High genomic grade index associated with poor prognosis for lymph node-negative and estrogen receptor-positive breast cancers and with good response to chemotherapy

Authors


Abstract

BACKGROUND:

The aim of the present study was to investigate the prognostic value of the genomic grade index for lymph node-negative and estrogen receptor (ER)-positive breast cancers of Japanese women treated with adjuvant hormonal therapy alone, as well as the relation between genomic grade index and pathological complete response (CR) to neoadjuvant chemotherapy.

METHODS:

Genomic grade index was determined by DNA microarray (U133plus2.0; Affymetrix, Santa Clara, Calif) in tumor tissues obtained from lymph node-negative and ER-positive breast cancers (n = 105) treated with adjuvant hormonal therapy alone or in breast tumor biopsy specimens (n = 84, Mammotome) obtained before neoadjuvant chemotherapy (paclitaxel followed by 5-fluorouracil/epirubicin/cyclophosphomide) to investigate the prognostic and predictive values of genomic grade index.

RESULTS:

Recurrence-free survival of patients with high genomic grade index tumors was significantly (P < .001) lower than that of patients with low genomic grade index tumors (55% vs 88%, 10 years after surgery). Multivariate analysis demonstrated that genomic grade index was the most important and significant predictive factor for disease recurrence (P = .013) independently of other prognostic factors, including tumor size, histological grade, progesterone receptor, human epidermal growth receptor 2, and Ki67. High genomic grade index tumors showed a significantly (P = .022) higher pathological CR rate for neoadjuvant chemotherapy than low genomic grade index tumors (31.9% [15 of 47] vs 10.8% [4 of 37]).

CONCLUSIONS:

Genomic grade index is a powerful prognostic factor for lymph node-negative and ER-positive tumors treated with adjuvant hormonal therapy alone, and high genomic grade index tumors are more likely to respond to chemotherapy. Genomic grade index also appears to be very useful for decision making regarding the need for adjuvant chemotherapy for lymph node-negative and ER-positive breast cancers. Cancer 2011. © 2010 American Cancer Society.

It is very important to determine highly accurate patient prognosis to implement personalized medicine for breast cancer patients. Lymph node-negative and estrogen receptor (ER)-positive breast cancer patients show a relatively favorable prognosis when treated with adjuvant hormonal therapy alone, but about 20% of them ultimately develop recurrence 10 years after surgery. To reduce the recurrence rate, the majority of such patients are therefore treated in daily practice with not only adjuvant hormonal therapy but also adjuvant chemotherapy, although adjuvant chemotherapy is considered to be unnecessary for most of them. Currently the most important clinical question in this regard, among breast cancer clinicians and researchers, is which patients with lymph node-negative and ER-positive tumors can be safely spared adjuvant chemotherapy. In clinical practice, patient prognosis is estimated based on the results of conventional histological examination of tumor size, histological grade, and biomarkers such as ER, progesterone receptor (PR), and human epidermal growth receptor 2 (HER2), of which histological grade has been repeatedly shown to be significantly prognostic.

There is no doubt that lymph node-negative and ER-positive patients with histological grade 3 tumors show a poor prognosis, and those with histological grade 1 tumors a good prognosis, so that histological grade 3 tumors are usually treated with adjuvant hormonal therapy plus chemotherapy and histological grade 1 tumors with adjuvant hormonal therapy alone. The clinically important problem concerning histological grade is that as many as 50% of tumors are classified as histological grade 2 tumors, which are associated with a prognosis intermediate between histological grade 1 and histological grade 3 tumors. Histological grade is thus unlikely to provide a clear answer as to the indication of adjuvant chemotherapy for about half of all lymph node-negative and ER-positive patients. Another problem with histological grade classification is that it is subjectively determined by pathologists and thus can differ depending on the pathologist.

To overcome such drawbacks of histological grade classification, the genomic grade index has been developed, which can classify breast tumors into high genomic grade index and low genomic grade index tumors by analyzing the expression of 97 genes. These genes were selected by using DNA microarray to compare the gene expression profile of histological grade 1 tumors with that of histological grade 3 tumors. With this index, the majority of histological grade 3 tumors are classified into high genomic grade index tumors and the majority of histological grade 1 tumors into low genomic grade index tumors, whereas about half of histological grade 2 tumors are classified into high genomic grade index tumors and the other half into low genomic grade index tumors. Several studies have shown that patients with high genomic grade index tumors show a significantly poorer prognosis than those with low genomic grade index tumors in the lymph node-negative and ER-positive subset treated with adjuvant hormonal therapy (tamoxifen) alone.1-3

These results seem to suggest that genomic grade index would be useful for decision making as to whether adjuvant chemotherapy should be added to adjuvant hormonal therapy for lymph node-negative and ER-positive breast cancers. In the study presented here, we therefore attempted to confirm the prognostic value of genomic grade index for Japanese women with lymph node-negative and ER-positive breast cancers. The prognostic value of genomic grade index has been examined for Caucasians only and not for Asians. The finding that the biological characteristics including patient prognosis are different for these 2 groups indicates the need to confirm the prognostic value of genomic grade index for Japanese women. In addition, we assessed the prognostic value of genomic grade index in comparison with that of HER2 and Ki67, because the previous studies of genomic grade index did not always include these biomarkers, although both are widely used in current practice. Furthermore, it has been reported very recently that high genomic grade index tumors are associated with a better response to neoadjuvant chemotherapy as compared with low genomic grade index tumors.4 In our study, we therefore also evaluated the relation between genomic grade index and pathological complete response (CR) to neoadjuvant chemotherapy.

MATERIALS AND METHODS

Patients and Tumors

During the period from 1996 to 2005, 250 patients with lymph node-negative and ER-positive invasive breast cancer underwent breast-conserving surgery followed by radiation therapy or mastectomy and were treated with adjuvant hormonal therapy alone. Of these patients, fresh tumor samples were obtained from 105 patients whose tumors were not too small for sampling and who consented to this study. Patient characteristics are shown in Table 1 (genomic grade index and prognosis study). Of these 105 patients, 68 were treated postoperatively with tamoxifen (20 mg/d) or toremifene (40 mg/d) for 5 years, 28 with goserelin (3.75 mg/4 weeks) plus tamoxifen (20 mg/d), and 9 with anastrozole (1 mg/d). Tamoxifen, toremifene, and anastrozole were administered for 5 years or until recurrence before 5 years, whereas goserelin was administered for 2 years or until recurrence before 2 years. Tumor samples obtained at surgery were snap frozen in liquid nitrogen and kept at −80°C until use. Informed consent for the study was obtained from each patient before surgery.

Table 1. Characteristics of Patients Enrolled in the GGI and Prognosis Study and GGI and Chemoresponse Study
CharacteristicGGI and Prognosis StudyGGI and Chemoresponse Study
TotalH-GGIL-GGIPaTotalH-GGIL-GGIP
  • GGI indicates genomic grade index; H-GGI, high-GGI; L-GGI, low-GGI; ER, estrogen receptor; PR, progesterone receptor; HER2, human epidermal growth receptor 2.

  • a

    Chi-square test (P values refer to all the comparisons in each of the categories).

  • b

    ER positive when positive tumor cells ≥10%.

  • c

    PR positive when positive tumor cells ≥10%.

  • d

    HER2 positive when +3 immunostaining or fluorescent in situ hybridization score ≥2.0.

  • e

    Ki-67 positive when positive tumor cells ≥20%.

Menopausal status   .233   .379
 Premenopausal482028422121
 Postmenopausal571641422616
Tumor size   .047   .976
 T1581642321
 T2451926573225
 T32201798
 T4000743
Lymph node status       .021
 Positive0  371522
 Negative105  473215
Histological grade   <.001   .025
 129425725
 2622042603129
 31412217143
ER       .136
 Positiveb1053768492325
 Negative000352412
PR   .024   .132
 Positivec872661301317
 Negative18117543420
HER2   .043   .195
 Positived1911823167
 Negative862660613130
Ki67   <.001   <.001
 Positivee19145433310
 Negative862363411427

For an additional study (genomic grade index and chemoresponse study), 84 patients with stage II-III primary breast cancer and characteristics as shown in Table 1 were enrolled. These patients were treated with neoadjuvant chemotherapy with P-FEC (paclitaxel 80 mg/m2 weekly for 12 cycles followed by 5-fluorouracil [500 mg/m2]/epirubicin [75 mg/m2]/cyclophosphamide [500 mg/m2] on Day 1 every 3 weeks for 4 cycles) during the period 2002 to 2008. Before neoadjuvant chemotherapy, all of these patients underwent tumor biopsy with a vacuum-assisted core biopsy instrument (8G; Mammotome HH Ethicon Endosurgery; Johnson and Johnson Company, Langhorne, Pa) under ultrasonographic guidance for histological examination and gene expression analysis. Tumor samples for histological examination were fixed in 10% buffered formaldehyde, and those for gene expression analysis were snap frozen in liquid nitrogen and kept at −80°C until use. Informed consent regarding the study was obtained from each patient before tumor biopsy.

RNA Extraction and Gene Expression Profiling

In the genomic grade index and prognosis study, RNA was extracted from 105 tumor samples obtained at surgery with the aid of the Qiagen RNeasy mini kit (QIAGEN Sciences, Germantown, Md). RNA (1 μg; RIN value >6) was used for generation of second-strand cDNA, and cRNA was amplified with the Oligo dT primer, then biotinylated and fragmented with One-Cycle Target Labeling and control reagents (Affymetrix, Santa Clara, Calif), followed by hybridization to U133 Plus 2.0 array overnight (17 hours) according to the manufacturer's protocol. Finally, the hybridized DNA microarray was fluorescence stained with GeneChipFluidics Station 450 (Affymetrics), and scanned with the GeneChip Scanner 3000 (Affymetrics).

In the genomic grade index and chemoresponse study, Trizol (Invitrogen, Carlsbad, Calif) was used to extract RNA from the 84 tumor biopsy samples obtained with Mammotome for the additional study of sensitivity to chemotherapy. Presence of tumor cells in these biopsy samples was estimated by histological confirmation of their presence in the adjacent tumor biopsy samples. RNA (50 ng) was used for generation of second-strand cDNA, and cRNA was amplified with a random primer (WT-Ovation FFPE RNA Amplification System V2; NuGEN, Cincinnati, Ohio). Next, the amplified cRNA was biotinylated and fragmented with the Flow Ovation cDNA Biotin Module V2 (NuGEN), and hybridized to a DNA microarray (Human Genome U133 Plus 2.0 Array; Affymetrics) overnight (17 hours) according to the manufacturer's protocol. The hybridized DNA microarray was then fluorescence stained with GeneChip Fluidics Station 450, and scanned with the GeneChip Scanner 3000. Informed consent was obtained from each patient before tumor biopsy with Mammotome.

Determination of Genomic Grade Index

Genomic grade index was calculated according to the method of Sotiriou et al1 using the gene expression data obtained by DNA microarray. In brief, all probe sets of the U133 Plus 2.0 array were normalized by the robust multiarray average method, and the 128 probe sets (representing 97 genes) listed in the article of Sotiriou et al1 were used to calculate the genomic grade index score for each sample by using the same formula described in their article. A threshold of zero was used for categorization of high genomic grade index and low genomic grade index. The same method was used for calculation of genomic grade index in the prognosis study and in the chemoresponse study.

Histological Evaluation of Response to Chemotherapy

After P-FEC, the patients underwent breast-conserving surgery or mastectomy. Pathological response to P-FEC was evaluated by using the specimens obtained at surgery, which were cut into 5 μm slices and stained with hematoxylin and eosin sections to determine the presence or absence of tumor cells. A complete loss of invasive tumor cells and absence of lymph node metastasis were considered pathological CR irrespective of the presence or absence of ductal carcinoma in situ components. Pathological examinations including the determination of response to chemotherapy, histological grade, and immunohistochemistry (ER, PR, Ki67, HER2) were done by well-trained pathologists, who were blinded to the results of genomic grade index and prognosis.

Immunohistological Examination

ER, PR, and Ki67 in tumor samples obtained with Mammotome before neoadjuvant chemotherapy were immunohistochemically examined with a previously described method.5 ER and PR were defined as positive when ≥10% of the tumor cells were positive for immunohistochemical staining (ER: Clone 6F11; Ventana Japan K.K., Tokyo, Japan; PR: Clone16; SRL Inc., Tokyo, Japan). Ki67 was defined as positive when ≥20% of tumor cells stained positive.6 The cutoff value of 20% for Ki67 was used, because this value was found to be optimal for the differentiation between high-risk and low-risk groups for disease recurrence as well as the differentiation between responders and nonresponders to neoadjuvant chemotherapy (P-FEC) in this study.

HER2 was determined by immunohistochemistry (Antihuman c-erbB-2 polyclonal antibody; Nichirei Biosciences, Tokyo, Japan) or FISH (PathVysion HER-2 DNA Probe Kit; Vysis/Abbott Molecular, Des Plaines, Ill). Fluorescent in situ hybridization (FISH) scoring was performed by counting fluorescence signals in at least 60 malignant cell nuclei per case, and for each specimen, the ratio of HER2 gene signals to chromosome 17 centromere signals (FISH ratio) was calculated. When a tumor showed +3 immunostaining or an FISH ratio ≥2.0, it was considered HER2 positive, and the histological grade was determined with the Scarff-Bloom-Richardson grading system.7

Statistical Analysis

Relations between genomic grade index and the various clinicopathological parameters were evaluated by chi-square test. The Cox proportional hazard model was used for univariate and multivariate analysis of the various prognostic parameters. All statistical analyses consisted of 2-sided tests. Significance was judged from a 5% standard. Except for 1 step in the preliminary treatment, for which Affymetrix Expression Console software was used, all analyses were performed with the aid of R software (http://www.R-project.org/) and the Bioconductor package (http://www.bioconductor.org/).

RESULTS

Genomic Grade Index and Prognosis

Lymph node-negative and ER-positive breast tumors (n = 105) treated with adjuvant hormonal therapy alone were analyzed in terms of their gene expression profile by DNA microarray and classified into high genomic grade index tumors and low genomic grade index tumors (Fig. 1). Of the 29 histological grade 1 tumors, 25 (86.2%) were classified as low genomic grade index tumors and 12 (85.7%) of the 14 histological grade 3 tumors as high genomic grade index tumors. Of the 62 histological grade 2 tumors, 42 (67.7%) were classified as low genomic grade index tumors and 20 (32.3%) as high genomic grade index tumors. Relations between genomic grade index and the various clinicopathological parameters are shown in Table 1, indicating that high genomic grade index tumors were more likely to be large, PR negative, HER2 positive, and Ki67 positive.

Figure 1.

A heat map of tumors in the genomic grade index (GGI) and prognosis study is shown. One hundred five tumors were subjected to gene expression analysis. (A) Columns of the heat map correspond to 97 genes used for GGI analysis, and rows of the heat map correspond to individual tumors, which were sorted first by histological grade (HG) 1, HG2, or HG3, and then by GGI within each histological grade category. In the heat map, high expression is red and low expression is green. (B) GGI score of each tumor is plotted on the right side of the corresponding row of heat map. (C) The prognosis of patients with each tumor is shown. Red represents patients with recurrence, and blue represents those without recurrence. L indicates low; H, high.

The recurrence-free survival (RFS) rate for patients with high genomic grade index tumors was significantly (P < .001) lower than for those with low genomic grade index tumors (55% vs 88%, 10 years after surgery) (Fig. 2A). After classification of the 62 histological grade 2 tumors into high genomic grade index and low genomic grade index tumors, the RFS rates for the patients with these tumors were compared (Fig. 2B). The rate for patients with high genomic grade index tumors was significantly (P = .012) lower than for those with low genomic grade index tumors (54% vs 83%, 10 years after surgery), indicating that genomic grade index can help to clearly differentiate high-risk from low-risk tumors even among histological grade 2 tumors.

Figure 2.

Recurrence-free survival curves are shown (A) for all patients (n = 105) and (B) for patients with histological grade 2 tumors (n = 62) according to genomic grade index (GGI). L indicates low; H, high; HG2; histological grade 2.

Comparison of Genomic Grade Index With Other Prognostic Factors for the Prediction of RFS Rates

Association of various conventional prognostic factors with RFS was analyzed by means of univariate analysis (Table 2), and it was found to be significant between high recurrence rates and large tumor size, high histological grade, or positive HER2. Moreover, high genomic grade index was significantly (P < .001) associated with a high recurrence rate. Multivariate analysis demonstrated that genomic grade index was the most important and significant predictive factor for disease recurrence (P = .013) independently from other, conventional prognostic factors.

Table 2. Univariate and Multivariate Analysis of Various Prognostic Factors
FactorUnivariateaMultivariatea
Hazard RatioLower 95% CIUpper 95% CIPHazard RatioLower 95% CIUpper 95% CIP
  • CI indicates confidence interval; Mens; menstruation; T, tumor size; HG, histological grade; PR, progesterone receptor; HER2, human epidermal growth receptor 2; GGI; genomic grade index.

  • a

    Cox proportional hazard model.

Mens1.790.804.04.1601.490.603.68.390
T2.491.225.05.0121.720.883.37.110
HG2.711.086.84.0351.490.494.17.520
PR0.510.201.30.1600.630.221.80.390
HER22.841.216.65.0161.800.724.46.210
Ki671.690.674.26.2700.650.221.89.430
GGI2.221.453.39<.0011.861.143.05.013

Genomic Grade Index and Response to Chemotherapy

The gene expression profile of the tumor biopsy samples (n = 84) obtained before neoadjuvant chemotherapy (P-FEC) were analyzed by means of DNA microarray for classification into high genomic grade index and low genomic grade index tumors (Fig. 3). Of the 7 histological grade 1 tumors, 5 (71.4%) were classified into low genomic grade index tumors, and of the 17 histological grade 3 tumors, 14 (82.4%) were classified into high genomic grade index tumors. Of the 60 histological grade 2 tumors, 29 (48.3%) were classified into low genomic grade index tumors and 31 (51.7%) into high genomic grade index tumors.

Figure 3.

A heat map of tumors in the genomic grade index and chemoresponse study is shown. Eighty-four tumors were subjected to gene expression analysis. (A) Columns in the heat map correspond to 97 genes used for genomic grade index (GGI) analysis, and rows in the heat map correspond to individual tumors, which were sorted first by histological grade (HG) 1, HG2, or HG3, and then by GGI within each HG category. In the heat map, high expression is red and low expression is green. (B) The GGI score of each tumor is plotted on the right side of the corresponding row of heat map. (C) The pathological response to neoadjuvant chemotherapy (paclitaxel followed by 5-fluorouracil/epirubicin/cyclophosphamide) of each tumor is shown. Red represents pathological complete response (pCR), and blue represents non-pCR. L indicates low; H, high.

The pathological CR rate to P-FEC was significantly (P = .022) higher for high genomic grade index (31.9%) tumors than low genomic grade index tumors (10.8%) (Table 3). The subset analysis according to ER status showed that the pathological CR rate was higher for high genomic grade index tumors than low genomic grade index tumors both in the ER-positive subset (17.4% vs 4.0%) and in the ER-negative subset (45.8% vs 25.0%), although the difference was not statistically significant.

Table 3. Relationship Between GGI and pCR Rates for Neoadjuvant Chemotherapy
 No. of TumorspCR RatePa
H-GGIL-GGI
  • GGI indicates genomic grade index; pCR, pathological complete response; H-GGI, high-GGI; L-GGI, low-GGI.

  • a

    Chi-square test.

Total8431.9%10.8%.022
ER positive4817.4%4.0%.129
ER negative3645.8%25.0%.227

Disease-free survival of these 84 patients according to genomic grade index is shown in Figure 4 (median follow-up period, 28 months; range, 1-56 months). There was no significant difference in RFS between high genomic grade index tumors and low genomic grade index tumors.

Figure 4.

Prognosis of breast cancer patients treated with neoadjuvant chemotherapy is shown. Recurrence-free survival curves of 84 patients treated with neoadjuvant chemotherapy (paclitaxel followed by 5-fluorouracil/epirubicin/cyclophosphamide) are shown according to genomic grade index (GGI). H indicates high; L, low.

DISCUSSION

In the study presented here, we evaluated the prognostic value of genomic grade index for lymph node-negative and ER-positive breast cancer patients treated with adjuvant hormonal therapy alone. We were able to demonstrate that in the case of Japanese breast cancers, most histological grade 3 tumors can be classified into high genomic grade index tumors and most histological grade 1 tumors into low genomic grade index tumors, as was previously reported for Caucasian breast cancers. High genomic grade index tumors were more likely to be ER negative, PR negative, HER2 positive, and Ki67 positive, and are thus considered to possess a biologically aggressive phenotype. Accordingly, prognosis for high genomic grade index tumors was significantly poorer than for low genomic grade index tumors, and more importantly, histological grade 2 tumors could be clearly differentiated into high genomic grade index tumors with poor prognosis and low genomic grade index tumors with good prognosis. All these findings for the association between genomic grade index and prognosis for Japanese breast cancers are consistent with those reported for Caucasian breast cancer patients, indicating that genomic grade index determined by the 97-gene signature for molecular classification of breast tumors is a statistically robust method regardless of ethnicity.

Because prognosis for lymph node-negative and ER-positive breast tumors is usually estimated by using tumor size, histological grade, PR, HER2, and more recently Ki67, it seems important to compare genomic grade index with these conventional prognostic parameters to determine the independent prognostic value of genomic grade index. However, such a comparison has never been done in previously reported studies. We therefore conducted a multivariate analysis of genomic grade index and conventional prognostic factors and were able to show that genomic grade index is a highly significant prognostic factor independent of the other, conventional prognostic factors. These results seem to indicate that genomic grade index would be very useful for decision making as to whether adjuvant chemotherapy should be added to adjuvant hormonal therapy for lymph node-negative and ER-positive breast cancers.

Prognostic signatures (MammaPrint and Oncotype DX) other than genomic grade index have also been reported, and all these signatures appear to have a similar prognostic performance, but are likely to be limited to ER-positive tumors.8 The common and most prominent characteristic among these prognostic signatures is the high expression of the genes responsive for cell proliferation. Ki67 is the most reliable marker for cell proliferation among those currently available, and we have been able to show that, as expected, high genomic grade index tumors are associated with Ki67-positive tumors. However, we believe that the prognostic value of genomic grade index is stronger than that of Ki67, because the multivariate analysis has clearly shown that genomic grade index is most strongly associated with prognosis independently of other parameters, including Ki67. These results seem to suggest that genomic grade index can identify cell proliferation more accurately than Ki67 or that genomic grade index can identify not only cell proliferation but also other biological features related to metastases.

Because high genomic grade index tumors are associated with high proliferation, and highly proliferating tumors can be expected to respond well to chemotherapy, we also investigated, as part of this study, the association of genomic grade index with pathological CR rates in the neoadjuvant setting, and were able to show that high genomic grade index tumors are significantly (P = .022) associated with a high pathological CR rate. This tendency was also observed in the ER-positive tumor subset. These results are essentially consistent with those reported very recently by Liedtke et al, who have shown that high genomic grade index breast cancer tumors are associated with a high pathological CR rate to neoadjuvant chemotherapy consisting of paclitaxel followed by 5-fluorouracil/doxorubicin/cyclophosphamide.4 All these results taken together indicate that, in the lymph node-negative and ER-positive subset, high genomic grade index tumors show a poor prognosis, but at the same time they are thought to have a relatively higher sensitivity to P-FEC. Therefore, it is speculated that prognosis of patients with such tumors can be improved by addition of adjuvant chemotherapy (P-FEC) to adjuvant hormonal therapy. This speculation needs to be proven in future study.

We found that there was no significant difference in RFS between high genomic grade index tumors and low genomic grade index tumors (Fig. 4). Because high genomic grade index tumors have a worse baseline prognosis as compared with low genomic grade index tumors, a modest increase in sensitivity to P-FEC associated with high genomic grade index tumors might be unable to translate into the improvement of their prognosis over low genomic grade index tumors. Interestingly, Liedtke et al reported that high genomic grade index tumors showed a worse prognosis than low genomic grade index tumors in the ER-positive subset.4 Therefore, it is considered that genomic grade index modestly predicts chemosensitivity in patients who receive neoadjuvant chemotherapy, but its impact on prognosis still remains to be clarified.

In the prognosis study of 105 patients, we used poly-T primers for RNA amplification, but in the chemoresponse study of 84 patients, we used the Ovation method (random primers) because of a small sample volume of Mammotome specimens. It is possible that determination of genomic grade index would be affected by the different chemical approaches to RNA amplification.9, 10 However, concordance between histological grade 1 and low genomic grade index and between histological grade 3 and high genomic grade index was similarly high both in the prognosis study (37 of 43, 86%) and in the chemoresponse study (19 of 24, 79%), suggesting that genomic grade index results obtained using the Ovation method would be comparable to those obtained using poly-T primers.

In conclusion, we were able to show that genomic grade index is a very strong prognostic factor for lymph node-negative and ER-positive tumors independently of the currently available prognostic biomarkers, including PR, HER2, and Ki67, and that high genomic grade index tumors are more likely to respond to chemotherapy (P-FEC). Thus, genomic grade index seems to represent a helpful diagnostic tool for decision making as to the addition of adjuvant chemotherapy to adjuvant hormonal therapy for this subset of breast tumors. Very recently, Toussaint11 et al reported that they succeeded in the conversion of DNA microarray-based determination of genomic grade index to a quantitative reverse-transcriptase polymerase chain reaction (PCR) assay (PCR-genomic grade index), which can be used for RNA derived from formalin-fixed paraffin-embedded tumor specimens. The methods they have developed seem to have the potential to be widely used for risk assessment of lymph node-negative and ER-positive breast cancers.

CONFLICT OF INTEREST DISCLOSURES

Grants were provided by the Knowledge Cluster Initiative and Scientific Research on Priority Areas programs of the Ministry of Education, Culture, Sports, Science, and Technology of Japan, and by the Comprehensive 10-Year Strategy for Cancer Control program of the Ministry of Health, Labor, and Welfare, Japan.

Ancillary