By Jonathan Bridges, chief innovation officer at Exponential-e.
Artificial intelligence (AI) is set to change the field of pathology, transforming how diseases are detected and diagnosed. With its ability to enhance efficiency, accuracy, and the speed of screening, AI can significantly improve the detection of conditions such as breast, bowel, and cervical cancer.
As pathology departments across the UK embrace digital slide processing, integrating AI into these workflows presents a substantial opportunity to modernise diagnostic processes.
However, alongside its benefits, implementing AI-powered pathology comes with security and governance challenges that must be addressed to ensure safe and effective deployment.
The healthcare sector is already witnessing AI’s transformative impact, from sophisticated language models to pioneering research tools.
In pathology, AI’s potential lies in its ability to analyse vast datasets and recognise patterns within tissue samples – providing a powerful way to identify diseases quickly and more accurately than ever before.
Traditional methods relying on glass slides are costly, time-consuming, and introduce security risks, sometimes making it difficult for pathologists to effectively collaborate. AI’s ability to assist pathologists opens the door to faster, more precise results, transforming routine analysis and tackling complex diagnostic tasks.
With the NHS aiming for a 75 per cent cancer diagnosis rate within 28 days, the current delays, which are affecting tens of thousands of patients, highlight the need to embrace AI solutions.
This urgency is further highlighted by the demographics of the pathology workforce in the UK, where a significant number of pathologists are over the age of 55 and 60 per cent are approaching retirement age. Enhancing diagnostic capabilities and streamlining workflows through AI integration is one way to meet these challenges and maintain quality care.
Enhancing diagnosis through AI
One of the most promising applications of AI in pathology is the analysis of histology samples to detect disease patterns based on specific biomarkers.
By automating the analysis of these samples, AI can enable earlier and more accurate diagnoses, reducing the time it takes from biopsy to results.
AI-driven image analysis algorithms have the potential to automate time-consuming tasks, allowing laboratories to operate more efficiently and focus on higher-value activities.
For patients, this could mean quicker access to treatment options, reduced financial impact, and, ultimately, life-saving interventions.
Despite these promising benefits, the adoption of AI in pathology has been slow, mainly due to the costs associated with implementing, calibrating, and maintaining the technology.
The digitisation of pathology is still in its infancy, but it presents a significant opportunity to accelerate diagnostic timelines and improve patient outcomes. The path forward involves embracing AI while addressing the hurdles that accompany its adoption.
The security and governance challenge
A major consideration for AI deployment in pathology is how to securely and efficiently manage data processing. Centralising AI capabilities within a dedicated data centre is generally more secure than distributing data across multiple external systems and suppliers.
This approach reduces security risks and ensures better control over data governance and business processes. In contrast, a distributed model requires careful selection of AI suppliers and a robust ecosystem management strategy to oversee various agreements and data processing protocols.
The sensitive nature of pathology data, combined with the unique workflows in each department, makes it crucial to integrate AI systems in a way that does not compromise security.
For example, sending images via secure links to external AI platforms introduces potential vulnerabilities, unless these interactions are rigorously managed.
By centralising AI within a controlled data environment, pathology departments can more effectively enforce data governance and security protocols, which reduces the risk of security breaches.
The importance of robust data security in AI adoption has been highlighted by recent cyber-attacks targeting NHS pathology testing organisations. These incidents not only disrupted operations but also caused delays in outpatient appointments, illustrating the impact that security breaches can have on healthcare delivery. Maintaining control over AI applications and data within a centralised system can mitigate such threats, ensuring that AI’s integration into pathology enhances rather than compromises patient care.
Working hand in hand with pathologists
It’s important to emphasise that AI is not intended to replace human pathologists. Instead, its role is to expand their capabilities by serving as a ‘second pair of eyes’ that automates the identification of specific biomarkers and analyses large volumes of slides.
By rapidly identifying and prioritising cases, AI can support pathologists in making more informed decisions, providing second opinions. Ultimately, enabling them to concentrate on complex diagnostic challenges that require nuanced expertise.
Additionally, with many pathologists nearing retirement, AI can play a role in addressing workforce shortages. By streamlining routine tasks and allowing pathologists to focus on high-priority cases, AI can take some of the pressure off understaffed departments and could help bridge the gap between the current pathology workforce and growing patient demand.
AI stands ready to play a role in reshaping pathology.
However, achieving this transformation requires a comprehensive approach that balances technological innovation with robust security measures and thoughtful governance. By embracing these principles, pathology departments can safely and effectively integrate AI into their diagnostic workflows, driving better accuracy, operational efficiency, and ultimately, improved patient care.