.png){kind=logo}
AI is reducing diagnosis times for rare diseases, benefiting millions of patients with limited access to adequate treatments
Rare diseases affect millions of people worldwide, yet their low prevalence often leaves patients undiagnosed for years. With over 7,000 known rare diseases, around 400 million people are impacted globally. The growing integration of artificial intelligence (AI) into healthcare systems has the potential to revolutionize the diagnosis and treatment of these complex conditions. In this article, we explore the world of rare diseases, the challenges patients face, and how AI-driven solutions are helping to improve healthcare outcomes.
A rare disease is defined as a condition affecting fewer than 1 in 2,000 people in any given region. Despite their individual rarity, collectively, these diseases affect a significant portion of the global population. 1 in 17 people worldwide may have a rare disease, with 80% of these conditions having a genetic origin. However, the average diagnosis time is 5 years, with some cases taking up to 17 years to diagnose, a delay that can have severe consequences.
These diseases represent a significant challenge for healthcare systems. The 75% of affected patients are children, and rare diseases are one of the leading causes of infant mortality. Despite the need for effective treatments, only 5% of rare diseases have an FDA-approved therapy.
A critical issue in the treatment and recognition of rare diseases is the inconsistent global criteria for defining them. For example, in the European Union, a disease is considered rare if it affects 1 in 2,000 people, while in the United States, the threshold is 1 in 1,600. In Taiwan, it’s 1 in 10,000. These disparities complicate research, funding, and the development of effective treatments, often leaving patients with limited options.
Living with a rare disease doesn’t just affect the patient—it also impacts their families and caregivers. A Canadian study showed that 3 out of 4 caregivers are parents, with 86% of them being women. Of these, 90% reported that their financial situation was negatively impacted. Furthermore, the 71% of caregivers had to leave their jobs or seek flexible working conditions, and 41% missed six or more days of work per month due to caregiving responsibilities.
The emotional and social toll is equally significant, with 80% of caregivers reporting mental health challenges, and 58% stating that they had lost relationships due to their caregiving role. Social isolation is common, as half of these caregivers can only engage in social activities once a month or less.
At Arkangel Ai, we explore the potential of artificial intelligence to assist in tackling the complexities of rare diseases. AI models are being developed to aid in both diagnosis and treatment, making healthcare systems more responsive to patients with rare conditions. There are two main types of AI models contributing to this effort:
Artificial intelligence is being applied in research on rare diseases from three primary perspectives: algorithms focused on specific diseases, groups of diseases, and the entire spectrum of rare diseases. These technologies are unlocking vast amounts of data generated from clinical and biological research, though one challenge remains: there is often a lack of data on patients with confirmed diagnoses.
However, AI improves over time. As new data becomes available, these models can adjust, improving diagnostic accuracy. AI is also being widely used in image analysis, where visual data tends to be more standardized and consistent than clinical records. This helps ensure that AI models have the high-quality data they need to succeed.
For example, CliniFace is a technology that allows non-experts to take 3D facial images and analyze them for clinical significance. This is particularly useful in diagnosing rare diseases with subtle facial features. AI models like CliniFace use anthropometric measures from these 3D images with a level of accuracy comparable to expert analyses.
Another AI tool, MediCutting, uses knowledge graphs to relate existing and validated concepts quickly. This method can speed up research that would traditionally take years, allowing clinicians to connect discoveries with their clinical applications. MediCutting helps identify potential therapies by drawing on a global network of findings, potentially finding personalized treatments for patients with rare conditions.
One remarkable example of AI’s power in rare disease diagnosis comes from Mayo Clinic, which announced in August 2023 that it had identified an ultra-rare condition affecting two brothers, Emery and Aiden. Mayo Clinic’s Renew algorithm compared their genetic sequences with new findings from around the world and found a match with a case in Norway from 2021.
This algorithm can search global databases and deliver results in less than a minute. While the percentage of diagnoses may be small, the impact on individual patients is monumental, giving families answers they’ve waited years for.
The growing use of AI in rare disease research and treatment marks a new era in healthcare. By analyzing complex data and making connections between symptoms and patient histories, AI is providing hope for millions of patients who have gone years without a diagnosis. As these technologies continue to evolve, we expect to see improved treatments, faster diagnoses, and better outcomes for patients worldwide.
Artificial intelligence offers a promising future, but it also highlights the need for greater research, funding, and data availability. If we invest in these areas, AI could be the key to unlocking solutions for the world’s most challenging rare diseases.
Listen to our full podcast here: https://youtu.be/-YLcpbbql3w?si=C_7q1nEjYXlUYAVE
Sources:
