A randomly selected dataset from 621 patients was used to develop the predictive model. This dataset included a wide range of demographic, clinical and surgical variables, enabling the AI to identify patterns associated with higher or lower blood loss. The remaining 100 patient cases were then used to validate the model’s performance
A newly developed artificial intelligence (AI) model could significantly improve safety in cosmetic surgery by accurately predicting blood loss in patients undergoing high-volume liposuction, according to a new study that researchers have described as a major step forward in surgical planning and patient care.
Liposuction is among the most commonly performed cosmetic procedures worldwide, with more than 2.3 million patients undergoing the surgery each year to remove stubborn fat from areas such as the abdomen, thighs, arms, face and neck. While the procedure is generally considered safe, excessive blood loss remains a potentially serious complication, particularly in large-volume liposuction where more than four litres of fat and fluid are removed.
The study, published in the journal Plastic and Reconstructive Surgery, reports that an AI-driven predictive model can estimate blood loss with remarkable precision, allowing surgeons to anticipate risks before the operation begins. Researchers have called the development a “groundbreaking advancement” with the potential to enhance both patient safety and surgical outcomes.
“By leveraging the power of AI-driven predictive models, surgeons can tailor their interventions to each patient’s unique needs, ensuring optimal outcomes and minimising the risk of complications such as excessive blood loss,” said the international research team, which included experts from the Department of Public Health at Ecuador’s Health Ministry and the Mayo Clinic in the United States.
To build the model, the researchers used machine learning techniques to analyse detailed data from 721 patients who underwent large-volume liposuction. Each procedure involved the removal of more than 4,000 millilitres of fat and fluid and was performed at one of two clinics—one in Colombia and the other in Ecuador. Importantly, both clinics followed identical surgical protocols, ensuring consistency in the data used to train and test the AI system.
A randomly selected dataset from 621 patients was used to develop the predictive model. This dataset included a wide range of demographic, clinical and surgical variables, enabling the AI to identify patterns associated with higher or lower blood loss. The remaining 100 patient cases were then used to validate the model’s performance.
The results were striking. The AI model predicted blood loss with an accuracy of 94 per cent, a level of precision that could meaningfully influence real-world surgical decision-making. According to the researchers, this high degree of accuracy underscores the model’s value as a decision-support tool in body contouring procedures, where unanticipated blood loss can quickly escalate into a medical emergency.
“Such accuracy reinforces the model’s potential as a decision-support tool in body contouring procedures, where anticipating intraoperative blood loss is crucial for patient safety and operative planning,” the researchers noted.
Knowing in advance how much blood a patient is likely to lose allows surgeons and anaesthesiologists to better prepare for surgery. Predicted blood loss estimates can guide decisions on fluid management, the possible need for blood transfusions, and other critical perioperative care measures. This proactive approach can reduce complications, shorten recovery times and improve overall surgical outcomes.
Beyond liposuction, the study highlights the expanding role of artificial intelligence in surgical medicine. AI models are increasingly being used to analyse complex datasets that exceed human capacity, transforming them into actionable insights for clinicians. As cosmetic procedures grow in popularity and complexity, such tools could help standardise care and reduce variability in outcomes across different settings.
While the researchers emphasise that AI tools are meant to support—not replace—clinical judgement, they believe this model represents a significant advance in making cosmetic surgery safer. With further validation and wider adoption, AI-driven predictions of blood loss could become a routine part of preoperative planning, offering patients greater reassurance and surgeons a powerful new ally in the operating room.
