This study aimed to evaluate the value of [18F]FDG PET/CT parameters and serological indices in predicting lymph node metastasis (LNM) in patients with intrahepatic cholangiocarcinoma (ICC) preoperatively.

The study included 102 patients with ICC diagnosed between January 2018 and December 2022. All patients underwent preoperative [18F]FDG PET/CT scanning and serological marker evaluations before surgical tumor resection. The [18F]FDG PET/CT scanning was jointly evaluated by two nuclear medicine physicians blinded to the clinical information and compared with postoperative pathologic results. The tumor SUVmax, tumor-to-background ratios (TBR), SUVpeak, and total lesion glycolysis (TLG) were obtained from the PET/CT scans. Presurgical serological testing included carbohydrate antigen 19–9 (CA19-9) and carcinoembryonic antigen (CEA). The results were analyzed by bivariate and multivariate logistic regression to develop a prediction model for LNM.

Among the 102 patients with ICC, 60 (58.8%) patients had LNM while 42 patients (41.2%) were free of LNM. TBR(without LNM: 6.9 ± 3.6 vs. LNM: 8.2 ± 2.9, P = 0.019), CA19-9(506.9 ± 1602.7 vs. 1901.2 ± 3195.1, P = 0.011) and CEA(5.1 ± 8.9 vs. 32.5 ± 82.7, P = 0.035) in the LNM group were significantly higher than in the without LNM group. Three additional, statistically significant features, tumor TBR > 6.4, CA19-9 > 149.7 U/mL and CEA > 3.2 ng/mL, were integrated into a prediction model, which substantially outperformed use of the single parameter in ROC analysis (AUC 0.802 vs. 0.676, 0.671 and 0.656).

TBR > 6.4 and CA19-9 > 149.7 U/mL were identified as effective preoperative predictors of LNM in patients with ICC. The prognostic model we propose, integrating imaging parameters and serological indicators, significantly enhances predictive capability.