ABSTRACT
Background
Early detection of hepatocellular carcinoma (HCC) is crucial in providing more effective therapies. As routine laboratory variables are readily accessible, this study aimed to develop a simple non-invasive model for predicting hepatocellular cancer.
Methods
Two groups of patients were recruited: an estimation group (n = 300) and a validation group (n = 625). Each comprised two categories: hepatocellular cancer and liver cirrhosis. Logistic regression analyses and receiver operating characteristic (ROC) curves were used to develop and validate the HCC-Mark model comprising AFP, high-sensitivity C-reactive protein, albumin and platelet count. This model was tested in cancer patients classified by the Barcelona Clinic Liver Cancer (BCLC), Cancer of Liver Italian Program (CLIP) and Okuda systems, and was compared with other non-invasive models for predicting hepatocellular cancer.
Results
HCC-Mark produced a ROC AUC of 0.89 (95% CI 0.85–0.90) for discriminating hepatocellular carcinoma from liver cirrhosis in the estimation group and 0.90 (0.86–0.90) in the validation group (both p < 0.0001). This AUC exceeded all other models, that had AUCs from 0.41 to 0.81. AUCs of HCC-Mark for discriminating patients with a single focal lesion, absent macrovascular invasion, tumour size <2 cm, BCLC (0-A), CLIP (0–1) and Okuda (stage Ι) from cirrhotic patients were 0.88 (0.85–0.90), 0.87 (0.85–0.89), 0.89 (0.85–0.93), 0.87 (0.84–0.89), 0.85 (0.82–0.87) and 0.86 (0.83–0.89), respectively (all p < 0.0001).
Conclusion
HCC-Mark is an accurate and validated model for the detection of hepatocellular cancer and certain of its clinical features.
Disclosure statement
The authors declare no conflict of interest.