Overview

Schema Version

https://atlas.rsna.org/schemas/2025-11/dataset.json

Name

Africa Neuroimaging Archive (AfNiA)

Link

https://pmc.ncbi.nlm.nih.gov/articles/PMC9344206

Indexing

Keywords: Africa Neuroimaging Archive, AfNiA, brain MRI, glioma segmentation, BraTS, sub-Saharan Africa, LMICs, global health

Content: IN, NR, MR, OI, HP

RadLex: RID11221, RID28917, RID10312, RID4026

SNOMED: 84757009, 115240006, 254956000, 393564001

Author(s)

Udunna C. Anazodo

Maruf Adewole

Farouk Dako

Organization(s)

McGill University, Montreal Neurological Institute (Montreal, Canada)

University of Lagos, Department of Radiation Biology, Radiotherapy and Radiodiagnosis (Lagos, Nigeria)

University of Pennsylvania, Department of Radiology (Philadelphia, USA)

Lacuna Fund

Brain Tumor Segmentation (BraTS) Challenge

Contact

Address correspondence to Farouk Dako (email as published): ude.nnepu.enicidemnnep@okad.kuoraf

Funding

Supported by the Lacuna Fund in Equity & Health (for AfNiA brain tumor data release).

Ethical review

Policies for ethical use are being developed to govern access to annotated data in lieu of region-specific data protection provisions; data preparation to be compliant with FAIR principles and conventional data privacy and protection regulations.

Comments

AfNiA is described as a publicly available repository of annotated brain MRI studies to enable generalizable AI solutions in sub-Saharan Africa; first data release will support glioma tumor segmentation/classification in partnership with BraTS, with future releases planned for pituitary adenoma, ischemic stroke, and epilepsy.

References

[1] Anazodo UC; Adewole M; Dako F. "AI for Population and Global Health in Radiology". Radiology: Artificial Intelligence. 2022-06-22. doi:10.1148/ryai.220107. PMID: 35923372. PMCID: PMC9344206.

[2] . "Announcing Awards for Health Datasets". Lacuna Fund. 2022-05-19. Available from: https://lacunafund.org/announcing-awards-for-health-datasets/

[3] Menze BH; Jakab A; Bauer S; et al.. "The Multimodal Brain Tumor Image Segmentation Benchmark (BRATS)". IEEE Trans Med Imaging. 2015. PMID: 25494501. PMCID: PMC4833122.

Dataset

Motivation

Enable generalizable AI solutions in LMICs by providing annotated brain MRI datasets with local data diversity; address infrastructure, data quality, policy, and capacity gaps.

Sampling

Data transfer process from a network of local clinics (including centers without PACS or internet) to enable participation.

Missing information

No specific counts, formats, or release dates provided in the article.

Relationships between instances

Imaging linked to ground truth labels for AI model validation and testing (e.g., tumor segmentation/classification).

Noise

Article notes that many SSA brain MRI studies have poor contrast, low resolution, and inherent noise.

External data

Repository aims to aggregate clinical brain MRI from a network of local clinics in sub-Saharan Africa; first release to enrich BraTS challenge data.

Confidentiality

Patient imaging data intended to be de-identified with controlled access policies for ethical use.

Re-identification

De-identification of personal information is planned as part of data preparation compliant with privacy regulations.

Sensitive data

Contains medical imaging data; policies for ethical use and access are being developed.