Prediction of clinically significant complication following neoadjuvant chemoimmunotherapy in resectable esophageal...

Neoadjuvant chemoimmunotherapy (NICT) shows promise in locally advanced esophageal squamous cell carcinoma (LA-ESCC), yet may induce complex inflammatory-immune changes that complicate postoperative risk stratification. Current static inflammatory indices lack generalizability across populations and timing.

This single-center retrospective cohort study analyzed 273 patients. Patients were stratified by Clavien-Dindo grade into a non-significant complication (NSC) group (CD < II) and a clinically significant complication (CSC) group (CD ≥ II). Dynamic (Δ) inflammatory indices were constructed from pre- and post-NICT hematologic markers. Optimal cutoffs were determined via ROC analysis and the Youden index, after which each Δ indicator was transformed into a dichotomous variable. Significant predictors from univariable logistic regression were incorporated into the multivariable logistic regression, which integrated ΔNPR-SII2 with clinical factors to build a final predictive model.

Of the cohort, 212 and 61 patients were classified into the NSC and CSC groups, respectively. Multivariate logistic regression identified age ≥ 70 years (odds ratio [OR] = 2.37, 95% CI: 1.01-5.55), low body mass index (< 18.5; OR = 2.20, 95% CI: 1.08-4.50), an elevated ECOG score (≥ 2; OR = 3.98, 95% CI: 1.13-14.05), and ΔNPR-SII2 (OR = 8.34, 95% CI: 3.92-17.74) as independent predictors of major complications. The model demonstrated good discriminatory power, with an area under the curve (AUC) of 0.784 (95% CI: 0.724-0.844) and a bootstrap-validated C-index of 0.772. Calibration performance was satisfactory (Brier score = 0.138).

We developed and internally validated a dynamic inflammatory biomarker-based model for preoperative prediction of major complications post-NICT. The ΔNPR-SII2 index captures treatment-induced immune-inflammatory dysregulation and provides a clinically translatable tool for risk-adapted perioperative management.