Industry Insights AI: 25/07/25
This week’s insights post focusses on the latest advancements in the application of artificial intelligence for disease detection.
AI can help detect postoperative infections with accuracy
Surgical site infections (SSIs) are a leading cause of postoperative complications. They often go undetected until symptoms worsen. A team of clinicians and scientists at the Mayo Clinic has recently developed an AI model that may improve SSI identification using data from patient records and clinical notes. The model was trained on more than 20,000 images from surgical cases across nine Mayo Clinic hospitals and was able to detect SSIs with 73% accuracy from patient-submitted photos. This real-time tool may potentially help reduce complications and improve patient outcomes.
Link to article abstract (Annals of Surgery)
Novel AI tool can detect Alzheimer’s disease via retinal imaging
The Hong Kong Hospital Authority has introduced the world’s first retinal imaging AI tool for detecting early signs of Alzheimer’s disease. The system analyses subtle changes in retinal blood vessels and tissue patterns using deep learning algorithms. It was trained on 13,000 fundus images from over 600 patients with Alzheimer’s disease. Unlike conventional neurological assessments, this tool promises to be a quick, low-cost and non-invasive alternative that can become part of routine eye exams. Early trials have demonstrated an 80-92% accuracy of the tool in identifying patients at risk.
Link to full study (The Lancet Digital Health)
The potential of AI in diagnosing autism and ADHD
Diagnosing and subtyping neurodivergent conditions like autism spectrum disorder (ASD) and attention-deficit/hyperactivity disorder (ADHD) is often delayed, time-consuming and resource-intensive. A new artificial intelligence (AI) tool developed by researchers at Indiana University may help make this process easier and more effective. The system uses machine learning and motion-tracking data to provide diagnostic support. In clinical trials, the tool demonstrated 71.8% accuracy in classifying participants based on their neurological traits. According to the researchers, their AI tool may complement other diagnostic tests to recognise and classify neurodivergent conditions.
Link to full study (Nature – Scientific Reports)
Want to get in touch? Contact us at https://elion.nz/get-in-touch/
