PT  - JOURNAL ARTICLE
AU  - Manoj Narayanan
AU  - Alexander Fischer
AU  - Timo Paulus
AU  - Andrew Shields
AU  - Ann Wilson
AU  - Joseph Rajendran
TI  - Impact of partial volume correction on FDG treatment response measures in head and neck and advanced ovarian cancer
DP  - 2011 May 01
TA  - Journal of Nuclear Medicine
PG  - 2101--2101
VI  - 52
IP  - supplement 1
4099  - http://jnm.snmjournals.org/content/52/supplement_1/2101.short
4100  - http://jnm.snmjournals.org/content/52/supplement_1/2101.full
SO  - J Nucl Med2011 May 01; 52
AB  - 2101 Objectives Partial volume (PV) effects can introduce quantitative biases in PET uptake values. This is important in FDG-based treatment response monitoring where a reduction in tumor size may falsely increase the magnitude of a decrease in metabolic activity. We evaluated the impact of 2 methods of partial volume correction (PVC) in serial FDG scans for patients with head and neck cancer (HNC) and advanced ovarian cancer (AOC). Methods FDG PET scans were acquired before and after 3 weeks of chemoradiation on 3 HNC patients, and before and after the third cycle of neoadjuvant chemotherapy on 3 AOC patients. Tumor ROIs based on a region-growing method with a lower threshold of 80% of maximum SUV were generated for each sequential scan. Using a Philips IMALYTICS research workstation, two PV correction methods—Richardson-Lucy deconvolution (RL) and blind deconvolution (BD)—were used to correct mean SUV values within the ROI. Relative percentage (%) changes in mean SUV between serial scans were calculated. Results Compared to no PVC, PVC increased mean SUV in pre-tx AOC, post-tx AOC, pre-tx HNC, and post-tx HNC scans by 6.9%, 11.8%, 13.9%, and 13%, respectively, but in these scans RL and BD differed minimally from each other (1%, 0%, 0%, and 0%, respectively). For 2 AOC patients showing partial response, mean SUV reduction was 53.9, 54.5 and 54% (no PVC, RL, and BD, respectively). For an AOC patient with stable disease, no PVC resulted in a much bigger change (10.5%) compared to PVC (RL=1.9%, BD=1.7%). For 3 HNC patients with complete or partial response, mean SUV reduction was 38.6, 40.6 and 40.7% (no PVC, RL, and BD). Conclusions Compared to no PVC, PVC resulted in higher mean SUV values for all FDG scans, regardless of the correction algorithm, and showed the greatest difference in a patient with stable disease, indicating a potential role. Additional studies are warranted to better understand the impact of PV effects in this setting. Research Support NIH 1R21CA135830, Philips Healthcar