Understanding Outcomes in Hepatocellular Carcinoma and Cirrhosis in England

Background: Hepatocellular carcinoma (HCC) is a major cause of death globally and its incidence is rising. It usually occurs on a background of cirrhosis. Survival statistics are limited outside clinical trials and population-based studies are complex because patients have competing risks of liver- and cancer-related mortality. This thesis aimed to developed methods to assess cirrhosis severity from population-based electronic heath records (EHRs), in order to understand these clinical outcomes for patients with HCC, and to describe regional variation in treatment allocation across England. Methods: Algorithms were developed to determine cirrhosis severity from inpatient EHRs, using case note review from two NHS centres for validation. Competing risk was modelled using the presence of hepatic decompensation to identify liver-related outcomes. Using cancer registry data linked to the Hospital Episodes Statistics database, all patients diagnosed with HCC between 2007 and 2016 in England were identified. These patients were characterised using the algorithms and the predictors of HCC treatment allocation were identified. The impact of cirrhosis severity on overall survival and competing risk was investigated. Regional variation in the proportion of patients receiving different HCC treatments was described. Results: The sensitivity for cirrhosis detection from EHRs was 86% using the validated algorithm. Cirrhosis severity was the main predictor of overall survival following HCC treatment, and this difference correlated with an increase in liver-related mortality in competing risk analyses. Surgical treatments for HCC were more common in those regions with specialist services. Conclusion: Overall survival in HCC in England is comparable with published estimates from clinical trials. Survival is strongly influenced by underlying cirrhosis severity and this is associated with increased liver-related mortality. Careful consideration of the effect of previous hepatic decompensation is essential for treatment selection in HCC. Exploring barriers to receiving specialist treatment may improve the observed regional variation in treatment allocation.