Blockchain-based information system for medical image transfer
The project aims to build a blockchain-based information system for medical image sharing between different entities, particularly facilitating image transfer to enable a data library of medical images for an AI/ML application to improve image processing, analysis, reconstruction, and enhancement. We plan to test the system with real image data and assess its performance from a socioeconomic perspective.
Conventional clinical data has been stored and managed in a fragmented manner, which creates frictions in information exchange at the point of care and hinders large-scaled research empowered by emerging technology such as AI/ML. Failure in timely access to health information could impede effective treatment decision-making, which will adversely affect patient health, and also incur unnecessary costs such as duplicated tests. Regulations on protecting patient privacy add a layer of complexity in data transfer. While the explosion in the number and capability of tools eases the process of data collection, data retrieval and information analysis have been slow and complicated in the field of medicine, which has become a global challenge faced by both the developed and developing countries. All the stakeholders have been calling for a method to securely and efficiently share healthcare data.
We propose a blockchain-based technology that can help address the above issues in the following three aspects: (1) improving security and protection of privacy, (2) maintaining flexibility, and (3) enforcing data sovereignty. Blockchain is a distributed system to store a series of time-stamped records in a decentralized network. It relies on “established cryptographic techniques to allow each participant in a network to interact without pre-existing trust between these parties.”
Medical image is a central part of diagnostics in today’s health care, as a diagnostic imaging service is “paramount in confirming, correctly assessing and documenting courses of many diseases as well as in assessing responses to treatment” (WHO 2019). The current imaging database and platforms that patient care and workflows are based on lend themselves naturally to the big data initiative and the application of AI/ML methodology. However, the integration of quality data becomes increasingly challenging but paramount in streamlining workflow and improving patient care. The project seeks to develop a distributed information system that helps to address these issues in image data transfer and provide a prototype for health information management in a broader context.
1. Develop a prototype system in IBM Hyperledger for image data sharing between various entities, from data acquisition, storage, and transportation between health care providers, patients, and the research community.
a. Develop privacy-preserving techniques for safeguarding the identity of the patients when acquiring the image data.
b. Develop novel algorithms for storing and retrieving image data facilitated by the blockchain.
2. Conduct a cost-effectiveness analysis on the proposed system, accounting for costs, and outcomes such as changes in provider workflow and healthcare quality.
3. The proposed working system is essential for a grant proposal, which will look very promising for sponsored research funds with the proposed work.
Ge Wang, Dept. of Biomedical Engineering, Rensselaer Polytechnic Institute
Lirong Xia, Dept. of Computer Science, Rensselaer Polytechnic Institute
Oshani W. Seneviratne, Rensselaer Polytechnic Institute