Premium
TH‐A‐WAB‐04: Respiratory Sorted Imaging Using Region‐Of‐Interest Robotic Multi‐Pinhole SPECT System
Author(s) -
Yan S,
Bowsher J,
Yin F
Publication year - 2013
Publication title -
medical physics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.473
H-Index - 180
eISSN - 2473-4209
pISSN - 0094-2405
DOI - 10.1118/1.4815698
Subject(s) - imaging phantom , pinhole (optics) , single photon emission computed tomography , physics , spect imaging , iterative reconstruction , nuclear medicine , collimated light , region of interest , image quality , image resolution , computer science , computer vision , optics , medicine , laser , image (mathematics)
Purpose: Single Photon Emission Computed Tomography (SPECT) may provide important information on tumors in the thorax and abdomen. Respiratory sorted imaging using conventional full cross‐section methods is generally degraded by noise. The purpose is to investigate noise reduction by region‐of‐interest (ROI) SPECT. Methods: The 4D XCAT phantom was computer simulated with a 10mm diameter tumor added in the right lung. Respiratory motion was simulated for 6 and 10 phases, accomplished by simulating 30 frames in one 5sec respiratory cycle. A non‐sorted phantom was generated by summing the 30 phantoms to mimic motion blur. A static phantom with no respiratory motion was simulated using one of the 30 phantoms. The maximum tumor motion was 2cm Sup/Inf and 1.2cm AP. Projections were computer simulated for 4‐minute scans with a reference conventional parallel‐hole collimated SPECT and a 49‐pinhole SPECT that was optimized to image a 7cm diameter region‐of‐interest(ROI) encompassing the tumor and its motion. Noisy projections were then generated with random Poisson noise. The SPECT images were reconstructed by OSEM with up to 7 iterations. Images were evaluated as a function of phase by visual inspection and root mean squared error (RMSE). Results: The tumor was visualized well above background in most phases for the 49‐pinhole SPECT scan. Visual quality was not as good for the parallel‐hole scan. The 49‐pinhole SPECT resulted in smaller RMSEs than the broader cross‐section parallel‐hole detector. RMSE values were around 0.11 for the 49‐pinhole SPECT scan and 0.18 for the parallel‐hole SPECT scan. The RMSEs for 10 respiratory phases were similar to that of 6 phases. Conclusion: The higher sensitivity and resolution of ROI imaging can alleviate noise effects of respiratory sorted imaging, provided the motion range is within the ROI. The specific ROI SPECT system could enable this gated imaging on‐board radiation therapy machines. PHS/NIH/NCI grant R21‐CA156390‐01A1