Hybrid Deep Learning Approach for Accurate Tumor Detection in Medical Imaging Data

• Published in: Diagnostics (Basel) 2023-03, Vol.13 (6), p.1025
• Main Authors: Cifci, Mehmet Akif, Hussain, Sadiq, Canatalay, Peren Jerfi
• Format: Article
• Language: English
• Subjects: Accuracy; Algorithms; Annotations; Clinical outcomes; Datasets; Deep learning; Diagnostic imaging; Efficiency; electronic medical records; Health aspects; Health informatics; joint extraction; Machine learning; medical event extraction; Medical records; Metastasis; Neural networks; Optimization techniques; Support vector machines; Technology application; transfer learning; Tumors
• ISSN: 2075-4418; 2075-4418
• DOI: 10.3390/diagnostics13061025

The automated extraction of critical information from electronic medical records, such as oncological medical events, has become increasingly important with the widespread use of electronic health records. However, extracting tumor-related medical events can be challenging due to their unique characteristics. To address this difficulty, we propose a novel approach that utilizes Generative Adversarial Networks (GANs) for data augmentation and pseudo-data generation algorithms to improve the model's transfer learning skills for various tumor-related medical events. Our approach involves a two-stage pre-processing and model training process, where the data is cleansed, normalized, and augmented using pseudo-data. We evaluate our approach using the i2b2/UTHealth 2010 dataset and observe promising results in extracting primary tumor site size, tumor size, and metastatic site information. The proposed method has significant implications for healthcare and medical research as it can extract vital information from electronic medical records for oncological medical events.