This dataset contains anonymised human lung computed tomography (CT) scans with COVID-19 related findings, as well as without such findings. A small subset of studies has been annotated with binary pixel masks depicting regions of interests (ground-glass opacifications and consolidations). CT scans were obtained between 1st of March, 2020 and 25th of April, 2020, and provided by municipal hospitals in Moscow, Russia. Permanent link: https://mosmed.ai/datasets/covid19_1110 . This dataset is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported (CC BY-NC-ND 3.0) License.

With the ongoing COVID-19 pandemic decreasing availability of polymerase chain reaction with reverse transcription and the snowballing growth of medical imaging, especially the number of chest computed tomography (CT) scans being performed, methods to augment and automate the image analysis, increasing productivity and minimizing human error are of particular importance. The creation of high-quality datasets is essential for the development and validation of artificial intelligence algorithms. Such technologies have sufficient accuracy in diagnosing COVID-19 in medical imaging. The presented large-scale dataset contains anonymized human CT scans with COVID-19 features as well as normal studies. Some studies were tagged by radiologists using binary pixel masks of regions of interest (e.g., characteristic areas of consolidation and ground-glass opacities). CT data were acquired between March 1, 2020, and April 25, 2020, and provided by municipal hospitals in Moscow, Russia. The presented dataset is licensed under Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported (CC BY-NC-ND 3.0).

The current COVID-19 pandemic overloads healthcare systems, including radiology departments. Though several deep learning approaches were developed to assist in CT analysis, nobody considered study triage directly as a computer science problem. We describe two basic setups: Identification of COVID-19 to prioritize studies of potentially infected patients to isolate them as early as possible; Severity quantification to highlight patients with severe COVID-19, thus direct them to a hospital or provide emergency medical care. We formalize these tasks as binary classification and estimation of affected lung percentage. Though similar problems were well-studied separately, we show that existing methods could provide reasonable quality only for one of these setups. We employ a multitask approach to consolidate both triage approaches and propose a convolutional neural network to leverage all available labels within a single model. In contrast with the related multitask approaches, we show the benefit from applying the classification layers to the most spatially detailed feature map at the upper part of U-Net instead of the less detailed latent representation at the bottom. We train our model on approximately 1500 publicly available CT studies and test it on the holdout dataset that consists of 123 chest CT studies of patients drawn from the same healthcare system, specifically 32 COVID-19 and 30 bacterial pneumonia cases, 30 cases with cancerous nodules, and 31 healthy controls. The proposed multitask model outperforms the other approaches and achieves ROC AUC scores of 0.87±0.01 vs. bacterial pneumonia, 0.93±0.01 vs. cancerous nodules, and 0.97±0.01 vs. healthy controls in Identification of COVID-19, and achieves 0.97±0.01 Spearman Correlation in Severity quantification. We have released our code and shared the annotated lesions masks for 32 CT images of patients with COVID-19 from the test dataset.

The objective: to predict lethal outcomes in patients with COVID-19 based on the result of chest computed tomography (chest CT) using a semi-quantitative visual scale of the pulmonary parenchyma lesion.Subjects and methods. Inclusion criteria: patients who underwent chest CT from March 2 to May 1, 2020, inclusively, upon referral of a general practitioner due to suspected community-acquired pneumonia caused by COVID-19. Chest CT was performed in 48 medical organizations providing primary medical care to the adult population in Moscow. Exclusion criteria: patients whose chest CT was not assessed by CT 0-4 score; patients who were not confirmed as COVID-19 positive.In the Russian Federation, CT 0-4 score is recommended to be used for assessment of the extent of lung parenchyma lesion in cases with suspected COVID-19.Data on fatal outcomes were received on May 4, 2020 inclusive.Results: data of 13,003 patients from the Unified Radiological Information Service were retrospectively included in the study according to inclusion and exclusion criteria. The test aimed to detect a trend of directional changes in the proportion of deceased patients among various categories using CT 0-4 score demonstrated a statistically significant result (p < 0.0001). The chance of a lethal outcome increases directionally from CT-0 to CT-4. The test for deviations from the linear trend also provided p < 0.0001, i.e. when moving to higher scores (CT-3 and CT-4), there was an accelerated increase in the risk of death. Analysis of overall survival using the Cox regression model showed that the assessed factors (age and CT 0-4 score) were statistically significantly associated with the time to death from COVID-19 (p < 0.05). The risk of death increased with age, on average, 8.6% for every 5 years (95% CI 0.8-17.0%). When transferring from one category of CT to the next one, the risk increased by 38% on the average (95% CI 17.1-62.6%). There was no statistically significant association of gender factor with overall survival (p = 0.408).The visual score of CT 0-4 recommended for use in the Russian Federation to assess lung parenchyma lesions according to chest CT data, is a predictor of a lethal outcome in patients with COVID-19. CT 0-4 score is convenient for practical use.

BACKGROUND:In the current epidemiological situation, computed tomography (CT) of chest organs plays an important role in disease diagnosis. Clinical and CT data allow physicians to quickly establish the probability of the presence and prognosis of patients with coronavirus disease 2019 (COVID-19). AIMS:This study aimed to predict outcomes in patients with laboratory-confirmed COVID-19 based on chest CT and a semi-quantitative visual pulmonary lesion grading system (CT 04). MATERIALS AND METHODS:A retrospective analysis of the Unified Medical Information and Analytical Service and Unified Radiological Information Service records from March 01, 2020 to July 30, 2020 was performed. The inclusion criteria were as follows: patients diagnosed with U07.1 (laboratory-verified coronavirus infection) from March 01, 2020 to July 30, 2020 and referred for a chest CT by a physician with suspected community-acquired pneumonia caused by COVID-19; the maximum period between laboratory verification and CT was not more than five days. The observation period for each patient was at least till 30 days from the date of CT. CT was performed in 48 medical organizations providing primary medical care to adults in Moscow. The exclusion criterion was a negative reverse transcription-polymerase chain reaction results by July30, 2020. The CT 04 scale is recommended for use in the Russian Federation to estimate the volume of lung parenchyma lesions when COVID-19 is suspected. RESULTS:The total sample volume was 38,051 patients. In this study, the risk of death was three times higher for CT-4 than for CT-0. In the KaplanMeier survival curve, the survival rate of patients in the CT-3 category was almost three times lower (hazard ratio = 2.94) than in the CT 02 categories; in addition, the higher the initial category of CT, the lower the risk of deterioration. The time for hospitalization decreased with the increase in the CT grade. CONCLUSION:The visual CT 04 scale can be used to predict outcomes, such as hospitalizations and deaths, in patients suspected of COVID-19 who underwent chest CT in primary health care.