Abstract
1615
Objectives: To propose standardized methods for quantifiably evaluating therapy response using PET metabolically active tumor volume (matV) and CT volume (ctV) as biomarkers in FDG PET/CT.
Methods: PET/CT data of 101 lesions before and after therapy were selected from routine scans (60-90mins post injecting 450–550MBq FDG on a Siemens HI-REZ. 3D VOIs were placed in the high, middle and low third of the aorta (1.5±0.5cm3) on CT, well ways from the edge, then were used to report SUV on corresponding PET. SUVmax was calculated for each lesion and setup to estimate the matV at cutoffs of (a) 50%, (b) SUV2.5 and (c) a correction after background subtraction (BG) separately. ctV was measured slice-by-slice for each lesion. Therapy response was compared between ctV and all matVs.
Results: Significant differences were found in matVs through (a), (b) and (c) (p<0.001);no pronounced variance was found for SUVs from three aorta regions and leaded matVs after BG. Comparing to ctV <2, 2-10 and >10 cm3, matVs from (a) and (b) yielded variances of 228±280%, -30±63% and -30±74% as well as 200±271%, 130±121% and 63±83% separately, while decreased to -3±85%, 37±71% and 12±60% if using (c). When lesion-background ratio less than 8, (a) and (b) resulted in overestimated matVs (180±261%) while (c) gave more accurate results (-14±70%). To monitor tumor response to therapy, matV from (c) held the same/similar trend as SUV response, while the trend was not true under (b) and much worse if using (a) without background subtraction.
Conclusions: The proposed cutoff approach (c) after background subtraction provided a standard way for more accurately contouring PET volumes and leaded to better interpretation of therapy response as SUV changes.
- Society of Nuclear Medicine, Inc.