NEUTROPHIL-TO-LYMPHOCYTE RATIO FOR PREDICTING POSTOPERATIVE RECURRENCE IN COMBINED HEPATOCELLULAR–CHOLANGIOCARCINOMA: A SECONDARY ANALYSIS OF A PUBLIC DATASET

Background: Combined hepatocellular–cholangiocarcinoma (cHCC-CCA) is a rare primary liver cancer with a high rate of postoperative recurrence. The neutrophil-to-lymphocyte ratio (NLR) is a simple, routinely available marker of systemic inflammation. We aimed to identify predictors of postoperative recurrence—focusing on NLR—by framing recurrence as a binary outcome, an approach distinct from prior survival-based analyses of the same tumor. Methods: This was a secondary analysis of a publicly available dataset of 42 patients who underwent hepatectomy for cHCC-CCA (Chiu et al., 2020). Recurrence (present/absent) was the primary outcome. Each clinicopathological variable was assessed by univariable logistic regression; the discriminative performance of NLR was evaluated by receiver operating characteristic (ROC) analysis with an optimal cut-off determined by the Youden index. Patients were dichotomized by this cut-off, and disease-free survival was compared by the Kaplan–Meier method and log-rank test. A parsimonious multivariable logistic model was built from univariable-significant variables. Results: Recurrence occurred in 23 of 42 patients (54.8%). NLR was the strongest predictor of recurrence (univariable odds ratio [OR] 1.60 per unit; 95% confidence interval [CI] 1.15–2.22; p=0.005). NLR discriminated recurrent from non-recurrent patients with an area under the ROC curve of 0.80 (95% CI 0.64–0.94); the Youden-optimal cut-off was 2.78 (sensitivity 78%, specificity 79%). Patients with high NLR (≥2.78) had markedly worse disease-free survival than those with low NLR (median 5.1 months versus not reached; log-rank p<0.0001). NLR remained independently associated with recurrence after adjustment for the surgical margin (OR 1.42; 95% CI 1.00–2.01; p=0.049) and, separately, for AJCC tumor stage (OR 1.60; 95% CI 1.15–2.23; p=0.005). Conclusions: A higher NLR was independently associated with postoperative recurrence in cHCC-CCA, with a data-derived optimal cut-off (2.78) closely matching the conventional threshold of 3. By modeling recurrence as a binary outcome with logistic regression and ROC analysis, this study complements earlier survival-based findings and supports NLR as a simple, accessible candidate marker of recurrence risk. Validation in larger, prospective, externally validated cohorts is warranted.