Abstract
Patients with pancreatic adenocarcinoma (PAC) are often treated with neoadjuvant chemoradiation (NACRT) in hopes of downstaging their disease for potential surgical resection. We hypothesized that increasing the radiation dose to the area of the tumor abutting the vessel(s) of concern would increase the rate of surgical resection in patients with borderline resectable PAC (BRPAC) and locally advanced PAC (LAPAC) treated with NACRT.
We retrospectively reviewed consecutive cases of BRPAC and LAPAC treated with NACRT from January 2006 to December 2013, with or without a vessel boost (VB), at a single institution. The primary endpoints were rate of R0/R1 potentially curative surgical resection and acute toxicity. Univariate analysis with the Fisher exact test was performed to evaluate the effect of each variable. Multiple logistic regression was used to adjust for the following covariates: year of diagnosis, age, sex, carbohydrate antigen 19-9 (CA19-9) level at diagnosis, and BRPAC or LAPAC.
Of the 104 patients identified, 22% (n = 23) received a VB (median, 54 Gy; range, 54-64 Gy), and 78% (n = 81) received no boost (median, 50.4 Gy; range, 48.6-52.2 Gy). More patients in the VB group were treated from 2010 to 2013 (P < .001) and with intensity modulated radiation therapy (P = .002). Other baseline characteristics were balanced. After adjustment for covariates, there was a statistical trend toward increased surgical resection in patients who received a VB (odds ratio [OR], 2.77; 95% confidence interval [CI], 0.89-8.57; P = .077). Age (≥70 years; OR, 0.42; 95% CI, 0.16-1.05; P = .064) and LAPAC (OR, 0.32; 95% CI, 0.09-1.09; P = .068) also trended toward significance. CA19-9 ≥47.9 U/mL (OR, 0.24; 95% CI, 0.08-0.71; P = .010) was significant on multivariate analysis. There was no significant difference in acute or late toxicity between groups.
In our retrospective series, dose escalation was associated with an improved surgical resection rate in BRPAC and LAPAC patients treated with NACRT, although this improvement was not statistically significant.
