Thursday, 28 March, 2024
HomeCoronavirusAI model for severity prognosis in COVID-19 patients — COVID-AI Brasil

AI model for severity prognosis in COVID-19 patients — COVID-AI Brasil

The first coronaviruses discovered in the world were responsible for respiratory and intestinal infections, of which the vast majority had a self-limited course and led mostly to symptoms of common cold, say a study by researchers at the Diagnósticos da América (Dasa) – São Paulo in Brazil. However, they can eventually develop into serious infections in groups at risk (heart diseases, diabetes, among others), in the elderly and also in children.

Before the beginning of this current pandemic, two highly pathogenic coronavirus species (SARS and MERS) were described and were responsible for outbreaks of severe acute respiratory syndromes. Regarding this new coronavirus (COVID-19) it was recognised as a causative agent of pneumonia that leads to severe acute respiratory syndrome (SARS-CoV-2).

One of its main challenges is its rapid transmission capacity and, in some cases, progression to severe pulmonary conditions that have demanded from the health system a care and combat strategy never seen before in the whole world.

In Brazil, the expectation is to be of exponential growth, which is why there is a need to implement drastic measures to control population circulation and prevention. Due to the speed of transmission, in most countries, including Brazil, early-stage preventive measures were not implemented, causing an explosion of symptomatic cases, many of them severe, with prolonged demand from tertiary health services.

Given this scenario, the inevitable emergence of a large contingent of critically ill patients with COVID-19, with different prognoses, made it crucial to search for early diagnostic mechanisms for better screening and treatment adequacy in each case.

In this context, even before confirmation of the infection, screening for patients with respiratory symptoms is carried out through clinical analysis and imaging tests such as chest computed tomography (CT).

In clinical evaluation, the main described symptoms of infection are fever (88.5%), cough (68.6%), myalgia or fatigue (35.8%), sputum (28, 2%) and dyspnoea (21.9%). Other symptoms also described include headache and dizziness (12.1%), diarrhoea (4.8%), nausea and vomiting (3.9%). In addition, some haematological changes were observed: lymphocytopenia (64.5%), increased C-reactive protein (CRP) (44.3%), increased lactic dehydrogenase (DHL) (28.3%), and leukopenia (29, 4%).

Chest CT is considered as the best imaging method for assessment of COVID-19, since conventional radiography has low sensitivity, notably in early stages. Typical findings described in the literature include ground-glass opacities (GGO) with a more peripheral distribution, associated with septal thickening and consolidations, usually affecting multiple lobes, although these findings can also be found in other viral pneumonias.

In this way, machine learning applied to diagnostic imaging can enable the development of tools that can standardise the diagnosis and provide potential findings suggestive of the presence of the disease, its severity, and therefore its prognosis.

The purpose of this study is to validate the results of AI algorithms in the stratification of pulmonary changes in chest CT with multi-centre data through scientific methodology to correlate pulmonary involvement with the clinical outcome of patients.

Study details
Machine learning model for predicting severity prognosis in patients infected with COVID-19: Study protocol from COVID-AI Brasil

Flávia Paiva Proença Lobo Lopes, Felipe Campos Kitamura , Gustavo Faibischew Prado, Paulo Eduardo de Aguiar Kuriki, Marcio Ricardo Taveira Garcia, COVID-AI-Brasil

Published in PLOS One on 1 February 2021

Abstract
The new coronavirus, which began to be called SARS-CoV-2, is a single-stranded RNA beta coronavirus, initially identified in Wuhan (Hubei province, China) and currently spreading across six continents causing a considerable harm to patients, with no specific tools until now to provide prognostic outcomes. Thus, the aim of this study is to evaluate possible findings on chest CT of patients with signs and symptoms of respiratory syndromes and positive epidemiological factors for COVID-19 infection and to correlate them with the course of the disease. In this sense, it is also expected to develop specific machine learning algorithm for this purpose, through pulmonary segmentation, which can predict possible prognostic factors, through more accurate results. Our alternative hypothesis is that the machine learning model based on clinical, radiological and epidemiological data will be able to predict the severity prognosis of patients infected with COVID-19. We will perform a multicenter retrospective longitudinal study to obtain a large number of cases in a short period of time, for better study validation. Our convenience sample (at least 20 cases for each outcome) will be collected in each center considering the inclusion and exclusion criteria. We will evaluate patients who enter the hospital with clinical signs and symptoms of acute respiratory syndrome, from March to May 2020. We will include individuals with signs and symptoms of acute respiratory syndrome, with positive epidemiological history for COVID-19, who have performed a chest computed tomography. We will assess chest CT of these patients and to correlate them with the course of the disease. Primary outcomes:1) Time to hospital discharge; 2) Length of stay in the ICU; 3) orotracheal intubation;4) Development of Acute Respiratory Discomfort Syndrome. Secondary outcomes:1) Sepsis; 2) Hypotension or cardiocirculatory dysfunction requiring the prescription of vasopressors or inotropes; 3) Coagulopathy; 4) Acute Myocardial Infarction; 5) Acute Renal Insufficiency; 6) Death. We will use the AUC and F1-score of these algorithms as the main metrics, and we hope to identify algorithms capable of generalizing their results for each specified primary and secondary outcome.

 

[link url="https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0245384"]PLOS One study (Open access)[/link]

MedicalBrief — our free weekly e-newsletter

We'd appreciate as much information as possible, however only an email address is required.