Zhang M, Lei M, Lin F, Chen Y, Xie Y, Liu J, et al . Deep learning algorithm for CT scanning in ankle trauma: Technical principles and prospects for reducing effective radiation dose. Int J Radiat Res 2025; 23 (3) :813-817
URL:
http://ijrr.com/article-1-6683-en.html
Department of Radiology, Zhuhai Hospital, Guangdong Provincial Hospital of Traditional Chinese Medicine, 53 Jingle Road, Zhuhai City, Guangdong Province, China , xmqzhuhai@163.com
Abstract: (34 Views)
Background: This study aims to compare the image quality and effective dose (ED) of deep learning algorithm computer tomography (DLA-CT) with standard CT (SD-CT) in managing ankle trauma. Materials and Methods: In this prospective study, 88 patients underwent random allocation to either the SD-CT group (utilizing a tube voltage of 120KV and an automatic tube current milliampere setting) or the DLA-CT group (employing automatic tube voltage and current adjustments). Senior radiologists assessed objective image quality using parameters such as noise, signal-to-noise ratio (SNR), and contrast-to-noise ratio (CNR). The influence of subjective image quality and its impact on treatment plans were evaluated using a 5-point Likert scale, with scores of 3 or higher indicating acceptable image quality and adherence to treatment plan requirements. Dose length product (DLP) (mGy×cm) was automatically recorded by the scanner software, DLP×k. Results: Objective image quality with DLA-CT was found to be inferior to that obtained with SD-CT (P<0.001). No significant differences were noted in subjective image quality scores between the two groups (P =0.60). Additionally, both subjective image quality scores and the effect of image quality on treatment decisions were scored at 3 or higher. The ED was significantly reduced by 48.83% in the DLA-CT group compared to the SD-CT group (21.42±2.62 µSv and 10.96±1.12 µSv, respectively, P=0.001). Conclusion: DLA-CT proves effective in the management of ankle trauma, significantly reducing ED while meeting clinical diagnostic and treatment standards.