MENU

Click here to

×

Are you sure ?

Yes, do it No, cancel

Impact of Lesion Locations in PET Scan Field On Image Quality

M Park1,2*, E Sassaroli1 , R Reddy1,2 , (1) Brigham and Women's Hospital, Boston, MA, (2) Harvard Medical School, Boston, MA

Presentations

(Tuesday, 7/31/2018) 10:00 AM - 10:30 AM

Room: Exhibit Hall | Forum 8

Purpose: To estimate the variability of image quality and activity quantitation on PET images when a lesion is located anywhere within the scan field.

Methods: A NEMA image quality phantom was filled with F-18 with 2:1 sphere-to-background activity ratio to simulate a tumor imaging. We scanned the phantom on a PET/CT system using SiPM and LYSO crystals with a 20-cm axial length. The center of the spheres in the phantom was positioned at 15 different locations and scanned in 3D mode. The scan duration was 4min for the first scan and longer for subsequent scans to compensate for decay; therefore, the total coincidence counts were similar in all 15 PET data. All images were reconstructed using 6 different reconstruction algorithms with combinations of ordered-subsets expectation maximization (OSEM), Bayesian penalized likelihood (BPL), point spread function (PSF) modeling, and time-of-flight (TOF). Reconstruction parameters were adjusted to achieve an equal noise level in the background. Contrast-to-noise ratio (CNR) for each sphere was calculated as a measure of image quality. Relative variability of CNR was also estimated using the coefficient of variation (COV) along transaxial (y) and axial (z) positions.

Results: CNR increases with sphere size in all reconstruction methods and was lowest for spheres near the end of the field and highest at the center. CNR variability was smaller in y locations than in z locations. The COV was 13% (y) and 28% (z) with OSEM, and 11% (y) and 22% (z) for OSEM+TOF reconstruction. Advanced reconstruction algorithms such as TOF, PSF, and BPL offer better image quality compared with OSEM, especially for smaller spheres.

Conclusion: CNR depends not only on sphere size but also on sphere location and reconstruction algorithm. Results demonstrate the importance of patient positioning and choice of reconstruction in tumor imaging.

Keywords

Not Applicable / None Entered.

Taxonomy

Not Applicable / None Entered.

Contact Email