Breast Cancer Grading Identification of the Genomic Grade

Pathologists have traditionally used histological grade to describe distinct breast cancer phenotypes: whereas, grade I or well-differentiated tumors are composed of polarized groups of cells that form tubular or duct-like structures; grade III or undifferentiated tumors are associated with a high mitotic activity, nuclear pleo-morphism, and no tubular formation; grade II tumors display intermediate characteristics. In addition, high- and low-grade tumors have been correlated with the expression of different biological markers, and several groups have also reported that mitotic/apoptotic activity is higher in high grade/poorly differentiated tumors (21). Finally, the recent results regarding the molecular classification of breast cancer have strongly pointed to an association of histological grade with particular gene expression tumor profiles (4,5).

There is growing evidence that high- and low-grade tumors should be viewed as distinct disease entities. Indeed, Roylance et al., found that the long arm of chromosome 16 is lost in 65% of grade I tumors compared to only 16% in grade III tumors, implying that the majority of well differentiated tumors do not evolve towards an undifferentiated state during tumor progression, as regain of genetic material is very unlikely (22). By investigating different markers in in situ and invasive breast cancer lesions, Warnberg et al. (23) suggested that evolution from in situ to invasive cancer likely occurs independently from tumor grade.

High- and low-grade tumors are also associated with a different clinical outcome, undifferentiated tumors being associated with the highest rate of recurrence and shorter recurrence time when compared to well-differentiated tumors (24). However, clinicians are confronted with a real problem with respect to patients carrying intermediate-grade tumors (grade II). These tumors, which represent 30% to 60% of breast cancer cases, are the major source of inter-observer discrepancy and display intermediate phenotype and survival: consequently treatment decisionmaking for these patients is a great challenge, with subsequent under- or over-treatment likely.

Our group decided to examine whether histological grading was associated with characteristic gene expression profiles and whether these profiles could be of use to refine histological grading in a reproducible manner (25). Using a training set of 33 histological grade I and 31 grade III ER-positive breast tumors, we identified 97 unique genes that were differentially expressed between low- and highgrade tumors. By restricting our training set to only ER-positive samples, we avoided selecting ER-related genes that would have been spuriously associated with grade. The majority of the 97 genes we identified were over-expressed in high-grade tumors and associated with cell cycle progression and proliferation. This is not surprising since histological grading is based on mitotic index, nuclear pleomorphism, and differentiation.

In order to summarize the expression of these 97 genes, we developed a score, the gene expression grade index, "GGI," in which a high index corresponds to high grade and vice-versa. In the validation sets, when comparing the gene expression profiles of intermediate-grade tumors with both high- and low-grade tumors, no distinct expression profile was observed, and the GGI of these tumors spanned the values for histological grade I and III tumors. Interestingly, when examining the prognostic value of the GGI in histological grade II tumors, we found a statistically significant difference in relapse-free survival between genomic grade I and grade III tumors, similar to the difference observed between histological grade I and III tumors. Importantly, we observed consistent results across multiple independent and heterogeneous validation sets and microarray platforms, emphasizing the reproducible behavior of the grade-associated genes.

Given the existence of other gene sets with prognostic information, we wondered how those could be compared to genomic grade. When we considered a less stringent threshold to identify genes associated with histological grade I and III tumors, we identified a larger list of 183 genes. Eleven and seven of these genes were present in the 70- and 76-gene prognostic signatures (10,12), respectively. By evaluating a predictor based on the genes common to our 97-gene genomic grade and the Amsterdam 70-gene signature (10) in their original data set, we observed a similar prognostic performance (data not shown). Also, when comparing the 70- and 76-gene signatures (10,12) with the genomic grade in the TRANSBIG validation series reported earlier (14,15), we observed consistent predictions of outcomes, both in terms of time to distant metastasis and overall survival (data not shown), indicating that proliferation-related genes appear to be an important—if not the most important—part of most prognostic gene sets.

10 Ways To Fight Off Cancer

10 Ways To Fight Off Cancer

Learning About 10 Ways Fight Off Cancer Can Have Amazing Benefits For Your Life The Best Tips On How To Keep This Killer At Bay Discovering that you or a loved one has cancer can be utterly terrifying. All the same, once you comprehend the causes of cancer and learn how to reverse those causes, you or your loved one may have more than a fighting chance of beating out cancer.

Get My Free Ebook


Post a comment