Similar assessment of tumor characteristics in localized intrahepatic cholangiocarcinoma with four-phase CT and MRI
Frida Karlsson, Marco Gerling, Christina Villard, Aristeidis Grigoriadis, Hannes JanssonAbstract
Introduction
For staging of localized intrahepatic cholangiocarcinoma (iCCA), Swedish guidelines recommend either four-phase computed tomography (CT) or magnetic resonance imaging (MRI), while international guidelines prioritize MRI. Few head-to-head comparisons of CT and MRI are available.
Methods
Consecutive patients with localized iCCA and tumor-directed treatment (2023–2025) at Karolinska University Hospital (Stockholm, Sweden) were retrospectively included. Imaging studies from patients with both CT and MRI were reassessed regarding tumor number (≥4 tumors), size (>5 cm), bilobar distribution, lymph node metastases and vascular invasion. Agreement between CT and MRI was evaluated with Cohen's kappa (k) and Gwet's agreement coefficient (AC).
Results
Among 63 patients with localized iCCA and tumor-directed treatment, 23 (37%) had ≥4 tumors, 39 (62%) had tumors >5 cm and 37 (59%) had a bilobar tumor distribution. Twenty-eight patients (44%) were investigated with both CT and MRI. Patients with prior liver disease were more likely to undergo both CT and MRI (P = 0.017). Tumors >5 cm were more frequent in the CT only group (79% versus 46%, P = 0.013). MRI detected ≥4 tumors in 9/28 studies and CT in 7/28 studies, with a non-significant inter-modality difference (k = 0.51, AC = 0.72, P = 0.317). Inter-modality agreement was high regarding tumor distribution (k = 0.94, AC = 0.95), lymph node metastasis (k = 0.71, AC = 0.70) and vascular invasion (k = 0.71, AC = 0.90).
Discussion
While MRI has been recommended based on improved detection of small tumor lesions, four-phase CT and MRI provided similar assessments of tumor characteristics in this current single-center cohort. A limited sample size restricted statistical power. Differences in patient characteristics indicate selection when ordering imaging.