Association of adjuvant chemotherapy with overall survival in resected pancreatic adenocarcinoma previously treated with neoadjuvant therapy

404 Background: Guidelines for adjuvant chemotherapy in patients with resected pancreatic adenocarcinoma (PDAC) who received neoadjuvant chemotherapy are equivocal. A lymph node ratio (LNR) ≥ 0.15 may predict lack of benefit, but conflicting results are reported. Methods: The National Cancer Database was searched to identify patients who were resected after neoadjuvant chemotherapy in 2006-2013. Exclusions: metastases at surgery, 90-day postoperative mortality, adjuvant radiation, and outlier interval from diagnosis to surgery (10 months). The association between adjuvant chemotherapy and overall survival (OS) from diagnosis was examined using multivariable Cox regression and inverse propensity of treatment weighted (IPTW) Cox regression. An IPTW based estimator of the average treatment effect (ATE) was used to quantify the population average survival benefit of treatment. Outcomes were examined in all patients and in those with LNR < 0.15 and ≥ 0.15. Results: 681/2488 patients (27%) received adjuvant chemotherapy. In multivariable Cox regression, adjuvant chemotherapy was associated with improved OS in the overall cohort and in patients with LNR < 0.15. A trend towards improved OS was also observed for those with LNR ≥ 0.15. After accounting for indication bias using IPTW, a significant survival benefit for was observed only for patients with LNR < 0.15. The ATE among LNR < 0.15 patients was 3.3 (95% CI 1.0, 5.7) months, indicating that the average survival of the population would be 3.3 months longer if all received treatment. Conclusions: Adjuvant chemotherapy in resected PDAC patients who received neoadjuvant therapy appears to be beneficial in patients with negative lymph nodes or minimal nodal burden. High LNR after neoadjuvant therapy may be an indicator of adverse tumor biology that is less likely to derive a therapeutic benefit. [Table: see text]