Table 2.
Intraobserver reproducibility of histological grading for invasive breast carcinoma by study pathologists who interpreted the same cases in Phase I and II, with data shown by phase and interpretive format (22 invasive cases)
| Intraobserver reproducibilitya | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Phase of study and format | Number of pathologists | Number of paired interpretations | κb (95% CIs) | Percentage agreementc,d (95% CIs) | ||||||||
| Phase I | Phase II | T | N | M | Nottingham grade | T | N | M | Nottingham grade | |||
| Same format | Glass slide format | Glass slide format | 49 | 254 | 0.73 (0.66-0.81) | 0.44 (0.34-0.54) | 0.52 (0.42-0.62) | 0.57 (0.48-0.66) | 84* (79-88) | 68 (62-73) | 79 (73-84) | 73 (68-78) |
| Digital format | Digital format | 41 | 214 | 0.48 (0.38-0.59) | 0.41 (0.30-0.53) | 0.37 (0.26-0.49) | 0.48 (0.37-0.58) | 72 (65-78) | 69 (62-75) | 72** (65-79) | 68 (61-75) | |
| Change in format | Glass slide format | Digital format | 45 | 242 | 0.50 (0.41-0.60) | 0.27 (0.16-0.38) | 0.37 (0.26-0.48) | 0.35 (0.25-0.46) | 72 (66-77) | 62* (55-69) | 74*,e (68-79) | 61 (55-67) |
| Digital format | Glass slide format | 37 | 193 | 0.50 (0.39-0.61) | 0.38 (0.26-0.50) | 0.44 (0.33-0.55) | 0.42 (0.30-0.53) | 73 (68-78) | 65 (58-72) | 77* (71-82) | 66 (59-73) | |
| Combinedf | 82 | 435 | 0.50 (0.43-0.57) | 0.32 (0.24-0.40) | 0.40 (0.32-0.48) | 0.38 (0.30-0.46) | 72 (68-76) | 63 (58-68) | 75 (71-79) | 63 (59-68) | ||
aThe exact same measurement was taken within each study phase with an intervening 9 months or more hiatus, bSimple κ coefficient, cGEE multivariable modeling, dContingency tables tested for homogeneity of marginal distribution (Bhapkar) and symmetry (Bowker), eFor the marginal homogeneity test only, fCombined represents both the glass to digital results in combination with the digital to glass results, *P<0.05, **P<0.01. CIs: Confidence intervals, T: Tubular score, N: Nuclear pleomorphism score, M: Mitotic score, GEE: General estimating equation