|
|
 |
| |
|
|
|
A Laboratory Specialist’s Guide to Integrating Biomarker Analysis and Equipment in Modern CT Radiation Safety |
|
|
|
PP: 291-306 |
|
|
|
doi:10.18576/amis/190206
|
|
|
|
Author(s) |
|
|
|
Waleed F. Alanazi,
Abdulmajeed M. M. Alanazi,
Marzouq H. M. Albalawi,
Abdulaziz M. R. Alruwaili,
Muath H. M. Alanazi,
Bayan A. O. Alanazi,
Abdullah M. Q. Alanazi,
Abdullah S. J. Alanazi,
|
|
|
|
Abstract |
|
|
| The integration of U-Net-based convolutional neural networks (CNNs) and transformers represents a significant advancement in patient-specific dose estimation for computed tomography (CT) imaging, addressing critical concerns regarding personalized radiation management. With CT imaging’s essential role in diagnostics, optimizing radiation exposure without compromising image quality is vital. This study proposes a hybrid model leveraging U-Net’s spatial feature extraction and transformers’ contextual analysis to deliver tailored dose estimations based on high-resolution CT images and patient-specific data, including demographics and medical history. Advanced preprocessing techniques enhance data quality, while the self-attention mechanisms in transformers capture long-range dependencies, improving dose prediction accuracy. The framework aligns with the ALARA principle (As Low as Reasonably Achievable), supporting safer imaging practices while ensuring diagnostic precision. Validation using clinical datasets demonstrates the model’s reliability and its capability to generate detailed dose distribution maps critical for radiological safety and treatment planning. By incorporating both physical and biological data, including blood-based biomarkers of radiation exposure, the method provides a robust, scalable solution for personalized dosimetry. The findings highlight the model’s potential to transform radiological practices, improve patient safety, and advance personalized medicine.
|
|
|
|
|
|
|
|
