Innovative Tool Protects Patient Privacy in Pediatric Brain Tumor Data
Posted on
Progress in any field requires dedication, time, and effort, and this is especially true in pediatric brain tumor research. Recently, a project led by Dr’s Ariana M. Familiar and Ali Nabavizadeh at the Children’s Brain Tumor Network (CBTN) published an article in the American Journal of Neuroradiology (AJNR) Imaging Publication about the development of a new tool to de-identify pediatric brain MRIs. This tool helps researchers share data more easily, which is crucial for advancing our understanding and treatment of pediatric brain tumors.
The Importance of Data Sharing
In recent years, the importance of open data sharing has become increasingly apparent, especially for rare diseases like pediatric brain tumors. Sharing data allows researchers to study the diversity and complexity of these tumors, leading to a better understanding of how to treat them. Government agencies, such as the National Institutes of Health (NIH), now require researchers to share their data.
However, sharing MRI scans presents a unique challenge: these images contain identifiable facial information. To protect patient privacy, this information must be removed, a process called “de-identification” or “defacing.”
Existing tools for de-identification were not designed for the specific needs of pediatric brain tumor research. They often struggled with the unique anatomical features of children's brains and the variety of MRI scans used in clinical settings.
The Challenge of De-Identification
Brain MRIs contain sensitive information, including the patient’s face. To share these images, researchers need to de-identify them, which means removing any information that could be used to identify the patient. While there are tools to de-identify adult brain MRIs, they often don’t work well for pediatric images due to the anatomical differences between children and adults.
CBTN researchers developed a new tool using artificial intelligence (AI) to address this gap. This tool automatically removes facial features from pediatric brain MRI scans, allowing the images to be shared while safeguarding patient privacy.
A New Solution
To address this challenge, Dr. Familiar and her team developed a new tool that uses artificial intelligence (AI) to remove facial features from pediatric brain MRIs automatically. This tool is designed to work with the unique characteristics of pediatric images and can handle the diversity of image types collected in clinical settings.
How the Tool Works
The AI tool uses “deep learning” to identify and remove facial features in MRI scans. Researchers trained the AI using a large set of CBTN MRI images from the Children’s Brain Tumor Network database, ensuring it could accurately recognize and remove faces across various scans and patient ages.
Benefits of the New Tool
This new tool offers several benefits to researchers:
- It is specifically designed for pediatric brain MRIs, ensuring accurate de-identification.
- It can handle various image types, making it useful in clinical settings.
- It is open-source and free to use, making it widely accessible to the research community.
- It is computationally efficient and can be used on a local laptop.
- It does not influence the usability of the data in further analyses.
The Future of Pediatric Brain Tumor Research
This innovative tool has the potential to significantly impact pediatric brain tumor research by making it easier to share data. This, in turn, can accelerate the pace of discovery and lead to new and improved treatments for children with brain tumors.
CBTN is committed to advancing pediatric brain tumor research, and this publication is an example of the amazing science happening across the network.
Reference
Familiar, Ariana M., Khalili, Neda, et al. “Empowering Data Sharing in Neuroscience: A Deep Learning Deidentification Method for Pediatric Brain MRIs.” American Journal of Neuroradiology, vol. 46, no. 5, May 2025, pp. 964–972.