A Prototype Artificial Intelligence Algorithm Versus Liver Imaging Reporting and Data System (LI-RADS) Criteria in Diagnosing Hepatocellular Carcinoma on Computed Tomography: a Randomized Trial

What this Trial Is About

Liver cancer is a common and deadly disease, ranking as the sixth most diagnosed cancer and the fourth leading cause of cancer death worldwide. In Hong Kong, it is the third most common cause of cancer death, with survival rates varying dramatically depending on the stage at diagnosis. Early and accurate diagnosis is crucial for improving survival. This trial evaluates a new artificial intelligence (AI) algorithm developed to diagnose liver cancer from CT scans, comparing it to the current standard radiological method called LI-RADS. The study is randomized and aims to confirm the AI's diagnostic accuracy in patients at risk for liver cancer. The trial compares two diagnostic approaches: a prototype AI algorithm developed by the University of Hong Kong and the LI-RADS criteria independently assessed by experienced abdominal radiologists. Participants undergo triphasic contrast-enhanced CT scans, and both methods analyze the liver images separately. The study is designed to assess whether the AI can improve early diagnosis by reducing inconclusive results and errors that can delay treatment. This evaluation is separate from routine clinical reporting to ensure unbiased results. Participants are adults at risk for liver cancer, including those with cirrhosis or chronic hepatitis B, who have new liver nodules detected via ultrasound. The study measures diagnostic accuracy over 12 months using a combination of clinical and radiological data as the standard reference. Additional outcomes include other diagnostic performance measures, time taken to interpret scans, and any technical issues. The research includes regular assessments and follows participants for one year to monitor results and safety.

Artificial Intelligence vs. LIRADS in Diagnosing HCC on CT