**AI Model Revolutionises Diagnosis of Rare Diseases in Children**
:max_bytes(150000):strip_icc():format(jpeg)/AI-Medical-Diagnosis-0622326-7999fc6ffa5548f8b360ace4ebd06e30.jpg)
A recent study conducted by OpenAI has showcased the significant potential of artificial intelligence in the medical field, particularly in diagnosing rare diseases among children. The research, published in the *New England Journal of Medicine’s* AI-focused section, NEJM AI on June 18, highlighted the capabilities of OpenAI’s o3 Deep Research model, which successfully identified diagnoses for 18 children at Boston Children’s Hospital.
Catherine Brownstein, from the Manton Centre for Orphan Disease Research, expressed the impact of these findings, stating it is “a total game changer.” The centre is dedicated to assisting over 3,500 patients worldwide who are affected by rare diseases, often struggling to find definitive answers regarding their conditions. Brownstein, a key researcher in the study, noted that while the centre routinely examines patient genomes to identify new potential diagnoses, traditional methods frequently yield no results.
Suyash Shringarpure, a technical researcher at OpenAI involved in this project, explained the limitations faced by researchers. He pointed out that even as new research emerges, a single case can become more complex over time, leading to potential oversights. “Maybe a case remained unsolved when it first came to them,” Shringarpure said, “but a year later, a paper was published that clarifies the link between the gene and the disease.”
The research involved the analysis of 376 genomes from children with undiagnosed rare diseases. Utilising the o3 Deep Research model, the team managed to successfully diagnose 10 patients with rare neurodevelopmental disorders, four patients with neuromuscular issues, two cases of early psychosis, and two instances of sudden unexplained deaths.
While a diagnostic yield of 5% may seem minimal at first glance, Brownstein indicated its significance given the extensive prior analysis of these cases. “Considering how many times these had already been analysed, that’s a huge number, and each one means an answer for a family,” she remarked. The findings suggest that even a small increase in diagnostic success can provide invaluable reassurance and clarity for affected families.
Dr. Adam Rodman, an expert in AI applications in medicine at Beth Israel Deaconess Medical Center, commented on the study’s implications. He stated that a 5% diagnostic yield is meaningful and could serve as a vital tool in addressing the significant backlog of complex medical cases that medical professionals face.
Despite the promising outcomes of the research, experts cautioned against the potential misuse of AI tools for independent diagnosis. The study reinforces that AI, while an excellent tool for medical professionals, is not intended for consumer-level diagnosis or treatment. “We definitely don’t want to overhype this,” Ashley Alexander, head of health for OpenAI, noted, stressing the importance of a cautious approach to the burgeoning technology.
However, Alexander also emphasised the need for awareness regarding the utility of AI in medicine, stating, “I want to make sure that people don’t miss what’s happening and what’s possible with even just the version of ChatGPT that’s in their pocket today.” This signals the immense potential for AI technology in transforming various sectors, medicine being a particularly critical one.
Overall, the success of the o3 Deep Research model exemplifies the promising intersection between technology and healthcare. As research in AI continues to evolve, its role in facilitating more efficient diagnoses and contributing to our understanding of complex medical phenomena carries significant hope for the future of medicine, particularly for individuals suffering from rare conditions.