Comparison of the Discrimination Performance of AI Scoring and the Brixia Score in Predicting COVID-19 Severity on Chest X-Ray Imaging: Diagnostic Accuracy Study

BACKGROUND

The artificial intelligence (AI) analysis of chest x-rays can increase the precision of binary COVID-19 diagnosis. However, it is unknown if AI-based chest x-rays can predict who will develop severe COVID-19, especially in low- and middle-income countries.

OBJECTIVE

The study aims to compare the performance of human radiologist Brixia scores versus 2 AI scoring systems in predicting the severity of COVID-19 pneumonia.

METHODS

We performed a cross-sectional study of 300 patients suspected with and with confirmed COVID-19 infection in Jakarta, Indonesia. A total of 2 AI scores were generated using CAD4COVID x-ray software.

RESULTS

The AI probability score had slightly lower discrimination (area under the curve [AUC] 0.787, 95% CI 0.722-0.852). The AI score for the affected lung area (AUC 0.857, 95% CI 0.809-0.905) was almost as good as the human Brixia score (AUC 0.863, 95% CI 0.818-0.908).

CONCLUSIONS

The AI score for the affected lung area and the human radiologist Brixia score had similar and good discrimination performance in predicting COVID-19 severity. Our study demonstrated that using AI-based diagnostic tools is possible, even in low-resource settings. However, before it is widely adopted in daily practice, more studies with a larger scale and that are prospective in nature are needed to confirm our findings.

Derivation and validation of novel integrated inpatient mortality prediction score for COVID-19 (IMPACT) using clinical, laboratory, and AI-processed radiological parameter upon admission: a multicentre study

Limited studies explore the use of AI for COVID-19 prognostication. This study investigates the relationship between AI-aided radiographic parameters, clinical and laboratory data, and mortality in hospitalized COVID-19 patients. We conducted a multicentre retrospective study. The derivation and validation cohort comprised of 512 and 137 confirmed COVID-19 patients, respectively. Variable selection for constructing an in-hospital mortality scoring model was performed using the least absolute shrinkage and selection operator, followed by logistic regression. The accuracy of the scoring model was assessed using the area under the receiver operating characteristic curve. The final model included eight variables: anosmia (OR: 0.280; 95%CI 0.095-0.826), dyspnoea (OR: 1.684; 95%CI 1.049-2.705), loss of consciousness (OR: 4.593; 95%CI 1.702-12.396), mean arterial pressure (OR: 0.928; 95%CI 0.900-0.957), peripheral oxygen saturation (OR: 0.981; 95%CI 0.967-0.996), neutrophil % (OR: 1.034; 95%CI 1.013-1.055), serum urea (OR: 1.018; 95%CI 1.010-1.026), affected lung area score (OR: 1.026; 95%CI 1.014-1.038). The Integrated Inpatient Mortality Prediction Score for COVID-19 (IMPACT) demonstrated a predictive value of 0.815 (95% CI 0.774-0.856) in the derivation cohort. Internal validation resulted in an AUROC of 0.770 (95% CI 0.661-0.879). Our study provides valuable evidence of the real-world application of AI in clinical settings. However, it is imperative to conduct prospective validation of our findings, preferably utilizing a control group and extending the application to broader populations.