Artificial intelligence could soon play a crucial role in improving safety during cosmetic surgery, with a new study showing that an AI-based model can accurately predict blood loss in patients undergoing liposuction.

The findings, published in the January issue of Plastic and Reconstructive Surgery, suggest that machine learning tools may help surgeons better plan procedures and manage risks, particularly in large-volume liposuction cases.

The study was led by Dr. Mauricio E. Perez Pachon of Total Definer, Bogotá, Colombia, along with his research team. The researchers analysed data from 721 patients who underwent large-volume liposuction at two medical centres between 2019 and 2023. The dataset was divided into a training group of 621 patients and a testing group of 100 patients to develop and validate the AI model.

Using a supervised machine learning approach, the model was trained to predict intraoperative blood loss based on multiple clinical parameters. These predictions were then compared with actual clinical outcomes to assess accuracy.

Most of the patients included in the study were women (79.2 per cent), with an average body mass index of 24.34 kg/m². The average blood volume (volemia) was about 3,924 mL, while the mean infiltrated volume was 5,800 mL and the aspirated volume was approximately 3,900 mL. Notably, nearly one-third of the patients (32 per cent) had undergone previous liposuction procedures. Researchers reported no significant differences between the training and testing groups, ensuring the reliability of the model’s evaluation.

The AI model demonstrated strong performance, achieving a mean absolute error of just 22.09 mL and a root mean square error of 34.1 mL. With an R² value of 0.974, the model showed an excellent fit between predicted and actual blood loss. Overall, it achieved an accuracy rate of 94.1 per cent in predicting blood loss during liposuction.

“Developing and implementing our AI model for predicting blood loss in liposuction is a groundbreaking advancement that promises to improve patient safety and surgical outcomes,” the authors noted in the study.

The researchers also disclosed that they are involved in the development and ongoing commercialisation of LISA (Liposuction Intelligent Safety Assistant), an application based on this AI model. The tool is designed to assist surgeons in real-time decision-making, potentially reducing complications related to excessive blood loss.

Experts say such innovations reflect the growing role of artificial intelligence in surgical planning and risk management. If widely adopted, AI-assisted tools like LISA could help standardise safety protocols and enhance outcomes in aesthetic and reconstructive surgery, particularly as demand for cosmetic procedures continues to rise globally.