Tracking and predicting COVID-19 radiological trajectory on chest X-rays using deep learning

Abstract

Radiological findings on chest X-ray (CXR) have shown to be essential for the proper management of COVID-19 patients as the maximum severity over the course of the disease is closely linked to the outcome. As such, evaluation of future severity from current CXR would be highly desirable. We trained a repurposed deep learning algorithm on the CheXnet open dataset (224,316 chest X-ray images of 65,240 unique patients) to extract features that mapped to radiological labels. We collected CXRs of COVID-19-positive patients from an open-source dataset (COVID-19 image data collection) and from a multi-institutional local ICU dataset. The data was grouped into pairs of sequential CXRs and were categorized into three categories: ‘Worse’, ‘Stable’, or ‘Improved’ on the basis of radiological evolution ascertained from images and reports. Classical machine-learning algorithms were trained on the deep learning extracted features to perform immediate severity evaluation and prediction of future radiological trajectory. Receiver operating characteristic analyses and Mann-Whitney tests were performed. Deep learning predictions between “Worse” and “Improved” outcome categories and for severity stratification were significantly different for three radiological signs and one diagnostic (‘Consolidation’, ‘Lung Lesion’, ‘Pleural effusion’ and ‘Pneumonia’; all P < 0.05). Features from the first CXR of each pair could correctly predict the outcome category between ‘Worse’ and ‘Improved’ cases with a 0.81 (0.74–0.83 95% CI) AUC in the open-access dataset and with a 0.66 (0.67–0.64 95% CI) AUC in the ICU dataset. Features extracted from the CXR could predict disease severity with a 52.3% accuracy in a 4-way classification. Severity evaluation trained on the COVID-19 image data collection had good out-of-distribution generalization when testing on the local dataset, with 81.6% of intubated ICU patients being classified as critically ill, and the predicted severity was correlated with the clinical outcome with a 0.639 AUC. CXR deep learning features show promise for classifying disease severity and trajectory. Once validated in studies incorporating clinical data and with larger sample sizes, this information may be considered to inform triage decisions.

Deep learning of chest X-rays can predict mechanical ventilation outcome in ICU-admitted COVID-19 patients

“…The collected data was first presented in another publication. where it was used for severity prediction 16 .…”

“…We will first define the various datasets on which the system was trained, validated and tested, before describing the algorithms employed. The model development process is similar to that of a previous publication 16 .…”

“…Using transfer learning, the learned CXR-specific domain knowledge from large scale datasets could be leveraged and applied to smaller COVID-19 CXR datasets. A previous work 16 demonstrated the feasibility of developing COVID-19 models on small datasets using transfer learning for COVID-19 radiological severity assessment and progression. Parts of the literature also used transfer learning in assessing radiological severity of the disease 16 – 18 and predicting the potential for adverse events 19 for hospital-admitted patients; none have tackled outcome once on ventilation.…”

“…A previous work 16 demonstrated the feasibility of developing COVID-19 models on small datasets using transfer learning for COVID-19 radiological severity assessment and progression. Parts of the literature also used transfer learning in assessing radiological severity of the disease 16 – 18 and predicting the potential for adverse events 19 for hospital-admitted patients; none have tackled outcome once on ventilation. Furthermore, most prognosis models are either based on imaging data or demographics/risk factors alone.…”

Still a role for the chest radiograph – humble but helpful!

“…The authors conclude that the chest radiograph scoring system is a useful screening tool in this setting that has important clinical implications, and as a consequence may limit morbidity and mortality.Several others have demonstrated and reported similar utility and benefit of the chest radiograph in this context, including the association with severity and need for intubation and mechanical ventilation. [2][3][4][5][6][7][8][9][10] Despite major technological advances in chest imaging, including elements such as pragmatic point-of-care ultrasound, the chest radiograph remains an essential tool in our daily clinical armamentarium. Mastering interpretation of the chest radiograph enhances one's clinical acumen, and its utility in the clinical context for the astute clinician is protean.…”