In reply: The letter by Larcos provides additional commentary regarding screening for hepatocellular carcinoma (HCC) in patients with cirrhosis. Evidence from an early Australian study showed that liver ultrasound screening was superior to α‐fetoprotein in detecting HCC.1 This finding, combined with other data, is the basis whereby not all international professional societies recommend α‐fetoprotein use for HCC screening. Survival following a diagnosis of HCC has increased during more recent years and likely represents improvement in contemporary treatments.2 It is now established that ultrasound screening detects earlier stage HCC and improves treatment outcomes. The accuracy of ultrasound scans for the detection of HCC is less than that for diagnostic tests such as computed tomography or magnetic resonance imaging; however, additional characteristics are required of a screening test that affects cost‐effectiveness. These include patient acceptability, test availability, likely uptake, the cost of the test, and the cost of false positive and false negative results. Further research and validation of promising serum markers of HCC may allow improved risk stratification of HCC and direct diagnostic tests to patients at highest risk.3,4 We agree that an Australian cost‐effectiveness analysis of ultrasound screening for HCC is required. Cost‐effectiveness analysis in a screening context commonly uses decision analytic‐simulation modelling, such as Markov models, to predict the aggregated resources used in screening and all downstream events, diagnostic accuracy and the consequences of false results, patient‐reported quality of life, and likely survival.5 While a randomised controlled trial in liver ultrasound screening would be useful, decision‐analytic modelling is critical to comprehensively assess the merits of a screening program, encompass all sources of evidence and inform decision makers about the overall costs, benefits and harms of screening.
Publication of your online response is subject to the Medical Journal of Australia's editorial discretion. You will be notified by email within five working days should your response be accepted.