“AI explodes: Taking the Pulse of artificial intelligence in medicine,” is the title of the latest issue of the Stanford Medicine Magazine. The issue includes a number of articles that explore the promise of AI in medicine, ranging from What’s driving this tumor, — which discusses the efforts to stymie breast cancer through gene testing and AI, — to Adding ethics to the mix when developing health care AI, — which deals with the potential hazards posed by algorithms developed by AI experts with no formal training in the ethical standards for treating patients.
“Imagine a future where your doctor has an AI medical assistant by their side – distilling, in seconds, a world’s worth of medical research into a personalized treatment plan for you,” wrote Dr. Lloyd Minor, Dean of the Stanford University School of Medicine in the issue’s Letter from the Dean. “What if, at the click of a button, a researcher could design a custom molecule with the potential to treat a previously untreatable disease? With artificial intelligence’s rapid emergence, we are barreling toward this reality. Academic medical centers around the world, including Stanford Medicine, have begun investigating how AI, including large language models such as ChatGPT, can help us improve patient care, reduce clinician workload, better understand complex biological systems and accelerate drug discovery.”
As I discussed in a recent blog, “The Promise of Generative AI in Healthcare,” there’s a lot of excitement that AI might help us address our highly complex healthcare industry. Healthcare is a system of coupled systems, encompassing medical and pharmaceutical research; the delivery of healthcare to patients by a variety of practitioners, including hospitals, physicians, nurses, and pharmacists; and the insurance companies and governments that pay for healthcare.
According to Wikipedia, the healthcare industry comprises over 10% of the GDP of most developed countries. “The per capita expenditure on health and pharmaceuticals in OECD countries has steadily grown from a couple of hundred in the 1970s to an average of US$4'000 per year in current purchasing power parities.” In 2022, US healthcare spending “grew 4.1% to $4.5 trillion in 2022, or $13,493 per person, and accounted for 17.3% of Gross Domestic Product (GDP).”
In addition to its inherent complexity, healthcare has several attributes that make the successful deployment of new technologies significantly more challenging than in other industries. HIPPA, for example, protects the privacy of individuals in the US, but it also restricts the data sharing that’s essential to achieving many of the benefits of generative AI. And because the stakes in healthcare are too high to tolerate flaws that could harm patients, it’s harder to introduce new technologies and applications compared to most other industries.
“Medicine’s AI Boom,” the lead article in the “AI explodes issue, is a good overview of the current efforts at Stanford Medicine. To begin with, Dr. Nigam Shah, professor of medicine and chief data scientist at Stanford Health Care, reminds us that the current AI boom is nothing new. AI was born in the mid-1950s as a promising new academic discipline that aimed to develop intelligent machines capable of handling human-like tasks like natural language and playing chess. The Stanford AI Lab was founded in 1963 when computer scientist John McCarthy moved from MIT to Stanford.
AI became one of the most exciting areas in computer sciences over the next couple of decades, but after years of unfulfilled promises and hype, a so called AI winter of reduced interest and funding ensued in the 1980s that nearly killed the field.
Shah described our current AI explosion as “a moment of both high frenzy and immense opportunity, with a venture-capital-fueled rush to deliver applications with lasting value — a goal he estimates only 5% to 10% of the applications are hitting in today’s influencer-inspired culture that seeks ‘breakthroughs every 24 hours.’ Through that haze of ambition, real-world innovations in medicine are emerging. They are just more subtle than sexy; more incremental than game-changing. … It’s a historic opportunity, and it raises questions about how to use AI in medicine responsibly, how to set realistic expectations for its potential and what part the humans behind the algorithms will play.”
“Fears about AI are real, particularly in the sensitive world of health care,” the article adds. “Could these new algorithms compound existing challenges such as bias in how people are treated based on their race and loss of privacy due to health data breaches? Could they ratchet up distrust in the health care system and those who provide care?” To address these fears, in June of 2023 the Stanford Institute for Human-Centered AI and the School of Medicine announced the launch of RAISE-Health (Responsible AI for Safe and Equitable Health), a pioneering initiative to bring together a diverse set of voices, — including medical ethicists, AI technologists, and social scientists, with the mission to “guide the responsible use of AI across biomedical research, education, and patient care.”
According to Shah, generative AI “will not factor prominently into medicine and healthcare any time soon because of the concerns around accuracy and trust. Nor, he said, will 90% of the algorithms being produced today. Ten years from now, we’ll be immensely grateful for the 10% that panned out and changed the science, the practice, or the delivery of care in medicine.”
The article mentions a number of promising AI research projects being pursued by member of the Stanford Medicine faculty. Let me summarize a few of these efforts.
Imaging is one of the areas where AI has shown the most promise over the past several years. “In radiology, imaging technologies such as X-ray and MRI are used to diagnose patients, and the field has produced one of the few robust and consistent datasets in medicine. … Of the 500-plus AI algorithms approved by the FDA, 75% are radiology-focused and 85% are imaging-focused. At Stanford Medicine, AI innovation skews heavily toward imaging as well.”
In 2018 Stanford established the Center for Artificial Intelligence in Medicine and Imaging (AIMI), bringing together 50 faculty across 20 departments to conduct research on the use of AI techniques in clinically important imaging problems that would benefit patients. “AI can be, in some ways, superhuman because of its ability to link disparate data sources,” said Dr. Curtis Langlotz, professor of medicine and radiology and director of AIMI.
Dr. Langlotz’ lab has developed AI algorithms to detect and classify disease on medical images by finding linkages between imaging and genomic information that humans couldn’t possibly make. More recently, his lab developed natural language processing methods that analyze the clinical reports from radiologists to create annotated image training sets in order to provide real-time decision support systems to help radiologists improve accuracy and reduce errors.
Dr. Euan Ashley, founder and director of the Stanford Center of Inherited Cardiovascular Disease, and of the Stanford Clinical Genomics Program, has been conducting research on AI-based algorithms at the intersection of genetics and cardiology in order to improve the detection of cardiac risk. His Ashley Lab is focused on the science of precision medicine, that is, on using genomics information and computational techniques like AI to tailor medical decisions and treatment based on the individual characteristics of each patients instead of a one-drug-fits-all model. The key challenge, he said, is to now conduct the necessary clinical trials, and if successful, deploy AI-based personalized algorithms in actual medical practice.
Christina Curtis, — professor of medicine, genetics, and biomedical data science, — has been conducting research on new ways to prevent, diagnose, and treat breast cancer. Her lab has been analyzing the molecular profiles of tumor samples and pathology images, to figure out how to integrate detailed genomic data that would enable clinicians to personalize the best treatment for each breast cancer patient.
“Currently, most cancer patients undergo sequencing only once they’ve developed treatment-resistant metastatic disease,” she said. “There is a missed opportunity to have such information earlier in the disease course, at the time of initial diagnosis, both to compare a given patient to other similar patients and to monitor how the disease changes over time. This could enable more precise and anticipatory care.”
“As we stand on the cusp of this revolution and imagine how our lives and roles will change, I believe that AI’s impact will rival that of some of the most transformative innovations of human history, including the printing press and the internet,” wrote Dr. Minor in his Letter from the Dean. “Progress will not be linear, but initiatives such as RAISE-Health will be critical in establishing best practices to secure a healthier and more equitable future for people around the world.”
Comments