The geriatric nutritional risk index predicts pathological complete response and anastomotic leakage in esophageal...

Esophageal squamous cell carcinoma (ESCC) poses a significant global health challenge. Pathological complete response (pCR) to neoadjuvant therapy followed by surgery is a crucial predictor of survival, but treatment response is highly variable. Nutritional status and systemic inflammation are key determinants of outcomes. This study aimed to evaluate the geriatric nutritional risk index (GNRI), alone and combined with body mass index (BMI), as a predictor for pCR and anastomotic leak (AL) in this patient population.

In this retrospective study, 143 ESCC patients undergoing neoadjuvant therapy and radical esophagectomy were enrolled. The GNRI was calculated using serum albumin and body weight. The optimal GNRI cutoff for predicting pCR was determined by receiver operating characteristic (ROC) curve analysis. Patients were stratified into high (≥106.7) and low (<106.7) GNRI groups. Univariate and multivariate logistic regression analyses were performed to identify independent predictors for pCR and AL. The synergistic value of GNRI and BMI was assessed through collinearity analysis, mutual adjustment in multivariate models, and model fit statistics [Akaike information criterion (AIC), likelihood ratio tests].

The pCR rate was significantly higher in the high GNRI group (39.3% vs. 10.2%, P<0.001). GNRI demonstrated good predictive accuracy for pCR [area under the curve (AUC) =0.706] with high sensitivity (87.2%). Multivariate analysis confirmed high GNRI as an independent predictor for pCR [odds ratio (OR)=3.90] and a protective factor against AL (OR =0.39). The AL rate was significantly lower in the high GNRI group (16.67% vs. 33.90%, P=0.02). Despite a strong correlation between BMI and GNRI (R=0.76, P<0.001), both retained independent predictive value for pCR after mutual adjustment. A combined model (GNRI + BMI + confounders) yielded the best fit (lowest AIC=139.8) and the highest discriminative ability (AUC=0.873).

Pretreatment GNRI is a powerful, independent predictor for pCR and AL in ESCC patients. Its high negative predictive value helps identify patients with a very low probability of responding to standard neoadjuvant therapy. The combination of GNRI and BMI provides superior predictive power for pCR, as they offer complementary information. These simple, cost-effective indices should be integrated into initial patient assessment to improve risk stratification and guide personalized treatment strategies.