Development and validation of a prognostic model for overall survival in pN0 esophageal cancer patients after neoadjuvant chemotherapy: a SEER database-based study.

Abstract

Background: Esophageal cancer (EC) is a common malignancy globally, with neoadjuvant chemotherapy plus surgery being the standard treatment. Clinically, accurate prognosis for pN0 patients after neoadjuvant therapy remains challenging, as the traditional tumor-node-metastasis (TNM) staging system may understate the impact of lymph node dissection extent and individual clinical variables on survival. However, the impact of lymph node dissection extent on survival in patients with pN0 status following neoadjuvant therapy remains unclear. This study aimed to clarify this relationship and develop a predictive model for overall survival (OS).

Methods: Using the Surveillance, Epidemiology, and End Results (SEER) database, we identified 1,393 patients with pN0 after neoadjuvant chemotherapy followed by surgery between 2007 and 2021. Baseline clinical factors (age, sex, race, tumor location, T/M stage, lymph node count, etc.) were extracted, and OS was defined as the time from diagnosis to death or last follow-up (minimum 1-month follow-up). Patients were randomly divided into a training set and validation set at a 7:3 ratio by histology [squamous cell carcinoma/adenocarcinoma (SCC/AC)]. Restricted cubic spline regression and multivariate Cox regression were used to identify lymph node dissection thresholds and independent prognostic factors. A nomogram was constructed and validated via concordance index (C-index), receiver operating characteristic (ROC) curves, calibration curves, and decision curve analysis (DCA).

Results: The study cohort had a median age of 63 years, with 82.1% male and 80.4% AC. For SCC, ≥13 lymph nodes dissected was associated with improved OS [hazard ratio (HR) =0.60, 95% confidence interval (CI): 0.43-0.85], and for AC, ≥15 lymph nodes (HR =0.71, 95% CI: 0.60-0.85). Independent predictors included sex, lymph node count (SCC), and sex, age, M stage, lymph node count (AC). The nomogram achieved C-indices of 0.593 (training) and 0.677 (validation) for SCC, and 0.599 (training) and 0.634 (validation) for AC, outperforming TNM staging in 1-, 3-, 5-year OS prediction (SCC validation AUCs: 0.773, 0.678, 0.651; AC: 0.696, 0.664, 0.640).

Conclusions: The findings suggest that dissecting an adequate number of lymph nodes is associated with improved survival in patients with pN0 EC. Most probably, this is due to more adequate staging of the pN-status of patients and avoidance of understaging. The nomogram provides more precise survival prediction than TNM staging, aiding personalized prognosis and treatment planning. Clinicians should prioritize achieving these lymph node dissection thresholds to optimize staging accuracy, though external validation in diverse populations is warranted.