Modified Abstract:

Liver cancer represent a significant global health burden, with the hepatocellular carcinoma (HCC) type of liver cancer ranking among the leading causes of cancer-related mortality. The progression from healthy liver function to liver malignancy and cancer follows a well-established pathway through distinct stages, with Stage I representing lifestyle indicators that offer the greatest potential for prevention. Current diagnostic approaches often identify liver ailments leading to cancer, at advanced stages when therapeutic options are limited, emphasizing the need for early detection strategies.

This cross-sectional study analyzed lifestyle and clinical data from 1,700 participants aged 18-80 years to develop and validate a comprehensive risk assessment framework for Stage I liver disease detection to prevent the developing of cancers at later stage. Eleven primary features were assessed, including lifestyle factors (alcohol consumption, smoking status, physical activity, BMI category) and clinical indicators (diabetes and liver function tests).

Machine learning techniques, specifically Gradient Boosting with hyperparameter optimization, were employed to develop predictive models. The results acheived serve as input variables for the intermediate stage, where the subjects oundergo the diagnostics and theraupetics. For the patients at an advanced stage of liver disease, these Machine learning algorithms confirmed on the basis of the comprehensive risk assessment framework. The confirmation of the advancement of the liver disease to a new level that is close to liver tumor, the irreversible liver damage, on the basis of the output from stage 1, serve as variables, along with new liver tests results.

The encouraging results only serve as an indication of the early cause of tumor, only if the medical tests of the subjects confirm the liver tumor for these subjects and other subjects with similar or close to similar data including life style, medical tests and the medicine usage.