Pilot study on the impact of F18-labeled thymidine PET/CT on gross tumor volume identification and definition for pancreatic cancer.

Abstract

Accurate target definition for radiation therapy planning in localized pancreatic cancer is critical, particularly when using strategies that omit elective coverage. Standard imaging modalities such as computed tomography (CT), magnetic resonance imaging, and endoscopic ultrasound have limited concordance with pathologic evaluation. Biologic imaging with [F18]-fluorodeoxyglucose (FDG)-positron emission tomography (PET) can also be difficult to interpret because increased activity is indicative of increased glucose metabolism, rather than cellular proliferation. [F18]-3'-deoxy-3'-fluorothymidine labeled thymidine (FLT) is a proliferative marker which exploits the expression of pyrimidine-metabolizing enzymes. We evaluate the impact of FLT-PET on pancreatic target definition for radiation planning. Patients with biopsy-proven, newly diagnosed, untreated pancreatic adenocarcinoma were enrolled on an institutional review board-approved prospective study. Patients were injected with FLT and scanned 20 to 30 minutes later. Two physicians (referred to as observer 1 and observer 2) independently contoured the gross tumor volume (GTV) and involved nodes on CT scan only and then again with the assistance of coregistered FLT-PET. Conformality index (CI), the ratio of the volumes of intersection and union, was used as the metric for volume comparison (where CI = 0 represents no overlap and CI = 1 represents perfect overlap). Nine patients were enrolled in this study. FLT-avidity was discerned in 8 of 9 patients. Average CT-GTV volume for observers 1 and 2 was 38.1 and 26.5 mL, respectively. Average FLT-GTV volume for observers 1 and 2 was 39.1 and 25.0 mL, respectively. For the 8 patients with FLT-avid tumors, addition of FLT data improved concordance of GTV definition between physicians in 6 of 8 tumors. Average CI for interobserver CT-GTV was 0.325. Addition of FLT-PET information improved the average CI to 0.400. FLT-PET improves interobserver concordance in GTV definition. Further studies will focus on verification of these findings, pathologic verification of the FLT-PET signal, and optimization of the FLT-PET signal threshold for autosegmentation.