Impact of Tumor Dosimetry on Clinical Outcomes after Stereotactic Body Radiation Therapy for Pancreatic Adenocarcinoma

Abstract

We hypothesized that escalated biological effective dose (BED) via stereotactic body radiation therapy (SBRT) would improve outcomes for patients with borderline resectable (BRPC) and locally advanced pancreatic cancer (LAPC).A retrospective review was performed on consecutive patients who received pancreas SBRT from June 2009 to June 2020, were non-metastatic at the time of SBRT, received neoadjuvant chemotherapy, and had available gross tumor volume (GTV) dosimetry. Variables of interest included: age, gender, tumor location (body/tail vs head/neck), chemotherapy type (5-fluorouracil-based vs gemcitabine-based), AJCC V8 stage, pre- and post-radiotherapy (RT) CA19-9, GTV D95, and GTV V55-40 Gy. GTV volume variables were analyzed as continuous (cont) and as binary (+) functions. All patients with BRPC at diagnosis were analyzed by a logistic regression model for variables that predicted a higher likelihood of undergoing surgical resection (Model A). A second model of all patients (BRPC and LAPC) unable to undergo surgery was constructed by Cox regression analysis investigating variables that predicted for improved overall survival (OS), progression free survival (PFS), time to loco-regional failure (TTLRF), and time to distant failure (TTDF; Model B). All clinical time points were calculated from the end of SBRT and data analysis was performed in SAS with backward elimination variable selection.A total of 365 patients had a median follow up from SBRT of 12.1 months and 251 patients (69%) were diagnosed with BRPC. Model A composed of 200 BRPC patients (56% resected) with 3 clinical variables: chemotherapy (P = .39), AJCC stage (OR 0.54, P = 0.01), post-RT CA19-9 (OR 0.27, P = < 0.001); and 3 dosimetric variables: GTV D95 (OR 3.81, P = 0.03), GTV V40+ (OR 2.52, P = 0.07), and GTV V40cont (OR 0.98, P = 0.03). Model B included 168 patients (51% BRPC) unable to undergo surgery. In this group, 5-fluorouracil-based chemotherapy was associated with improved OS (HR 0.51, P = < 0.001), PFS (HR 0.48, P = < 0.001), TTLRF (HR 0.47, P = 0.004), and TTDF (HR 0.54, P = 0.006). GTV V40+ was significantly associated with improved PFS (HR 0.59, P = 0.005) and TTLRF (HR 0.58, P = 0.036), but not with OS (HR 1.06, P = 0.78) or TTDF (HR 0.68, P = 0.113). Pre-RT CA19-9 was associated with worse OS (HR 1.43, P = 0.010) and TTLRF (HR 1.59, P = 0.022). The proportion of patients receiving 5-flurouracil-based chemo and achieving GTV V40+ was evenly distributed across the cohort (χ2 = 0.55, P = 0.45).In the largest series of pancreatic cancer SBRT to date, dose-escalated tumor coverage specifically to 40 Gy (BED10 = 72 Gy) increases the likelihood of surgical resection in patients with upfront BRPC and is associated with improved PFS that is driven by improvement in loco-regional disease control for unresected patients. These data may help to inform the evolving landscape on the use of neoadjuvant and definitive radiotherapy in pancreatic cancer.