Akademik Veri Yönetim Sistemi
BARIATRIC SURGICAL PRACTICE AND PATIENT CARE, 2024 (SCI-Expanded)
Background: Bariatric surgery is vital for combating obesity, but inconsistent weight loss outcomes pose a challenge. We explored the use of deep learning (DL) models, to forecast weight loss success. Methods: We created a dataset using our surgical database and postoperative images from bariatric surgery patients. DL model based on our dataset was trained to predict weight loss outcomes. Results: Our models achieved mean average precision scores of 0.954 for the failure group and 0.909 for the successful weight loss group. In evaluating the model's diagnostic performance to predict weight loss failure (defined as excess weight loss <50 patients) using an external test dataset, we observed a sensitivity of 94.12% and specificity of 95.92%. Conclusion: DL shows promise in predicting weight loss success in bariatric surgery, with significant potential to enhance patient care, optimize surgical strategies, and combat obesity. Further research and validation are essential for broad clinical applicability.