AI Tool Identifies Surgical Site Infections From Patient Photos

Researchers at Mayo Clinic have developed an artificial intelligence (AI) system capable of analyzing patient-submitted photographs of postoperative wounds to identify surgical site infections (SSIs).

The study, published in Annals of Surgery, describes a multi-step pipeline trained on more than 20,000 images collected from over 6,000 patients treated across 9 Mayo Clinic hospitals.

The AI system is trained to perform three functions: it first determines whether a submitted image contains a surgical incision, then assesses the quality of the image and finally evaluates the incision for signs of infection. 

Supporting outpatient recovery with automated screening

With the increasing shift to outpatient surgeries and virtual follow-up care, clinicians are often required to assess postoperative recovery remotely. This approach can delay diagnosis if images are not reviewed promptly. 

"We were motivated by the increasing need for outpatient monitoring of surgical incisions in a timely manner," said Cornelius Thiels, D.O., a hepatobiliary and pancreatic surgical oncologist at Mayo Clinic and co-senior author of the study. "This process, currently done by clinicians, is time-consuming and can delay care. Our AI model can help triage these images automatically, improving early detection and streamlining communication between patients and their care teams."

The model’s operates using a two-stage model. First, it begins with incision detection. If an incision is confirmed, the wound features are then assessed to evaluate whether there are any signs of infection.

The model has achieved 94% accuracy in identifying incision presence and achieved an area under the curve (AUC) of 0.81 in detecting infections. Critically, the model continued to perform at consistently high levels across diverse patient demographics, mitigating concerns over potential bias.

"Our hope is that the AI models we developed — and the large dataset they were trained on — have the potential to fundamentally reshape how surgical follow-up is delivered," said Hojjat Salehinejad, Ph.D., a senior associate consultant of health care delivery research within the Kern Center for the Science of Health Care Delivery and co-senior author. "Prospective studies are underway to evaluate how well this tool integrates into day-to-day surgical care."

Future applications in clinical workflows

Although the tool currently serves as a proof of concept, the research team is exploring how it could be used in real-world surgical care workflows.

"For patients, this could mean faster reassurance or earlier identification of a problem," said Hala Muaddi, M.D., Ph.D., a hepatopancreatobiliary fellow at Mayo Clinic and first author. "For clinicians, it offers a way to prioritize attention to cases that need it most, especially in rural or resource-limited settings."

The team are hopefully that this technology could help support patients who are recovering from surgery at home. With further validation, they believe it could be used as a frontline screening tool to alert physicians to potentially concerning incisions.

Reference: Hala Muaddi, Choudhary A, Lee F, et al. Imaging Based Surgical Site Infection Detection Using Artificial Intelligence. Ann Surg. 2025. doi: 10.1097/sla.0000000000006826

This article has been republished from the following materials. Note: material may have been edited for length and content. For further information, please contact the cited source. Our press release publishing policy can be accessed here.

This content includes text that has been generated with the assistance of AI. Technology Networks' AI policy can be found here.