Segmentation and quantification of COVID-19 infections in CT using pulmonary vessels extraction and deep learning

Abstract

At the end of 2019, the World Health Organization (WHO) reported pneumonia that started in Wuhan, China, as a global emergency problem. Researchers quickly advanced in research to try to understand this COVID-19 and sough solutions for the front-line professionals fighting this fatal disease. One of the tools to aid in the detection, diagnosis, treatment, and prevention of this disease is computed tomography (CT). CT images provide valuable information on how this new disease affects the lungs of patients. However, the analysis of these images is not trivial, especially when researchers are searching for quick solutions. Detecting and evaluating this disease can be tiring, time-consuming, and susceptible to errors. Thus, in this study, we aim to automatically segment infections caused by COVID19 and provide quantitative measures of these infections to specialists, thus serving as a support tool. We use a database of real clinical cases from Pedro Ernesto University Hospital of the State of Rio de Janeiro, Brazil. The method involves five steps: lung segmentation, segmentation and extraction of pulmonary vessels, infection segmentation, infection classification, and infection quantification. For the lung segmentation and infection segmentation tasks, we propose modifications to the traditional U-Net, including batch normalization, leaky ReLU, dropout, and residual block techniques, and name it as Residual U-Net. The proposed method yields an average Dice value of 77.1% and an average specificity of 99.76%. For quantification of infectious findings, the proposed method achieves results like that of specialists, and no measure presented a value of ρ < 0.05 in the paired t-test. The results demonstrate the potential of the proposed method as a tool to help medical professionals combat COVID-19. fight the COVID-19.

Keywords: COVID-19; CT findings; Infection quantification; Infection segmentation; Lung segmentation; Medical imaging.

Fig. 1

Flowchart of materials and method

Fig. 2

LCTSC database: (a) slice of institute 1; (b) slice of institute 2; (c) slice of institute 3

Fig. 3

HUPE database: Example of three patients with different degrees of infections

Fig. 4

Architecture of residual U-Net

Fig. 5

Segmentation refinement: Contour of the lungs segmented by Residual U-Net (a) and result of segmentation refinement

Fig. 6

Enhancement of pulmonary vessels: (a) Result of lung segmentation used to demarcate the region of application of the vessel enhancement filter, which results in (b)

Fig. 7

Extraction of pulmonary vessels from the markings made by the specialist: (a) original image; (b) image with emphasis on pulmonary vessels; (c) ground truth built by a specialist; and (d) ground truth without pulmonary vessels

Fig. 8

Examples of CT findings of COVID-19 infections: (a) early stage patient and (b) advanced stage patient

Fig. 9

Infections classification. (a) Original image. (b) GGO (green color) and consolidation (red color) can be seen in the left lung by axial, coronal and sagittal views, respectively

Fig. 10

Separation of the lungs. (a) Volume resulting from the lung segmentation step with the range of the projection histogram calculation illustrated by orange, (b) resulting histogram with the cutoff point illustrated by red, (c) left lung, and (d) right lung. Ground glass (GgV), and consolidated (CdV) region volume of the lesions and lung volume (LV). The volume of the lung was calculated by subtracting the volume of the pulmonary vessels. All measures were calculated for the left lung (LL), right lung (RL), and both lungs (BL) in milliliters (mL)

Fig. 11

Patient case study MM0015 01 in blue marked by the specialist and green marked by the method (a) slice 80, (b) slice 97, and (c) slice 114

Fig. 12

Infections classification from case studies MM0015 01: (a) original image. (b) GGO (green color) and consolidation (red color) can be seen by axial, coronal and saggittal views, respectively

Fig. 13

Patient case study MM0016 01 in blue marked by the specialist, green marked by the method (a) slice 93, (b) slice 107, and (c) slice 160

Fig. 14

Infections classification from case studies MM0016 01: (a) original image. (b) GGO (green color) and consolidation (red color) can be seen by axial, coronal and sagittal views, respectively