AI Model Accurate for Predicting Blood Loss in Liposuction

TUESDAY, Jan. 6, 2026 (HealthDay News) — An artificial intelligence (AI) model is accurate for predicting blood loss in patients undergoing liposuction, according to a study published in the January issue of Plastic and Reconstructive Surgery.

Mauricio E. Perez Pachon, M.D., from Total Definer in Bogota, Colombia, and colleagues evaluated data from 721 large-volume liposuction patients at two centers between 2019 and 2023 to assess AI models for blood loss prediction and management. The dataset was split into training and testing sets (621 and 100 patients, respectively). A supervised machine learning model was trained to predict blood loss, with predictions compared to clinical data.

Overall, 79.2 percent of patients were women (body mass index, 24.34 kg/m²; volemia, 3,924.41 mL; infiltrated volume, 5,800 mL; and aspirated volume, 3,900 mL). The researchers found that 32 percent had previous liposuction. The training and testing cohorts did not differ significantly. The model achieved a mean absolute error of 22.09 mL, with root mean square error of 34.1 mL, and an R² value of 0.974, demonstrating excellent model fit and high predictive accuracy. The model showed 94.1 percent accuracy for predicting blood loss.

“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 write.

The authors disclosed their involvement in the development and current commercialization of the LISA (Liposuction Intelligent Safety Assistant) app.

Abstract/Full Text (subscription or payment may be required)