Classification of COVID-19 from community-acquired pneumonia: Boosting the performance with capsule network and maximum intensity projection image of CT scans.

BACKGROUND: The coronavirus disease 2019 (COVID-19) and community-acquired pneumonia (CAP) present a high degree of similarity in chest computed tomography (CT) images. Therefore, a procedure for accurately and automatically distinguishing between them is crucial. METHODS: A deep learning method for distinguishing COVID-19 from CAP is developed using maximum intensity projection (MIP) images from CT scans. LinkNet is employed for lung segmentation of chest CT images. MIP images are produced by superposing the maximum gray of intrapulmonary CT values. The MIP images are input into a capsule network for patient-level pred iction and diagnosis of COVID-19. The network is trained using 333 CT scans (168 COVID-19/165 CAP) and validated on three external datasets containing 3581 CT scans (2110 COVID-19/1471 CAP). RESULTS: LinkNet achieves the highest Dice coefficient of 0.983 for lung segmentation. For the classification of COVID-19 and CAP, the capsule network with the DenseNet-121 feature extractor outperforms ResNet-50 and Inception-V3, achieving an accuracy of 0.970 on the training dataset. Without MIP or the capsule network, the accuracy decreases to 0.857 and 0.818, respectively. Accuracy scores of 0.961, 0.997, and 0.949 are achieved on the external validation datasets. The proposed method has higher or comparable sensitivity compared with ten state-of-the-art methods. CONCLUSIONS: The proposed method illustrates the feasibility of applying MIP images from CT scans to distinguish COVID-19 from CAP using capsule networks. MIP images provide conspicuous benefits when exploiting deep learning to detect COVID-19 lesions from CT scans and the capsule network improves COVID-19 diagnosis.

Earlier diagnosis improves COVID-19 prognosis: a nationwide retrospective cohort analysis.

BACKGROUND: Risk of adverse outcomes in COVID-19 patients by stratifying by the time from symptom onset to confirmed diagnosis status is still uncertain. METHODS: We included 1,590 hospitalized COVID-19 patients confirmed by real-time RT-PCR assay or high-throughput sequencing of pharyngeal and nasal swab specimens from 575 hospitals across China between 11 December 2019 and 31 January 2020. Times from symptom onset to confirmed diagnosis, from symptom onset to first medical visit and from first medical visit to confirmed diagnosis were described and turned into binary variables by the maximally selected rank statistics method. Then, survival analysis, including a log-rank test, Cox regression, and conditional inference tree (CTREE) was conducted, regarding whether patients progressed to a severe disease level during the observational period (assessed as severe pneumonia according to the Chinese Expert Consensus on Clinical Practice for Emergency Severe Pneumonia, admission to an intensive care unit, administration of invasive ventilation, or death) as the prognosis outcome, the dependent variable. Independent factors included whether the time from symptom onset to confirmed diagnosis was longer than 5 days (the exposure) and other demographic and clinical factors as multivariate adjustments. The clinical characteristics of the patients with different times from symptom onset to confirmed diagnosis were also compared. RESULTS: The medians of the times from symptom onset to confirmed diagnosis, from symptom onset to first medical visit, and from first medical visit to confirmed diagnosis were 6, 3, and 2 days. After adjusting for age, sex, smoking status, and comorbidity status, age [hazard ratio (HR): 1.03; 95% CI: 1.01-1.04], comorbidity (HR: 1.84; 95% CI: 1.23-2.73), and a duration from symptom onset to confirmed diagnosis of >5 days (HR: 1.69; 95% CI: 1.10-2.60) were independent predictors of COVID-19 prognosis, which echoed the CTREE models, with significant nodes such as time from symptom onset to confirmed diagnosis, age, and comorbidities. Males, older patients with symptoms such as dry cough/productive cough/shortness of breath, and prior COPD were observed more often in the patients who procrastinated before initiating the first medical consultation. CONCLUSIONS: A longer time from symptom onset to confirmed diagnosis yielded a worse COVID-19 prognosis.

Changes of quantitative CT-based airway wall dimensions in patients with COVID-19 during early recovery.

BACKGROUND: As the coronavirus disease 19 (COVID-19) pandemic evolves, the need for recognizing the structural pulmonary changes of the disease during early convalescence has emerged. Most studies focus on parenchymal destruction of the disease; but little is known about whether the disease affects the airway. This study was conducted to investigate the changes in airway dimensions and explore the associated factors during early convalescence in patients with COVID-19. METHODS: We retrospectively analyzed quantitative computed tomography (CT)-based airway measures of 69 patients with COVID-19 from 5 February to 17 March 2020, and 32 non-COVID-19 participants from 1 January 2018 to 31 December 2019 from Guangzhou, China. The well-established measures of wall area fraction and the square root of the wall area of a hypothetical bronchus with an inner perimeter of 10 mm, were used to describe airway wall dimensions. We described the characteristics of the dimensions and inner area of airways in 66 patients with COVID-19 at the initial and convalescent stages of the disease, and compared them with the non-COVID-19 group. Linear regression models were constructed to investigate the association of airway dimensions with duration of hospitalization or disease severity after recovery. Partial correlation coefficients were calculated to investigate whether inflammatory markers were related to airway dimensions. RESULTS: Among 66 patients with COVID-19, airway dimensions were greater during disease initiation than early convalescence, which was significantly greater than in non-COVID-19 participants. No significant difference was found between the patients with COVID-19 at the initial stage and the non-COVID-19 controls regarding the first to eighth generations of the inner area. In adjusted regression models, duration of hospitalization was negatively associated with wall area fraction of the first to the sixth generation of airways. No significant associations exist between airway dimensions and disease severity, or airway dimensions with inflammatory markers. CONCLUSIONS: Airway dimensions in patients with COVID-19 during disease initiation are greater than those in non-COVID-19 participants. Such structural airway changes continue to remain significantly greater during early convalescence. No evidence shows that disease severity or inflammatory markers are associated with airway dimensions, implying that the primary lesion attacked by COVID-19 might not be the airways.