Protein expression profile and prevalence pattern of the molecular classes of breast cancer--a Saudi population based study.

backgroundBreast cancer is not a single entity but a diverse group of entities. Advances in gene expression profiling and immunohistochemistry as its surrogate marker have led to the unmasking of new breast cancer molecular subtypes, resulting in the emergence of more elaborate classification systems that are therapeutically and prognostically more predictive. Molecular class distribution across various ethnic groups may also reveal variations that can lead to different clinical outcomes in different populations.

methodsWe aimed to analyze the spectrum of molecular subtypes present in the Saudi population. ER, PR, HER2, EGFR and CK5/6 were used as surrogate markers for gene expression profiling to classify 231 breast cancer specimens. Correlation of each molecular class with Ki-67 proliferation index, p53 mutation status, histologic type and grade of the tumor was also carried out.

resultsOut of 231 cases 9 (3.9%) were classified as luminal A (strong ER +ve, PR +ve or -ve), 37 (16%) as luminal B (weak to moderate ER +ve, and/or PR +ve), 40 (17.3%) as HER2+ (strong or moderately positive HER 2 with confirmation by silver enhanced in-situ hybridization) and 23 (10%) as basal (CK5/6 or EGFR +ve). Co-positivity of different markers in varied patterns was seen in 23 (10%) of cases which were grouped into a hybrid category comprising luminal B-HER2, HER2-basal and luminal-basal hybrids. Ninety nine (42.8%) of the tumors were negative for all five immunohistochemical markers and were labelled as unclassified (penta negative). A high Ki-67 proliferation index was seen in basal (p=0.007) followed by HER2+ class. Overexpression of p53 was predominantly seen in HER2+(p=0.001) followed by the basal group of tumors. A strong correlation was noted between invasive lobular carcinoma and hormone receptor expression with 8 out of 9 lobular carcinoma cases (88.9%) classifiable as luminal cancers. Otherwise, there was no association between the molecular class and the histologic type or grade of the tumor.

conclusionsSubtyping by use of this immunohistochemical panel revealed a prevalence pattern that is unique to our population; luminal tumors comprised only 19.9%, and the unclassified group (penta negative) 42.8%, a distribution which is distinctive to our population and in contrast with all Western studies. The presence of a predominant unclassified group also suggests that the currently used molecular analytic spectrum may not completely encompass all molecular classes and there is a need to further refine and develop the existing classification systems.