Prevalence of basal-like phenotype in triple-negative invasive breast carcinomas and a review of the literature

INTRODUCTION

Breast carcinomas are a highly heterogeneous group of cancers with a wide range of differences in its morphology, molecular genetics, biology, and clinical outcomes.^[1,2] It is the most frequently diagnosed cancer among women worldwide with associated dismal outcomes.^[3] The classification of breast carcinoma histologically is of the essence in prognostication and predicting response to therapy with proven limitations.^[4,5] These limitations are, however overcome to some extent by the introduction of immunohistochemical (IHC) classification, which involves microscopic localization of specific antigens or proteins in tissues by staining with labeled antibodies.^[6,7] The expression of antigens, for example, hormone receptors (estrogen receptor [ER], progesterone receptor [PR]), and human epidermal receptor-2 (HER-2) assign individuals to groups that can be treated with specific drugs.^[6,7]

The differential staining of breast cancer for ER, PR, and HER-2 in immunohistochemistry allows pathologists to classify breast cancer into distinct subtypes which include; hormone-receptor-positive (ER+/PR+/HER-2−, ER+/PR+/HER-2+), triple-negative (ER−/PR−/HER-2−), and HER-2 positive breast cancers (ER−/PR−/HER-2+).^[1,7,8]

Triple-negative breast carcinoma (TNBC) on its own is a heterogeneous disease and represents 10–17% of all breast cancers.^[9,10] It is a highly aggressive carcinoma that lacks targeted therapy and is characterized by poor prognosis, high risk of distant recurrence, and death as well as a tendency to present as interval cancers due to their high proliferation rates.^[11]

With the recent advances in research, breast carcinomas are now classified using gene expression profiling into four to five molecular subtypes, including Luminal (A and B), HER-2, basal-like, and normal breast subtypes that have different biology and clinical outcomes.^[2,11]

Although TNBC is not one of the molecular subtypes initially identified by gene expression profiling, majority of the TNBCs fall into the basal-like subtype when subjected to molecular studies and can be further stratified into basal-like, immunomodulatory, mesenchymal, mesenchymal stem cell-like, and luminal subtype expressing androgen receptor, all of which differ in racial distribution, clinicopathologic features and therapeutic requirements, hence the need to identify such subtypes.^[12]

Basal-like breast carcinoma derived its name from its propensity to express genes of normal basal myoepithelial cells of the breast and represents 10–25% of all breast cancers (depending on the population) and makes up about 50–75% of the triple-negative subtype.^[13,14] The basal-like TNBC is the most aggressive subtype and appears to be more common in premenopausal African and African-American women.^[2,15,16] Histologically, such cancers are characterized by a high mitotic rate, pushing margin, solid architecture, lymphocytic infiltrates, increased nuclear grade, central fibrosis, and geographic necrosis.^[10,14] Basal-like breast carcinomas are less likely to metastasize to the axillary lymph nodes and bone but have a propensity to involve the brain and lungs through the hematogenous spread.^[14] In addition, several studies have documented immunohistochemistry as a surrogate to identify these subtypes of breast carcinomas.^[14,16] This study is aimed at analyzing the clinicopathological features and IHC expression of basal-like phenotype in TNBC at Uyo, as well as stratifying of invasive breast carcinomas into main molecular subtypes and correlating the basal-like breast carcinoma subtype with the size of tumor, grade, and age of patient in the University of Uyo Teaching Hospital (UUTH), Uyo, Akwa Ibom State, Nigeria.

MATERIAL AND METHODS

RESULTS

IHC characteristics

Out of the 79 cases, ER and PR were positive in 42 (53.2%) and 26 (32.9%) of cases, respectively. HER-2 was positive in 10 (12.7%) of cases. Four (5.1%) were equivocal for HER-2 and, therefore, were excluded in subsequent phenotypic classification since they were not analyzed further with FISH. CK5/6 and EGFR were also positive in 13 (16.5%) and 9 (11.4%) of cases, respectively. Ki67 was high (>14%) in 21 (26.6%) of cases [Table 2, Figures 1-4].

Table 2: The immunohistochemical characteristics of breast carcinoma.

Markers	Negative (%)	Positive (%)	Equivocal (%)
ER	37 (46.8)	42 (53.2)
PR	53 (67.1)	26 (32.9)
HER-2	65 (82.3)	10 (12.7)	4 (5.1)
EGFR	70 (88.6)	9 (11.4)
CK5/6	66 (83.5)	13 (16.5)
	<14	>14
Ki67	58 (73.4)	21 (26.6)

PR: Progesterone receptor, ER: Oestrogen receptor, HER2: Human epidermal receptor, EGFR: Epidermal growth factor receptor

Close

Figure 1:

Photomicrograph of invasive carcinoma no special type showing estrogen receptor nuclear-positive staining (Original magnification ×100).

Export to PPT

Close

Figure 2:

Photomicrograph of invasive carcinoma no special type, showing progesterone receptor-positive reaction (Original magnification ×40).

Export to PPT

Close

Figure 3:

Photomicrograph of invasive ductal carcinoma showing Her-2+++ (Original magnification ×40).

Export to PPT

Close

Figure 4:

Photomicrograph of invasive breast carcinoma no special type showing ki67 nuclear reactivity (Original magnification ×100).

Export to PPT

TNBCs and basal-like subtype

Twenty-seven (36.0%) of the cases were TNBCs. The basal markers (CK5/6 and EGFR) were positive in 13 out of the 27 TNBCs (48.1%) representing predominant subtypes of TNBC and represented 17.3% of total number of non-equivocal immuno reportable breast carcinomas. This entailed 4 TN/EGFR-positive breast carcinoma cases (14.8%) and 8 TN/CK5/6-positive cases (29.6%). One (3.7%) of the TNBCs was positive for both CK5/6 and EGFR tumor markers [Table 4, Figures 5 and 6].

Table 4: Proportion of triple-negative breast carcinomas.

HER2	ER	PR		Total
		Negative (%)	Positive (%)
Negative	Negative	27 (36.0)	3 (4.0)	30 (40.0)
	Positive	18 (24.0)	17 (22.7)	35 (46.6)
Positive	Negative	3 (4.0)	2 (2.7)	5 (6.7)
	Positive	2 (2.7)	3 (4.0)	5 (6.7)
	Total	50 (66.7)	25 (22.3)	75 (100)

PR: Progesterone receptor, ER: Estrogen receptor, HER2: Human epidermal receptor

Close

Figure 5:

Photomicrograph of invasive breast carcinoma no special type showing epidermal growth factor receptor membrane reactivity (Original magnification ×100).

Export to PPT

Close

Figure 6:

Photomicrograph of invasive breast carcinoma no special type showing CK5/6 cytoplasmic reactivity (Original magnification ×100).

Export to PPT

DISCUSSION

Breast carcinoma is the leading cancer in females with a highly heterogeneous behavior regarding its morphology, molecular genetics, biology, as well as clinical outcomes with or without treatment.^[1-3] Hence, the classification of breast carcinomas histologically and immunohistochemically is imperative to prognosticating and predicting the response to therapy, though with limitations that might be overcome by the use of gene expression profiling.^[4-16]

The index study showed that the frequency of breast cancers in Uyo, South-South Nigeria, was 48.8%; this compares relatively with a finding of 44.3% in a previous study in Uyo, as well as 45.8% in Makurdi, North Central Nigeria, 43.7% in Ebonyi, South East Nigeria and 47.9% in Sokoto, North West Nigeria, respectively.^[22-25] Outside Nigeria, very low and high rates of 11.8% and 54.8% were reported in Pakistan and India, respectively.^[26,27] From the foregoing, it is obvious that the prevalence of breast cancers differs with geographical location, probably due to the interplay of racial and cultural factors in the development of this disease.

On the basis of immunohistochemistry, this study showed that most of the breast carcinomas expressed ER positivity (53.2%) and PR positivity (32.9%). This is similar to the increased positivity expression of ER (65.1%), PR (54.7%) and ER (54.2%), PR (50.0%) reported by Adebamowo et al. and Nwafor and Keshinro in Ibadan and Lagos, respectively.^[28,29] Concordant results of increased hormone receptor expression were also reported by Salhia et al. in Egypt and Carey et al. in North Carolina, USA, where positivity rates of ER (65.0%), PR (43.8%) and ER (60.0%), PR (50.0%) were reported, respectively;^[15,30] although these values were higher than what was obtained in the index study. In contrast, studies conducted in Ghana and India showed low rates of ER+ tumors of 32.1% and 36.5%, respectively.^[3,31] On the other hand, the index study showed that HER-2 was positive in 12.7% of cases; this compares with the results of 2.4%, 10.6%, 11.0%, and 25.5% reported by Rao et al. in India, Ukah et al. in Nnewi, South East Nigeria, Roy and Othieno. in Uganda, and Seshie et al. in Ghana, respectively.^[31-34]

These variations in immunohistochemistry results from region to region could probably be explained by improper specimen collection and preservation, improper tissue processing, variable IHC techniques, interpretation, and scoring. Five molecular breast carcinoma subtypes were identified in this study. In decreasing order, the 75 breast carcinomas were distributed into Luminal A, Luminal B, basal-like, normal like/unclassified, and HER-2 enriched subtypes in 34 (45.3%), 10 (13.3%), 13 (17.3%), 14 (18.7%), and 4 (5.3%), respectively. This indicates that the Luminal A breast carcinoma phenotype accounted for the majority of the molecular phenotypes in this study. A concordant result of 43.9% was obtained by Ukah et al. in Nnewi, South East Nigeria.^[34] Similar trends were also reported by Omoruyi et al. in Calabar, South South Nigeria, Bennis et al. in Morocco, Ihemelandu et al. in the USA, and Ozmen. In Turkey, where Luminal A constituted 52.4%, 53.6%, 55.4%, and 62.0% of breast carcinomas,^[35-38] although the values are higher than the 45.3% in the index study. Preat et al., in their study on immigrant Arab women and European women, noted that the Luminal A subtype was more common in European women (46.0%).^[39] In contrast, lower prevalence rates of 14.5% and 3.9%, 20.0%, and 25.6% for Luminal A subtype have been reported in Brazil, Saudi Arabia, Ilorin North Central Nigeria, and Ghana, respectively.^[22,40-42] Luminal B breast carcinoma phenotype accounted for 13.3% of all cases in the index study. This differs from rates of 2.6%, 6.0%, 12.9%, 11.0%, 12.2%, and 11.8% recorded in Ibadan, Nnewi, Calabar, Ilorin, Ghana, and USA, respectively.^{[29,32,34,35,37,42]} but contrasts with higher prevalence rates of 44.6%, and 56.0% reported in Brazil and Morocco, respectively.^[39,40] The wide variations in the prevalence of the Luminal breast carcinoma phenotypes could be due to the use of different combinations of IHC markers in defining them. These include deployment of positivity or negativity of ER, PR, and HER-2 protein with or without positivity of Ki-67 ranging from 1% to 20%, as well as variable IHC pre-analytic and analytic study protocols influencing the immunohistochemistry result.^[24,31] In addition, the biology of the tumors, as well as the genetic differences among different ethnic and racial, groups, has been shown to influence the expression pattern of tumors.^[30,38] HER-2-enriched breast carcinoma phenotype accounted for 5.3% of the cases in the index study. This is, however less than the observed varying rates from 8.2% to 22.1% in some previous studies.^[28,32] The difference in prevalence of HER-2-enriched breast carcinoma subtype could be adduced to the strong influence of pre-analytical variables such as cold ischemic time, period of fixation, and type of fixative as well as differences in the biology of tumors and criteria used in the IHC interpretation of the tumor markers. The basal-like breast carcinoma phenotype constituted 17.3% in this present study. This is lower than the rates of 21.2%, 26.5%, 33.1%, and 40.7% reported in the USA as well as in Calabar, Lagos, Ife and Nnewi, and Nigeria, respectively.^{[28,34,35,37]} However, other studies have reported lower prevalence rates of 11.3%, 12.6%, and 10% in Egypt, Morocco, and Saudi Arabia, respectively.^[30,36,41] Carey et al. reported that basal-like breast carcinoma phenotypes are seen more in premenopausal African-American (39.1%) when compared to postmenopausal African-American (14.1%).^[15] The basal-like breast carcinoma phenotypes sometimes equate TNBCs, which were originally defined by gene expression profiling.^[2,34,37,43]

Immunohistochemistry is used as a surrogate, employing different basal markers in addition to ER, PR, and HER-2. In this study, TNBC accounted for 36.0% of the cases. This is higher than 29.2%, 15.1%, 17.1%, and 15.0% reported in Lagos, Southwest Nigeria, Thailand, Brazil, and India, respectively.^{[28,31,38,40]} In contrast, much higher rates were reported in Nnewi, Ghana, and India, where rates of 40.7%, 49.4%, and 50.0% were recorded, respectively.^[31,32,34] The differences in the prevalence of TNBC could be due to the effect of pre-analytical variables as well as to the scoring system applied in interpreting the immunohistochemistry result. Furthermore, the ethnic, racial, and genetic differences in individuals and the biology of tumors can account for the differences in prevalence.

Although TNBC may be regarded as Basal-like subtype which could be distinguished by a wide variety of basal markers like CK5, CK5/6, EGFR, and CK14 ^[19,44,45], there is still no universally accepted immunohistochemical (IHC) definition of Basal-like breast carcinoma. A five marker model as proposed by Nielsen et al. (ER, PR, HER-2, CK5/6, and EGFR) is seen as the best combination of markers to further characterize this subtype where basal-like cancers are defined as those lacking both ER, PR, and HER-2 expression and expressing CK5/6 and/or EGFR.^[19] This panel has a specificity of 100% and a sensitivity of 76% for the identification of basal-like tumors.^[19] The index study showed 13 out of the 27 TNBC to be either EGFR+ or CK5/6 or both, giving a basal-like breast carcinoma frequency of 48.1%. Gazinska et al. reported a frequency of 47.9%, which is similar to this index study.^[14] However, Rakha et al in UK and Reang et al. in India reported a higher frequency of 72.4% and 71%, respectively.^[46,47] In contrast to the above, Cheang et al. recorded a low prevalence of 9.0%.^[48] This wide variation may be due to the nonstandardization of the various combinations of tumor markers that are used in identifying the basal-like subtype as well as the difference in the biology of tumors.

This study showed that nine out of 13 basal-like breast carcinomas were grade 3, with a positive association between this tumor phenotype and high tumor grade. This is comparable to similar studies by Zakaria et al., Gazinska et al., and Cheang et al., which showed that basal-like breast cancers are associated with high-grade carcinomas.^[14,20,21] In contrast, a study conducted in Egypt did not find any association between the tumor grade and breast cancer molecular subtype.^[30]

CONCLUSION

Our study confirmed that the Luminal A subtype of breast carcinomas was the most commonest, and there was an appreciable number of the basal-like breast carcinoma subtype in Uyo with a frequency of 48.1% among the TNBCs.