Delving deeper into AI and blockchain

We are all becoming used to hearing about AI and blockchain, two technology trends that go hand in hand – the brains and the brawn, the automation specialist and its security detail. While some of us might have begun to wrap our heads around what blockchain actually is by this point, here we ask Richard Norman and Vladimir Makarov to delve a little deeper into how these technology trends are beginning to impact life science and healthcare and how we can prepare to leverage these new tools ethically and effectively.

Q: How is blockchain already making a difference in patient data management and sharing and what is its future potential?

Richard: “There are many examples of applying blockchain technology in healthcare and to patient data management and sharing, all the way from proof-of-concept studies to pre-production and production solutions. When examining potential use cases, we must keep in mind the core uses of blockchain, i.e., that it provides a single source of truth, immutability and trust. Thus, blockchain technology can help manage the consent (and its revocation) given by a patient to share their data; be it medical records, or data from a clinical trial. In this space, Embleema are one of the companies pioneering the use of blockchain to drive this concept of a ‘data marketplace’ in which blockchain offers proof of data provenance and destination and puts individuals in control of access. I expect that, in future, patients will be far more conscious about the value of their identities and associated personal data and will assert that these are used solely based on their wishes and preferences. Blockchain will help to enable this and will become much more integrated with other technologies and applications such as Decentralised Identities, IoT devices, tokenisation, Machine Learning, etc.”

Q: What types of skills are core to the implementation of blockchain in life sciences and healthcare and how might skills gaps be filled?

Richard: “When implementing blockchain in life sciences and healthcare there are four key groups of skills that must be fulfilled

1. Life science and healthcare understanding – especially awareness of the complex and varied regulatory environment
2. Identification and recruitment of expert third parties with the necessary technology expertise
3. Business to business relationship management
4. Project management

We must also note that while strides have been taken, blockchain implementation in healthcare and life sciences is behind other sectors, especially when compared to the finance industry. There are opportunities to learn from other industries, especially around standards, implementation and integration. It must also be considered whether blockchain is the ‘right’ technology for the problem or application at hand, and whether as an organisation you want to develop in-house knowledge and expertise or just bring in an off-the-shelf solution.

Finally, when looking to fill skills and knowledge gaps the focus should be on training and knowledge transfer. To this end, the Pistoia Alliance is running a cross-disciplinary project to investigate the use of blockchain in the clinical trial informed consent process, and is developing a proof of concept.”

Q: What is required and who is responsible for ensuring standards are in place to cover the use of blockchain and AI across these sectors?

Richard: “The industry itself is ultimately responsible for its choice of tools and choice of standards, and where necessary, their implementation in a regulatory compliant manner. Not-for-profit organisations like the Pistoia Alliance are good candidates to develop standards, because they bring together relevant players from the industry to define standards that work for companies of all sizes.

However, when a technology is developing rapidly, standards tend not to be the first objective. When a technology starts to mature and use cases are defined, the opportunity for standards emerges. Such standards can bring enhanced efficiencies, and effectiveness. For blockchain specifically, organisations like W3C and IEEE are good candidates, along with large projects with dedicated funding like the IMI PharmaLedger project.”

Vladimir: “In the fast-growing field of AI, standards are emerging right now. It would be advantageous if the scientific community comes together to agree on what makes for suitable standards, before too many different initiatives are set up. The Pistoia Alliance supports the process of development and adoption of standards and best practices, using the FAIR Implementation Toolkit, Best Practices in AI and ML, and standards for reporting of bioassay protocols.”

Q: How can AI support faster R&D?

Vladimir: “The opportunities for AI to enhance R&D are legion – from ab initio drug design, to optimising compound synthesis. It also helps to find the right principal investigators with the right patients, with the right inclusion criteria and in the right locations, to support the processing of adverse events monitoring. Robotic Process Automation (RPA) helps here. Think how adverse event systems will be straining under the extra workload imposed by the worldwide vaccine rollout programmes. AI and automation also help in the analysis of real-world-evidence, from understanding the effectiveness of a drug in a real world setting like social media, to the analysis of genome sequencing. AI is particularly suited to automating time-consuming, low-value tasks, allowing drug developers to use their time on creative work that humans excel at. A good example of this approach is Project BRADSHAW at GSK, a Pistoia Alliance member company. BRADSHAW is a system for automated molecular design that integrates methods for chemical structure generation, experimental design, active learning, and cheminformatics tools.

Q: How does blockchain support computational drug repurposing and are there any other specific uses of this technology in the pipeline?

Richard: “Though it is early days for the use of blockchain in drug repurposing, there are opportunities to apply the technology to assist with traceability. This means that scientists working on repurposing projects can more quickly identify discovery data associated with a particular molecule, ascertain ownership of the IP, and establish if and when a molecule has been licensed. Blockchain’s immutability supports computational drug repurposing because it offers assurance of data integrity. This gives scientists a greater degree of confidence that (discovery) data for a particular molecule under study has not been tampered with. This is important when conducting a molecule data audit (an essential step in repurposing) – so that a scientist can differentiate where they have generated additional data between what are old and new data.”

Q: Please can you provide examples where computational drug repurposing has been implemented successfully?

Vladimir: “An example is the application of deep data mining for drug repurposing that was undertaken in a repurposing datathon organised by the Pistoia Alliance and Elsevier, in collaboration with Cures Within Reach and Mission:Cure. Our scientists were able to discover four drug candidates for repurposing within one month that have potential to become treatments for chronic pancreatitis – a rare disease afflicting around 1 million people globally – which currently doesn’t have an approved treatment.”