in (c).
The transformation will have a degree of complexity c~L~",Llate to the images being aligned, and may include:
( i ) simple rr 1 ~ ~; t of the images;
( ii ) scaling of one o~ the images; and ( iii ) warping one of the images .
The functional STET image associated with the SPTCT image is ` ~ 2~77'7~4 transformed uslng the same transformatlon as that used for the 8PTCT lmage.
In a preferred embodlment of the lnventlon, reglsterlng of two STET lmages 6 and 6' 18 achleved by reglstering the two respectlve assoclated SPTCT lnages uslng the above descrlbed method . The re~lst rat lon of STET lmages 6 and 6 ' f o l l ow8 aut omat l ca l l y .
In an addltlonal preferred ~ of the lnvent lon a STET lmage 6 18 to be reglstered to a st ructural lmage such as X-ray CT lmage, a MRI lmage or an ultrasound lmage. Flg. 3 shows a CT lmage 70, such as 18 to be reglstered to STET lmage 6. The reglstratlon 18 performed by uslng the ~bove descrlbed process to reglster SPTCT lmage 7, that 18 assoclated wlth STET lmage 6, to CT lmage 70. The registratlon of STET lmage 6 to CT lmage 70 follows automatlcally, uslng the same transformatlon used to reglster the two structural lmages.
In yet another preferred . ~i t of the lnventlon, a SPECT lmage 18 reglstered to a structural lmage, ~uch as an X-ray CT lmage, uslng a second SPECT lmage as a structural lmage lnstead of uslng a SPTCT lmage. A SPECT
devlce 18 used to slmultaneously acgulre two lmages, wlth one lmage showlng enough structure to be used as a structural lmage. The two lmages are ac~ulred uslng a dual lsotope gamma camera and a dlfferent radiopharmaceutlcal for each lmage.
Slnce the functlonal and the structural SPECT lmages are automatlcally re~lstered, reglsterlng the structural SPECT
lmage wlth the X-ray CT lmage or other structural lmage 21777~
automatically reglsters the functlonal 8PECT lmage wlth the X-ray CT lmage or other structural lmage. Accordlngly, the re~lst rat lon between the st ructural SPECT lmage and the structural image i9 performed by uslng the &bove descrlbed reglstration procesE. The registratlon of the functlonal SPECT lmage to the structural lmage follows automatlcally, uslng the same transformatlon used to reglster the two structural lmages.
For example, to detect and locate malignant llver lesions, two SPECT lmages and one CT lmage are acgulred o~ the liver. A flrst SPECT lmage, which 18 acqulred uslng FDG, highlights only malignant tumors and shows little body structure. A second 8PECT image, acquired slmultaneously using lntravenously in~ected Tc99m collold, clearly shows the anatomlc boundaries of the liver and leslons. A CT lmage of the llver and surroundlng tlssue also clearly shows the anatomlc boundarles of the llver and leslons. Therefore, the CT lmage (the structural lmage) 18 reglstered to the second SPECT lmage (the structural SPECT lmage) uslng the registratlon process descrlbed hereln. Consequently, the ~lrst SPECT image is reglstered to the CT lmage (because the two SP3CT lmages are acqulred slmultaneously and, therefore, automatlcally reglstered to each other) 80 that the mallgnant leslons can be polnted out on the CT lmage.
Typlcally a three dlmenslonal lmage 18 acqulred and processed as a serles of two dlmenslonal sllces. In order to properly reglster sllces of three dlmenslonal lmages, as ` 217778~
descrlbed above, sllce palrs that have the same locatlon along the pat lent ' 8 longltudlnal ~ Z ) axls must be chosen .
In the case of matching two 6TET lmages, COLL~ VI1r71n~ sllces from the two SPTCT lmages must be chosen.
Two preferred methods for matchlng sllces are:
(1) the operator chooses the approprlate sllces, based on hls/her understandlng of the lmages and hls/her knowledge of human anatomy; and ~ 11) slnce the lmage modallty 18 the same for both BPTCT
lmages, a computer can search for the closest matchlng sllce palr uslng a correlatlon algorlthm.
Once the closest matchlng sllces are found, the process contlnues as descrlbed above. Alternatlvely, uslng lmage matchlng technlques known ln the art of lmage processlng, the two SPTCT lmages can be matched ln the axlal dlrection wlth a preclslon hlgher than the wldth of a sllce.
Slnce the STET lmage 1B a true three dlmenslonal lmage, one of the two lmages can be "re-sllced", 80 that the lmage sllces of one STET lmage are exactly allgned to the sllces of the other STET lmage.
In the case of reglsterlng a STET lmage to a X-ray CT lmage, the preferred way to flnd the correct matchlng CT
and SPTCT sllces 18 to have the physlclan choose the sllce palr, based on hls understandlng of the lmages and hls knowledge of human anatomy. Once the closest matchlng sllces are found, the STET lmage can be re-sllced 80 that the ST~T
lmage sllces fall on boundarles of the CT sllces. For lmages derlved from dlfferent modalltles, the Z scale may be 217778~
different. A sllce scale factor may be derived belsed on matchlng a plurality of structural features ln dlfferent s 1 ices .
In an addltional preferred ~ 1. of the invention, steps (a) and (b) of the reglstration process are replaced by a slngle step of correlating the two images as a whole. Additlonally, three dlmensional images may also be correlated as wholes, without first slicing them and correlating the slices.
In order to facilitate manual finding and matching or marking of prominent body structures between images, it is useful to dlsplay the images as three-dimensional images on a computer screen and mark the pLI 1n~nt structure on the three-dimensional lmages, 80 that the attending doctor will not have to work directly with lmage slices.
Once the transformation between the two images is known, many image processing techniques are applicable, for example~ im~ge subtraction, rapid flipping of two or more images, superpositioning of outlines of the active areas from one 8TET image on another STET image or on a CT image and pseudo coloring of different areas. Fig. gA shows the superpositioning of the outline of an active area from the STET image 6 on the STET image 6'. Fig. 4B shows the superpositioning of the outline of the active area from the 8TET image 6 on the CT image 70.
In addition, the present invention enables simultaneous processing and viewing of several images which are registered to each other using the methods described 21777~
hereln. For example, two lmages are dlsplayed slde by slde on a computer screen, a portlon of one lmage 18 marked off and radlatlon emltted by that portlon ls computed. The radlatlon emltted by the matchlng portlon of the other lmage 18 calculated and dlsplayed automatlcally by the computer.
In general, the correlatlon algorlthms used for matchlng lmages and sllces, between and wlthln modalltles and the subsequently derlved transformatlons are any of a varlety of methods known ln the art of lmage reglstratlon. The followlng lmage reglstratlon methods are useful ln carrylng out preferred ~ 1r- ' 8 of the lnventlon.
1. Landmark matchlng. CollP~I,ol ~1n~ anatomlcal or external markers are ldentlfled ln the sets of data to be matched. A mlnlmum root mean square all~nment transformatlon 18 then calculated to allgn one set of markers wlth the other set. Preferably, the markers are ~dentlfled by an operator.
2. 8ur~ace matchlng. The surface representatlons of two data sets are correlated by flndlng the transformatlon whlch ylelds the minlmum root mean square dlstance between the two surfaces. Thls method 18 descrlbed ln "Accurate Three-Dlmenslonal Reglstratlon of CT, PET and/or MR Images of the Braln", by Pellzzarl C . A., et al ., Journal of Computer Asslsted Tl -"L~lly, volume 13, 1989.

3. Volume matchlng. The two data sets are correlated by flndlng the transformatlon whlch ylelds the maxlmum cross correlatlon value between the sets. Thls method 18 descrlbed ln "MRI-P~T Reglstratlon wlth Automated 14a 21777~
Algorlthm", by Woods R.P. et al., Journal of Computer Asslsted T~ yL~plly, volume 17, 1993.

4. 8patlal parameters matchlng. The two data sets are correlated by matchlng spatlal parameters such as the moments of the data sets. The moments can be matched by flnding the prlnclple axls for whlch they attaln thelr mlnlmal value. Thls method 18 descrlbed ln "The prlnclple Axes Transformatlon - a Method for Image Reglstratlon", by Alpert N.M., et al., Journal of Nuclear Medlclne, volume 31, 1990.

5. Invarlant geodesic llnes and polnts matchlng.
The data sets are analyzed uslng a dlfferentlal analysls of thelr surfaces discrete representatlon, yleldlng llnes and polnts whlch correspond to local maxlma and/or mlnlma of surface curvature. A global afflne transformatlon 18 then found that dellvers the best matchlng of the coLL~ n~lng llnes and polnts from the two data sets. Thls method 18 descrlbed ln "The External Mesh and the Understandlng of 3D
6urfaces", research report number 1901 from 14b ~ 2~ 77~
the Institute National de Recherche en Informatique et en Automatique ( INRIA), May 1993, and "New Feature Points Based on Geometrical Invariants for 3D Image Registration", research report number 2149 from the INRIA, both by Jean-Phillipe Thirion.
In an additional preferred embodiment of the invention, f iduciary marks may be added to the STET image by f irst adding f; d~Ct A~y marks to a structural image that is registered to the STET image, and then transforming those marks to the STET image.
Additionally, these marks may be added from a template once the tranxroL.Ilal,lon is known. Fig. 5 shows a STET image with f;~t~ y marks thereon.
In a further preferred embodiment of the invention, image acquisition is gated to body rhythms and motions. Preferably, the xtructural images are also synchronized in the same manner.
For example, gated CT images are used as 3tructural images instead of regular CT images when the STET images are gated. An advantage of c~mhtntns STET imaging with gating is the ability to correct binned data for patient motion during data acquisition by realignment based on registration of the images. This corrects for smearing otherwise produced by patient motion and enables the use of longer acquisition times. Additionally, data from separate bins is more easily combined.
Another advantage is the ability to correct organ motion caused by the gated rhythm, by applying a geometric transformation to data acquired based on the phase of the gated rhythm. Yet another advantage is the ability to register transmission images to emisgion images even when they are not acquired simultaneously . A tr~nc m~ Sst on image of a patient which is gated to body rhythms can be automatically registered to it corresponding gated emisslon image, since most of the misalignment between the two images is caused by body rhythms which are, in general, repetitive. ~
Fig. 6 indicates in simplified block diagram form a STET
system 21 equipped to accomplish either cardiac or respiratory gating or both . System 21 generally comprises a detector 22 f or detecting radiation. The radiation can be emanating from a 995 ~:15 ~ 1~
"~ ~ 2177~8~
patient 23 or from a raaiation source 24, typically comprising a radioisotope material. When sourca 24 is a radioisotope, detector 22 is preferably an Anger type camera.
The output of detector 22 is processed by a signal processor 26 . Processor 26 detPrmi n~c the location and energy of photons striking detectors 22.
The output of signal processor 26 is further processed by image procassor Z7 to provide lmage data using a memory 28. The processed images are shown on display 29.
Gating controls are provided for system 21. More particularly, respiratory gating uses a position sensor 31 which senses the thorax position of patient 23 during the STET process.
The sensed ~; cpl ~ t is operated on to provide windows or bins using a displacment detector 32. A position gate signal unit 33 provides gating signals to signal processor 26 based on the thorax position determined by detector 32. The cardiac gating system senses the heart beat with a sensor 3 6 . The R-wave is detected by a wave detector 37. A cardiac gating signal is provided to signal processor 26 by a wage gate slgnal unit 38 responsive to detection of the R-wave by detector 37 . U. S . Patent 4, 617, 938, the disclosure of which is incorporated herein by ref erence, describes a gating system .
STET system 21 is shown to be under the control of a controller 41 which supplies the appropriate control and timing signals .
The present invention was described in the context of nuclear medicine imaging ~ However~ the present invention is applicable to other types of imaging systems, provided that functional images (as described herein) have structural images that are registered to them where needed. Additionally, structural images of modalities other than X-~ay CT, MRI, ultra sound and SPECT can be registered to nuclear ~fif -~ n~ images by u1~ n~ the present invention.
It will ~e appreciated by persons skilled in the art that the present invention is not limited by what has been particularly shown and described herein. Rather, the scope of the 5' 1~5 J:15 p.t. 2~5~TI:X
~ 2~777~4 present invention is defined only by the claims which ~ollow:

Claims

1. A method of registering a plurality of functional images comprising:
providing a plurality of functional images;
providing a plurality of structural images each of which has a known positional relationship to at least one of said plurality of functional images;
finding a first mapping transformation between pairs of structural images; and determining a second mapping transformation between pairs of functional images based on said first mapping transformation and said positional transformation.

2. A method of registering a functional image to a structural diagnostic image comprising:
providing a functional image;
providing a first structural image;
providing a second structural image having a known positional relationship to said functional image;
finding a first mapping transformation between the two structural images; and determining a second mapping transformation between the functional image and the first structural image, based on said first mapping transformation and on said known positional relationship.

3. A method according to claim 2, wherein said first functional image is a STET image.

4. A method according to claim 3, wherein said second structural image is a SPTCT image.

5. A method according to claim 3, wherein said second structural image is a SPECT image.

6. A method according to claim 4, wherein the first structural image is an X-ray CT image.

7. A method according to claim 4, wherein the first structural image is an MRI image.

8. A method according to claim 4, wherein the first structural image is an ultrasound image.

9. A method according to claim 4, wherein said transformation between structural images includes a warping transformation.

10. A method according to claim 4, wherein said functional and structural images are provided as sets of two-dimensional slices and further comprising finding corresponding slices by matching slices between sets.

11. A method according to claim 10, wherein finding corresponding slices comprises manually matching slices.

12. A method according to claim 10, wherein finding corresponding slices comprises correlating slices.

13. A method according to claim 10, further comprising determining a new set of slices for said first functional image to improve correspondence between slices of said functional and structural images.

14. A method according to claim 10, wherein finding a first mapping transformation comprises correlating the two structural images.

15. A method according to claim 10, wherein finding a first mapping transformation comprises:
finding prominent structural details in the images, and matching the details between the images.

16. A method according to claim 15, wherein matching is done manually.

17. A method according to claim 15, wherein matching is done by correlation.

18. A method according to claim 10, comprising displaying emphasized features from one registered image on a second registered image.

19. A method according to claim 10, comprising displaying the difference between two registered images.

20. A method according to claim 10, comprising sequentially displaying a series of images.

21. A method according to claim 10, comprising displaying overlaid registered images.

22. A method for adding fiduciary markings to a functional image comprising:
providing a functional image;
providing a structural image having a known mapping transformation to said functional image;
determining reference positions on said structural image;
and marking the functional image at points associated with the reference positions using said known mapping transformation.

23. A method according to claim 22, wherein marking is done with fiduciary marks provided from a template.

24. A method according to claim 22, comprising matching to a different image using fiduciary markings that are registered to the functional image.

25. A method according to claim 10, comprising displaying at least one of the images as a three-dimensional image.

26. A method according to claim 9, wherein said known positional relationship includes a warping transformation.

27. A method according to claim 4, wherein said STET image is of a patient and wherein said STET image is gated to at least one of said patient's body rhythms.

28. A method according to claim 27, wherein said body rhythm is the cardiac rhythm.

29. A method according to claim 27, wherein said body rhythm is the respiratory rhythm.

30. A method according to claim 27, wherein said gating comprises binning.

31. A method according to claim 4, further comprising providing a second functional image, wherein said second functional image has a known positional transformation to said first structural image and wherein said functional images are binned.

32. A method according to claim 27, further comprising providing a second functional image, wherein said second functional image has a known positional transformation to said first structural image and wherein said functional images are acquired in different phases of said rhythm.

33. A method according to claim 32, further comprising combining said functional images into a third functional image.

34. A method according to claim 27, wherein said gating comprises windowing.

35. A method according to claim 1, wherein said plurality of functional images are acquired in sequence from a single patient and wherein said plurality of structural images are acquired in sequence from said patient during the same time period and wherein said plurality functional images is greater than said plurality of structural images.

36. A method according to claim 35, further comprising correcting motion distortion in said plurality of functional images by applying said second mapping transformation to some of said plurality of functional images.

37. A method of correcting motion smear in a plurality of binned functional images comprising:
acquiring binned data for a set of functional images during a plurality of sequential imaging periods;
acquiring a plurality of structural images each of which has a known positional relationship to the binned data acquired during at least one of said imaging periods;
finding a first mapping transformation between pairs of structural images;
determining a second mapping transformation between data acquired during two separate imaging periods based on said first mapping transformation and said positional transformation; and reconstructing said set of functional images from said binned data, wherein said second mapping transformation is applied to said binned data.

38. A method according to claim 37, wherein said plurality of imaging sequences is greater than said plurality of structural images.

39. A method according to claim 38, wherein said structural images are acquired in the same time period as said binned data.

40. A method of acquiring an absorption corrected image comprising:
acquiring a transmission image with gating;
acquiring an emission image with gating, wherein said transmission image and said emission image are gated so the same rhythm and wherein said transmission image and said emission image are acquired at the same phase of said rhythm; and creating an absorption corrected image by correcting said emission image with said transmission image.

41. A method according to claim 40, wherein said transmission image is a SPTCT image.

42. A method according to claim 41, wherein said emission image is a SPECT image.

43. A method of acquiring a gated absorption corrected image comprising:
acquiring an emission image of a region of a patient's body;
simultaneously acquiring a transmission image of said region;
measuring a body rhythm which affects said emission or said transmission images; and creating a gated absorption corrected image by correcting said emission image with said transmission image and adding measurements of said body rhythm.

44. A method according to claim 43, wherein said transmission image is a SPTCT image.

45. A method according to claim 44, wherein said emission image is a SPECT image.

46. A method according to claim 45, further comprising applying a geometric transformation to said acquired images responsive to said measured body rhythm.