BACKGROUND: Skull radiography, an assessment method for initial diagnosis and post-operative follow-up, requires substantial retaking of various types of radiographs. During retaking, a radiologic technologist estimates a patient’s rotation angle from the radiograph by comprehending the relationship between the radiograph and the patient’s angle for adequate assessment, which requires extensive experience. OBJECTIVE: To develop and test a new deep learning model or method to automatically estimate patient’s angle from radiographs. METHODS: The patient’s position is assessed using deep learning to estimate their angle from skull radiographs. Skull radiographs are simulated using two-dimensional projections from head computed tomography images and used as input data to estimate the patient’s angle, using deep learning under supervised training. A residual neural network model is used where the rectified linear unit is changed to a parametric rectified linear unit, and dropout is added. The patient’s angle is estimated in the lateral and superior-inferior directions. RESULTS: Applying this new deep learning model, the estimation errors are 0.56±0.36° and 0.72±0.52° in the lateral and superior-inferior angles, respectively. CONCLUSIONS: These findings suggest that a patient’s angle can be accurately estimated from a radiograph using a deep learning model leading to reduce retaking time, and then used to facilitate skull radiography.

Radiography is used for initial diagnosis and postoperative follow-up. If a radiograph is deemed unsuitable for diagnosis, it is rejected. Retaking a radiograph is disadvantageous for the patient because it prolongs the examination time and increases the radiation dose. Skull radiography is the position in which retaking occurs most frequently. In skull radiography, the patient’s rotational direction is estimated from minute changes in the inner ear’s structure in the lateral skull radiograph. When retaking, the amount of correction for patient positioning is generally estimated from the errors in the rejected image through empirical evidence. Therefore, considerable expertise is needed to correct the positioning appropriately, and inexperienced radiologic technologists take considerable time to estimate this error. This study aimed to estimate the patient’s angle from lateral skull radiographs to compensate for radiologic technologists’ lack of experience and reduce the burden on patients. The radiograph is a simulated 2-D line-integral projection of a 3-D CT image, and we developed an estimation method using deep learning with supervised training. The network is based on a re-scaled ResNet. The patient’s angle was estimated in the lateral and superior-inferior directions. We evaluated the accuracy of estimation with the projected images of 256 simulated cases. The estimation errors were 0.48 ± 0.41° and 0.55 ± 0.50° in the lateral and superior-inferior angles, respectively. These findings suggest that a patient’s angle can be accurately estimated from a radiograph using deep learning, compensating for the lack of experience and reducing retaking time.