Pretreatment metabolic tumor volume (MTV) as a predictor of survival in borderline resectable pancreatic cancers treated with neoadjuvant therapy.

Abstract

310 Background: A higher initial metabolic tumor burden is associated with lower median survival in locally advanced pancreatic cancer (LAPC), yet the prognostic utility of PET/CT is not defined in the setting of borderline resectable pancreatic cancer (BRPC). Methods: We performed a retrospectivereview of our institutional experience treating BRPC. Initial staging included endoscopic ultrasound as well as pancreatic protocol CT and PET/CT scans. All patients underwent neoadjuvant gemcitabine-based chemotherapy and radiation therapy (RT). RT was delivered using standard fractionation intensity modulated radiation therapy (IMRT) or stereotactic body radiation therapy (SBRT). Restaging CT and PET/CT scans were obtained approximately 4 weeks following RT completion. We measured the significance of the pre-treatment and post-treatment SUV maximum and metabolic tumor volume (MTV) 2.5 to 5.0, which was defined as the MTV above a threshold SUV of 2.5, 3.0, 4.0 and 5.0. Cox regression models were used to evaluate the significance between these parameters and disease free survival (DFS) and overall survival (OS). Results: We evaluated a total of 72 BRPC patients. Median follow up was 12.7 months. 56 patients (77%) received induction chemotherapy with gemcitabine, docetaxel and capecitabine (GTX). 43 (59.7 %) underwent surgical resection. Significant predictors for OS in the whole cohort included pre-treatment SUV maximum (p=0.0042), post-treatment SUV maximum (p=0.0183), pre-treatment MTV 2.5 (p=0.0016) and pre-treatment MTV 4.0 (p=0.0111). In addition, the difference between the MTV 4.0 pre-treatment and post-treatment was significant (p=0.0285). In patients who underwent surgical resection, there was a significant correlation between OS with pre-treatment SUV max (p=0.0229) and post-treatment SUV maximum (p=0.0325) but not pre-treatment MTV 2.5 (p=0.0654) nor MTV 4.0 (p=0.0928) nor the differences between each variable pre (0.1482) or post-treatment (0.0959). Conclusions: This is the first study to suggest that pre and post treatment PET activity is prognostic for BRPC.