Data and code for: Position-dependent bias and uncertainty in stiffness and viscosity estimates from renal shear wave elastography: a finite element study of focal fibrosis

Three regions are analysed: before the lesion (45 to 58 mm), inside it (59 to 81 mm), and beyond it (82 to 90 mm). The tissue is nearly incompressible, with a density of 1000 kg/m3 in the background and 1050 kg/m3 in the lesion. Stiffness and viscosity are recovered three ways: time of flight from a windowed peak, time of flight from cross-correlation, and shear wave dispersion fitted to a Kelvin-Voigt model.

Noise is then added and the dispersion analysis repeated 200 times at each of 0.5, 1 and 2 per cent of the peak velocity, from a fixed seed. That is 9000 fits for the focal model and 3000 for the diffuse one. The deposit also holds the checks behind the analysis: a linear elastic sweep over seven stiffness values, dispersion in two gel phantoms, a mesh convergence study, a peak-detection window check, a ridge threshold sweep, and a control run in uniform tissue that fixes the level below which regional differences cannot be put down to a lesion.

Contents: code/ (25 scripts) - the three recovery methods, the noise sweeps, the validation and control scripts, and the figure scripts. data/ (16 files) - the COMSOL exports for both models at every stage, the probe-table export, the linear elastic sweep, and the mesh convergence and phantom data. output/ (44 files) - the result tables as CSV, the figures as PDF and PNG, and the noise sweep log. run_all.m - runs everything in the right order.

README.md - requirements, run order, the data file layouts, and what produces each table and figure. Every table and figure in the paper comes from a named script listed in README.md, and every reported number can be traced to a CSV in output/. The scripts need MATLAB R2023b with the Optimization Toolbox and the Signal Processing Toolbox.

One figure is also given as a Python script. The finite element solution was produced in COMSOL Multiphysics 6.2.0.415; the model file is not included. What the data show: at the last CKD stage, recovered stiffness runs from 1.6 per cent below to 29.5 per cent above the assigned value, depending only on where the measurement is taken and which method is used.

Beyond the lesion the dispersion fit stays repeatable but reads about 90 per cent low.