Medicine Meets Machine: The Emerging Role of AI in Healthcare

Medicine Meets Machine: The Emerging Role of AI in Healthcare

Tags: , , ,

BY: Alaa Youssef

 

Today, artificial intelligence (AI) is the big hype in science. It is shaping the world’s economy, education, health, and policy. AI, a broad term used to describe a wide range of machine learning approaches, is used to process large datasets to discern meaningful knowledge. While advances in AI bring infinite possibilities to understand inherently complex phenomena, these technologies pose new questions for society to address. This year, on May 7th, Raw Talk Podcast hosted ‘Medicine Meets Machine’, a live podcasting event at JLABS Toronto, attracting audiences and experts from diverse backgrounds to discuss the challenges AI poses to healthcare.

AI is transforming medicine in a myriad of ways by solving complex problems across a wide range of clinical domains. In this context, speakers in Raw Talk’s first panel discussion provided a glimpse into some of the ways AI is transforming medicine.  Dr. Oren Karus, Co-Founder Phenomic AI, described how advances in computer vision and deep learning techniques had transformed clinical imaging to a new realm, to enable image reconstruction, noise reduction, quality assurance, triage, segmentation, computer-aided detection, computer-aided classification, and radiogenomics. Notwithstanding these advances, there exist some challenges to harnessing the low hanging fruit AI affords. As Dr. Jason Lerch, Senior Scientist at the Hospital for Sick Children, highlighted “The ripe of AI in brain imaging and neuroscience research is yet to rise. While complex AI algorithms are slowly replacing classic models, the capacity of AI to elevate our understanding of complex imaging structures is yet to flourish as the process of data interpretability remains unknown.” Ongoing research efforts to understand the decision-making processes behind these complex algorithms promise to expand our understanding of complex and rare diseases to an unprecedented level.

Application of AI algorithms in other clinical domains, such as critical, neonatal, and geriatric care, promises to bridge wide existing gaps in healthcare systems. Growing evidence demonstrates that developing AI algorithms utilized in acute care outperform current clinical scoring systems, showing a higher level of accuracy monitoring and predicting changes in patients’ vitals. Thus, AI may offer practical solutions to many prevailing care delivery dilemmas, enhancing care efficiency and quality. Most importantly, as Dr. Marzyeh Ghassemi, Assistant Professor in the Department of Computer Science and Medicine, noted that “AI will transform medicine by allowing clinicians to focus on patient care” as these systems aid professionals with better evidence-based recommendations, and treatment options—augmenting clinician decision-making processes. Dr. Sunit Das, Neurosurgeon and Scientist at the St. Michael’s Hospital and The Hospital for Sick Children, noted that “AI is simply a tool to assist patients in making their decision…… and that the weight of that decision still lies in the communication between myself and my patient.”

The conversation in the second panel tackled some of the ethical and safety concerns AI may impose on society. Specifically, the discussion centered around the potential benefits and risks of using AI algorithms to leveraging accessible population health data. A key take away from this panel was that purposefully integrating AI systems, at the population level, might be the breakthrough to bridge the ‘Quality Chasm’ in healthcare, addressing the complexity of comorbid disease management and promoting preventative medicine. Conversely, the oblivious generalization of non-representative consented data could exacerbate societal health disparities and inequities. Thus, as Dr. Alison Paprica, Vice President of the Health Strategy and Partnerships at the Vector Institute put forth, harnessing existing population health data and mitigating societal ethical concerns will require  “engaging with the public in genuine ways [to translate] data into knowledge of what they value and what they agree with.”

The message from this event was clear–AI in medicine is to enhance evidence-based decision making to answer critical questions, ‘why’ and ‘what if’’, and not to replace experts. Thus, it is incumbent on us to foster interdisciplinary collaborations to enact AI systems safely and equitably to promote societal benefits and growth.