<< Back to the abstract archive
Predicting Breast Cancer Related Lymphedema After Immediate Lymphatic Reconstruction: An Artificial Intelligence Approach with Synthetic Data
Berk B. Ozmen, Graham Schwarz
Cleveland Clinic Department of Plastic Surgery
2024-02-01
Presenter: Berk B. Ozmen
Affidavit:
This abstract has been presented at the American Society of Reconstructive Microsurgery 2024 Annual Meeting and the abstract has been published in the Plastic and Reconstructive Surgery–Global Open journal ASRM 2024 Annual Meeting Abstracts section. After contacting Darlene, we were told to submit with this disclosure and it would be committee's decision to whether to accept or not to accept the abstract. All of the work is original work of the authors.
Director Name: Graham Schwarz
Author Category: Fellow Plastic Surgery
Presentation Category: Clinical
Abstract Category: Breast (Aesthetic and Recon.)
Background: Lymphedema is a significant concern for patients undergoing breast cancer surgery, with profound affect on quality of life and functional outcomes. Early identification of high-risk patients allows for targeted interventions and potentially mitigates this debilitating condition. Artificial intelligence (AI) predictive models hold great potential to improve predicting the onset and management of lymphedema, enabling personalized care strategies based on individual risk profiles. In this study, we propose an AI-based approach, utilizing a deep learning model with a synthetic dataset to harness the power of big data, to predict the risk of lymphedema based on immediate lymphatic reconstruction surgical outcomes data.
Methods: We utilized a synthetic dataset generated with General Adversarial Networks (GAN) with 10,000 cases based on a single institution immediate lymphatic reconstruction outcomes dataset. This dataset included pre-op, intra-op and post-op follow-up parameters including data on lymphedema development. A Feedforward Neural Network (FNN), a form of deep learning model, was trained using this dataset to predict the development of lymphedema.
Results: The AI model demonstrated an exceptional performance, achieving a prediction accuracy of 95.25% on the test set for the prediction of lymphedema.
Conclusion: Our findings suggest that AI, and specifically deep learning models trained on big data, hold significant promise in predicting and managing lymphedema risk following breast cancer surgery. These models could potentially enable clinicians to identify patients at high risk and implement personalized, preventative measures, thereby improving patient outcomes.
