Table 3.
Conventional model | Revised model | |||
---|---|---|---|---|
Magee Score #3 coefficients | Linear regression coefficients | Linear regression coefficients | Standardized weight | |
ER | −0.022 | −0.025 | −26.75 | 2.28 |
PR | −0.029 | −0.040 | −10.83 | 3.59 |
Ki-67 | 0.186 | 0.196 | 11.64 | 3.15 |
Her2 | 1.47a | 7.97a | 9.14 | 2.06 |
BCL-2 | – | – | −7.39 | 2.17 |
a Her2 score was defined differently in our calculation of Her2 score, reflecting the change in Her2 recommendations from Hammond, et al., 2013
The relative contributions of each marker to the IHC-based scores are presented using metrics that examine the weighting of each quantity in the equations. The conventional model refers to the model based on the technique described in Klein, et al. using ER and PR H-score, Ki-67% positive cells, and Her2 score. The conventional model uses a variable weight based on the categorical interpretation of Her2; for clarity, only the coefficient that is applied to a Her2 score of 2 is shown. The equation’s coefficients derived from the Klein data set are shown in the first column and the coefficients that we derived from our cohort are shown in the second column. Notably, BCL-2 is excluded from this analysis because it was not a variable included in the authors’ formulation of the equation. The coefficients from the revised model, which includes BCL-2 as a prognostic variable, ER/PR percent positive cells as a replacement for H-score, and follows the logistic transformation applied to ER, PR, and Ki-67, and a linear treatment of the Her2 score, are shown in the third column. The standardized weights of the markers, defined as the product of the coefficients and the standard deviations of the values in our cohort, are shown in the fourth column