Fully automated accurate patient positioning in computed tomography using anterior–posterior localizer images and a deep neural network: a dual-center study - Summary - MDSpire
Advertisement
Fully automated accurate patient positioning in computed tomography using anterior–posterior localizer images and a deep neural network: a dual-center study
To automate the detection of the patient’s body centerline distance from the gantry isocenter for accurate patient positioning in chest CT scans using specific deep learning algorithms.
Key Findings:
High prevalence of patient mis-centering in clinical practice, leading to increased radiation doses and degraded image quality, which poses risks to patient safety.
Deep learning algorithms can effectively automate patient positioning using AP localizer images.
Interpretation:
The study demonstrates the potential of deep learning to improve patient positioning accuracy in CT scans, which could enhance image quality and reduce radiation exposure compared to traditional methods.
Limitations:
The study was retrospective and limited to specific imaging centers, which may introduce biases.
Exclusion of cases with truncation artifacts may limit generalizability.
Conclusion:
Automating patient positioning in chest CT scans using deep learning can significantly reduce mis-centering errors, potentially improving diagnostic outcomes.
