A nomogram integrating postoperative pathology and systemic inflammation for predicting early recurrence after conversion surgery in initially unresectable intrahepatic cholangiocarcinoma

About 70% of intrahepatic cholangiocarcinoma (ICC) patients have advanced, unresectable disease at first diagnosis. Preoperative conversion therapy may enable subsequent curative-intent hepatectomy in selected patients, but early recurrence (ER) may compromise the survival benefit achieved by conversion surgery. Currently, studies on ER following conversion surgery for ICC are limited. This study aims to develop a nomogram for predicting ER by integrating postoperative pathological factors and systemic inflammatory markers. This retrospective cohort study analyzed 95 patients with initially unresectable ICC who received preoperative conversion therapy and subsequently underwent R0 conversion surgery at a single center (Aug 2021–Mar 2024). ER was defined as recurrence within 12 months post-surgery. Univariate and multivariate logistic regression identified ER predictors, with treatment regimen included as a covariate. A nomogram was developed and internally validated using 1000 bootstrap resamples. Discrimination (AUC), calibration, and clinical utility (DCA) were assessed. 40 patients (42.1%) experienced ER. Multivariate analysis identified microvascular invasion (MVI) (OR = 7.394, P = 0.003), high platelet-to-lymphocyte ratio (PLR) (OR = 5.128 for high vs. low, P = 0.0035), and poor pathological response (Grade I vs. Grade III: OR = 15.625, P = 0.014) as independent predictors. The nomogram achieved an AUC of 0.852 (corrected: 0.824) with good calibration and clinical utility. Using a risk-score cutoff of 122, patients were stratified into high-risk ( n = 29) and low-risk ( n = 66) groups. High-risk patients had shorter median RFS (8 vs. 24 months, p < 0.001). A nomogram incorporating MVI, PLR, and pathological response may facilitate postoperative risk stratification for early recurrence in patients undergoing conversion surgery for initially unresectable ICC. After internal validation, the model showed promising discrimination and may help identify patients who could benefit from closer postoperative surveillance. However, because this model was internally validated only, external validation in independent cohorts is required before routine clinical application.