Skip to main content

Table 4 Multiple disease detection. The datasets, manifestations, assessment measures and results are shown in each column, respectively

From: Computer-aided detection in chest radiography based on artificial intelligence: a survey

Study Datasets Conditions Measurements Results
Avni et al. [104] Custom Left and right pulmonary pleural effusion, cardiomegaly, and septum enlargement AUC Left and right pulmonary pleural effusion: 80%; cardiomegaly: 79.2%; septum enlargement: 88.2%
Noor et al. [105] Custom Lobar pneumonia, tuberculosis, and lung cancer Accuracy 70%, 97%, and 79%, respectively
Bar et al. [75] Custom Right pleural effusion, cardiomegaly, health, and abnormal disease AUC 93%, 89%, and 79%, respectively
Cicero et al. [106] Custom Normal, cardiomegaly, pleural effusion, pulmonary edema, and pneumothorax AUC 96.4%, 87.5%, 85%, 96.2%, 86.8%, and 86.1%, respectively
Wang et al. [23] Chest-Xray14 14 common diseases in CXRs AUC Mean: 73.8%
Yao et al. [107] Chest-Xray14 14 common diseases in CXRs AUC Mean: 80.3%; however, limited training focuses on biased interdependence and cannot accurately represent the actual distribution of morbidities
Rajpurkar et al. [13] Chest-Xray14 14 common diseases in CXRs AUC Mean: 84.2%; pneumonia (76.8%) exceeded the human level
Kumar et al. [109] Chest-Xray14 14 common diseases in CXRs AUC Mean: 79.5%; cardiomegaly (91.33%) beyond the previous method
Guan et al. [111] Chest-Xray14 14 common diseases in CXRs AUC Mean: 87.1%