Association of Survival with Radiologic-Pathologic Discordance in Patients with Hepatocellular Carcinoma: A...

No studies have investigated the effects of radiologic-pathologic discordance on the outcomes of patients who have undergone resection for hepatocellular carcinoma (HCC). Therefore, we investigated the effects of these discrepancies on the outcomes of such patients.

This study included patients diagnosed with HCC on magnetic resonance imaging who underwent resection, including 1,790 from the Korean Primary Liver Cancer Registry (nationwide cohort) and 185 from Chungnam National University Hospital (hospital cohorts). Radiologic-pathologic discrepancies in five factors (maximum tumor diameter, tumor number, vascular invasion, bile duct invasion, and lymph node metastasis) were evaluated using Kaplan-Meier and Cox regression analyses.

The survival rate in the nationwide cohort was lower when all five factors were discordant than when all were concordant (p<0.001). A similar trend was observed in the hospital cohort; however, it was not statistically significant (p=0.260). In multivariate analyses, radiologicpathologic discrepancies in more than two factors (hazard ratio [HR], 3.251) and vascular invasion (HRs, 2.044 and 2.596), and lymph node metastasis (HRs, 8.157 and 7.209) on pathology or both imaging and pathology, respectively, were independent predictors of survival (all p<0.001). Similarly, lymph node metastasis on imaging emerged as an independent predictor (HR, 3.386; p=0.009). Age, an alpha-fetoprotein ≥200 ng/mL, and a modified Union for International Cancer Control stage were additional independent predictors.

This is the first study to demonstrate that radiologic-pathologic discordance in patients with HCC who have undergone resection are significantly associated with worse survival. More accurate and appropriate preoperative evaluations are essential for optimizing treatment and improving prognosis.