Table 4.
Inter‐rater agreement and agreement between each rater and Ariol automated quantitative HER2 scores for cores in the virtual TMA
| HER2 semi‐quantitative score (0/1, 2,3) | N | Observed agreement | Kappa (95% CI) |
|---|---|---|---|
| Comparisons | |||
| Rater 1 vs rater2 | 660 | 92.7 | 0.71 (0.65, 0.78) |
| Ariol vs rater 1 | 693 | 90.7 | 0.62 (0.56, 0.68) |
| Ariol vs rater 2 | 716 | 93.7 | 0.71 (0.65, 0.77) |
| HER2 dichotomous score | N | % Pos. | Observed agreement | Kappa (95% CI) | Se (%) | Sp (%) | PPV (%) | NPV (%) |
|---|---|---|---|---|---|---|---|---|
| HER2 2+ (0/1 vs 2/3) | ||||||||
| Rater 1 vs rater2 | 660 | 20.9 | 91.4 | 0.73 (0.65, 0.81) | 74.6 | 95.8 | 82.4 | 93.5 |
| Ariol vs rater 1 | 693 | 21.2 | 90.0 | 0.69 (0.62, 0.77) | 72.1 | 94.9 | 79.1 | 92.7 |
| Ariol vs rater 2 | 716 | 19.6 | 91.6 | 0.73 (0.66, 0.81) | 77.9 | 95.0 | 79.0 | 94.6 |
| HER2 3+ (0/2 vs 3) | ||||||||
| Rater 1 vs rater2 | 660 | 12.6 | 94.1 | 0.68 (0.61, 0.76) | 59.0 | 99.1 | 90.7 | 94.4 |
| Ariol vs rater 1 | 693 | 12.4 | 91.3 | 0.46 (0.40, 0.53) | 34.9 | 99.3 | 88.2 | 91.5 |
| Ariol vs rater 2 | 716 | 8.4 | 95.8 | 0.67 (0.60, 0.74) | 55.0 | 99.5 | 91.7 | 96.0 |
% Pos.=% positive cores for reference rater
Se, sensitivity; Sp, Specificity; PPV, positive predictive value; NPV, negative predictive value.