Application of the Doylestown algorithm for the early detection of hepatocellular carcinoma

• DOI: 10.1371/journal.pone.0203149
• Publication Date: 2018-08-31T17:27:51Z
• Authors: Anand S. Mehta; Daryl T.-Y. Lau; Mengjun Wang; Aysha Islam; Bilal Nasir; Asad Javaid; Mugilan Poongkunran; Timothy M. Block
• Keywords: Etiology; Liver Cancer

Background We previously developed a logistic regression algorithm that uses AFP, age, gender, ALK and ALT levels to improve the detection of hepatocellular carcinoma (HCC). In 3,158 patients from 5 independent sites, this algorithm, referred as "Doylestown" increased AUROC AFP 4% 12% had equal benefit regardless tumor size or etiology liver disease. Aims Analysis Doylestown using samples individuals taken before their diagnosis HCC. Methods Here, was tested at multiple time points (a) with established chronic disease, without HCC (120 patients) (b) 116 (85 early stage 31 recurrent HCC), of, up 12 months prior cancer diagnosis. Results Among who HCC, comparing fixed cut-off 20 ng/mL, True Positive Rate (TPR) in identification 36 50%, point conventional detection. Similar results were obtained those where TPR 18% 59%, recurrence. Conclusions This significantly improves prediction by alone may have value