Purpose
Prospectively assess relationships between dosimetric parameters and histopathologic/clinical duodenal toxicities in patients on a phase I trial for pancreatic cancer.
Methods
Forty-six borderline resectable/unresectable patients were enrolled on a prospective trial testing neoadjuvant gemcitabine/5-fluorouracil followed by SBRT (5 daily fractions of 5–8 Gy) and concurrent nelfinavir. Post-SBRT surgery was performed in 13 resectable patients, which constituted the patient population herein. Pathologic duodenal damage was assessed using predetermined criteria: 1, no/minimal; 2, moderate; and 3, marked damage. Clinical toxicities were assessed per CTCAE. Duodenal dosimetric parameters included V5–V40 and mean/maximum doses. Spearman correlation and linear regression evaluated associations between dosimetric parameters and clinical/pathologic duodenal toxicity.
Results
The median duodenal mean and maximum doses were 20 and 37 Gy. Median duodenal V5–V40 were 64, 62, 52, 39, 27, 14, 5 and 0 cc, respectively. The median duodenal damage score was 2 (four 1, eight 2, and one 3). Higher duodenal damage scores correlated with higher duodenal mean doses (r = 0.75, p = 0.003), V35 (r = 0.61, p = 0.03), V30 (r =0.67, p = 0.01), V25 (r =0.68, p = 0.01), V20 (r = 0.56, p = 0.05), and the planning target volume (PTV) mean (r =0.59, p = 0.03) and maximum (r = 0.61, p = 0.03) doses. Clinical toxicities did not correlate with dosimetric parameters or duodenal pathologic damage.
Conclusions
Duodenal histologic damage correlates with mean duodenal dose, V20–V35, and PTV mean/−maximum doses.