Purpose: Immune checkpoint inhibition has been demonstrated to be an effective anticancer strategy. Several lines of evidence support the study of immunotherapy in triple-negative breast cancer (TNBC). We assessed the safety and antitumor activity of the programmed cell death protein 1 (PD-1) inhibitor pembrolizumab in patients with advanced TNBC. Methods: KEYNOTE-012 ( ClinicalTrials.gov identifier: NCT01848834) was a multicenter, nonrandomized phase Ib trial of single-agent pembrolizumab given intravenously at 10 mg/kg every 2 weeks to patients with advanced PD-L1–positive (expression in stroma or ≥ 1% of tumor cells by immunohistochemistry) TNBC, gastric cancer, urothelial cancer, and head and neck cancer. This report focuses on the TNBC cohort. Results: Among 111 patients with TNBC whose tumor samples were screened for PD-L1 expression, 58.6% had PD-L1–positive tumors. Thirty-two women (median age, 50.5 years; range, 29 to 72 years) were enrolled and assessed for safety and antitumor activity. The median number of doses administered was five (range, 1 to 36 doses). Common toxicities were mild and similar to those observed in other tumor cohorts (eg, arthralgia, fatigue, myalgia, and nausea), and included five (15.6%) patients with grade ≥ 3 toxicity and one treatment-related death. Among the 27 patients who were evaluable for antitumor activity, the overall response rate was 18.5%, the median time to response was 17.9 weeks (range, 7.3 to 32.4 weeks), and the median duration of response was not yet reached (range, 15.0 to ≥ 47.3 weeks). Conclusion: 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.

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 to data fusion, in order to validate a new breast tumor intrinsic list. Results: A 105-tumor training set containing 26 sample pairs was used to derive a new breast tumor intrinsic gene list. This intrinsic list contained 1300 genes and a proliferation signature that was not present in previous breast intrinsic gene sets. We tested this list as a survival predictor on a data set of 311 tumors compiled from three independent microarray studies that were fused into a single data set using Distance Weighted Discrimination. When the new intrinsic gene set was used to hierarchically cluster this combined test set, tumors were grouped into LumA, LumB, Basal-like, HER2+/ER-, and Normal Breast-like tumor subtypes that we demonstrated in previous datasets. These subtypes were associated with significant differences in Relapse-Free and Overall Survival. Multivariate Cox analysis of the combined test set showed that the intrinsic subtype classifications added significant prognostic information that was independent of standard clinical predictors. From the combined test set, we developed an objective and unchanging classifier based upon five intrinsic subtype mean expression profiles (i.e. centroids), which is designed for single sample predictions (SSP). The SSP approach was applied to two additional independent data sets and consistently predicted survival in both systemically treated and untreated patient groups. Conclusion: This study validates the "breast tumor intrinsic" subtype classification as an objective means of tumor classification that should be translated into a clinical assay for further retrospective and prospective validation. In addition, our method of combining existing data sets can be used to robustly validate the potential clinical value of any new gene expression profile.

The immune system is regulated by a delicate balance of factors that initiate antitumor immune responses and inhibit excessive inflammation and autoimmunity. Cells from both the innate and adaptive immune systems work to eradicate pathogens and other threats, including cancer. A number of investigations^1-3 over the past decade suggest that a proportion of breast cancers are immunogenically active and that some breast tumors have a substantial lymphocytic infiltrate. Lymphocyte-predominant breast cancers are characterized by tumor-infiltrating lymphocytes comprising 50% or more of the tumor bed. Primary breast tumors with a robust immune infiltrate are associated with a better response to neoadjuvant chemotherapy.

The Cancer Genome Atlas (TCGA) is one of the largest biorepositories of digital histology. Deep learning (DL) models have been trained on TCGA to predict numerous features directly from histology, including survival, gene expression patterns, and driver mutations. However, we demonstrate that these features vary substantially across tissue submitting sites in TCGA for over 3,000 patients with six cancer subtypes. Additionally, we show that histologic image differences between submitting sites can easily be identified with DL. Site detection remains possible despite commonly used color normalization and augmentation methods, and we quantify the image characteristics constituting this site-specific digital histology signature. We demonstrate that these site-specific signatures lead to biased accuracy for prediction of features including survival, genomic mutations, and tumor stage. Furthermore, ethnicity can also be inferred from site-specific signatures, which must be accounted for to ensure equitable application of DL. These site-specific signatures can lead to overoptimistic estimates of model performance, and we propose a quadratic programming method that abrogates this bias by ensuring models are not trained and validated on samples from the same site.

The AURORA US Metastasis Project was established with the goal to identify molecular features associated with metastasis. We assayed 55 females with metastatic breast cancer (51 primary cancers and 102 metastases) by RNA sequencing, tumor/germline DNA exome and low-pass whole-genome sequencing and global DNA methylation microarrays. Expression subtype changes were observed in ~30% of samples and were coincident with DNA clonality shifts, especially involving HER2. Downregulation of estrogen receptor (ER)-mediated cell–cell adhesion genes through DNA methylation mechanisms was observed in metastases. Microenvironment differences varied according to tumor subtype; the ER⁺/luminal subtype had lower fibroblast and endothelial content, while triple-negative breast cancer/basal metastases showed a decrease in B and T cells. In 17% of metastases, DNA hypermethylation and/or focal deletions were identified nearHLA-Aand were associated with reduced expression and lower immune cell infiltrates, especially in brain and liver metastases. These findings could have implications for treating individuals with metastatic breast cancer with immune- and HER2-targeting therapies.

PURPOSE: The ATEMPT trial was designed to determine if treatment with trastuzumab emtansine (T-DM1) caused less toxicity than paclitaxel plus trastuzumab (TH) and yielded clinically acceptable invasive disease-free survival (iDFS) among patients with stage I human epidermal growth factor receptor 2–positive (HER2+) breast cancer (BC). METHODS: Patients with stage I centrally confirmed HER2+ BC were randomly assigned 3:1 to T-DM1 or TH and received T-DM1 3.6 mg/kg IV every 3 weeks for 17 cycles or T 80 mg/m² IV with H once every week × 12 weeks (4 mg/kg load →2 mg/kg), followed by H × 39 weeks (6 mg/kg once every 3 weeks). The co-primary objectives were to compare the incidence of clinically relevant toxicities (CRTs) in patients treated with T-DM1 versus TH and to evaluate iDFS in patients receiving T-DM1. RESULTS: The analysis population includes all 497 patients who initiated protocol therapy (383 T-DM1 and 114 TH). CRTs were experienced by 46% of patients on T-DM1 and 47% of patients on TH ( P = .83). The 3-year iDFS for T-DM1 was 97.8% (95% CI, 96.3 to 99.3), which rejected the null hypothesis ( P < .0001). Serially collected patient-reported outcomes indicated that patients treated with T-DM1 had less neuropathy and alopecia and better work productivity compared with patients on TH. CONCLUSION: Among patients with stage I HER2+ BC, one year of adjuvant T-DM1 was associated with excellent 3-year iDFS, but was not associated with fewer CRT compared with TH.