Neutrophil to lymphocyte ratio (NLR) predicts pathological response to neoadjuvant immunochemotherapy in resectable esophageal cancer: a retrospective study

Background: Neoadjuvant immunochemotherapy (nICT) has demonstrated promising efficacy in locally advanced esophageal cancer. The pretreatment neutrophil to lymphocyte ratio (NLR) has shown prognostic value across various solid tumors. This study aimed to evaluate the predictive value of pretreatment NLR for pathological response in patients with resectable esophageal cancer receiving nICT.

Methods: A total of 295 patients with resectable esophageal cancer who underwent nICT followed by surgery were retrospectively analyzed. Pretreatment NLR was calculated from peripheral blood counts. Pathological response was assessed using the Becker tumor regression grading (TRG) system. Receiver operating characteristic (ROC) curve analysis was conducted to evaluate the predictive performance of NLR. Multivariable logistic regression and bootstrap internal validation were performed to assess its independent predictive value.

Results: Pretreatment NLR was significantly higher in non-pathological complete response (non-pCR) patients than in pathological complete response (pCR) patients (P=0.04). A weak positive correlation was observed between NLR and TRG (r=0.130, P=0.03). The area under the ROC curve (AUROC) of NLR for predicting non-pCR was 0.586 [95% confidence intervals (CI): 0.510–0.663, P=0.04], with an optimal cutoff value of 3.59 (sensitivity 37.6%, specificity 79.3%, positive predictive value 89.0%, negative predictive value 24.1%). The combination of NLR with clinical stage improved the AUROC to 0.633 for predicting non-pCR. In the esophageal squamous cell carcinoma (ESCC) subgroup (n=271), NLR demonstrated improved predictive performance for both non-pCR (AUROC =0.609, 95% CI: 0.530–0.687) and non-major pathological response (non-MPR) (AUROC =0.570, 95% CI: 0.512–0.642). Integration with clinical T and N categories further increased the AUROC to 0.654 for non-pCR and 0.626 for non-MPR. Multivariable logistic regression confirmed that NLR was an independent predictor of non-pCR [odds ratios (OR) =1.325, 95% CI: 1.005–1.747, P=0.046], and bootstrap validation supported its robustness.

Conclusions: Elevated pretreatment NLR is associated with poorer pathological response to nICT in patients with resectable esophageal cancer, particularly in ESCC. When combined with clinical staging, NLR may serve as an accessible adjunctive tool to identify patients with a low probability of pCR.