AI and Computational Pathology

FROM OPRHelath: I attended this week a webinar on Computational Pathology presented by Dr. Andrew Janowczyk, Assistant Professor at Emory University, and facilitated by Stanford’s Center for Artificial Intelligence in Medicine and Imaging (AIMI). Dr. Janowczyk “helped co-found and was elected secretary of the Swiss Digital Pathology Consortium (SDiPath). He leads the development of the HistoTools suite, a set of open-source computational pathology tools for quality control (histoqc.com), annotation (quickannotator.com), labeling (patchsorter.com), validation (cohortfinder.com). His newer research focuses on the tasks of predicting prognosis and therapy response.”

The webinar captured the challenges and promises of these innovative AI-based technologies. It was a fascinating presentation that illuminated the capabilities and limitations of artificial intelligence, i.e., AI is powerful, but it is not technical pixie dust. I appreciated the philosophy behind these systems, which aligned with our perspectives on innovative clinical technologies. Specifically, they are not designed and intended to replace pathologists, but rather to perform tasks that are tedious, time-consuming, and beyond practical human performance.

These are important considerations because AI is frequently framed as technologies that eliminate human engagement, reduce employment, and diminish valuable skill sets. Biobeat’s innovative technologies including sensors and AI-based systems follow a similar philosophy by empowering doctors with unprecedented insights without interfering with their clinical roles and doctor-patient relationships.

I had an opportunity to ask Dr. Janowczyk about medicine’s challenges in coping with and integrating innovative technologies. We shared concerns over clinical time frames and the pace of technological innovations. It’s an ongoing challenge without simple solutions. Overall, this was one of the best webinars I’ve attended on AI, technological innovation, and clinical applications. Thank you Dr. Janowczyk and Stanford AIMI for making it possible.