Optimizing Hepatocellular Carcinoma Screening: A Comparative Simulated Abbreviated MRI Protocol Study

Background Patients with risk factors such as viral hepatitis-induced liver cirrhosis, advanced-stage primary biliary cirrhosis, hereditary hemochromatosis, metabolic-associated fatty liver disease, and alcoholic liver disease are more likely to develop hepatocellular carcinoma (HCC). Most HCC patients have advanced-stage disease unresponsive to treatment. Therefore, avoiding or treating viral infections and early detection through routine surveillance, such as repeated liver ultrasonography, are the most effective ways to reduce HCC-related mortality. However, as the sensitivity of ultrasound is low, many centers use contrast-enhanced CT and/or MRI for HCC surveillance. This study aimed to evaluate the per-lesion performance of simulated abbreviated non-contrast MRI (aNC-MR) compared to conventional complete MRI (cMR) for screening high-risk patients for HCC. In addition, it aimed to assess the per-lesion performance of simulated abbreviated dynamic multiphase contrast-enhanced MRI (aDCE-MRI) protocol compared to cMR for screening high-risk patients for HCC. Methodology This retrospective study conducted at a single tertiary care center included patients with liver disease at high risk for HCC over five years. These patients underwent cMR for screening. The aNC-MR protocol included diffusion-weighted imaging and T2-weighted imaging, while the aDCE-MR protocol included only the non-contrast T1 fat-saturated and dynamic post-contrast sequence. Two independent radiologists, blinded to the original cMR diagnosis, assessed two abbreviated protocol image sets, namely, an aDCE-MR set and an aNC-MR set. Each lesion was categorized as suspicious for HCC or suitable for interval screening on the aNC-MR protocol set. On the aDCE-MR image set, the lesions were designated using Liver Imaging Reporting and Data System categorization. Results Of a total of 125 lesions, 110 were evaluated and characterized using the simulated aNC-MR protocol, as 15 lesions could not be seen on aNC-MR, giving the aNC-MR protocol an overall sensitivity of 88%. On a per-lesion basis, the aNC-MR protocol compared with cMR demonstrated a negative predictive value of 96.4% and a specificity of 96.4%. Overall, 125 lesions were evaluated and characterized using the simulated aDCE-MR protocol. On a per-lesion basis, the aDCE-MR protocol compared with cMR demonstrated an negative predictive value of 100% and a specificity of 98.6%. Comparable results were observed in patients with hepatic steatosis. Conclusions The aNC-MR protocol can significantly reduce the cost and time required for MRI screening in patients at high risk for HCC. This protocol was equally effective in patients with hepatic steatosis and suboptimal ultrasound results, making it a viable and efficient screening modality.