Introduction Screening for liver fibrosis continues to rely on laboratory
panels and non-invasive tests such as FIB-4-score and transient elastography. In
this study, we evaluated the potential of machine learning (ML) methods to predict
liver steatosis on abdominal ultrasound and liver fibrosis in individuals
participating in a screening program for colorectal cancer.
Methods We performed ultrasound on 5834 patients admitted between 2006 and
2020, and transient elastography on a subset of 1240 patients. Steatosis on
ultrasound was diagnosed if liver areas showed a significantly increased
echogenicity compared to the renal parenchyma. Liver fibrosis was defined as a liver
stiffness measurement≥8kPa in transient elastography. Extreme gradient
boosting (XGBoost) algorithms were prospectively evaluated for prediction.
Results The mean age was 58±9 years with 3036 males (52%), and
77% suffered from metabolic syndrome. Modelling laboratory parameters,
clinical parameters, and data on eight food types/dietary patterns, good
accuracy in predicting liver steatosis on ultrasound (AUC-ROC 0.87) moderate
accuracy in predicting liver fibrosis with XGBoost (AUC-ROC of 0.71) could be
achieved. Limiting variables to non-self-reported (non-subjective) variables did not
significantly alter performance. Gender-specific analyses showed significantly
higher performance in male (AUC-ROC 0.70) compared to females (AUC-ROC 0.60) in
predicting liver fibrosis.
Conclusion ML based on point-prevalence laboratory and clinical information
predicts liver steatosis with high and liver fibrosis with moderate accuracy. It is
conceivable that a model that includes parameters at different time points might
perform better. The observed gender differences suggest the need to develop
sex-specific models.