metricas
covid
Buscar en
Radiología (English Edition)
Toda la web
Inicio Radiología (English Edition) Development of severity and mortality prediction models for covid-19 patients at...
Journal Information
Vol. 64. Issue 3.
Pages 214-227 (May - June 2022)
Share
Share
Download PDF
More article options
Visits
1498
Vol. 64. Issue 3.
Pages 214-227 (May - June 2022)
Original articles
Development of severity and mortality prediction models for covid-19 patients at emergency department including the chest x-ray
Elaboración de modelos predictivos de la gravedad y la mortalidad en pacientes con COVID-19 que acuden al servicio de urgencias, incluida la radiografía torácica
Visits
1498
P. Calvillo-Batllésa,
Corresponding author
calvillo_mar@gva.es

Corresponding author.
, L. Cerdá-Alberichb, C. Fonfría-Esparciaa, A. Carreres-Ortegaa, C.F. Muñoz-Núñeza, L. Trilles-Olasoa, L. Martí-Bonmatía,b
a Servicio de Radiología, Hospital Universitario y Politécnico La Fe, Valencia, Spain
b Grupo de Investigación Biomédica en Imagen (GIBI230), Instituto de Investigación Sanitaria La Fe, Valencia, Spain
This item has received
Article information
Abstract
Full Text
Bibliography
Download PDF
Statistics
Figures (6)
Show moreShow less
Tables (4)
Table 1. Characteristics of COVID-19 patients. Demographic variables and comorbidities investigated as potential predictors. * Predictors in the final prognostic prediction models.
Table 2. Clinical presentation of COVID-19 patients. Clinical and laboratory variables investigated as potential predictors. Gastrointestinal symptoms: diarrhea, vomiting or abdominal pain. SatO2/FiO2: Peripheral oxygen saturation /inspiratory oxygen fraction (room air or oxygen therapy). * Predictors in the definitive prognostic prediction models.
Table 3. Lung involvement on CXR, distribution and extension. CXR features investigated as potential predictors. ExtScoreCXR: Extent score of lung involvement on CXR. * Predictor in the final prognostic prediction models.
Table 4. Metrics of the severity and mortality predictive models. Performance metrics of the internal validation performed with an unseen dataset for each of the selected severity and in-hospital mortality predictive models built with three different combinations of parameters. The Youden index was used for the optimal threshold selection of the classification models. RF: Random Forest, GB: Gradient Boosting.
Show moreShow less
Additional material (1)
Abstract
Objectives

To develop prognosis prediction models for COVID-19 patients attending an emergency department (ED) based on initial chest X-ray (CXR), demographics, clinical and laboratory parameters.

Methods

All symptomatic confirmed COVID-19 patients admitted to our hospital ED between February 24th and April 24th 2020 were recruited. CXR features, clinical and laboratory variables and CXR abnormality indices extracted by a convolutional neural network (CNN) diagnostic tool were considered potential predictors on this first visit. The most serious individual outcome defined the three severity level: 0) home discharge or hospitalization ≤ 3 days, 1) hospital stay >3 days and 2) intensive care requirement or death. Severity and in-hospital mortality multivariable prediction models were developed and internally validated. The Youden index was used for the optimal threshold selection of the classification model.

Results

A total of 440 patients were enrolled (median 64 years; 55.9% male); 13.6% patients were discharged, 64% hospitalized, 6.6% required intensive care and 15.7% died. The severity prediction model included oxygen saturation/inspired oxygen fraction (SatO2/FiO2), age, C-reactive protein (CRP), lymphocyte count, extent score of lung involvement on CXR (ExtScoreCXR), lactate dehydrogenase (LDH), D-dimer level and platelets count, with AUC-ROC = 0.94 and AUC-PRC = 0.88. The mortality prediction model included age, SatO2/FiO2, CRP, LDH, CXR extent score, lymphocyte count and D-dimer level, with AUC-ROC = 0.97 and AUC-PRC = 0.78. The addition of CXR CNN-based indices did not improve significantly the predictive metrics.

Conclusion

The developed and internally validated severity and mortality prediction models could be useful as triage tools in ED for patients with COVID-19 or other virus infections with similar behaviour.

Keywords:
COVID-19
Chest X-Ray
Prognosis
Mortality
Predictive models
Artificial intelligence
Resumen
Objetivos

Desarrollar modelos de predicción de pronóstico para pacientes con COVID-19 que acuden a urgencias, basados en la radiografía de tórax inicial (RXT), parámetros demográficos, clínicos y de laboratorio.

Métodos

Se reclutaron todos los pacientes sintomáticos con COVID-19 confirmada, que ingresaron en urgencias de nuestro hospital entre el 24 de febrero y el 24 de abril de 2020. Los parámetros de la RXT, las variables clínicas y de laboratorio y los índices de hallazgos en RXT extraídos por una herramienta diagnóstica de inteligencia artificial en esta primera visita se consideraron potenciales predictores. El desenlace individual más grave definió los tres niveles de gravedad: 0) alta domiciliaria u hospitalización de 3 días o inferior, 1) hospitalización más de 3 días y 2) necesidad de cuidados intensivos o muerte. Se desarrollaron y validaron internamente modelos de predicción multivariable de gravedad y mortalidad hospitalaria. El índice de Youden se utilizó para la selección del umbral óptimo del modelo de clasificación.

Resultados

Se registraron 440 pacientes (mediana de 64 años; 55,9% hombres); el 13,6% de los pacientes fueron dados de alta, el 64% hospitalizo más de 3 días, el 6,6% requirió cuidados intensivos y un 15,7% falleció. El modelo de predicción de gravedad incluyó saturación de oxígeno/fracción de oxígeno inspirado (SatO2/FiO2), edad, proteína C reactiva (PCR), linfocitos, puntuación de la extensión de la afectación pulmonar en la RXT (ExtScoreRXT), lactato deshidrogenasa (LDH), dímero D y plaquetas, con AUC-ROC = 0,94 y AUC-PRC = 0,88. El modelo de predicción de mortalidad incluyó edad, SatO2/FiO2, PCR, LDH, ExtScoreRXT, linfocitos y dímero D, con AUC-ROC = 0,97 y AUC-PRC = 0,78. La adición de índices radiológicos obtenidos por inteligencia artificial no mejoró significativamente las métricas predictivas.

Conclusión

Los modelos de predicción de pronóstico desarrollados podrían ser útiles para clasificar en urgencias a los pacientes con COVID-19 u otras infecciones víricas con comportamiento similar.

Palabras clave:
COVID-19
Radiografía torácica
Pronóstico
Mortalidad
Modelos predictivos
Inteligencia artificial

Article

These are the options to access the full texts of the publication Radiología (English Edition)
Subscriber
Subscriber

If you already have your login data, please click here .

If you have forgotten your password you can you can recover it by clicking here and selecting the option “I have forgotten my password”
Purchase
Purchase article

Purchasing article the PDF version will be downloaded

Price 19.34 €

Purchase now
Contact
Phone for subscriptions and reporting of errors
From Monday to Friday from 9 a.m. to 6 p.m. (GMT + 1) except for the months of July and August which will be from 9 a.m. to 3 p.m.
Calls from Spain
932 415 960
Calls from outside Spain
+34 932 415 960
E-mail
Article options
es en pt

¿Es usted profesional sanitario apto para prescribir o dispensar medicamentos?

Are you a health professional able to prescribe or dispense drugs?

Você é um profissional de saúde habilitado a prescrever ou dispensar medicamentos

Quizás le interese:
10.1016/j.rxeng.2021.05.001
No mostrar más