Contrast Study of Low Tube Current Combined with Deep Learning Algorithms in Paranasal Sinus CT Imaging

Objective: To explore the application effect of a low tube current combined with a deep learning algorithm in paranasal sinus computed tomography (CT) imaging and to evaluate its advantages in terms of image quality and radiation dose. Methods: Patients who underwent paranasal sinus CT examinations at Beijing Friendship Hospital, Capital Medical University, between March and November 2024 were retrospectively collected and divided into three groups: conventional dose group, low tube current with deep learning algorithm (CI) group, and iterative algorithm (CV) group. The CT values, noise values (SD), signal-to-noise ratios (SNR), and contrast-to-noise ratios (CNR) of the inferior turbinate mucosa, medial pterygoid muscle, and temporal fossa fat were measured and calculated for each group to objectively assess image quality. In addition, two head and neck radiologists subjectively scored the image quality of the thinnest slice on a 4-point scale. The radiation doses in the conventional and low tube current groups were also compared. Results: A total of 80 patients were included in this study, with 40 in each group. There were no statistically significant differences in the CT values among the three groups for the inferior turbinate mucosa, medial pterygoid muscle, and temporal fossa fat. There were no statistically significant differences in SD, SNR, and CNR between the conventional-dose and CI groups. However, statistically significant differences were observed in SD and SNR between the conventional and CV groups, as well as between the CI and CV groups for the inferior turbinate mucosa, medial pterygoid muscle, and temporal fossa fat. For CNR, statistically significant differences were also found between the conventional and CV groups and between the CI and CV groups in the inferior turbinate mucosa and medial pterygoid muscle regions. In terms of subjective image quality scores, the conventional and CI groups scored 3.93±0.26 and 3.88±0.33, respectively, which were significantly higher than the CV group’s score of 2.70±0.46. Additionally, the radiation dose in the low tube current group was reduced by approximately 74.8% compared to that in the conventional group, with a statistically significant difference. Conclusion: Low tube current combined with a deep learning algorithm (ClearInfinity) in paranasal sinus CT imaging can significantly reduce the radiation dose while maintaining image quality.