18F-FDG PET as a predictor of resectability and clinical outcomes in locally advanced pancreatic cancer patients treated with radiotherapy.

Abstract

378 Background: Surgical resection is the preferred approach for localized pancreatic adenocarcinoma, yet likelihood of margin positive resection precludes its use in patients with locally advanced (LAPA) disease. Whereas CT imaging is the standard for defining resectability, studies suggest reduced accuracy following radiotherapy (RT). Given the potential clinical implications, improved methods to define resectability are needed. We investigate whether PET and other clinical factors can be used to predict likelihood of margin negative resection in LAPA patients. Methods: We reviewed LAPA patients treated with RT from 2011-2013 who also underwent pre- and 6-week post-RT PET/CT. Mann-Whitney test and Cox modeling were used to identify predictors of margin negative resection and progression-free survival (PFS), respectively. Variables assessed included: pre- and post-RT SUV max, mean, coefficient of variation (CV), total lesion glycolysis, CA19-9 and tumor size. Results: Eighteen LAPA patients treated with RT were identified. Median age was 66 years. Seven were considered resectable following RT. 6 patients underwent margin negative resection, while the 7thpatient refused surgery. An additional 2 patients were technically resectable, but developed metastatic disease. Mean value for post-RT SUV max was 3.5 in resected patients versus 4.9 in unresected patients. This difference was borderline significant (p=.08) for predicting resectability. Similarly, pre- to post-RT change in SUV max was -4.37 in resected patients and -1.26 in unresected patients, which was also not significant (p=.17). Median PFS was 13 months for resected patients and 8.5 months for unresected patients (p=.06). Predictors of PFS included post-RT SUV CV (p=.05) and pre-RT CA19-9 (p=.03). Conclusions: Mean post-RT SUV max was borderline significant for predicting which LAPA patients may undergo successful margin negative resection. Post-RT metabolic response was also predictive of improved PFS, further validating PET/CT as an important prognostic imaging tool. Studies are ongoing to further validate these predictors of resectability given the potential clinical significance.