DoctorateOpen Access

A deep learning-enhanced clinical decision support system for MRI-based brain tumor diagnosis in healthcare

Is this your thesis?

This record came from a bulk archive import. If it’s yours, link it to your profile.

Abstract (EN)

In healthcare, time and errors are critical factors; improvements in these areas contribute significantly to healthcare institutions. In this context, this doctoral thesis aims to develop an enhanced deep learning-based clinical decision support system for magnetic resonance imaging (MRI)-based brain tumor diagnosis, focusing on the transformative potential of artificial intelligence (AI) and health informatics in clinical decision support systems. The central hypothesis of the thesis is that AI-based deep learning models can provide faster, more accurate, and consistent results by supporting radiologists in brain tumor diagnosis. This study proposes a hybrid deep learning model based on EfficientNetV2 and integrating Global Attention Mechanism (GAM) and Efficient Channel Attention (ECA), which outperforms 45 state-of-the-art models in the literature with 99.76% test accuracy in brain tumor classification (glioma, meningioma, pituitary and no tumor). Enhanced with Grad-CAM visualization for transparency, the model demonstrates high precision, recall, and F1 scores, offering clinicians greater confidence in the diagnostic process. Emphasizing clinical applicability, this hybrid approach is expected to make a significant contribution to health informatics due to its computational efficiency and adaptability to diverse patient populations. However, considering certain limitations such as data diversity and generalizability, future studies are recommended to incorporate larger datasets and advanced interpretability methods. In conclusion, this thesis provides a theoretical and practical framework for the integration of health informatics and AI into healthcare, offering a valuable contribution to the future of clinical decision support systems.

Author

Ömer Çelik

How to Cite

Ömer Çelik (Doctorate thesis). A deep learning-enhanced clinical decision support system for MRI-based brain tumor diagnosis in healthcare, 2025, Süleyman Demirel University.

Keywords

License

Tüm Hakları Saklıdır

This work is shared under the specified license terms.

More theses from Süleyman Demirel University