Hepatic Fat Quantification Using Beta Distribution and a Probabilistic Neural Network in a Prepubertal Male Cohort.

Area of Science:

• Radiology
• Medical Imaging
• Machine Learning in Medicine

Background:

• High liver fat content (hepatic steatosis) is linked to serious health issues like cancer, diabetes, and cardiovascular disease.
• Accurate quantification of hepatic steatosis is crucial for patient management, but current MRI methods have limitations.
• Existing MRI approaches can suffer from data variability and lack uncertainty estimates for borderline cases.

Purpose of the Study:

• To develop and validate a probabilistic approach for precise hepatic steatosis quantification.
• To combine a neural network with a beta distribution for predicting hepatic fat percentage with confidence intervals.
• To improve the accuracy and interpretability of MRI-based hepatic fat assessment.

Main Methods:

• Utilized single in-phase Dixon MRI liver images from 84 prepubertal males.
• Employed a probabilistic neural network integrated with a beta distribution framework.
• Estimated hepatic fat content and associated confidence intervals, comparing predictions to MRI-derived ground truth.

Main Results:

• The probabilistic method demonstrated low prediction error and strong agreement with ground truth.
• Achieved a mean absolute error (MAE) of 0.44 percentage points and R² of 0.98.
• Showcased good performance on the test set with an empirical standard deviation of 0.0609.

Conclusions:

• The probabilistic framework offers an interpretable measure of variability alongside point estimates for hepatic steatosis.
• This uncertainty quantification enhances the clinical utility of MRI-based fat estimation.
• Further validation in diverse populations is recommended for broader applicability.