Premium
Modeling the benefits and harms of surveillance for hepatocellular carcinoma: Information to support informed choices
Author(s) -
Taylor Eleanor J.,
Jones Rebecca L.,
Guthrie J. Ashley,
Rowe Ian A.
Publication year - 2017
Publication title -
hepatology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 5.488
H-Index - 361
eISSN - 1527-3350
pISSN - 0270-9139
DOI - 10.1002/hep.29315
Subject(s) - medicine , overdiagnosis , hepatocellular carcinoma , harm , watchful waiting , credible interval , false positive paradox , confidence interval , cancer , radiology , psychology , social psychology , prostate cancer , machine learning , computer science
Surveillance by ultrasonography for hepatocellular carcinoma (HCC) for individuals with cirrhosis is recommended. There is debate regarding the effectiveness of surveillance in reducing mortality, and there is little information on the harms available to patients considering surveillance. The aim of this study was to provide estimates of both the benefit and harms of surveillance. A Markov model was built to simulate outcomes of individuals aged 50 years with well‐compensated cirrhosis entering surveillance. Following identification of a focal lesion by ultrasound surveillance, further investigations were defined by the European Association for the Study of the Liver/European Organization for Research and Treatment of Cancer recall policy. Benefit and harm outcomes are expressed per 1,000 patients over 5 years. For every 1,000 patients in surveillance over 5 years, there are 13 fewer deaths (95% confidence interval [CI], 12‐14) compared with no surveillance, equating to a number needed to screen to prevent one death from HCC of 77. In comparison, many more individuals experienced harm through surveillance. For every 1,000 patients, 150 (95% CI, 146‐154) had one or more false‐positive tests equating to a number needed to harm from surveillance of 7. As a consequence of a false‐positive test, 65 individuals required at least one additional unnecessary computed tomography scan or magnetic resonance imaging and 39 required an unnecessary liver biopsy according to the recall policy. Surveillance benefits were sensitive to the incidence of HCC and the mortality benefit achieved by treatment. Harms were sensitive to the rates of false‐positive testing and the frequency of liver biopsy. Conclusion : There is a balance between the small absolute mortality benefit to surveillance for HCC and the numerically more frequent harms resulting from false‐positive testing. Implementation of the recently revised American Association for the Study of Liver Diseases recommendations is predicted to reduce harms from unnecessary liver biopsy. (H epatology 2017;66:1546–1555).