MRI-PDFF evaluation of frequency difference electrical impedance tomography for low-cost, non-invasive liver fat quantification

Motivation: To inform the development of low-cost wearable diagnostic tools for MASLD and its associated comorbidities. Goal(s): To use well-established MRI-PDFF to evaluate fd-EIT for the quantification of liver steatosis. Approach: MRI-PDFF and fd-EIT liver data were obtained from 26 volunteers and used to improve and validate postprocessing of our fd-EIT approach for liver steatosis assessment. The performance of fd-EIT and fd-EIT-based prediction models was evaluated against MRI-PDFF as reference. Results: EIT conductivity difference values showed moderate to good correlation (Pearson=0.45) with liver MRI-PDFF values. Demographic modeling further improved prediction performance in logistic and linear regression analysis with AUC = 0.89 and Pearson=0.78, respectively. Impact: fd-EIT could have a large impact on early diagnosis and monitoring of liver disease. Further development and validation in larger cohorts is needed to support the generalizability of our findings.