Deep learning model for pleural effusion detection via active learning and pseudo-labeling: a multisite study

BMC Medical Imaging

Table 3 Performance of the AI algorithm by each subpopulation

	Sensitivity (95% Wilson CI)	Specificity (95% Wilson CI)	AUC (95% DeLong’s CI)
Overall	0.95 (0.92, 0.97)	0.97 (0.95, 0.99)	0.97 (0.95, 0.99)
Gender
Female	0.93 (0.87, 0.96)	0.98 (0.94, 0.99)	0.96 (0.92, 0.99)
Male	0.96 (0.92, 0.98)	0.97 (0.93, 0.99)	0.98 (0.96, 1.00)
Age Group
18–49 y/o	0.93 (0.79, 0.98)	0.99 (0.95, 1.00)	0.96 (0.90, 1.00)
50–64 y/o	0.99 (0.95, 1.00)	0.96 (0.90, 0.99)	0.98 (0.95, 1.00)
Above 65 y/o	0.93 (0.88, 0.96)	0.97 (0.91, 0.99)	0.96 (0.94, 0.99)
Data Source
US	0.92 (0.87, 0.95)	0.96 (0.91, 0.98)	0.95 (0.92, 0.98)
Taiwan	0.99 (0.96, 1.00)	0.99 (0.96, 1.00)	1.00 (0.99, 1.00)
Manufacturer
Samsung Electronics	0.97 (0.89, 0.99)	0.99 (0.92, 1.00)	1.00 (0.99, 1.00)
Shimadzu	1.00 (0.95, 1.00)	0.99 (0.94, 1.00)	0.99 (0.98, 1.00)
Toshiba	0.98 (0.91, 1.00)	0.99 (0.94, 1.00)	1.00 (1.00, 1.00)
Others^a	0.88 (0.80, 0.93)	0.92 (0.82, 0.96)	0.91 (0.86, 0.96)
Size of Pleural Effusion
Small	0.96 (0.93, 0.98)	-
Moderate	0.90 (0.82, 0.95)	-
Large	1.00 (0.81, 1.00)	-
Location of Pleural Effusion
Right	0.92 (0.86, 0.95)	-
Left	0.99 (0.94, 1.00)	-
Bilateral	0.97 (0.89, 0.99)	-

^a Other X-ray manufacturers include Konica Minolta, GE Healthcare, Drtech, Canon Inc., Siemens, Oehm und Rehbein GmbH, Philips Medical Systems, Swissray, Kodak, Agfa, Fujifilm, and unknown.

ISSN: 1471-2342