Abstract: Objective To explore the clinical value of quantitative analysis of multi-slice spiral CT (MSCT) pixel index (PI) in the first CT and disease course assessment of COVID-19. Methods The chest CT date of 48 patients diagnosed with COVID-19 in our hospital from January 22,2020 to March 1,2020 were collected. The Thoracic VCAR software was used for the pulmonary lobes segmentation and PI quantitative evaluation, PI values of pulmonary lope of <-700Hu, -700Hu — -300Hu, >-300Hu were measured. The characteristics of PI in the initial CT examination and the changes of PI among different stages（in 1^st, 2^nd, 3^rd, 4^th and beyond weeks）were analyzed. Results ①The difference of PI<-700Hu, PI-700Hu — -300Hu and PI>-300Hu between the lung lobe groups were found statistically significant (P>0.05). ②Comparison between each lung lobe group: the PI>-300Hu of right middle lobe was higher than that of right upper lobe (P<0.05); PI-700Hu — -300Hu and PI>-300Hu of bilateral lower lobe were higher than those of bilateral upper lobe and lower right middle lobe, PI<-700Hu of bilateral lower lobe were lower than that of bilateral upper lobe and right middle lobe (P<0.05), and the rest of the lung lobe pairwise comparison were not statistically significant(P>0.05). ③PI -700Hu — -300Hu and PI>-300Hu of whole lung were found highest in the second weeks after onset [11.41%(8.66%—19.24%)、2.47%(1.85%—4.36%)], while the PI -1024Hu — -300Hu of the whole lung was fount lowest[86.31%(74.70%—89.54%)]. Conclusion The PI of MSCT CT can quantitatively analyze the pulmonary exudation and consolidation of COVID-19 patients,from visual assessment to accurate quantification. Combined with conventional CT, it can be used for clinical diagnosis and follow-up treatment evaluation of COVID-19 patients.

Key words: Computed tomography, Pixel index, Coronavirus disease 2019, Quantitative analyze