Abstract
Background: One of the most challenging aspects related to Covid-19 is to establish the presence of infection in an early phase of the disease. Texture analysis might be an additional tool for the evaluation of Chest X-ray in patients with clinical suspicion of Covid-19 related pneumonia.
Objective: To evaluate the diagnostic performance of texture analysis and machine learning models for the diagnosis of Covid-19 interstitial pneumonia in Chest X-ray images.
Methods: Chest X-ray images were accessed from a publicly available repository(https://www.kaggle. com/tawsifurrahman/covid19-radiography-database). Lung areas were manually segmented using a polygonal region of interest covering both lung areas, using MaZda, a freely available software for texture analysis. A total of 308 features per ROI was extracted. One hundred-ten Covid-19 Chest X-ray images were selected for the final analysis.
Results: Six models, namely NB, GLM, DL, GBT, ANN, and PLS-DA were selected and ensembled. According to Youden’s index, the Covid-19 Ensemble Machine Learning Score showing the highest area under the curve (0.971±0.015) was 132.57. Assuming this cut-off the Ensemble model performance was estimated by evaluating both true and false positive/negative, resulting in 91.8% accuracy with 93% sensitivity and 90% specificity. Moving the cut-off value to -100, although the accuracy resulted lower (90.6%), the Ensemble Machine Learning showed 100% sensitivity, with 80% specificity.
Conclusion: Texture analysis of Chest X-ray images and machine learning algorithms may help in differentiating patients with Covid-19 pneumonia. Despite several limitations, this study can lay the ground for future research works in this field and help to develop more rapid and accurate screening tools for these patients.
Keywords: X-ray, COVID-19, pneumonia, thorax, interstitial pneumonia, radiomics, texture analysis.
Graphical Abstract
[http://dx.doi.org/10.1080/22221751.2020.1725399] [PMID: 32020836]
[http://dx.doi.org/10.1056/NEJMoa2001017] [PMID: 31978945]
[http://dx.doi.org/10.1136/bmj.m1036] [PMID: 32165426]
[http://dx.doi.org/10.1001/jama.2020.1585] [PMID: 32031570]
[http://dx.doi.org/10.1038/s41586-020-2196-x] [PMID: 32235945]
[http://dx.doi.org/10.1001/jama.2020.3786] [PMID: 32159775]
[http://dx.doi.org/10.1001/jama.2020.8259] [PMID: 32374370]
[PMID: 32216717]
[http://dx.doi.org/10.1177/0846537120916419] [PMID: 32233876]
[http://dx.doi.org/10.2214/AJR.20.22954] [PMID: 32130038]
[http://dx.doi.org/10.4239/wjd.v10.i6.362] [PMID: 31231459]
[http://dx.doi.org/10.1016/S0140-6736(20)30627-9] [PMID: 32178769]
[http://dx.doi.org/10.1148/radiol.2015151169] [PMID: 26579733]
[http://dx.doi.org/10.2214/AJR.18.20624] [PMID: 30645163]
[http://dx.doi.org/10.1067/j.cpradiol.2019.10.009] [PMID: 31761413]
[http://dx.doi.org/10.1097/RTI.0000000000000268] [PMID: 28346329]
[http://dx.doi.org/10.1007/s11432-020-2849-3]
[http://dx.doi.org/10.1007/s11547-020-01195-x] [PMID: 32367319]
[http://dx.doi.org/10.1109/TMI.2020.2993291] [PMID: 32396075]
[http://dx.doi.org/10.1016/j.mri.2003.09.001]
[http://dx.doi.org/10.1007/s11306-017-1274-z]
[http://dx.doi.org/10.1007/s11306-018-1370-8] [PMID: 30830338]
[http://dx.doi.org/10.1021/acs.jproteome.7b00503] [PMID: 29235868]
[http://dx.doi.org/10.2307/2531595] [PMID: 3203132]
[http://dx.doi.org/10.1007/s00330-020-06898-3] [PMID: 32377812]
[http://dx.doi.org/10.1002/jum.15284] [PMID: 32198775]
[http://dx.doi.org/10.1002/jum.15285] [PMID: 32227492]