Molecular qRT-PCR grade index: a new tool for breast cancer (BC) patient grading improvement
Toussaint J, Sieuwerts A, Durbecq V, Haibe-Kains B, Berns E, Harris AL, Larsimont D, Piccart MJ, Foekens J, Sotiriou C
Background:
We have recently shown that proliferation captured by the GGI is one of
the most important prognostic indicators in BC and may encompass a
significant portion of the predictive power of several previously
published prognostic signatures in particular for ER+ disease. The aims
of this study were 1) to convert this microarray index to an index
using qRT-PCR and 2) to assess its prognostic and predictive value for
tamoxifen response.
Methods: A qRT-PCR genomic
grade index (PCR-GGI) was developed based on the expression of 4 genes
selected from the GGI microarray signature and 4 reference genes. The
accuracy and concordance with the original microarray-derived GGI was
assessed using a BC set from which frozen, FFPE tissues and microarray
data were available (N = 19). The evaluation of the prognostic value of
the PCR-GGI was assessed using a consecutive series of 212 systemically
treated early BC patients according to the standard of care at the time
of diagnosis using FFPE material. The predictive performance for
tamoxifen response was assessed using an ER+ BC population treated
either with adjuvant tamoxifen only (n = 141) or first-line tamoxifen
for advanced disease (n = 279).
Results: A statistically
significant correlation was observed between GGI derived from
microarray and qRT-PCR assay using frozen (ρ = 0.95, p < 10-6) as
well as FFPE material (ρ = 0.89, p < 10-6). Similarly to our
previous microarray results, PCR-GGI redistributed histological grade 2
(HG2) tumors into two subgroups with statistically distinct clinical
outcome similar to those of HG1 and HG3 tumors, respectively (HR =
2.27; 95CI: 0.94–5.48, p = 0.068). Of notice, PCR-GGI was more
informative than Ki67 (IHC) in discriminating HG2 patients into good
and bad prognosis group. Additionally, PCR-GGI identified two distinct
ER+ subgroups with statistically different DMFS and response to
tamoxifen treatment (PFS) in both early (HR = 2.26; 95CI: 1.075–4.751,
p = 0.03) and advanced setting (HR = 1.95; 95CI: 1.49–2.544, p <
10-6) respectively. Interestingly, among the 66% of ER+ node-negative
early BC patients assigned as low risk, only 7% exhibited distant
recurrence compared to 47% for the high-risk patients at 10 years
follow-up.
Conclusions: GGI using a
qRT-PCR assay based on a limited number of genes recapitulated in an
accurate and reproducible manner the performances of the GGI developed
by microarray using both frozen and FFPE tumor samples. PCR-GGI can be
used as a powerful tool for BC management.