A dosimetric model of duodenal toxicity after stereotactic body radiotherapy for pancreatic cancer.

INTRODUCTION: Dose escalation for pancreas cancer is limited by the tolerance of adjacent normal tissues, especially with stereotactic body radiotherapy (SBRT). The duodenum is generally considered to be the organ at greatest risk. This study reports on the dosimetric determinants of duodenal toxicity with single-fraction SBRT. METHODS AND MATERIALS: Seventy-three patients with locally advanced unresectable pancreatic adenocarcinoma received 25 Gy in a single fraction. Dose-volume histogram (DVH) endpoints evaluated include V(5) (volume of duodenum that received 5 Gy), V(10), V(15), V(20), V(25), and D(max) (maximum dose to 1 cm(3)). Normal tissue complication probability (NTCP) was evaluated with a Lyman model. Univariate and multivariate analyses were conducted with Kaplan-Meier and Cox regression models.
RESULTS: The median time to Grade 2-4 duodenal toxicity was 6.3 months (range, 1.6-11.8 months). The 6- and 12-month actuarial rates of toxicity were 11% and 29%, respectively. V(10)-V(25) and D(max) all correlated significantly with duodenal toxicity (p<0.05). In particular, V(15)≥9.1 cm(3) and V(15)<9.1 cm(3) yielded duodenal toxicity rates of 52% and 11%, respectively (p=0.002); V(20)≥3.3 cm(3) and V(20)<3.3 cm(3) gave toxicity rates of 52% and 11%, respectively (p=0.002); and D(max)≥23 Gy and D(max)<23 Gy gave toxicity rates of 49% and 12%, respectively (p=0.004). Lyman NTCP model optimization generated the coefficients m=0.23, n=0.12, and TD(50)=24.6 Gy. Only the Lyman NTCP model remained significant in multivariate analysis (p=0.001).
CONCLUSIONS: Multiple DVH endpoints and a Lyman NTCP model are strongly predictive of duodenal toxicity after SBRT for pancreatic cancer. These dose constraints will be valuable in future abdominal SBRT studies.