Data and code for: Does MRI Radiomics Add to Tumor Location for Predicting Extent of Resection in Intracranial Meningioma?

Data and code accompanying the study "Does MRI Radiomics Add to Tumor Location for Predicting Extent of Resection in Intracranial Meningioma?" (submitted to Cancer Imaging). The study retrospectively evaluated, in 236 surgically treated intracranial meningiomas with complete preoperative MRI, whether radiomics adds predictive value beyond tumor location for gross-total versus subtotal resection. Tumor location was the strongest available preoperative correlate (multivariable odds ratio 4.2 per tier; AUC ≈ 0.70); radiomics from the enhancing tumor and peritumoral edema achieved comparable but not additive discrimination. This archive contains the de-identified radiomic feature matrix (236 subjects, keyed by a study pseudonym), the pre-ComBat feature matrix used to reproduce the leakage-controlled in-fold ComBat harmonization, supporting scanner and volume tables, the full analysis code (in-fold harmonization, repeated nested cross-validation, DeLong and patient-level bootstrap inference, sensitivity analyses, the location-and-age nomogram, and figure scripts), the result tables reported in the paper, a data dictionary, and a pinned software environment. Reproduction starts from the provided feature matrices (see README); the underlying MR images are not included owing to patient-privacy regulations, and the archive contains no protected health information. Code is released under the MIT License and the data under CC-BY-4.0.