BMC Pulmonary Medicine

Table 2 Prediction performance of DL models for COPD based on CXR images and clinical data in internal test

From: Screening and staging of chronic obstructive pulmonary disease with deep learning based on chest X-ray images and clinical parameters

	Based on CXR images only				Based on clinical data only				Multi-mode
		EfficientNet	ResNet50	DenseNet		Random Forest	Support Vector	Decision Tree		Ensemble
AUC		0.946	0.942	0.934		0.963	0.953	0.887		0.969
ACC		0.890	0.880	0.880		0.910	0.910	0.890		0.920
PPV		0.860	0.860	0.840		0.890	0.880	0.860		0.870
NPV		0.910	0.890	0.920		0.940	0.950	0.920		0.960
Sensitivity		0.910	0.890	0.920		0.940	0.960	0.920		0.960
Specificity		0.860	0.860	0.830		0.880	0.860	0.850		0.860
F1		0.890	0.880	0.880		0.910	0.920	0.890		0.920

Abbreviations DL, deep learning; CXR, chest x ray; COPD, chronic obstructive pulmonary disease; AUC, area under the curve; ACC, accuracy; PPV, positive predictive value; NPV, negative predictive value; F1, false positive rate

Back to article page

ISSN: 1471-2466

Contact us

Submission enquiries: bmcpulmonarymedicine@biomedcentral.com
General enquiries: ORSupport@springernature.com