Proposing Intelligent Approach to Predicting Air Kerma within Radiation Beams of Medical X-ray Imaging Systems

Area of Science:

• Medical Physics
• Radiological Sciences
• Artificial Intelligence in Healthcare

Background:

• Air kerma is crucial for assessing organ doses and patient hazards in diagnostic radiology.
• Nonuniformity of air kerma within X-ray beams, influenced by the heel effect and tube voltage, poses challenges for accurate dose evaluation.
• Predicting air kerma distribution is essential for radiation safety and quality assurance in medical imaging.

Purpose of the Study:

• To develop an intelligent technique for predicting air kerma at any point within X-ray beams.
• To utilize a radial basis function neural network (RBFNN) for accurate air kerma estimation based on discrete measurements.
• To enhance the assessment of organ doses and patient hazards in medical diagnostic radiology.

Main Methods:

• Modeled a diagnostic imaging system using the Monte Carlo N Particle X (MCNPX) code.
• Calculated air kerma at discrete points across various X-ray tube voltages (40-140 kV) to generate a training dataset.
• Trained an RBFNN using X-ray tube voltage and spatial location as inputs to predict air kerma output.

Main Results:

• The trained RBFNN model demonstrated capability in estimating air kerma at arbitrary positions within the X-ray beam's field of view.
• The model accurately predicted air kerma for X-ray tube voltages relevant to medical diagnostic radiology (20-140 kV).
• The technique effectively addresses the non-uniformity of air kerma caused by the heel effect and voltage variations.

Conclusions:

• An RBFNN is a viable and effective method for predicting air kerma distribution in medical diagnostic radiology.
• This approach allows for precise estimation of radiation intensity, improving the evaluation of organ doses and patient risks.
• The developed intelligent technique offers a valuable tool for enhancing radiation safety and quality control in medical imaging.