Skip to main content
. 2019 Apr 3;10:11. doi: 10.4103/jpi.jpi_29_18

Table 3.

Interobserver concordance of histological grading for invasive breast carcinoma among different pathologists interpreting the same cases (interobserver concordance) by study phase and interpretive format (22 invasive cases)

Interobserver concordancea

Phase of study and format Number of pathologists Number of paired interpretations κb (95% CIs) Agreementc (95% CIs)


T N M Nottingham grade T N M Nottingham grade
Phase I
 Glass slide format 115 17,162 0.51 (0.50-0.52) 0.22 (0.21-0.24) 0.42 (0.40-0.43) 0.48 (0.47-0.49) 71 (69-73) 58 (56-59) 74 (72-77) 68 (66-70)
 Digital format 93 10,562 0.40 (0.39-0.42) 0.22 (0.20-0.23) 0.25 (0.23-0.27) 0.32 (0.31-0.34) 67 (65-70) 58 (56-61) 70 (67-73) 60 (57-62)
Phase II
 Glass slide format 86 9362 0.47 (0.45-0.48) 0.25 (0.23-0.27) 0.45 (0.44-0.47) 0.49 (0.48-0.51) 71 (68-73) 56 (54-59) 76 (73-78) 69 (67-71)
 Digital format 86 9472 0.37 (0.35-0.38) 0.17 (0.15-0.19) 0.23 (0.21-0.25) 0.36 (0.34-0.37) 65 (62-67) 56 (53-58) 68 (65-70) 62 (60-64)

aAll pairwise combinations excluding overlapping pathologists, bSimple κ coefficient, cGEE multivariable modeling. CIs: Confidence intervals, T: Tubular score, N: Nuclear pleomorphism grade, M: Mitotic score, GEE: General estimating equation