Novel machine learning models outperform risk scores in predicting hepatocellular carcinoma in patients with chronic viral hepatitis
Grace Lai–Hung Wong, Vicki Wing‐Ki Hui, Qingxiong Tan, Jingwen Xu, Hye Won Lee, Terry Cheuk‐Fung Yip, Baoyao Yang, Yee‐Kit Tse, Chong Yin, Fei Lyu, Jimmy Che‐To Lai, Grace Lui, Henry Lik‐Yuen Chan, Pong C. Yuen, Vincent Wai‐Sun Wong
Abstract
•Accurate hepatocellular carcinoma (HCC) risk prediction is helpful in reducing mortality.•Existing HCC risk scores usually include a few known risk factors and preselected parameters.•Machine learning allows for direct selection of predictive parameters without subjective preselection.•HCC ridge score (HCC-RS) built from machine learning modelling has higher accuracy than existing HCC risk scores.•HCC-RS may be incorporated into electronic medical health systems to facilitate real-time update of HCC risk. Background & AimsAccurate hepatocellular carcinoma (HCC) risk prediction facilitates appropriate surveillance strategy and reduces cancer mortality. We aimed to derive and validate novel machine learning models to predict HCC in a territory-wide cohort of patients with chronic viral hepatitis (CVH) using data from the Hospital Authority Data Collaboration Lab (HADCL).MethodsThis was a territory-wide, retrospective, observational, cohort study of patients with CVH in Hong Kong in 2000–2018 identified from HADCL based on viral markers, diagnosis codes, and antiviral treatment for chronic hepatitis B and/or C. The cohort was randomly split into training and validation cohorts in a 7:3 ratio. Five popular machine learning methods, namely, logistic regression, ridge regression, AdaBoost, decision tree, and random forest, were performed and compared to find the best prediction model.ResultsA total of 124,006 patients with CVH with complete data were included to build the models. In the training cohort (n = 86,804; 6,821 HCC), ridge regression (area under the receiver operating characteristic curve [AUROC] 0.842), decision tree (0.952), and random forest (0.992) performed the best. In the validation cohort (n = 37,202; 2,875 HCC), ridge regression (AUROC 0.844) and random forest (0.837) maintained their accuracy, which was significantly higher than those of HCC risk scores: CU-HCC (0.672), GAG-HCC (0.745), REACH-B (0.671), PAGE-B (0.748), and REAL-B (0.712) scores. The low cut-off (0.07) of HCC ridge score (HCC-RS) achieved 90.0% sensitivity and 98.6% negative predictive value (NPV) in the validation cohort. The high cut-off (0.15) of HCC-RS achieved high specificity (90.0%) and NPV (95.6%); 31.1% of patients remained indeterminate.ConclusionsHCC-RS from the ridge regression machine learning model accurately predicted HCC in patients with CVH. These machine learning models may be developed as built-in functional keys or calculators in electronic health systems to reduce cancer mortality.Lay summaryNovel machine learning models generated accurate risk scores for hepatocellular carcinoma (HCC) in patients with chronic viral hepatitis. HCC ridge score was consistently more accurate than existing HCC risk scores. These models may be incorporated into electronic medical health systems to develop appropriate cancer surveillance strategies and reduce cancer death. Accurate hepatocellular carcinoma (HCC) risk prediction facilitates appropriate surveillance strategy and reduces cancer mortality. We aimed to derive and validate novel machine learning models to predict HCC in a territory-wide cohort of patients with chronic viral hepatitis (CVH) using data from the Hospital Authority Data Collaboration Lab (HADCL). This was a territory-wide, retrospective, observational, cohort study of patients with CVH in Hong Kong in 2000–2018 identified from HADCL based on viral markers, diagnosis codes, and antiviral treatment for chronic hepatitis B and/or C. The cohort was randomly split into training and validation cohorts in a 7:3 ratio. Five popular machine learning methods, namely, logistic regression, ridge regression, AdaBoost, decision tree, and random forest, were performed and compared to find the best prediction model. A total of 124,006 patients with CVH with complete data were included to build the models. In the training cohort (n = 86,804; 6,821 HCC), ridge regression (area under the receiver operating characteristic curve [AUROC] 0.842), decision tree (0.952), and random forest (0.992) performed the best. In the validation cohort (n = 37,202; 2,875 HCC), ridge regression (AUROC 0.844) and random forest (0.837) maintained their accuracy, which was significantly higher than those of HCC risk scores: CU-HCC (0.672), GAG-HCC (0.745), REACH-B (0.671), PAGE-B (0.748), and REAL-B (0.712) scores. The low cut-off (0.07) of HCC ridge score (HCC-RS) achieved 90.0% sensitivity and 98.6% negative predictive value (NPV) in the validation cohort. The high cut-off (0.15) of HCC-RS achieved high specificity (90.0%) and NPV (95.6%); 31.1% of patients remained indeterminate. HCC-RS from the ridge regression machine learning model accurately predicted HCC in patients with CVH. These machine learning models may be developed as built-in functional keys or calculators in electronic health systems to reduce cancer mortality.