Concordance among FDG PET, CT and bone scan in men with metastatic prostate cancer

Abstract

1576

Objectives: As an interim data analysis of an ongoing clinical imaging trial, we compared and correlated the imaging findings of FDG PET-CT and bone scan in men with metastatic prostate cancer.

Methods: We performed 41 FDG PET-CT and 19 bone scans in 11 (5 with androgen-sensitive disease and 6 with androgen refractory disease) men with metastatic prostate cancer. The abnormal findings on PET, CT, and bone scan were dichotomized as “seen” and “not seen’ and cross-tabulated combining all scan dates and all lesions for all patients. The serial scans in the same patient were treated as independent studies. Concordance among the imaging modalities was evaluated using the Spearman partial correlation coefficients that were calculated among mean SUV, max SUV and the lesion size on CT, adjusting for patient, lesion location (bone, lymph node, visceral) and scan date.

Results: Among the 140 bone lesions, 95 were seen on bone scan but not on PET, 116 were seen on bone scan but not on CT, and 5 were seen on both PET and CT but not on bone scan. There were 19% (14 of 72) of bone lesions, 2% (1 of 43) of lymph nodes, and 16% (4 of 25) of visceral lesions that were seen on PET but not on CT. The partial correlation coefficients for those lesions that were seen on both PET and CT was weak for bone (n=58, r=0.23, p=0.09), moderate for lymph nodes (n=42, r=0.66, p<0.001) and moderate for visceral lesions (n=15, r=0.60, p=0.031).

Conclusions: The concordance among FDG PET, CT, and bone scan is higher for lymph node and visceral lesions than for osseous lesions in metastatic prostate cancer. Further studies are currently underway on the effects of therapy and androgen responsiveness on diagnostic performance of FDG PET-CT and concordance with bone scintigraphy.

Research Support: Supported by NIH/NCI R01-CA111613-01.