IPD-Brain: An Indian histopathology dataset for glioma subtype classification

Area of Science:

• Neuropathology
• Medical Imaging
• Computational Pathology

Background:

• Accurate brain tumor classification is essential for effective patient management.
• Existing datasets may lack diversity in terms of demographics and comprehensive annotations.

Purpose of the Study:

• To introduce the IPD-Brain Dataset, a large, high-resolution neuropathological resource for glioma research.
• To facilitate research on glioma subtypes, immunohistochemical biomarkers, and ethnic variations in brain tumors.
• To support machine learning applications in neuropathology.

Main Methods:

• Compilation of 547 high-resolution H&E stained whole slide images from 367 patients.
• Inclusion of detailed clinical and histopathological annotations, including CNS WHO grade and IHC biomarker status (IDH1R132H, ATRX, TP53, Ki67).
• Dataset scanned at 40x magnification, focusing on Indian demographics.

Main Results:

• The IPD-Brain Dataset is one of the largest of its kind in Asia, with comprehensive data.
• Preliminary validation using Multiple Instance Learning demonstrated potential for glioma subtype classification and biomarker identification.
• The dataset provides a foundation for exploring regional and ethnic variations in glioma.

Conclusions:

• The IPD-Brain Dataset is a valuable open-access resource for the neuropathological community.
• It has the potential to significantly advance brain tumor research, diagnostic precision, and global collaboration.
• Further research using this dataset can enhance understanding of glioma heterogeneity.