Diagnostic Accuracy of Computed Tomography for Identifying Hospitalization in Patients with Suspected COVID-19
Authors:
Sergey P. Morozov,
Roman V. Reshetnikov,
Victor A. Gombolevskiy,
Natalia V. Ledikhova,
Ivan A. Blokhin,
Vladislav G. Kljashtorny,
Olesya A. Mokienko,
Anton V. Vladzymyrskyy
Abstract:
The controversy of computed tomography (CT) use in COVID-19 screening is associated with ambiguous characteristics of chest CT as a diagnostic test. The reported values of CT sensitivity and specificity calculated using RT-PCR as a reference standard vary widely. The objective of this study was to reevaluate the diagnostic and prognostic value of CT using an alternative approach. This study includ…
▽ More
The controversy of computed tomography (CT) use in COVID-19 screening is associated with ambiguous characteristics of chest CT as a diagnostic test. The reported values of CT sensitivity and specificity calculated using RT-PCR as a reference standard vary widely. The objective of this study was to reevaluate the diagnostic and prognostic value of CT using an alternative approach. This study included 973 symptomatic COVID-19 patients aged 42 $\pm$ 17 years, 56% females. We reviewed the disease dynamics between the initial and follow-up CT studies using a "CT0-4" grading system. Sensitivity and specificity were calculated as conditional probabilities that a patient's condition would improve or deteriorate relative to the initial CT study results. For the calculation of negative (NPV) and positive (PPV) predictive values, we estimated the COVID-19 prevalence in Moscow. We used several ARIMA and EST models with different parameters to fit the data on total cases of COVID-19 from March 6, 2020, to July 20, 2020, and forecast the incidence. The "CT0-4" grading scale demonstrated low sensitivity (28%) but high specificity (95%). The best statistical model for describing the pandemic in Moscow was ETS with multiplicative trend, error, and season type. According to our calculations, with the predicted prevalence of 2.1%, the values of NPV and PPV would be 98% and 10%, correspondingly. We associate the low sensitivity and PPV values with the small sample size of the patients with severe symptoms and non-optimal methodological setup for measuring these specific characteristics. The "CT0-4" grading scale was highly specific and predictive for identifying admissions to hospitals of COVID-19 patients. Despite the ambiguous accuracy, chest CT proved to be an effective practical tool for patient management during the pandemic, provided that the necessary infrastructure and human resources are available.
△ Less
Submitted 31 July, 2020; v1 submitted 29 July, 2020;
originally announced July 2020.
MosMedData: Chest CT Scans With COVID-19 Related Findings Dataset
Authors:
S. P. Morozov,
A. E. Andreychenko,
N. A. Pavlov,
A. V. Vladzymyrskyy,
N. V. Ledikhova,
V. A. Gombolevskiy,
I. A. Blokhin,
P. B. Gelezhe,
A. V. Gonchar,
V. Yu. Chernina
Abstract:
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 hospital…
▽ More
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. Key words: artificial intelligence, COVID-19, machine learning, dataset, CT, chest, imaging
△ Less
Submitted 13 May, 2020;
originally announced May 2020.
