Skip to main content
. 2021 Jan 25;11:10. doi: 10.1186/s13550-021-00751-4

Table 3.

The side-based sensitivity, specificity, and accuracy values of the human (clinicians') diagnoses and AI-assisted diagnoses on the 414 samples

Graded as positive Clinicians with/without AI assistance Sensitivity (%) Specificity (%) Accuracy (%)
3, 4, 5 Clinician A w/o AI 74.0 96.7 85.5
Clinician A w/ AI 76.5 94.3 85.5
Clinician B w/o AI 68.6 99.0 84.1
Clinician B w/ AI 68.6 99.0 84.1
4, 5 Clinician A w/o AI 59.8 99.0 79.7
Clinician A w/ AI 68.6 99.0 84.1
Clinician B w/o AI 57.4 99.5 78.7
Clinician B w/ AI 64.2 99.5 82.1
5 Clinician A w/o AI 37.3 100 69.1
Clinician A w/ AI 54.9 99.5 77.5
Clinician B w/o AI 33.8 100 67.4
Clinician B w/ AI 54.4 100 77.5