STEDIB de-identified participant-level dataset for liver-score and multiparametric ultrasound analyses in type 2 diabetes

STEDIB (Biomarkers of Steatohepatitis in Type 2 Diabetes Patients) public analysis dataset. This dataset contains de-identified participant-level data from 40 adults with type 2 diabetes mellitus and suspected liver involvement included in the STEDIB observational cohort. The dataset underlies the statistical analyses reported in the associated Scientific Reports manuscript evaluating blood-based liver fibrosis and steatosis scores, bioimpedance-derived body composition, selected circulating biomarkers, and multiparametric liver ultrasound phenotypes. Ultrasound outcomes include two-dimensional shear-wave elastography (2D-SWE), shear-wave dispersion (SWD), and attenuation imaging (ATI). The dataset contains the source variables required to calculate the 17 liver scores evaluated in the study and includes three exploratory circulating biomarkers: irisin, FABP2, and soluble CD14. IL-6 is not included in the finalized analysis. Multiparametric liver ultrasound and blood sampling were completed in all 40 analyzed participants. Standardized InBody assessment was available in 38 participants; two participants did not undergo InBody assessment because fasting conditions were not maintained. Names, personal identification numbers, dates of birth, examination dates, addresses, contact information, and the participant linkage key are not included in this public dataset in order to minimize the risk of re-identification. The accompanying reproducibility code is archived separately in Zenodo and reproduces the statistical analyses, multiple-testing corrections, liver-score calculations, and the score-ultrasound correlation heatmap reported in the manuscript.