Ready for Lift Off: Harnessing AI for Better Patient Care in Renal Pathology

​In the evolving landscape of health care, the Innovarium Launchpad offers essential seed funding for innovative ideas that can enhance patient outcomes. One of this year’s recipients is focused on rethinking renal pathology reporting using advanced artificial intelligence​; this project is led by Dr. Maziar Riazy MD PhD FRCPC, Clinical Associate Professor, and Mahdi Mobini, PHSA technical consultant.  

A Leap Forward in Report Accuracy and Efficiency 

Renal pathology reports inform clinicians about biopsy findings, which go on to inform the important decisions in patient care. Traditionally, these reports are dictated by pathologists and transcribed by transcriptionists, a process prone to typos and inconsistencies – human error is commonplace after all. However, small mistakes can lead to potential discrepancies in patient care and increased workload for health care professionals. Moreover, there are repetitive tasks during report generation that lend themselves well to automation.  ​

Enter the Multi-modal Language Models: this project leverages Language Models to automate the initial drafting of renal pathology reports. “Our goal is to reduce typographical errors, enhance consistency, and streamline workflow,” explains Dr. Riazy. “By automating the drafting process, we save time for pathologists and allow administrative staff to focus on other critical tasks.” 

The Future of Renal Pathology Reporting 

The use of AI not only promises to enhance report consistency and quality—ensuring all necessary elements are included and free of unnecessary information—but is poised to improve efficiency overall. By minimizing the time spent on report generation, pathologists will be able to fully focus on their complex cases, ultimately leading to better patient care and informed decision-making. 

 Currently, the project is in the cloud infrastructure design phase, using private language models and secure data storage. This framework ensures that all data remains confidential while enabling the scalable compute power for fine-tuning and hosting the solution to meet the unique preferences of individual pathologists. 

Envisioning the Future 

To gauge the success of the project, the team has established key performance indicators (KPIs) focusing on report quality, internal consistency, production time, and revision rates. These metrics will help assess improvements and guide future enhancements. ​

In an ideal future there could be a seamless integration of AI-generated reports and digital biopsy images. After the first phase, the technology could evolve to generate reports directly from whole-slide scanned images of renal biopsies. This would allow the automation of quantitative and particularly repetitive tasks to help pathologists start their workdays with preliminary reports already available and spend more time on the challenging interpretative tasks, further streamlining workflows and enhancing overall productivity. 

A Commitment to Innovation and Collaboration

This project aims to optimize renal pathology reporting while fostering collaboration between pathologists and AI specialists. By encouraging a multidisciplinary approach, it aligns with the Ministry of Health’s objectives to enhance diagnostic processes and promote a culture of innovation in healthcare. As the use of AI in renal pathology continues to gain momentum, this initiative positions the BC Renal Pathology Laboratory, St. Paul's Hospital, and Providence Health Care at the forefront of technological advancement in patient care. 

 Looking ahead, Innovarium is committed to driving programs forward that will continue to expand and flourish at the new St. Paul's and the Clinical Support & Research Centre. Projects like these will be supported by state-of-the-art facilities and programming, fostering an environment where innovation and collaboration can thrive.    

The journey of innovation at PHC doesn’t stop here. We invite you to return soon to learn about our other two funding recipients from the Innovarium Launchpad, coming soon to Connect.