Prognostic value of proliferation assay in the luminal, HER2-positive, and triple-negative biologic classes of breast cancer

Abstract

Introduction: Although the prognostic significance of proliferation in early invasive breast cancer has been recognized for a long time, recent gene-expression profiling studies have reemphasized its biologic and prognostic value and the potential application of its assessment in routine practice, particularly to define prognostic subgroups of luminal/hormone receptor-positive (HR+) tumors. This study aimed to assess the prognostic value of a proliferation assay by using Ki-67 immunohistochemistry as compared with mitotic count scores.

Method: Proliferation was assessed by using Ki-67 labeling index (Ki-67LI) and mitotic scores in a large (n = 1,550) and well-characterized series of clinically annotated primary operable invasive breast cancer with long-term follow-up. Tumors were phenotyped based on their IHC profiles into luminal/HR+, HER2+, and triple-negative (TN) classes. We used a split-sample development and validation approach to determine the optimal Ki-67LI cut-offs.

Results: The optimal cut-points of Ki-67LI were 10% and 50% for the luminal class. Both Ki7LI and MS were able to split luminal tumors into subgroups with significantly variable outcomes, independent of other variables. Neither mitotic count scores nor Ki-67LI was associated with outcome in the HER2+ or the TN classes.

Conclusions: Assessment of proliferation by using Ki-67LI and MS can distinguish subgroups of patients within luminal/hormone receptor-positive breast cancer significantly different in clinical outcomes. Overall, both Ki-67 LI and mitotic-count scores showed comparable results. The method described could provide a cost-effective method for prognostic subclassification of luminal/hormone receptor-positive breast cancer in routine clinical practice.

Figure 1

Box plot of Ki-67LI in different BC molecular classes, and their corresponding levels of significance in post hoc analysis.

Figure 2

Association between BCSS and Ki-67LI expressed in 10% increments (10% each.) However, tumors showing 50% to 69% and 80% to 100% Ki-67LI were considered as two groups, as the number in each 10% subgroup was small). Labels 1 through 8 represent patients' subsets based on tumor Ki-67LI, where 1 is 0 to 9%; 2 is 10% to 19%; 3 is 20% to 29%; 4 is 30% to 39%; 5 is 40% to 49%; 6 is 50% to 69%; 7 is 70% to 79%; and 8 is 80% to 100%.

Figure 3

Kaplan-Meier survival plot for luminal BC training set using Ki-67LI and MS. (a) Breast cancer-specific survival (BCSS). (b) Metastasis-free survival at 10% and 50% Ki-67LI cut-off point. (c, d) BCSS and DMFS for mitosis-frequency scores.

Figure 4

Kaplan-Meier survival plot for luminal BC validation set using Ki-67LI and MS. (a) Breast cancer-specific survival (BCSS) and (b) metastasis-free survival at 10% and 50% Ki-67LI cut-off point. (c, d) BCSS and DMFS for mitosis frequency scores (by using the validation set only).

Figure 5

Kaplan-Meier survival plot of luminal tumors showing association between HER2 status and proliferation. (Ki-67LI, a, and MS, b) and BCSS (LR = 88; P < 0.0001; HR, 1.683; 95% CI, 1.492 to 1.898, and LR, 90.239; P < 0.0001; HR, 1.749; 95% CI, 1. 543 to 1.982, respectively; validation set only). Number of patients at risk is shown above the curves.