Re-evaluating the Role of Neoadjuvant Chemotherapy in Early-Stage Pancreatic Ductal Adenocarcinoma: Unveiling the Impact of Immortal Time Bias on Survival.

Abstract

BACKGROUND: The survival benefit of neoadjuvant chemotherapy (NAC) in early-stage pancreatic ductal adenocarcinoma (PDAC) remains uncertain. Although retrospective studies often suggest improved outcomes, these findings may be confounded by immortal time bias (ITB), the interval between diagnosis and treatment during which patients must survive to receive therapy, potentially inflating survival estimates. This study applied multiple bias-adjusted analytic methods to re-evaluate the association between NAC and overall survival in early-stage PDAC. METHODS: Using the National Cancer Database (2012-2017), we identified adults with resectable clinical T1 and T2 PDAC who underwent pancreatectomy. Overall survival was compared between multiagent NAC and upfront surgery by using Kaplan-Meier and multivariable Cox models. We used three statistical approaches to adjust for immortal time bias: 1) a 9-month landmark analysis; 2) a time-varying Cox regression model; and 3) a propensity-matched time-varying Cox model. RESULTS: Among 13,466 patients, 15.8% received NAC. Before bias adjustment, NAC was associated with longer median survival (33.4 vs. 25.7 months; HR 0.78, p < 0.001). However, after correcting for immortal time bias, this apparent survival advantage disappeared: landmark analysis (HR 0.94, p = 0.072), time-varying model (HR 1.06, p = 0.2), and matched cohort (HR 0.94, p = 0.2). CONCLUSIONS: This study shows that ITB significantly influences survival estimates in retrospective analyses of NAC for early-stage resectable PDAC. When appropriately adjusted, NAC was not associated with a survival advantage compared to upfront surgery. These findings underscore the need for methodological rigor in retrospective studies and caution against overinterpreting unadjusted survival advantages.