Machine learning as a subset of artificial intelligence (AI) has increasingly become the subject of interest by many industries, including in the field of healthcare. For instance, AI and machine learning can play a key role in the diagnosis of rare medical conditions. AI and machine learning in the context of medicine and disease diagnosis use large sets of data to train algorithms and patterns in computers which can then be applied to new input in order to make a prediction such as a disease diagnosis. Machine learning in the context of diagnosis of rare genetic diseases would enable healthcare professionals to sift through large volumes of research and medical literature in order to draw conclusions that would have taken them years to reach had they gone through the research manually.
Different countries have alternative ways of defining a disease as ‘rare’. In Canada, a rare disease is defined as a “condition affecting fewer than 1 person within 2000 in their lifetime.” There are a total of 7000 known rare diseases and with the discovery of new genetic disorders every year, it is projected that 1 in every 12 Canadian will be affected by a rare disease in their lifetime. Although there are policy incentives for pharmaceutical companies to invest in research and development of treatments for rare diseases that affect small populations (see Canada’s regulatory approach to drugs for rare diseases), such research and development may not be profitable, and many conditions go undiagnosed and many individuals end up having to live with the symptoms of their chronic diseases for years. Additionally, rare genetic conditions are extremely difficult to diagnose since even the most experienced physicians may never come across a single patient with one during their years of practice.
Patients with rare genetic diseases can greatly benefit from the implementation of machine learning in healthcare. AI’s ability to “memorize” large amounts of data and extrapolate from them in a meaningful way in order to categorize patients or reach new diagnoses should give sufferers of rare diseases hope that with the help of rapidly improving technology, their conditions may be better understood and treated in the near future. There are several initiatives worldwide that are aimed at gathering information about rare diseases and making them accessible to healthcare professionals for use in diagnosis- Orphanet, CORD-MI in Germany, Undiagnosed Diseases Network (UDN) in the United States to name a few.
Although AI in the context of healthcare certainly offers significant benefits, it is by no means a ‘cure-all’ for the immense challenges of disease diagnosis. After all, the quality of the output from a computer algorithm- coded by a programmer or ‘learned’ by the machine itself- is only as good as the input. In a review published by the Journal of Rare Diseases in 2020, it was found that not all rare diseases are studied to an equal extent. Rare neurologic, rheumatologic, cardiac and gastroenterological diseases were more broadly studied and hence appeared in the literature more frequently. On the other hand, rare skin diseases were highly understudied and it was difficult for a computer to form meaningful algorithms from the limited data that was available in order to better understand the conditions and apply the existing expertise to new cases. That is to say, unless more funding, effort and incentives are invested in the study of rare genetic diseases, no amount of help from AI can help save patients and improve their quality of life. Per President and CEO of Rady Children’s Institute for Genomic Medicine, Dr. Stephen Kingsmore’s statement regarding artificial intelligence and medicine, “Patient care will always begin and end with the doctor.” Technology will only help professionals in connecting the ‘dots’ where there is existing data and research.
Written by Bonnie Hassanzadeh, IPilogue editor and Clinic Fellow at Osgoode Innovation Clinic.
2 Responses
The thought of applying AI and machine learning to scientific and medical research is incredibly exciting. In an ideal scenario, the use of AI and machine learning would help us, within a fraction of the time, sift through thousands of pages of research and provide new insight into therapeutic treatment options that were never considered by the medical community. In a more current and practical application, AI would still be able to complete a systematic review of our literature on specific diseases, perform meta-analyses, and extrapolate information that can guide our research toward particular gaps in our knowledge.
Studying rare diseases is inherently problematic. As you’ve mentioned and perhaps with good reason, funding for rare diseases is underwhelmingly small. Pharmaceutical companies are less inclined to invest in research and development for rare diseases due to smaller markets and profitability. Government and non-profit grants are also less likely to fund research for rare diseases due to its relatively small impact per dollar spent. Further, rare diseases by definition would also mean less people with the disease to study it in the first place. If we can use AI to help improve the efficiency of our investments both in time and money, it would already be a substantial development.
More importantly, the application of AI to help with disease diagnosis would not just be for rare diseases but for all diseases. This will hopefully better incentivize AI development in the context of healthcare and medical research as a whole.
Thank you, Bonnie, for your thought-provoking article on machine learning’s role in diagnosing rare genetic diseases. I think it is such a significant development for humanity that the medical community can utilize machine learning to help with the diagnosis of rare genetic diseases, especially given the fact that there is an apparent lack of research and funding in this area. With the deployment of all emerging technologies, novel legal challenges, such as the appropriate liability regimes, will surface as happened with self-driving cars and human-machine interface technologies. Hence, I wanted to get your insight on the legal question: what should be the liability regime for false positive or negative diagnoses for these rare genetic diseases? For instance, when a doctor gets an injury or illness wrong, the doctor can be liable for the damages and expenses an individual incurred due to a negligent misdiagnosis. The patient can sue the doctor and their employees for medical malpractice and be compensated for his or her suffering and injuries. With AI systems helping doctors diagnose, can an individual get compensation from the doctor, or would the use of an AI system cause difficulties in establishing the liable party’s negligence or causation? I agree with Dr. Kingsmore’s statement, “Patient care will always begin and end with the doctor.” In my opinion, this statement, assuming negligence and causation, points to liability in the doctor since the doctor and their employees will always be the primary people accountable for a diagnosis.