Interfractional dose variations in the stomach and the bowels during breathhold intensity-modulated radiotherapy for pancreatic cancer: Implications for a dose-escalation strategy

Abstract

This study aims to evaluate the interfractional dose variations in the organs-at-risk (OARs) during pancreatic breathhold intensity-modulated radiotherapy (IMRT) and to assess the impacts of "planning organs-at-risk volume" (POV) structures generated by isotropically expanding the dose-limiting OARs, based on the comparison of the interfractional doses to the OARs between IMRT plans and conventional three-dimensional-conformal radiotherapy (3D-CRT) plans. Thirty repeat CT scans were acquired from ten consecutive patients who were receiving chemoradiotherapy for pancreatic cancer. Six IMRT plans for each patient with two levels of prescription (45 and 51 Gy in 15 fractions) and 3 POV margin sizes (5, 7, and 10 mm) were generated based on the initial CT scan under predetermined constraints. Two 3D-CRT plans (39 and 42 Gy in 15 fractions) were simultaneously generated. The dose distribution of all of the treatment plans was recalculated with the repeat CT scans. The interfractional dose variations in the three OARs (stomach, duodenum, and small intestine) were evaluated, and the absolute volumes ≥39 Gy (V39Gy) of the OARs in the IMRT plans were compared to those in the 3D-CRT plans. Regression analyses were performed to assess the relative impact of the factors of interest on the interfractional dose variations of the OARs. Substantial dose excesses to the three OARs were observed at all of the prescription dose levels and the POV margin sizes on the repeat CT scans. The safety threshold based on the mean stomach V39Gy on the recalculated 39 Gy-3D-CRT plans was 1.9 ml. Statistically significant and marginally insignificant mean V39Gy values above the safety thresholds were observed in the stomach in the 51 Gy-IMRT plans (2.6 and 2.1 ml with the 5- and 7-mm PRV margins, respectively (P = 0.015 and 0.085)). Only in the case of the 10-mm POV margin did the metric fall below the safety threshold to 1.5 ml (P = 0.634). The duodenum and the small intestine did not violate the safety thresholds (1.4 and 3.8 ml, respectively). From the multiple regression analyses, only the margin size (P < 0.001) and the POV V39Gy (P < 0.001) were significantly associated with the distribution of recalculated V39Gy for the stomach. Multiple factors, including the margin size (P = 0.020) and the POV V39Gy (P < 0.001) were associated with the recalculated V39Gy for the duodenum. However, none of the POV parameters for the small intestine were associated with the recalculated V39Gy. Considerable interfractional dose variation was observed in three critical OARs. At the escalated prescription dose of breathhold IMRT, the dose variations could exceed the dose variations using 3D-CRT at the safe prescription dose level, indicating that a dose-escalation strategy based solely on the initial advantageous dose distribution in a breathhold IMRT can be problematic. Given the current limitations for predicting or coping with variation throughout the treatment course, the use of POV should be considered for safely delivering escalated doses to patients with pancreatic cancer.