New artificial intelligence tools are helping doctors identify serious diseases more quickly and accurately by analyzing medical images.
Mayo Clinic researchers found AI can predict brain tumor recurrence risk without expensive genetic testing.
A separate study found AI can improve detection of retinal diseases and may even reveal risks for heart, kidney and vascular conditions through routine eye scans.
Artificial intelligence is increasingly showing up in the doctor’s office, helping physicians identify disease risks, improve diagnoses and potentially personalize treatment plans without adding costly tests or procedures.
Two newly published studies highlight how AI is transforming patient care in specialties ranging from cancer treatment to ophthalmology, offering doctors new tools to uncover critical health information hidden in routine medical images.
Researchers at Mayo Clinic and collaborating institutions have developed an AI model that can analyze standard pathology slides to classify meningiomas — the most common primary brain tumor in adults — and predict the likelihood that a tumor will return after treatment.
A major shortcut
Currently, determining recurrence risk often requires advanced molecular or genetic testing that may not be available at every medical center. The new AI system extracts similar insights from routine pathology images, potentially making sophisticated tumor analysis more accessible and affordable.
The researchers trained the model using tissue samples, pathology images and clinical data from 672 patients. The findings suggest that AI can identify molecular patterns associated with tumor behavior and recurrence risk that may not be readily visible to the human eye.
For physicians, that information could help guide decisions about follow-up care, imaging schedules and whether additional treatments such as radiation therapy should be considered.
Helping eye doctors diagnose disease faster
In a separate study, researchers at Washington University School of Medicine developed an AI system called OCTCube-M that analyzes three-dimensional retinal scans commonly used in eye clinics. The technology is designed to help physicians process large volumes of imaging data more quickly and identify subtle signs of disease that might otherwise be overlooked.
The system was trained using more than 26,000 retinal scans containing approximately 1.62 million individual image slices. Compared with older AI approaches, OCTCube-M improved detection accuracy for six of eight retinal diseases, including age-related macular degeneration, one of the leading causes of blindness among older adults.
Researchers estimate the improvement could help identify dozens of additional disease cases for every 1,000 patients screened. The model also significantly improved predictions about how quickly geographic atrophy, a severe form of macular degeneration, is likely to progress.
Looking Beyond the Eye
One of the most intriguing findings from the retinal imaging study is AI's ability to identify signs of health conditions beyond vision problems.
Researchers found that retinal images contain clues associated with cardiovascular disease, kidney disease, stroke and heart attack risk. Because the retina's tiny blood vessels closely mirror those found elsewhere in the body, changes visible in eye scans can provide insights into overall vascular health.
The ability to extract that information from a routine eye exam could eventually help doctors identify high-risk patients earlier and intervene before serious health problems develop.
A tool for doctors, not a replacement
While both studies demonstrate the growing capabilities of AI in healthcare, researchers emphasize that the technology is designed to support physicians rather than replace them.
By rapidly analyzing large amounts of imaging data and highlighting patterns that may be difficult for humans to detect, AI can help doctors make more informed decisions, prioritize high-risk patients and deliver more personalized care.
As healthcare systems face growing patient demands and increasing volumes of medical data, these emerging AI tools could become valuable assistants in improving both the speed and quality of patient care.