Nonalcoholic fatty liver disease (NAFLD) affects 1 in 4 adults and may modulate drug metabolism, increasing likelihood of drug induced liver injury (DILI). We aim to develop a tool to predict NAFLD in trial volunteers. Here we report our interim results in healthy and patient volunteers (NCT04873258).Volunteers undergo bioimpedance vector analysis, anthropometric measurement (BMI, waist circumference), and lab bloods. A FibroScan is performed and E (fibrosis) and CAP (steatosis) scores recorded. Multivariable logistic regression is performed with ‘significant’ disease level set at E 7.5kPa and CAP 260dB. Individuals were identified as ‘healthy volunteers’ (HV) or ‘patients’ (PV) according to their cardiometabolic past medical history. Result:In the HV group (n=111), waist circumference (WC) was an accurate predictor of steatosis (AUROC 0.86). GGT was the strongest predictor of fibrosis, but this did not perform as well (AUROC 0.55).In the PV group (n=119) WC was a significant predictor of steatosis, with reasonable accuracy (AUROC 0.75). WC was a predictor for the presence of fibrosis (AUROC 0.67). Nomograms were produced (Fig 1,2,3).We found BMI performed less well compared to WC in HV group for predicting steatosis (AUROC 0.75 vs 0.86) and equally poor vs GGT for predicting fibrosis (AUROC 0.58 vs 0.55). In PVs performance between BMI and waist circumference was similar for steatosis (AUROC 0.74 vs 0.75) and fibrosis prediction (0.65 vs 0.67).Conclusion: We have produced prototype tools for the prediction of steatosis and fibrosis amongst patient volunteers. Accurate detection of NAFLD is difficult and may cast suspicion of potential for DILI on a new trial medicine. When compared to current standards, waist circumference outperformed BMI in most analyses, suggesting this may be a better tool for screening than traditional BMI, without study budget or operational burden. Further recruitment will improve these tools for deployment in clinical trials.