Digital Pathology and AI: An Evolving Conversation
Technology is quickly revolutionizing how healthcare understands diseases, both from a pharmaceutical and an academic lens, to help provide prompt identification and diagnosis of a disease. AI has been a leading technology which is prompting digital pathologists to see how they would use this in their day-to-day work.
Artificial Intelligence (AI)-based Digital Pathology — also known as computational pathology — is an emerging discipline and sub-specialty that promises to increase the accuracy and availability of high-quality healthcare to patients across many medical fields. While the future is bright for AI-based Digital Pathology, transitioning from the status quo to that future state presents both challenges and opportunities.
Slalom’s Healthcare and Life Sciences experts recently collaborated with leaders from the Digital Pathology Association in fields of research & development, healthcare, and technology for a roundtable discussion of the evolving role of Artificial Intelligence (AI).
Together we explored the current state of AI in Digital Pathology; the challenges and opportunities encountered by pathologists seeking to implement and/or leverage AI in their work; and, finally, their vision for the future. Here are a few high-level takeaways from our conversation:
The Current State of AI in Digital Pathology
- AI-based Digital Pathology currently lacks standardization across formats, protocols, systemic quality control, and workflows. This creates interoperability issues, increases potential for error, and can even hinder regulatory progress.
- Automated AI systems can integrate large amounts of complex data, but the legacy Laboratory Information Systems (LIS) and Picture Archiving and Communication Systems (PACS) found in hospitals today typically do not feature the technical capabilities needed to process AI data.
- The regulatory path for AI-based Digital Pathology is still being defined and ethics are still being considered, resulting in a lack of policies and regulations.
Challenges Associated with AI-based Digital Pathology
- Because policies and regulations are unclear, pathologists often fear personal liability and the legal implications associated with the possibility of AI-based diagnostic error.
- Pathology is not typically considered a technology department by C-Suite decision makers. This perception creates hurdles for funding the digital equipment and infrastructure needed to support AI-based Digital Pathology.
- Data ownership and patient privacy were also discussed by our thought leaders. For instance — in the case of pathologist consultations, which clinical site owns the data? And who bears the cost and accountability of maintaining the data pipeline, data transfer and secure storage?
Opportunities for AI-based Digital Pathology
- Clearly defining AI and clarifying whether it will be used for primary or secondary diagnosis present valuable opportunities to demystify the use of AI in the context of Digital Pathology.
- Establishing FDA-approved intended use boundaries would create the opportunity for AI to be treated like any other medical device — with a clear application, intended purpose, and specific performance.
- Developing minimum viable technical requirements and providing real-life use case examples enables the pathologist with an opportunity to define exactly what needs to be purchased, and why — which could help make a clearer business case for decision makers.
The Future of AI-based Digital Pathology
Although many technical and ethical challenges need to be addressed, enabling AI-based Digital Pathology to work effectively as a synergistic system will lead to improved workflows, and will enable clinical teams to share and analyze image data in a broader platform. Slalom had the following suggestions for helping organizations transition to the future state of AI-based Digital Pathology:
- Clarify that AI is a medical product, not just an algorithm. This distinction is key when it comes to considerations required for applying a regulatory framework and how we think about AI in the context of this specific area.
- Share lessons learned about enacting and enabling AI practices in digital pathology. Silos hinder the information sharing needed to make accelerated progress, standardization of best practices, and consensus needed to drive adoption.
- Take on the issue of health disparities and ensure that the population at large is being included in AI models and algorithms. All patients, regardless of their location or socioeconomic status, should benefit from AI-based Digital Pathology. This technology has the potential to improve access to and the distribution of insights.
- Help accelerate regulatory approval of AI through standardized risk classification. Having a consistent framework for evaluating the risks of AI-based digital pathology will accelerate the ability for the resulting products or treatments to go-to-market more effectively and improve patient outcomes.
One leader summarized their future vision as:
“The future of AI in digital pathology is not just about increasing the diagnostic accuracy of the pathologist. The future of AI is to refine predictable diagnostic tools for the clinician in service of the patient. This should be the focus of digital pathology.”
We have a ways to go in order to ensure organizations are prepared to utilize AI-based Digital Pathology in their daily workstreams. By ensuring appropriate use cases leverage AI-based technologies, establishing FDA boundaries, making the technology accessible, and educating and enabling the broader community, there is a bright future and opportunity to improve outcomes for all.
Meet the Authors:
Kamayani Gupta is Head of HCLS Research & Development and a Senior Principal at Slalom San Francisco. Find her here on LinkedIn.
Slalom is a modern consulting firm focused on strategy, technology and business transformation. Our healthcare and life sciences industry teams partner with healthcare, biotech and pharmaceutical leaders to strengthen their organizations, improve their systems, and help with some of their most strategic business challenges. Find out more about our people, our company and what we do.
