For people with rare diseases, finding the right diagnosis can become a medical mystery that lasts for years. Symptoms may resemble more common illnesses, while doctors may have limited experience recognizing conditions that affect relatively few people.
Researchers in Abu Dhabi are developing an artificial intelligence platform designed to help physicians work through those difficult cases.
Called iGenRARE, the system is being developed at Khalifa University to support doctors investigating more than 7,000 rare diseases. Rather than simply producing a single prediction, it is designed to bring together different types of medical information and present physicians with possible diagnoses, supporting evidence, and potential next steps.
Its goal is not to replace doctors, but to help them connect clues that might otherwise be difficult to see.
What Is iGenRARE and How Could It Help Doctors?
iGenRARE is an AI assistant for the diagnosis and management of rare genetic diseases. The project is led by Dr. Aamna AlShehhi, an assistant professor at Khalifa University whose research focuses on artificial intelligence, bioinformatics, and healthcare.
Unlike a system that relies on one type of medical information, iGenRARE is designed to work across multiple sources. These can include symptoms, clinical notes, genetic information, laboratory results, medical imaging, and scientific literature.
The platform uses multiple AI agents to examine different pieces of information and a central reasoning system to bring those findings together. It can then rank possible diagnoses and provide evidence supporting those possibilities.
That approach could be particularly useful when a patient's symptoms do not immediately point to an obvious condition. A single symptom may be nonspecific, but its significance can change when considered alongside a patient's medical history, laboratory findings, or genetic information.
Instead of requiring a physician to manually search through thousands of rare conditions and the research associated with them, an AI system could help narrow the field.
Why Rare Diseases Can Become Medical Mysteries
Rare diseases present a difficult diagnostic challenge because there are thousands of different conditions, many of which can produce symptoms that overlap with more familiar illnesses.
A physician may also rarely encounter a particular rare disease. That can make an unusual diagnosis difficult to recognize, especially when the early symptoms do not appear connected.
The result can be a lengthy diagnostic journey involving repeated appointments, tests, and referrals. Khalifa University's description of iGenRare notes that the diagnostic journey for rare-disease patients can take years, with patients seeing multiple physicians and undergoing numerous tests before receiving an answer.
For patients and their families, the delay can mean prolonged uncertainty and delayed access to appropriate treatment.
The challenge is not necessarily a lack of medical information. In many cases, the difficulty lies in connecting information scattered across a patient's medical history and recognizing that an unusual combination of findings could point toward a rare condition.
What the Researchers Found
The researchers have put iGenRARE through a retrospective evaluation using medical data from 20,238 patients, including 5,067 people with rare diseases.
When the system was given information from a patient's first hospital visit, it was able to distinguish rare-disease cases from controls with 68% accuracy. Its performance increased to 91% when researchers provided the patients' complete hospital histories.
The findings suggest that access to a broader medical history can make a significant difference when AI is trying to identify patterns associated with rare diseases.
The system's design also goes beyond simply naming a possible condition. It can use information from different areas of a patient's medical record and relevant scientific evidence to build a more comprehensive picture of the case.
That could give physicians another way to approach patients whose symptoms have not yet led to a clear diagnosis.
However, the results should be interpreted carefully. The evaluation was retrospective, meaning the researchers analyzed existing medical data rather than testing the system as part of routine patient care. iGenRARE has not yet demonstrated that it can reliably diagnose rare diseases in real-world clinical settings.
Could AI Shorten the Search for a Diagnosis?
The potential value of iGenRARE lies in what could happen earlier in a patient's diagnostic journey.
A person with an unusual combination of symptoms may initially be evaluated for more common conditions. If those explanations do not fit, the patient may undergo additional tests or be referred to different specialists.
An AI system that recognizes a possible rare disease earlier could potentially help a physician decide when further investigation or specialist referral is warranted.
This does not mean AI would automatically provide the correct diagnosis. Instead, it could help bring less obvious possibilities to a doctor's attention at a point when they might otherwise be overlooked.
With more than 7,000 rare diseases to consider, having a system designed specifically to navigate that diagnostic landscape could be particularly useful.
AI Can Offer Clues, But Doctors Still Make the Diagnosis
iGenRARE is intended to assist physicians rather than replace them.
An AI-generated possibility is not the same as a confirmed diagnosis. Doctors still need to evaluate a patient's symptoms, history, and test results and determine which additional investigations are appropriate.
The researchers also plan to conduct a clinical evaluation in which the system's recommendations can be compared with doctors' decisions. That step will be important in determining whether iGenRARE's performance in historical medical data translates to actual clinical care.
For now, the most intriguing possibility is not that AI could become a doctor. It is that AI could help doctors recognize patterns in cases that do not make sense at first glance.
Rare diseases can hide behind ordinary symptoms and scattered clues. If systems such as iGenRARE can help physicians connect those clues earlier, some of medicine's most difficult diagnostic mysteries could eventually become easier to solve.
