Search

EP-3746984-B1 - SCATTER CORRECTION FOR POSITRON EMISSION TOMOGRAPHY (PET)

EP3746984B1EP 3746984 B1EP3746984 B1EP 3746984B1EP-3746984-B1

Inventors

  • YE, JINGHAN
  • SONG, XIYUN
  • BAI, CHUANYONG
  • ANDREYEV, ANDRIY
  • TUNG, CHI-HUA
  • HU, ZHIQIANG

Dates

Publication Date
20260506
Application Date
20190124

Claims (15)

  1. A non-transitory computer-readable medium storing instructions readable and executable by a workstation (18) including at least one electronic processor (20) to perform an image reconstruction method (100), the method comprising: generating (102, 404, 504, 604), from received imaging data, a plurality of intermediate images reconstructed without scatter correction from data partitioned into different energy windows; generating (104, 506, 608) a fraction of true counts or events and a fraction of scatter counts or events in the generated intermediate images; generating (106, 406) a scatter estimate image from the intermediate images, the fraction of true counts or events, and the fraction of scatter counts or events in the intermediate images; forward-projecting (108, 408) the scatter estimate image to generate a scatter projections estimate; reconstructing (110) a final image from the imaging data with scatter correction using the scatter projections estimate; and at least one of controlling (112, 412) the non-transitory computer readable medium to store the final image and control a display device (24) to display the final image.
  2. The non-transitory computer-readable medium of claim 1, wherein the final image is reconstructed from all imaging data with scatter correction using the scatter projections estimate.
  3. The non-transitory computer-readable medium of claim 1, wherein the final image is reconstructed from the counts of the imaging data with energies above an energy window threshold with scatter correction using the scatter projections estimate for the imaging data.
  4. The non-transitory computer-readable medium of claim 1, wherein the plurality of intermediate images reconstructed without scatter correction comprises a high energy window image and a low energy window image, and where the method further includes: iteratively compensating the high energy window image for scatter using the scatter estimate image and repeating the generating of the scatter estimate image.
  5. The non-transitory computer-readable medium of claim 4, wherein the iteratively compensating is repeated until a stopping criterion is met.
  6. The non-transitory computer-readable medium of claim 1, wherein the generating of the fractions of true counts and fractions of scatter counts includes: simulating (606) acquisition of the received imaging data from an estimated true image with an energy distribution having a known contribution of true and scatter counts.
  7. The non-transitory computer-readable medium of any one of claims 1-6, wherein generating the final reconstructed image further includes: subtracting the scatter estimate image from the intermediate images to generate the final image.
  8. The non-transitory computer-readable medium of any one of claims 1-7, wherein the received imaging data comprises positron emission tomography (PET) imaging data; and the intermediate images include a high energy window (HEW) for an intermediate image reconstructed without scatter correction from counts of the PET imaging data with energies above the HEW threshold and a low energy window (LEW), for an intermediate image reconstructed without scatter correction from counts of the PET imaging data with energies below the LEW threshold.
  9. The non-transitory computer-readable medium of claim 8, wherein the method further includes: separating the PET imaging data that is less than 511 keV into the low energy window image and imaging data that is greater than 511 keV into the high energy window image.
  10. The non-transitory computer-readable medium of any one of claims 1-7, wherein the received imaging data comprises single positron emission computed tomography (SPECT) imaging data; and the intermediate images include a high energy window (HEW) for an intermediate image reconstructed without scatter correction from counts of the SPECT imaging data with energies above a HEW threshold and a low energy window (LEW) for an intermediate image reconstructed without scatter correction from counts of the SPECT imaging data with energies below a LEW threshold.
  11. The non-transitory computer-readable medium of claim 10, wherein the method further includes: separating the SPECT imaging data that is less than a true energy value of the SPECT radiopharmaceutical isotope administered to the patient prior to the imaging session into the LEW image and imaging data that is greater than the true energy value into the HEW image.
  12. The non-transitory computer-readable medium of claim 1, wherein the generating of the fractions of true counts and the fractions of scatter counts includes: fitting the PET imaging data to a one-dimensional true counts energy distribution and a one-dimensional scatter counts distribution; and generating the fractions of true counts and the fractions of scatter counts based on the fitted one-dimensional true counts and scatter counts distributions.
  13. The non-transitory computer-readable medium of claim 12, wherein the one-dimensional true counts energy distribution comprises a Gaussian distribution and the one-dimensional scatter counts distribution comprises a polynomial distribution.
  14. The non-transitory computer-readable medium of either one of claims 12 and 13, wherein the scatter estimate image is given by s = s H + s L = r S + 1 r T − r S r T y L − y H where s represents the scatter estimate image, s H represents scatter contribution events from the high energy window; s L represents scatter contribution events from the low energy window; r T represents the ratio of true events in the high energy window image versus the low energy window image; r S represents the ratio of scatter events in the high energy window image versus the low energy window image; y H represents the high energy window image, and y L represents the low energy window image.
  15. The non-transitory computer-readable medium of claim 1, the method comprising: generating an energy histogram of the imaging data; fitting the energy histogram to a one-dimensional true counts energy distribution and a one-dimensional scatter counts energy distribution; and wherein, generating the fraction of true counts and the fraction of scatter counts is based on the fitted one-dimensional true counts and the fitted one-dimensional scatter counts distributions.

Description

FIELD The following relates generally to the medical imaging arts, medical image interpretation arts, image reconstruction arts, and related arts. BACKGROUND In positron emission tomography (PET) imaging, scattered coincidences sometimes contribute a significant proportion to the total detected coincidences. The scatter coincidences should be corrected for during PET reconstruction. Currently, a single-scatter simulation (SSS) method is widely used for the purpose of scatter correction in PET. The method simulates the scatter distribution by calculating the probability that the pair of coincidence gammas undergo a single scattering event before being detected. SSS alone does not determine the relative amount of scatter contribution to the total detected coincidences. The relative amount of scatter contribution is typically determined by a tail-fitting method or a Monte Carlo simulation based method. When the tail-fitting based method is used for scatter estimation, the resulting image may have quite significant artifacts sometimes, such as in the cases if the patient encountered motion between computed tomography (CT) and PET scans, or if the patient is large so that the tail part has high noise or truncation, or random correction is not accurate for the tail part so that the scatter estimate accuracy is compromised. With the Monte Carlo simulation scatter amount estimate, although artifacts due to noise in the tail part are no longer an issue, scatter contribution from out-of-axial-field of view (FOV) activities cannot be accurately estimated for single-bed acquisition or end bed frames since the out-of-axial FOV activity is unknown. In document "Scatter correction for 3D PET using beam stoppers combined with dual-energy window acquisition: a feasibility study; Scatter correction using beam stoppers" by Jay Wu et al., PHYSICS IN MEDICINE AND BIOLOGY, INSTITUTE OF PHYSICS PUBLISHING, BRISTOL GB, vol. 50, no. 19, 7 October 2005, pages 4593-4607, XP020084360, ISSN:0031-9155, DOI:10.1088/0031-9155/50/19/012, the authors propose a scatter correction method for 3D PET based on beam stoppers and dual-energy window acquisition. The beam stoppers are placed surrounding the object to attenuate primary beams. The scatter fractions are directly estimated at those blocked lines of response and then the entire scatter fraction distribution is recovered using the dual-energy window ratio as reference. N C FERREIRA ET AL: "A hybrid scatter correction for 3D PET based on an estimation of the distribution of unscattered coincidences: implementation on the ECAT EXACT HR $plus$",PHYSICS IN MEDICINE AND BIOLOGY, vol. 47, no. 9, 7 May 2002 (2002-05-07), pages 1555-1571; LARS-ERIC ADAM* ET AL: "Energy-Based Scatter Correction for 3-D PET Scanners Using NaI(Tl) Detectors",IEEE TRANSACTIONS ON MEDICAL IMAGING, IEEE SERVICE CENTER, PISCATAWAY, NJ, US, vol. 19, no. 5, 1 May 2000; CHUANG K S ET AL: "Novel scatter correction for three-dimensional positron emission tomography by use of a beam stopper device",NUCLEAR INSTRUMENTS & METHODS IN PHYSICS RESEARCH. SECTION A, ELSEVIER BV * NORTH-HOLLAND, NL, vol. 551, no. 2-3, 11 October 2005 (2005-10-11), pages 540-552; and US 2005/072929 A1 disclose as well some similar systems. The following discloses new and improved systems and methods to overcome these problems. SUMMARY In one disclosed aspect, a non-transitory computer-readable medium stores instructions readable and executable by a workstation including at least one electronic processor to perform an image reconstruction method. The method includes: generating, from received imaging data, a plurality of intermediate images reconstructed without scatter correction from data partitioned into different energy windows; generating a fraction of true counts or events and a fraction of scatter counts or events in the generated intermediate images; generating a scatter estimate image from the intermediate images, the fraction of true counts or events, and the fraction of scatter counts or events in the intermediate images; forward-projecting the scatter estimate image to generate a scatter projections estimate; reconstructing a final image from the imaging data with scatter correction using the scatter projections estimate; and at least one of controlling the non-transitory computer readable medium to store the final image and control a display device to display the final image. In another disclosed aspect, a non-transitory computer-readable medium storing instructions readable and executable by a workstation including at least one electronic processor to perform an image reconstruction method. The method includes: generating, from received imaging data, a plurality of intermediate images reconstructed without scatter correction from data partitioned into different energy windows; generating a fraction of true counts and a fraction of scatter counts in the generated intermediate images by operations including: generating an energy histogram o