<p dir="ltr"><b>Dataset and Source Code for "Boundary-Resolved Multispectral Optical Heterogeneity Reconstruction in Tumor-Like Media: A Multiregion Boundary Element Methodology with Spectral Inversion and Sensitivity Analysis"</b></p>

Verification includes matched-model consistency tests, boundary-mesh refinement, Gaussian-quadrature refinement, finite-difference verification of the sensitivity operator, held-out detector prediction, and comparison with an independently implemented finite-element forward model. The reported matched-model reconstruction recovers the prescribed regional absorption spectra to numerical precision, while the independent FEM/BEM comparison reveals non-negligible forward-model discrepancy, emphasizing the distinction between code consistency and independent validation.</p><p dir="ltr">This Figshare deposit provides a <b>repository-ready reproducibility package</b> containing machine-readable numerical data, benchmark configuration files, reference Python modules for geometry, optical kernels and numerical metrics, scripts for reproducing reported figures and tables, verification datasets, unit tests, documentation, software-environment files, citation metadata, and checksum information.</p><p dir="ltr">The package is intended to support transparent numerical verification, reuse, and extension of the proposed methodology in <b>diffuse optical tomography, computational biophotonics, tissue optics, inverse problems, and interface-resolved optical modeling</b>.

It should be interpreted as a computational research resource rather than a clinically validated diagnostic or biopsy-replacement system.</p>