Predictive value for advanced lung cancer inflammation index within short- /long-term outcomes of esophageal squamous cell carcinoma after minimally invasive esophagectomy cases： a propensity score matching study

Abstract

Abstract Introduction: Advanced lung cancer inflammation index (ALI) within esophageal squamous cell carcinoma (ESCC) importance is unclear. We aimed to investigate whether ALI is linked to poor short-term outcomes and long-term prognosis within cases of ESCC after minimally invasive esophagectomy (MIE). Methods: Kaplan-Meier survival assessment was applied for comparing cancer-specific survival (CSS) across different cohorts. Clinicopathological features across the two cohorts were eliminated by propensity score matching (PSM). We established a new model for predicting CSS by combining ALI and tumor-node-metastasis (TNM) staging according to Cox multivariate results. Time-dependent area under the curve (t-AUC) and decision curve analyses (DCA) evaluated predictive /clinical relevance capacities for this model. Results: Severe postoperative complication manifestations within low ALI cohort were significantly elevated compared to within high ALI cohort (25.3% vs 16.7, P=0.01), nil variations were identified across both cohorts after PSM (25.3% vs 18.3%, P = 0.06). Both, within overall cohort and the matched cohort, low ALI only significantly reduced the 5-year CSS in locally advanced ESCC patients (all P < 0.05) relative to high ALI. Further analysis demonstrated that patients within high ALI cohort were at increased risk for adverse postoperative CSS in most clinicopathological subgroups. Cox multivariate analysis demonstrated that TNM staging and ALI were variables that independently influenced adverse CSS in both cohorts (P < 0.05). Therefore, a new prediction model was established by combining these two factors. The t-AUC and DCA demonstrated that this model had a more accurate prediction effect and better clinical use value than the TNM stage alone. Conclusion: ALI proved to be an effective biological indicator of CSS after MIE in locally advanced ESCC patients. The combined application of the ALI and TNM model can thus improve the clinical prediction ability.