Background and Aims
With ultrasound sensitivity limited in hepatocellular carcinoma (HCC) surveillance and few prospective studies on non-contrast abbreviated MRI (NC-AMRI), this study aimed to assess its diagnostic performance in detecting HCC.
Methods
This prospective study involved cirrhotic patients with contrast-enhanced MRI (CE-MRI) Liver Imaging Reporting and Data System (LI-RADS) LR-3 and LR-4 observations detected during HCC surveillance. Patients underwent average three complete CE-MRI rounds at 3-6 months interval, with approximately 12-month follow-up. NC-AMRI included diffusion-weighted(DWI), T2-weighted imaging(T2WI), and T1-weighted imaging(T1WI). NC-AMRI protocol images were analyzed for diagnostic performance, with subgroup analyses. CE-MRI and NC-AMRI images were independently reviewed by two experienced radiologists, with inter-reader agreement assessed with Kappa coefficient. The reference standard was the AASLD-defined presence of arterial hypervascularity and washout during the portal-venous or delayed phases on CE-MRI.
Results
In 166 CE-MRI follow-ups of 63 patients (median age: 63 years; 60.3% male, 39.7% female), 12 patients developed HCC, with average size of 19.6 mm. The NC-AMRI (DWI+T2WI+T1WI) showed 91.7% sensitivity (95%CI : 61.5–99.8) and 91.6% specificity (95%CI : 86.0–95.4), area under receiver operating characteristic 0.92 (95%CI : 0.83–1.00). Across different Body Mass Index categories, lesion size, Child-Turcotte-Pugh classes, Albumin-Bilirubin (ALBI) grades, and Model for End-Stage Liver Disease classes, sensitivity remained consistent. However, specificity differed significantly between ALBI grade 1 and 2 (86.7% vs. 98.4%,p = 0.010), and between viral and non-viral cirrhosis (93.8% vs. 80.8%,p = 0.010).
Conclusions
NC-AMRI proved clinically feasible, and exhibits high diagnostic performance in HCC detection.
Advances in knowledge
This study highlights efficacy of NC-AMRI in detecting HCC among cirrhotic patients with LR-3 and LR-4 observations, representing significant progress in HCC surveillance.