2026_StellatoEtAl_Dataset

ARTICLE (when using these files, please cite the following article):Mariachiara Stellato, ..., and Alessandro Stefano. MatRad-OIDS: an open-source tool for segmenting microscopy 3D images and extracting IBSI-compliant radiomic features, 2026.ABSTRACT OF THE ARTICLE:Radiomic analyses of 3D multicellular cultures, such as spheroids and organoids, are becoming increasingly popular. The list of features that can be extracted from 3D images, alongside their specific mathematical equations, has been validated since 2016 by the International Image Biomarker Standardisation Initiative (IBSI) commission. The aim of this work is to provide an integrated, user-friendly, and reproducible framework for radiomic analysis of 3D cell culture images. By bridging the gap between image acquisition, segmentation, feature extraction, dimensionality reduction, and machine learning, our tool enables both imaging specialists and non-expert users to generate high-quality, IBSI-compliant radiomic data and to explore its predictive potential in preclinical and translational research. We propose MatRad-OIDS, an open-source tool written in MATLAB to analyse z-stack images of 3D cell cultures acquired with a confocal or light-sheet fluorescence microscope. MatRad-OIDS allows users to: (I) visualise multilayer 3D TIF/TIFF; (II) segment the single slices according to different manual, semi- and fully automatic modalities; (III) compute and (IV) reduce IBSI-compliant features for further radiomic analyses; (V) train classification networks tailored to the input data. MatRad-OIDS segmentations have been validated using different microscopy image z-stacks, whilst extracted features have been compared with IBSI-compliant references. Segmentation masks, extracted and reduced features, and trained models are saved alongside the user-defined parameters to ensure full repeatability of the studies. MatRad-OIDS MATLAB source code, standalone Windows, MacOS and Linux applications, video tutorial, user documentation, and sample datasets are publicly available at: https://github.com/research-giovap/MatRad-OIDS/. The source code requires Python version 3.9 and MATLAB version 2024a.DESCRIPTION OF THE FILES IN THE COLLECTION:Files, images and data used in the above-mentioned scientific work.Single_Cells: The "Single_Cells" folder contains 3 8-bit, no LUT, z-stack composed of 67 slices, referring to 3 single mesenchymal stem cell acquired with a confocal microscope (Cell1_8bitNoLUT, Cell2_8bitNoLUT, Cell2_8bitNoLUT); The manually segmented 3D mask of the z-stacks (Cell1_Masks_GT, Cell2_Masks_GT, Cell3_Masks_GT); The automatic segmentation 3D masks obtained using MatRad-OIDS (Cell1_Masks_MatRad-OIDS, Cell2_Masks_MatRad-OIDS, Cell3_Masks_MatRad-OIDS).Spheroids: The "Spheroids" folder contains 3 8-bit, no LUT, z-stack composed of 262 slices, referring to 3 multicellular A549 lung cancer spheroid acquired with an LSFM (MA1_F05_ang1_8bitnoLUT); The manually segmented 3D mask of the z-stack (MA1_F05_ang1_Masks_GT, MA2_G09_ang1_8bitNoLUT, MA3_G04_ang1_8bitNoLUT); The automatic segmentation 3D mask obtained using MatRad-OIDS (MA1_F05_ang1_Masks_MatRad-OIDS, MA2_G09_ang1_Masks_MatRad-OIDS, MA3_G04_ang1_Masks_MatRad-OIDS).