An Interpretable Chest CT Deep Learning Algorithm for Quantification of COVID-19 Lung Disease and Prediction of Inpatient Morbidity and Mortality

Chamberlin, Jordan; Aquino, Gilberto; Schoepf, U. Joseph; Nance, Sophia; Godoy, Franco; Carson, Landin; Giovagnoli, Vincent; Gill, Callum E.; McGill, Liam; O’Doherty, Jim; Emrich, Tilman; Burt, Jeremy R.; Baruah, Dhiraj; Varga-Szemes, Ákos; Kabakus, Ismail

doi:10.1016/j.acra.2022.03.023

Cited by 10 publications

(5 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, because CT images were reviewed by radiologists and scored according to the degree of involvement, it was difficult to regard the CT severity score as a quantitative indicator. Recently, various studies have reported that the clinical outcome in COVID-19 patients can be predicted using QCT through AI software or deep learning machines [23,[40][41][42]. Our study showed that the QCT COVID scores at admission could predict rapid progression in patients with COVID-19.…”

Section: Discussionsupporting

confidence: 50%

Quantitative Computed Tomography Lung COVID Scores with Laboratory Markers: Utilization to Predict Rapid Progression and Monitor Longitudinal Changes in Patients with Coronavirus 2019 (COVID-19) Pneumonia

Kang,

Kim,

Park

et al. 2024

Biomedicines

View full text Add to dashboard Cite

Coronavirus disease 2019 (COVID-19), is an ongoing issue in certain populations, presenting rapidly worsening pneumonia and persistent symptoms. This study aimed to test the predictability of rapid progression using radiographic scores and laboratory markers and present longitudinal changes. This retrospective study included 218 COVID-19 pneumonia patients admitted at the Chungnam National University Hospital. Rapid progression was defined as respiratory failure requiring mechanical ventilation within one week of hospitalization. Quantitative COVID (QCOVID) scores were derived from high-resolution computed tomography (CT) analyses: (1) ground glass opacity (QGGO), (2) mixed diseases (QMD), and (3) consolidation (QCON), and the sum, quantitative total lung diseases (QTLD). Laboratory data, including inflammatory markers, were obtained from electronic medical records. Rapid progression was observed in 9.6% of patients. All QCOVID scores predicted rapid progression, with QMD showing the best predictability (AUC = 0.813). In multivariate analyses, the QMD score and interleukin(IL)-6 level were important predictors for rapid progression (AUC = 0.864). With >2 months follow-up CT, remained lung lesions were observed in 21 subjects, even after several weeks of negative reverse transcription polymerase chain reaction test. AI-driven quantitative CT scores in conjugation with laboratory markers can be useful in predicting the rapid progression and monitoring of COVID-19.

show abstract

Section: Discussionsupporting

confidence: 50%

Quantitative Computed Tomography Lung COVID Scores with Laboratory Markers: Utilization to Predict Rapid Progression and Monitor Longitudinal Changes in Patients with Coronavirus 2019 (COVID-19) Pneumonia

Kang,

Kim,

Park

et al. 2024

Biomedicines

View full text Add to dashboard Cite

show abstract

“…Scoring systems are a different method to assess AI results, which we consider to be less precise compared to the percentage of the volume of affected lungs. Also, the patient population in the above study was smaller and less homogenous compared to ours (27). In our study, we defined the hazard rates using statistical methods, namely, regression analysis and the Youden index.…”

Section: Discussionmentioning

confidence: 89%

“…Several published studies demonstrated the feasibility of using CT findings as well as AI-based CT evaluation to predict COVID-19 pneumonia patients' outcome in various ways (17,18,19,20,21,22,23,24,25,26,27). However, to the best of our knowledge, the hazard rates based on AI-driven CT percentages of the extent of COVID-19 pneumonia on such a large dataset of patients have not been defined yet.…”

Section: Introductionmentioning

confidence: 99%

CT quantification of COVID-19 pneumonia extent to predict individualized outcome

BERECOVA,

JUSKANIC,

HAZLINGER

et al. 2024

BLL

View full text Add to dashboard Cite

OBJECTIVES: This study aimed to predict individual COVID-19 patient prognosis at hospital admission using artificial intelligence (AI)-based quantification of computed tomography (CT) pulmonary involvement. BACKGROUND: Assessing patient prognosis in COVID-19 pneumonia is crucial for patient management and hospital and ICU organization. METHODS: We retrospectively analyzed 559 patients with PCR-verified COVID-19 pneumonia referred to the hospital for a severe disease course. We correlated the CT extent of pulmonary involvement with patient outcome. We also attempted to define cut-off values of pulmonary involvement for predicting different outcomes. RESULTS: CT-based disease extent quantification is an independent predictor of patient morbidity and mortality, with the prognosis being impacted also by age and cardiovascular comorbidities. With the use of explored cut-off values, we divided patients into three groups based on their extent of disease: (1) less than 28 % (sensitivity 65.4 %; specificity 89.1 %), (2) ranging from 28 % (31 %) to 47 % (sensitivity 87.1 %; specificity 62.7 %), and (3) above 47 % (sensitivity 87.1 %; specificity, 62.7 %), representing low risk, risk for oxygen therapy and invasive pulmonary ventilation, and risk of death, respectively. CONCLUSION: CT quantification of pulmonary involvement using AI-based software helps predict COVID-19 patient outcomes (Tab. 4, Fig. 4, Ref. 38).

show abstract

“…We observed that machine learning analyses of the variables of the severity of lung involvement in The existing literature on Covid-19 and the estimation of length of stay (LoS) was found to be limited. In a study involving 254 patients conducted by Chamberlin et al, a score other than CORADS was used to predict hospitalization and an accuracy of 0.77 was determined (24). In a different study by Purkayastha et al that used a different scoring system than CORADS, the accuracy levels of the ML analysis results using the "boosting" and "bayesian" methods were 0.68 (p=0.023) and 0.77 (p=0.023), respectively (25).…”

Section: Resultsmentioning

confidence: 99%

Prediction of Short or Long Length of Stay COVID-19 by Machine Learning

et al. 2023

View full text Add to dashboard Cite

Purpose: To predict with high accuracy, using machine learning, if the length of stay of Covid-19 patients will be short or not, based on the basic clinical parameters. Method: Seven primary variables, including age, gender, length of stay, c-reactive protein, ferritin, lymphocyte, and the COVID-19 Reporting and Data System (CORADS), were scanned and analyzed for 118 adult patients hospitalized with the diagnosis of Covid-19 between November 2020 and January 2021. Randomly, the data set is split into a training set of 80% and a test set of 20%. Using the caret package of the R programming language, machine learning models attempted to predict whether the length of stay was short or long, and the performance values of these models were recorded. Results: Among the models, the k-nearest neighbor model produced the best results. According to this model, the following were the estimation performance results for hospitalizations of 5 days or less or more than 5 days: The accuracy rate was 0.92 (95% CI, 0.73-0.99), the no-information rate was 0.67, the Kappa rate was 0.82, and the F1 score was 0.89 (p=0.0048). Conclusion: By applying machine learning into Covid-19, length of stay estimations may be made more accurately, providing for more effective patient management.

show abstract

An Interpretable Chest CT Deep Learning Algorithm for Quantification of COVID-19 Lung Disease and Prediction of Inpatient Morbidity and Mortality

Cited by 10 publications

References 33 publications

Quantitative Computed Tomography Lung COVID Scores with Laboratory Markers: Utilization to Predict Rapid Progression and Monitor Longitudinal Changes in Patients with Coronavirus 2019 (COVID-19) Pneumonia

Quantitative Computed Tomography Lung COVID Scores with Laboratory Markers: Utilization to Predict Rapid Progression and Monitor Longitudinal Changes in Patients with Coronavirus 2019 (COVID-19) Pneumonia

CT quantification of COVID-19 pneumonia extent to predict individualized outcome

Prediction of Short or Long Length of Stay COVID-19 by Machine Learning

Contact Info

Product

Resources

About