Is genomic grading killing histological grading ? 

Sotiriou C, Wirapati P, Loi S, Desmedt C, Haibe-Kains B, Piette F, Buyse M, Bontempi G, Delorenzi M, Piccart M

The histologic grade of breast carcinomas has long provided clinically important prognostic information. However, despite recommendations by the College of American Pathologists that tumor grade be used as a prognostic factor in breast cancer, the latest Breast Task Force of the American Joint Committee on Cancer did not include histologic tumor grade in its staging criteria, because of insurmountable inconsistencies in histologic grading between institutions. Concordance between two pathologists has been investigated and found to range from 50% to 85%. With the advent of new unified methods, such as the Elston and Ellis modification of the Bloom and Richardson method, the reproducibility of histologic grading has been improved. Although about half of all breast cancers are assigned histologic grade 1 or 3 status (with a low or high risk of recurrence, respectively), a substantial percentage of tumors (30%-60%) are classified as histologic grade 2, which is not informative for clinical decision making because of its intermediate risk of recurrence. This high percentage histologic grade 2 tumors is still observed when grading is performed by a single pathologist.

Recently, gene expression profiling has resulted in a paradigm shift in the way that researchers view breast cancer biology. In a previous work we have demonstrated, for example, that the ER status of the tumor was, indeed, the most important discriminator of expression subtypes, and that tumor grade came in second. Interestingly, other clinical features, namely positive lymph node status, menopausal status, and tumor size were not strongly reflected in the expression patterns.

Following our previous observation that tumor grade was an important discriminator of expression subtypes we sought whether grade could be refined by using gene expression profiling. In a recent work we have demonstrated that ``genomic'' tumor grade, which reflects differentiation and tumor progression on the basis of gene expression profiles (GEP), is effectively associated with distinct GEP and disease outcome in breast cancer far beyond the currently used clinico-pathological parameters. For that purpose we established a scoring system, referred to as the ``gene-expression grade index'' (GGI), and tested it on various independent validation datasets. We found that poorly differentiated compared with well-differentiated tumors are associated with distinct GEP and GGI and have statistically different clinical outcomes. Many of the markers are genes involved in cell cycle progression and proliferation, including CCNB2, CDC2, BUB1B, CDC25A and TPX2. We further demonstrated that intermediate grade tumors contain a mixture of well-differentiated and poorly differentiated expression patterns rather than a distinct or intermediate profile. This observation challenges the existence and clinical relevance of an intermediate grade classification. Interestingly, we also found that grade-related genes may encompass a significant portion of the predictive power of previously published prognostic signatures. Notably, we also found that genomic grade was also associated with the different molecular subtypes (previously identified by our group and others): basal-like, erbB2-like and luminal A, B and C subgroups. While the luminal A subgroup showed lower GGI levels, the basal-like, erbB2-like and luminal B and C subgroups had the worst clinical outcome in keeping with higher GGI levels. These results may suggest that the genomic grade, which essentially captures the degree of differentiation, may reflect the origin of the different cell lineages involved in breast cancer development.

We are currently validating our findings in the TRANSBIG series of 300 tumor samples from 5 different European institutions from which grading was determined based on a central pathology review. Additionally, we are in the process to convert genomic grade into a user-friendly RT-PCR tool which will assist clinicians and patients in optimizing treatment of early breast cancer.