Background: Validation of a novel gene expression signature in independent data sets is a critical step in the development of a clinically useful test for cancer patient risk-stratification. However, validation is often unconvincing because the size of the test set is typically small. To overcome this problem we used publicly available breast cancer gene expression data sets and a novel approach

This phase Ib study describes preliminary evidence of clinical activity and a potentially acceptable safety profile of pembrolizumab given every 2 weeks to patients with heavily pretreated, advanced TNBC. A single-agent phase II study examining a 200-mg dose given once every 3 weeks (ClinicalTrials.gov identifier: NCT02447003) is ongoing.

Microarray profiling of invasive breast carcinomas has identified five distinct subtypes of tumors (luminal A, luminal B, normal breast-like, HER2 overexpressing, and basal-like) that are associated with different clinical outcomes. The basal-like subtype is associated with poor clinical outcomes and is the subtype observed in BRCA1-related breast cancers. The aim of this study was to characterize the histologic and immunophenotypic properties of breast basal-like carcinomas that were first positively identified using DNA microarray analysis. Detailed histologic review was performed on 56 tumors with known microarray profiles (23 basal-like, 23 luminal, and 12 HER2 þ ). Immunohistochemistry for estrogen receptor (ER), HER2, EGFR, smooth muscle actin (SMA), p63, CD10, cytokeratin 5/6, cytokeratin 8/18, and vimentin was performed on 18 basal-like, 16 luminal, and 12 HER2 þ tumors. The basal-like tumors were grade 3 ductal/NOS (21/23) or metaplastic (2/23) carcinomas that frequently showed geographic necrosis (17/23), a pushing border of invasion (14/23), and a stromal lymphocytic response (13/23). Most basal-like tumors showed immunoreactivity for vimentin (17/18), luminal cytokeratin 8/18 (15/18), EGFR (13/18), and cytokeratin 5/6 (11/18), while positivity for the myoepithelial markers SMA (4/18), p63 (4/18) and CD10 (2/18) was infrequent. All basal-like tumors tested were ERÀ and HER2À. Morphologic features significantly associated with the basal-like subtype included markedly elevated mitotic count (Po0.0001), geographic tumor necrosis (P ¼ 0.0003), pushing margin of invasion (P ¼ 0.0001), and stromal lymphocytic response (P ¼ 0.01). The most consistent immunophenotype seen in the basal-like tumors was negativity for ER and HER2, and positivity for vimentin, EGFR, cytokeratin 8/18, and cytokeratin 5/6. The infrequent expression of myoepithelial markers in basal-like carcinomas does not support a direct myoepithelial cell derivation of these tumors. These findings should further assist in the identification of basal-like carcinomas in clinical specimens, facilitating treatment and epidemiologic studies of this tumor subtype.

ClinicalTrials.gov identifier: NCT01375842.

Background: Treatment options for previously treated metastatic triple-negative breast cancer (mTNBC) are limited. In cohort A of the phase II KEYNOTE-086 study, we evaluated pembrolizumab as second or later line of treatment for patients with mTNBC.Patients and methods: Eligible patients had centrally confirmed mTNBC, !1 systemic therapy for metastatic disease, prior treatment with anthracycline and taxane in any disease setting, and progression on or after the most recent therapy. Patients received pembrolizumab 200 mg intravenously every 3 weeks for up to 2 years. Primary end points were objective response rate in the total and PD-L1-positive populations, and safety. Secondary end points included duration of response, disease control rate (percentage of patients with complete or partial response or stable disease for !24 weeks), progression-free survival, and overall survival.Results: All enrolled patients (N ¼ 170) were women, 61.8% had PD-L1-positive tumors, and 43.5% had received !3 previous lines of therapy for metastatic disease. ORR (95% CI) was 5.3% (2.7-9.9) in the total and 5.7% (2.4-12.2) in the PD-L1-positive populations. Disease control rate (95% CI) was 7.6% (4.4-12.7) and 9.5% (5.1-16.8), respectively. Median duration of response was not reached in the total (range, 1.2þ-21.5þ) and in the PD-L1-positive (range, 6.3-21.5þ) populations. Median PFS was 2.0 months (95% CI, 1.9-2.0), and the 6-month rate was 14.9%. Median OS was 9.0 months (95% CI, 7.6-11.2), and the 6-month rate was 69.1%. Treatment-related adverse events occurred in 103 (60.6%) patients, including 22 (12.9%) with grade 3 or 4 AEs. There were no deaths due to AEs.Conclusions: Pembrolizumab monotherapy demonstrated durable antitumor activity in a subset of patients with previously treated mTNBC and had a manageable safety profile.