Deep learning can predict PNPLA3 I148M homozygous carriers in SLD patients based on MRIs from UK Biobank

Background: Steatotic liver disease (SLD) is the most common liver disease worldwide, affecting 30% of the global population, which is strongly associated with genetic factors. However, identification of variant carriers is not part of routine clinical care and required infrastructure for genetic testing.

Methods: We analyzed MRI images and genetic variants PNPLA3 I148M, TM6SF2 rs58542926_T, MTARC1 rs2642438_A, HSD17B13 rs72613567_T and GCKR rs1260326_T with 45,603 individuals from the UK Biobank. Proton density fat fraction (PDFF) values were derived by using deep learning models and water fat separation toolbox. Individuals with (PDFF≥5%) and without SLD (PDFF<5%) were used to train and test a Vision Transformer classification model with five-fold cross validation, respectively.

Results: The predictive performance was generally higher in SLD group. Homozygosity for the PNPLA3 I148M variant demonstrated the best predictive performance among five variants with AUROC of 0.68 (95% CI: 0.64-0.73) in SLD group. The AUROC for predicting PNPLA3 I148M was higher in females (0.61-0.71) and younger individuals (0.64-0.69). Additionally, attention maps for PNPLA3 I148M carriers showed that fat deposition in regions adjacent to the hepatic vessels, near the liver hilum, plays an important role in predicting the presence of the I148M variant.

Conclusion: Our study marks progress in the non-invasive detection of homozygosity for PNPLA3 I148M through the application of deep learning models on MRI images. Our findings suggest that PNPLA3 I148M might affect the liver fat distribution and could be used to predict the presence of PNPLA3 variants in patients with fatty liver.

Publication History

Article published online:
20 January 2025