Application Of Deep Learning Image Reconstruction Technique In Improving Low-dose CT Image Quality

Objective: Deep learning Image reconstruction（DLIR）is a new image reconstruction technique.In this study,the application value and advantages of DLIR technique in low-dose chest CT were discussed through clinical cases.Material and methods: A total of 48 patients（30 women and 18 men,mean age ±SD,49.8±14.3years）who underwent chest CT scanning in the First Affiliated Hospital of Anhui Medical University from January 2021 to March 2021.The age,height（unit: m）and weight（unit: kg）of the patients were recorded,and body mass index（BMI）was calculated,the formula was: BMI= weight/height 2（unit: kg/m2）.All patients should have a BMI range of 18.5 to 24.0.Patients underwent both the standard-dose CT（SDCT）and low-dose CT（LDCT）on a GE Revolution CT scanner.All patients gave written informed consent.All scans were reconstructed with ASIR-V40%.Additionally,LDCT scans were reconstructed with DLIR with high-setting（DLIR-H）and medium-setting（DLIR-M）.Image noise and contrast-noise-ratio（CNR）of thoracic aorta with different reconstruction modes were measured and compared.Using the GE Revolution CT machine,the scan parameters are as follows:（1）,standard dose scanning: voltage,120 k V;and automatic tube current;gantry rotation time,0.5 seconds;helical pitch,0.992:1.（2）,low dose scanning: voltage,80 k V;tube current,50 m A;gantry rotation time,0.5 seconds;helical pitch,0.992:1.Both LDCT and SDCT images were reconstructed at a slice thickness of 1.25 mm and with ASIR-V at a strength level of 40%（ASIR-V40%）.In addition,the LDCT scan datasets were reconstructed with DLIR at the medium（DLIR-M）and high（DLIR-H）levels.All image data are transmitted to PACS system,and finally image noise signal-to-noise ratio and contrast noise ratio under different reconstruction modes are measured and compared.In all cases,the scan range was from the lower neck to the lower upper quadrant of the diaphragm,ensuring that the entire lung was included.Objective analysis of image noise,signal-to-noise ratio and contrast noise ratio,subjective analysis of images by radiologists（including overall image quality and different morphology of nodules）,and comparison of radiation dose（volume CT dose index CTDIvol,dose length product DLP and estimated effective dose ED）.LDCT reconstructed images at DLIR-H and DLIR-M levels were compared with SDCT images.All the measured data were sorted out and analyzed,and P < 0.05 was statistically significant,while P > 0.05 was not.Results: LDCT reduced radiation dose by 96% compared with SDCT（CTDIvol: 0.54 m Gy vs.12.46 m Gy）.In LDCT,DLIR significantly reduced image noise compared with the state-of-theart ASIR-V40% with DLIR-H provided the lowest image noise and highest image quality score.In addition,the image noise,SNR,CNR of aorta and overall image quality of the low-dose DLIR-H images did not have significant difference compared with the SDCT ASIR-V40%images（all p>0.05）.Conclusion: DLIR can significantly reduce the image noise of LDCT chest scan and provide image quality similar to SDCT ASir-v image at only 4%SDCT radiation dose.