Convolutional neural network for detecting rib fractures on chest radiographs: a feasibility study

BMC Medical Imaging

Table 3 Comparison of sensitivity and precision in the independent testing group, based on fracture level

Data	Marked fractures	Detected fractures	Correctly detected fractures	Sensitivity	Precision
CNN model	402	437	351	87.3%	80.32%
Senior radiologist	402	392	323	80.3%	82.40%
Junior radiologist	402	361	295	73.4%	81.72%
P₁	NA	NA	NA	0.15	0.57
P₂	NA	NA	NA	0.13	0.43
P₃	NA	NA	NA	0.01	0.43

Sensitivity is the number of fractures detected correctly divided by the number of fractures marked. Precision is the number of fractures detected correctly divided by the number of fractures detected. P₁ is the P value for the senior radiologists versus the junior radiologists. P₂ is the P value for the CNN versus the senior radiologist. P₃ is the P value for the CNN versus the junior radiologist. Comparisons were conducted by using the chi-squared test
NA not available, CNN convolutional neural network

ISSN: 1471-2342