Can CA 19-9 predict inferior surgical and pathological outcomes for resectable pancreatic cancer?

Abstract

e16228 Background: Previous data have established that patients with pancreatic surgical specimens showing lymphovascular invasion (LVI) and positive surgical resection margins are associated with poor survival outcomes. However, there are no biomarkers to identify these high-risk patients preoperatively. In this study, we hypothesized that increased carbohydrate antigen 19-9 (CA 19-9) levels could be used to predict high-risk resectable pancreatic cancer features, including LVI and positive residual margins, thereby guiding the surgical decision-making process. Methods: We sampled the National Cancer Database for patients undergoing pancreatic cancer resection between 2004-2018. Logistic regression analysis was performed to analyze sociodemographic and disease characteristics associated with elevated CA 19-9 levels (> 98 U/ml) versus low CA 19-9 levels (< 98 U/ml). Additionally, Cox regression analyses were used to compare hazard ratios for demographic and disease characteristics. A p < 0.05 was considered statistically significant. Results: 12208 patients (50.76% males) with pancreatic cancer were sampled. Pathological findings of LVI (n = 5963), microscopic residual tumor margins (R1) (n = 1547), and residual margins (R), not otherwise specified (NOS) (n = 885) had a stronger association with high preoperative CA 19-9 levels when compared to those with low CA 19-9 levels (LVI: OR 1.212, 95% CI 1.125–1.305, p < 0.001, R1: OR 1.260, 95% CI 1.129–1.408, p = 0.001 and R NOS: OR 1.212, 95% CI 1.052–1.395, p = 0.008). Macroscopic residual margins (R2) (n = 85) were also associated with high CA 19-9 but statistically not significant (OR 1.242, 95% CI 0.802–1.924, p = 0.332) probably due to the low sample size for this subpopulation. Furthermore, all the factors listed below were independently associated with poor prognosis: (A) Elevated CA 19-9 levels: HR 1.411, 95% CI 1.350–1.475, p < 0.001, (B) LVI: HR 1.333, 95% CI 1.275–1.394, p < 0.001, (C) R NOS: HR = 1.517, 95% CI = 1.401–1.643, p < 0.001, (D) R1: HR 1.594, 95% CI 1.499–1.695, p < 0.001, (E) R2: HR 2.015, 95% CI 1.597–2.543, p < 0.001. Conclusions: Elevated CA 19-9 can be used as a biomarker to identify a subset of pancreatic cancer patients with high-risk surgical pathological features such as LVI or residual tumor margins. In these patients, perhaps neoadjuvant chemotherapy can facilitate downstaging and provide superior surgical and oncological outcomes.