Evolving Strategies for Management of Early-Stage and Metastatic Triple-Negative Breast Cancer
William J. Gradishar, MD, FASCO, FACP
Betsy Bramsen Professor of Breast Oncology & Professor of Medicine
Director, Maggie Daley Center for Women's Cancer Care
Robert H. Lurie Comprehensive Cancer Center
Northwestern University Feinberg School of Medicine
Introduction to TNBC
Triple-negative breast cancer (TNBC) is characterized by lack of estrogen receptor (ER) and progesterone receptor (PR) expression and lack of human epidermal growth factor receptor-2 (HER2) gene amplification or overexpression. The American Cancer Society estimates more than 232,000 new cases of breast cancer and 40,000 deaths from breast cancer in women will occur in 2014.1 TNBC accounts for 15% to 20% of all breast cancer diagnoses2-4 and is a highly aggressive form of breast cancer, leading to disproportionately higher rates of mortality and metastasis.5
Risk factors associated with the development of TNBC include race, age, genetic factors, waist/hip ratio, breastfeeding history, and parity. TNBC typically presents in younger women; diagnoses in women younger than 40 years of age are more common with this subtype.6 TNBC is more common in women who are African-American or of black ethnicity than in other ethnicities.2,6-9
Associations have been identified between typical hormonal changes and increased risk of developing TNBC. Higher parity and younger age at first full-term pregnancy increase the risk of TNBC; however, risk is reduced by breastfeeding, greater number of children breastfed, and increased duration of breastfeeding.10 Obesity may also increase the risk of developing TNBC in both pre- and post-menopausal women, particularly those with BRCA1 or BRCA2 mutations.10,11
The identification of clinicopathological factors associated with TNBC has been hindered by a lack of data on large populations.11 In particular, collection of HER2 status data has been a relatively recent addition to data collection protocols for most cancer registries, first appearing as a routine element in the National Comprehensive Cancer Network (NCCN) Breast Cancer Outcomes Database in 1999 (immunohistochemistry [IHC]) and 2001 (fluorescence in situ hybridization [FISH]).
TNBC has been associated with a higher rate of BRCA mutations compared with other breast cancer subtypes.4,12 Among patients less than 50 years of age who are diagnosed with TNBC and have no family history or known predisposition, 10% to 30% are BRCA1 mutation carriers.13,14 The basal-like breast cancer subtype has also been associated with germline BRCA1 mutations, with 75% of BRCA1-related breast cancers falling into the basal-like subtype by either IHC or microarray.15,16 Disruptions in DNA repair pathways and BRCA genomic instability appear to be important in both basal-like and BRCA-related breast cancers; thus, there is interest in determining the prevalence of BRCA mutations in TNBC and in predicting outcomes for BRCA-associated and non-associated TNBC. Although BRCA1 mutation prevalence has been well studied among women with known high-risk ancestry, such as those of Eastern European or Ashkenazi Jewish origins, BRCA mutation prevalence among women of other ethnicities and without any associated family history is less well understood, particularly as these mutations relate to TNBC.17
Studies evaluating BRCA mutation prevalence in TNBC, though different in design and immunohistochemical definitions, confirm an increased rate of BRCA mutation in TNBC. In a retrospective study of 491 patients with breast cancer referred for BRCA genetic testing, 11.4% had BRCA1 mutations, 6.1% had BRCA2 mutations, and 79.6% were BRCA-negative.12 Of the 93 women with TNBC, 34.4% had BRCA1 mutations, 7.5% had BRCA2 mutations, and 58.1% were BRCA-negative. A larger retrospective review of 469 patients with TNBC who underwent BRCA testing showed significant differences in mutation prevalence by race/ ethnicity and age.17 The BRCA1 and BRCA2 prevalences in this TNBC cohort were 23.5% and 7.1%, respectively.
A potential limitation of studies in selected populations, particularly those that focus only on BRCA1, is the likelihood of enrichment for TNBC, which may not accurately reflect a true mutation prevalence. Two recent studies addressed BRCA mutation prevalence in unselected cohorts. Gonzalez-Angulo and colleagues evaluated germline and somatic BRCA1/2 mutation incidence in 77 patients by sequencing tumor and normal DNA samples at the BRCA1/2 exons and flanking regions.4 The incidence of BRCA mutations in this cohort was 19.5%, with a 15.6% incidence of BRCA1 mutations, including one somatic mutation, and a 3.9% incidence of BRCA2 mutations. Patients with BRCA mutations were younger than those without mutations (P = .005), but no association between mutation status and tumor grade, stage, or histology was shown. Interestingly, BRCA mutation carriers had significantly better outcomes and lower risk of relapse than non-BRCA mutation carriers, with 5-year overall survival (OS) of 73.3% versus 52.8% (P = .225) and 5-year recurrence-free survival of 86.2% versus 51.7% (P = .031). Similarly, Hartman and colleagues evaluated an unselected cohort of TNBC patients without significant family history of breast cancer (N = 199).18 The authors reported a comparatively low BRCA1 mutation prevalence (6.5%) and a higher than expected prevalence of BRCA2 mutations (4%), in part due to low participation by women of Ashkenazi ethnicity.
About 20% of TNBC patients will have impairment of DNA repair mechanisms resulting from genetic and epigenetic changes that ultimately hinder the BRCA1 pathway, even in the absence of somatic BRCA1 mutations (BRCAness).10 Genomic instability in BRCA1/2 carriers may enhance the activity of DNA damaging drugs, such as platinum compounds, or drugs that prevent DNA repair, such as poly-ADP ribose polymerase (PARP) inhibitors. A comprehensive analysis of breast tumors by The Cancer Genome Atlas Network, integrating data from 6 different high-throughput genomic, proteomic, and epigenomic technology platforms, showed that 20% of basal-like tumors harbor germline and/or somatic BRCA mutations, suggesting that platinum compounds or PARP inhibitors could potentially benefit 1 in 5 patients with basal-like breast cancers.19 This analysis also highlights the possibility of common driving molecular events between basal-like breast tumors and serous ovarian carcinomas, as BRCA1 inactivation, TP53 and RB1 loss, and cMYC amplification were common to both of these cancers. Studies such as this point to a therapeutic approach illuminated by molecular profile, rather than strictly by tissue of origin, which is supported by the activity of taxanes and platinum analogs in both serous ovarian and basal-like breast cancers.10
Relapse, metastatic disease, prognosis
Outcomes are favorable and comparable to other breast cancer subtypes for patients with TNBC who respond well to neoadjuvant chemotherapy, as determined by achievement of a pathologic complete response (pCR).20,21 Unlike in other subtypes, relapse rates for TNBC are especially high (10%-15%) in the first several years following surgery, with a peak at about 3 years.22-25 Late relapses occur less frequently, and most patients with TNBC who eventually present with metastatic disease will have relapsed soon after previous treatment.5,26
Recurrence patterns are also different in TNBC than in hormone receptor (HR)-positive tumors. TNBC relapses tend to occur at distant sites, with preferential metastasis to visceral sites including lung, brain, and liver20,23,27 compared with ER-positive breast cancers, which typically metastasize to skin or bone.28 The risk of locoregional relapse in TNBC is a matter of ongoing debate, but data from several studies suggest that patients with TNBC have an increased risk of locoregional relapse following mastectomy.29-32
Survival after metastatic relapse is shorter for TNBC than for other subtypes, with a median OS of approximately 13 months after diagnosis of metastatic TNBC.11,33 Thus, treatment goals for advanced TNBC center upon prolonging survival, relieving symptoms, and enhancing quality of life.34-36
Molecular subtypes of TNBC
TNBC is a highly heterogeneous disease; correspondingly, outcomes are not entirely explained by its clinical characteristics or pathological markers. Most TNBCs are invasive ductal carcinomas; in approximately 90% of cases, they are poorly differentiated, highly proliferative tumors of relatively large size.2,10
In recent decades, a focus on molecular analysis has provided insight into the biology of breast cancers beyond standard clinical and pathological features. These approaches should enhance understanding of breast cancer biology and provide insight into the development of targeted therapeutic approaches.
Through global gene expression analyses, 5 intrinsic subtypes of breast cancer and a normal breast-like group were established (Table 1).37-39 Importantly, the intrinsic subtypes (basal-like, luminal A, luminal B, HER2-enriched, and claudin-low) differ according to developmental origin, incidence, treatment response, and survival.40 TNBC is most commonly associated with basal-like and claudin-low intrinsic subtypes37; however, the clinical utility of these designations in TNBC remains to be determined.41
More recently, TNBC was further delineated into 7 molecular subtypes characterized according to differential gene expression on microarray.42 These include basal-like 1 and 2 (BL1 and BL2), mesenchymal, mesenchymal stem-like, immunomodulatory, luminal androgen receptor (LAR), and unstable (UNS). Most TNBCs are of the basal-like molecular subtypes, characterized by expression of keratins 5, 6, and 17 and overexpression of genes involved in proliferation.38
The clinical relevance of the 7-subtype molecular classification was recently evaluated in a study of 146 patients with TNBC who underwent neoadjuvant chemotherapy.43 Samples classified by TNBC molecular subtype using gene expression microarray were then correlated with pCR status and compared with intrinsic subtypes. A significant association between TNBC subtype and pCR status was observed (P = .043), and TNBC subtype was shown to be independently predictive of pCR status (P = .022). The highest pCR rates were observed among patients with basal-like 1 TNBCs (52%), with lowest pCR rates for luminal androgen receptor (10%) and basal-like 2 subtypes (0%) (Table 2). These findings suggest that TNBC molecular subtypes are clinically relevant and useful as an independent predictor of pCR status in patients who have undergone standard neoadjuvant chemotherapy.