An AI model could help make liposuction safer

An AI model could help make liposuction safer

A newly developed artificial intelligence (AI) model is highly accurate at predicting blood loss in patients undergoing large-volume liposuction, a study reports in the January issue ofPlastic and Reconstructive Surgery®,the official medical journal of the American Society of Plastic Surgeons (ASPS). The journal is published in the Lippincott portfolio of Wolters Kluwer.

Developing an AI model to predict blood loss during liposuction is a "groundbreaking advance" with the potential to improve patient safety and surgical outcomes, according to the new study led by Mauricio E. Perez Pachon, MD, of Mayo Clinic, Rochester, Minnesota, and Jose T. Santaella, MD, of CIMA Clinic-Loja, Ecuador. “By harnessing the power of AI-driven predictive models, surgeons can tailor their procedures to the individual needs of each patient, ensuring optimal outcomes and minimizing the risk of complications such as excessive blood loss, the researchers write.

Using AI to predict blood loss during liposuction

Liposuction is the most common cosmetic surgery procedure worldwide and is performed on more than 2.3 million patients annually. Although liposuction is generally safe, excessive blood loss is a potentially serious complication, especially when larger amounts of fat are being removed. AI-based tools have been developed to prevent blood loss in various medical specialties and surgical procedures such as spine, orthopedic and trauma surgery.

Dr. Perez Pachon, Santaella and colleagues used machine learning technologies to analyze data from 721 patients who underwent large-volume liposuction, with a total volume of over 4,000 milliliters (four liters) of fat and fluid removed. All procedures were performed in two clinics, one in Colombia and one in Ecuador, following identical liposuction protocols.

Data from a random sample of 621 patients were used to develop a model to predict estimated blood loss that incorporated a wide range of demographic, clinical, and surgical data. The researchers then tested their model's performance in predicting blood loss volume in the remaining 100 patients.

With an accuracy of 94%, the model can help make liposuction safer

The results showed “excellent agreement” between predicted and estimated blood loss volumes, with a standard deviation (variation around the average) of 26 milliliters. The maximum difference between predicted and actual blood loss was about 188 ml, while the minimum difference was only 0.22 ml.

Overall, the AI ​​tool was 94% accurate in predicting blood loss. "This accuracy enhances the model's potential as a decision support tool in body contouring procedures, where prediction of intraoperative blood loss is critical to patient safety and surgical planning," the researchers write. “[S]urgeons can use the predicted blood loss estimates to make informed decisions about perioperative management, such as the need for blood transfusions, fluid management and other critical care measures.”

“This proactive approach can significantly reduce the occurrence of adverse events, shorten recovery times and contribute to better patient education and educational processes,” said Dr. Perez Pachon and Santaella finish.

The researchers plan further studies to refine their AI model, including further training using data from surgeons worldwide. Dr. Perez Pachon comments: “We believe that future research in AI technology has unlimited potential to improve patient safety and we look forward to further development in this area.”

Sources: