Does metabolic tumor volume predict tumor regression grade after neoadjuvant therapy for borderline resectable pancreatic cancer?

Abstract

214 Background: Metabolic tumor volume (MTV) based on FDG-PET has been shown to be prognostic for overall survival (OS) in patients with resectable, borderline resectable (BRPC) and locally advanced pancreatic cancer (LAPC). Tumor regression grade (TRG) following neoadjuvant therapy has also been shown to correlate with OS. No data currently exists evaluating whether metabolic response defined by MTV can predict pathologic response in BRPC. Methods: This IRB approved retrospective review evaluated 40 non-metastatic BRPC patients. All patients underwent staging imaging including FDG PET/CT. They then received induction multi-agent chemotherapy followed by five fraction stereotactic body radiation therapy (SBRT) with a median maximum dose of 35 Gy. Repeat pre-operative FDG PET/CT was performed for re-staging. Patients then underwent pancreatectomy. For this study, the diagnostic images were transferred to a software system that analyzed the pre-treatment, re-staging, and change in (differential) MTV values. We analyzed SUVmax and each of MTV2.5, MTV3.0, MTV4.0, and MTV5.0, as defined as the tumor volume with SUV above each threshold. At our center, each surgical specimen is assigned TRG using the College of American Pathology (CAP) criteria. The TRG scores regression from 0 – 3, with 0 indicating complete response and 3 indicating no response. Univariate ordinal logistic regression was used to correlate TRG with pre-, post-, absolute differential, and percent differential MTV 2.5-5.0 and SUVmax. Results: Of the sample population, TRG was as follows: TRG 0 n = 4, TRG 1 n = 17, TRG 2 n = 14, TRG 3 n = 5. TRG correlated with rradiologist reported re-staging SUVmax (p = 0.001) and percent differential SUVmax (p = 0.044). TRG did not correlate with pre-treatment SUVmax, absolute differential SUVmax, or any MTV measurements (all p > 0.05). Conclusions: We identified a correlation between SUVmax and percent change in SUVmax after neoadjuvant therapy with TRG (CAP scale). This analysis did not identify an association between any MTV threshold and TRG before or after neoadjuvant therapy. Future work seeks to identify alternate methods of neoadjuvant response assessment.