Dual isotope scintigraphic image enhancement

• US 4,846,187 A
• Filed: 12/01/1987
• Issued: 07/11/1989
• Est. Priority Date: 12/01/1987
• Status: Expired due to Fees

• Current Assignee: Center For Molecular Medicine and Immunology
• Original Assignee: Center For Molecular Medicine and Immunology
• Inventors: Siegel, Jeffry A.
• Primary Examiner(s): JAWORSKI, FRANCIS J
• Application Number:
  US07/127,230
• Time in Patent Office: 588 Days
• Field of Search: 128/1.1, 128/659, 424/1.1, 424/9, 436/804
• US Class Current: 600/436
• CPC Class Codes: G01T 1/1647 Processing of scintigraphic data not related to a particular imaging system G01T1/2992
• CPC Class Codes: G06T 5/50 using two or more images, e.g. averaging or subtraction

Abstract

An improved method of dual isotope subtraction scintigraphic image processing uses a band pass spatial frequency filter to smooth and enhance the image obtained with the less well resolved isotope and then applies the same filter function to the image from the better resolved isotope, after which the images are subtracted to produce an enhanced image with reduced edge artifacts and better resolution. The method can be applied to images where the target/background ratio approaches unity, and to single isotope imaging.

Claims

1. In a dual isotope subtraction method of imaging a tumor, lesion, organ or tissue, wherein an amount sufficient for imaging of a radiolabeled antibody or antibody fragment which specifically binds a marker produced by or associated with said tumor, lesion, organ or tissue is injected into a patient having such tumor, lesion, organ or tissue, an amount sufficient for background subtraction of a reference substance labeled with a different and separately detectable radioisotope is injected into the patient prior to scanning, the patient is scanned with a scintigraphic camera equipped with means to separately acquire emission data in energy windows corresponding to radioactive emissions from the two different radioisotope labels, as a function of the scanning position of the camera, and to store the data in digital form, said scanning being effected after a time sufficient to localize said radiolabeled antibody or antibody fragment, and the scans from the two energy windows are processed and subtracted such that the san data from the reference label are subtracted from the scan data from the specific antibody or antibody fragment label, to produce background-compensated image data for the tumor, lesion, organ or tissue, suitable for display in a form permitting visualization of the image,

the improvement which comprises:

(a) processing the raw digital data from the scans in the two energy windows, in the spatial frequency domain, using a band pass spatial frequency filter;

wherein the raw scan data for each of the radioisotopes, after normalization to the same number of total counts per image and transformation into the spatial frequency domain, are multiplied by the same filter function, determined for the less-well-resolved radioisotope (LWRR);

wherein the filter function is a two-dimensional circularly symmetric Gaussian filter function, the low frequency portion of which is substantially equal to the inverse of the modulation transfer function (MTF) for the scanning camera and collimator system, the MTF is obtained from the line spread function of a line source of the LWRR, the high frequency cutoff portion of the filter is a smoothly decreasing function going to zero over a short but finite range of frequencies, a twodimensional power spectrum of the spatial frequency domain-transformed scan data from the LWRR is formed, a one-dimensional power spectrum is formed from the two-dimensional power spectrum by averaging over annuli, the noise level is determined by averaging over the high frequency components, and the rolloff frequency of the filter is the point on the one-dimensional power spectrum which is about 1.5-3 times the noise level; and

wherein the processed data are transformed back to the spatial domain; and

(b) subtracting the processed data for the reference radioisotope from those of the specific antibody or antibody fragment label, to produce enhanced background-compensated image data suitable for display and visualization.

2. The method of claim 1, wherein the specific antibody or antibody fragment is labeled with Iodine-131, and the reference substance is labeled with Technetium-99m.

3. The method of claim 1, wherein an antibody fragment is used for imaging, said fragment being a F(ab'"'"')₂, Fab'"'"' or Fab fragment.

4. The method of claim 3, wherein said fragment is a Fab'"'"' or Fab fragment.

5. The method of claim 1, wherein said reference substance is the normal immunoglobulin or immunoglobulin fragment corresponding to the specific antibody or antibody fragment.

6. The method of claim 1, wherein the image is a two-dimensional planar image.

7. The method of claim 1, wherein the target imaged is a tumor.

8. The method of claim 1, wherein the target/background ratio of raw counts in the image is less than about 2.

9. The method of claim 8, wherein the target/background ratio of raw counts in the image is about 1.6.

10. The method of claim 8, wherein the target/background ratio of raw counts in the image is about 1.2.

11. The method of claim 1, wherein the rolloff frequency of the filter is the point on the one-dimensional power spectrum which is about twice the noise level.

12. In a method of imaging a tumor, lesion, organ or tissue, wherein an amount sufficient for imaging of a radiolabeled antibody or antibody fragment which specifically binds a marker produced by or associated with said tumor, lesion, organ or tissue is injected into a patient having such tumor, lesion, organ or tissue, the patient is scanned with a scintigraphic camera equipped with means to acquire emission data corresponding to radioactive emissions from the radioisotope label, as a function of the scanning position of the camera, and to store the data in digital form, said scanning being effected after a time sufficient to localize said radiolabeled antibody or antibody fragment, to produce image data for the tumor, lesion, organ or tissue, suitable for display in a form permitting visualization of the image.

the improvement which comprises:

(a) effecting said scan after a time when the ratio of (i) the radioactivity emitted by radioisotope localized at the site of said tumor, lesion, organ or tissue, to (ii) the background radioactivity emitted by non-localized radiolabel, is less than about 2; and

(b) processing the raw digital data from the scan in the spatial frequency domain, using a band pass spatial frequency filter;

wherein the raw scan data are transformed into the spatial frequency domain and multiplied by a filter function;

wherein the filter function is a two-dimensional circularly symmetric Gaussian filter function, the low frequency portion of which is substantially equal to the inverse of the system modulation transfer function (MTF) for the scanning camera, the MTF is obtained from the line spread function of a line source of radioisotope, the high frequency cutoff portion of the filter is a smoothly decreasing function going to zero over a short but finite range of frequencies, a two-dimensional power spectrum of the spatial frequency domain-transformed scan data is formed, a one-dimensional power spectrum is formed from the two-dimensional power spectrum by averaging over annuli, the noise level is determined by averaging over the high frequency components, and the rolloff frequency of the filter is the point on the one-dimensional power spectrum which is about twice the noise level; and

wherein the processed data are transformed back to the spatial domain to produce image data suitable for display and visualization.

13. The method of claim 12, wherein an antibody fragment is used for imaging, said fragment being a F(ab'"'"')₂, Fab'"'"' or Fab fragment.

14. The method of claim 13, wherein said fragment is a Fab'"'"' or Fab fragment.

15. The method of claim 12, wherein the image is a two-dimensional planar image.

16. The method of claim 12, wherein the target imaged is a tumor.

17. The method of claim 12, wherein the target/background ratio of raw counts in the image is less than about 2.

18. The method of claim 17, wherein the target/background ratio of raw counts in the image is about 1.6.

19. The method of claim 17, wherein the target/background ratio of raw counts in the image is about 1.2.

20. The method of claim 12, wherein the rolloff frequency of the filter is the point on the one-dimensional power spectrum which is about twice the noise level.