Kingdom-wide AI Tool Aims to Accelerate Rare Disease Diagnosis by Linking Symptoms and Genomes
Researchers at King Abdullah University of Science and Technology in Saudi Arabia have trained an artificial intelligence system to predict rare genetic diseases. The findings are detailed in a study shared with the journal BMC Medicine. This work highlights how machine intelligence can support clinicians who encounter patients with unusual or unknown genetic presentations across North America and beyond.
The newly developed tool, named STARVar, uses artificial intelligence to diagnose conditions by examining a patient’s symptoms alongside the genome. It then compares these insights with information from the scientific literature. STARVar stands for Symptom Based Vehicle, and its key distinction is its emphasis on the patient’s actual symptoms rather than how those symptoms are described or documented in existing literature. For US and Canadian clinicians, this symptom-first approach may offer a practical path to connect clinical observations with possible genetic causes, especially in cases where documentation is sparse or variable in terminology (King Abdullah University of Science and Technology, 2024).
Early testing has shown promising results. STARVar performed strongly against many existing methods when it analyzed the genomes of patients not only from Saudi Arabia but from other nations as well, including cases where traditional approaches focused primarily on clearly stated symptoms. In one instance, the system detected a rare genetic mutation in a patient who reported joint stiffness, swelling beneath the skin, and damage to bone structure. Among 800 candidate gene variants, STARVar highlighted a single mutation that could explain the clinical picture. This example underscores how the tool can narrow the field of possible causes by prioritizing the actual clinical presentation rather than relying solely on predefined symptom descriptions (King Abdullah University of Science and Technology, 2024).
The researchers expect STARVar to become more widely used in clinical genetics, offering a data-driven way to connect symptoms to potential genetic causes and to guide further testing and management for patients with unknown or unusual presentations. In North American practice, such a framework could complement genome-wide analyses and support multidisciplinary teams in Canada, the United States, and beyond by providing a practical, evidence-backed method for prioritizing testing and referral decisions (King Abdullah University of Science and Technology, 2024).
It is worth noting that artificial intelligence has previously been employed to assist in diagnosing heart disease, illustrating the broader potential for AI to support medical decision making in a range of genetic and non-genetic conditions. The expanding role of AI in medicine points to a future where clinicians in North America may rely on data-driven tools to interpret complex genetic data together with patient-reported symptoms, improving diagnostic speed and accuracy while maintaining clinician oversight and patient safety (King Abdullah University of Science and Technology, 2024).