Table 4.
Statistical measures of the pathologists’ performance on glass slides.
Senior Pathologist 1A | Senior Pathologist 1B | Senior Pathologist 2A | Senior Pathologist 2B | Pathologist A | Pathologist B | Resident A | Resident B | Our model | |
---|---|---|---|---|---|---|---|---|---|
Sensitivity | 0.9130 | 0.8695 | 0.9130 | 0.9565 | 0.5217 | 0.9565 | 0.4782 | 0.3478 | 0.9565 |
Specificity | 1 | 1 | 0.9729 | 1 | 0.9729 | 0.9189 | 1 | 0.9459 | 1 |
Precision | 1 | 1 | 0.9545 | 1 | 0.9230 | 0.88 | 1 | 0.8 | 1 |
Negative predictive value | 0.9487 | 0.925 | 0.9473 | 0.9736 | 0.7659 | 0.9714 | 0.7551 | 0.7 | 0.9737 |
False negative rate | 0.0869 | 0.1304 | 0.0869 | 0.0434 | 0.4782 | 0.0435 | 0.5217 | 0.6521 | 0.0435 |
False positive rate | 0 | 0 | 0.0270 | 0 | 0.0270 | 0.0811 | 0 | 0.0540 | 0 |
Accuracy | 0.9667 | 0.95 | 0.95 | 0.9833 | 0.8 | 0.9333 | 0.8 | 0.7166 | 0.9833 |
F1 | 0.9545 | 0.9302 | 0.9333 | 0.9778 | 0.6667 | 0.9167 | 0.6471 | 0.4848 | 0.9778 |
Experience (decades) | 1.2 | 1.4 | 0.7 | 0.9 | 0.4 | 0.4 | 0 | 0 | - |