Association of Immunophenotype With Pathologic Complete Response to Neoadjuvant Chemotherapy for Triple-Negative Breast Cancer: A Secondary Analysis of the BrighTNess Phase 3 Randomized Clinical Trial | Breast Cancer | JAMA Oncology | JAMA Network
[Skip to Navigation]
Sign In
Figure 1.  CONSORT Flow Diagram
CONSORT Flow Diagram
Figure 2.  Expression-Based Subtypes, Proliferation, and Immune Signatures and Response to Neoadjuvant Chemotherapy
Expression-Based Subtypes, Proliferation, and Immune Signatures and Response to Neoadjuvant Chemotherapy

Treatment for arms A and B was paclitaxel followed by doxorubicin and cyclophosphamide (T-AC) plus carboplatin with or without veliparib, and for arm C was T-AC alone. A, Percent pathologic complete response (pCR) of basal-like vs nonbasal-like PAM50 subtypes; P value for interaction test of basal-like vs nonbasal-like with arm (A and B vs C) is nonsignificant (P = .80), as indicated. B, Error bars denote 95% CIs based on normal approximation. Number of individual subtypes was too small to warrant statistical comparison. Triple-negative breast cancer (TNBC) subtypes are denoted as BL1, basal-like 1; BL2, basal-like 2; IM, immunomodulatory; LAR, luminal androgen receptor; M, mesenchymal; MSL, mesenchymal stem–like; and UNS, unselected TNBC type. C and D, Proliferation signature was calculated from a subset of PAM50-related genes as part of PAM50 classification.2 GeparSixto immune signature was calculated as the median expression of 7 genes (CXCL9, CD8A, CCL5, CXCL13, CD21, FOXP3, CD80) positively associated with tumor-infiltrating lymphocytes in that trial.10 Distribution of tumors by proliferation signature score and immune signature score, with pCR and residual disease (RD) indicated for patients who received T-AC plus carboplatin and/or veliparib (arms A and B) or T-AC alone (arm C). Dashed lines indicate median for each signature score, and proportions of pCR for each quartile above/below median are indicated in each corner.

Table.  Bivariable and Multivariable Logistic Models for Pathologic Complete Response to Neoadjuvant Chemotherapy, Overall and by Arm
Bivariable and Multivariable Logistic Models for Pathologic Complete Response to Neoadjuvant Chemotherapy, Overall and by Arm
1.
Perou  CM, Sørlie  T, Eisen  MB,  et al.  Molecular portraits of human breast tumours.   Nature. 2000;406(6797):747-752. doi:10.1038/35021093PubMedGoogle ScholarCrossref
2.
Parker  JS, Mullins  M, Cheang  MC,  et al.  Supervised risk predictor of breast cancer based on intrinsic subtypes.   J Clin Oncol. 2009;27(8):1160-1167. doi:10.1200/JCO.2008.18.1370PubMedGoogle ScholarCrossref
3.
Lehmann  BD, Bauer  JA, Chen  X,  et al.  Identification of human triple-negative breast cancer subtypes and preclinical models for selection of targeted therapies.   J Clin Invest. 2011;121(7):2750-2767. doi:10.1172/JCI45014PubMedGoogle ScholarCrossref
4.
Lehmann  BD, Jovanović  B, Chen  X,  et al.  Refinement of triple-negative breast cancer molecular subtypes: implications for neoadjuvant chemotherapy selection.   PLoS One. 2016;11(6):e0157368. doi:10.1371/journal.pone.0157368PubMedGoogle Scholar
5.
Loibl  S, O’Shaughnessy  J, Untch  M,  et al.  Addition of the PARP inhibitor veliparib plus carboplatin or carboplatin alone to standard neoadjuvant chemotherapy in triple-negative breast cancer (BrighTNess): a randomised, phase 3 trial.   Lancet Oncol. 2018;19(4):497-509. doi:10.1016/S1470-2045(18)30111-6PubMedGoogle ScholarCrossref
6.
Sikov  WM, Berry  DA, Perou  CM,  et al.  Impact of the addition of carboplatin and/or bevacizumab to neoadjuvant once-per-week paclitaxel followed by dose-dense doxorubicin and cyclophosphamide on pathologic complete response rates in stage II to III triple-negative breast cancer: CALGB 40603 (Alliance).   J Clin Oncol. 2015;33(1):13-21. doi:10.1200/JCO.2014.57.0572PubMedGoogle ScholarCrossref
7.
Rugo  HS, Olopade  OI, DeMichele  A,  et al; I-SPY 2 Investigators.  Adaptive randomization of veliparib-carboplatin treatment in breast cancer.   N Engl J Med. 2016;375(1):23-34. doi:10.1056/NEJMoa1513749PubMedGoogle ScholarCrossref
8.
Curtis  C, Shah  SP, Chin  SF,  et al; METABRIC Group.  The genomic and transcriptomic architecture of 2,000 breast tumours reveals novel subgroups.   Nature. 2012;486(7403):346-352. doi:10.1038/nature10983PubMedGoogle ScholarCrossref
9.
Denkert  C, Loibl  S, Noske  A,  et al.  Tumor-associated lymphocytes as an independent predictor of response to neoadjuvant chemotherapy in breast cancer.   J Clin Oncol. 2010;28(1):105-113. doi:10.1200/JCO.2009.23.7370PubMedGoogle ScholarCrossref
10.
Denkert  C, von Minckwitz  G, Brase  JC,  et al.  Tumor-infiltrating lymphocytes and response to neoadjuvant chemotherapy with or without carboplatin in human epidermal growth factor receptor 2-positive and triple-negative primary breast cancers.   J Clin Oncol. 2015;33(9):983-991. doi:10.1200/JCO.2014.58.1967PubMedGoogle ScholarCrossref
11.
Newman  AM, Liu  CL, Green  MR,  et al.  Robust enumeration of cell subsets from tissue expression profiles.   Nat Methods. 2015;12(5):453-457. doi:10.1038/nmeth.3337PubMedGoogle ScholarCrossref
12.
Li  B, Severson  E, Pignon  JC,  et al.  Comprehensive analyses of tumor immunity: implications for cancer immunotherapy.   Genome Biol. 2016;17(1):174. doi:10.1186/s13059-016-1028-7PubMedGoogle ScholarCrossref
13.
Cortazar  P, Zhang  L, Untch  M,  et al.  Pathological complete response and long-term clinical benefit in breast cancer: the CTNeoBC pooled analysis.   Lancet. 2014;384(9938):164-172. doi:10.1016/S0140-6736(13)62422-8PubMedGoogle ScholarCrossref
14.
Schmid  P, Cortes  J, Pusztai  L,  et al; KEYNOTE-522 Investigators.  Pembrolizumab for early triple-negative breast cancer.   N Engl J Med. 2020;382(9):810-821. doi:10.1056/NEJMoa1910549PubMedGoogle ScholarCrossref
Limit 200 characters
Limit 25 characters
Conflicts of Interest Disclosure

Identify all potential conflicts of interest that might be relevant to your comment.

Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.

Err on the side of full disclosure.

If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.

Not all submitted comments are published. Please see our commenting policy for details.

Limit 140 characters
Limit 3600 characters or approximately 600 words
    Brief Report
    February 18, 2021

    Association of Immunophenotype With Pathologic Complete Response to Neoadjuvant Chemotherapy for Triple-Negative Breast Cancer: A Secondary Analysis of the BrighTNess Phase 3 Randomized Clinical Trial

    Author Affiliations
    • 1Department of Medical Oncology, Dana-Farber Cancer Institute, Harvard Medical School, Boston, Massachusetts
    • 2Department of Medicine, The Ohio State University College of Medicine, Columbus
    • 3Division of Medical Oncology, The Ohio State University Comprehensive Cancer Center, Columbus
    • 4AbbVie, Inc, North Chicago, Illinois
    • 5Washington University, St Louis, Missouri
    • 6German Breast Group, Neu-Isenburg, Germany
    • 7Massey Cancer Center, Virginia Commonwealth University, Richmond
    • 8Now with Houston Methodist Cancer Center, Houston, Texas
    • 9Texas Oncology–Baylor Charles A. Sammons Cancer Center, Dallas
    • 10Helios Klinikum Berlin-Buch, Berlin, Germany
    • 11University of California, San Francisco
    • 12University Medical Center Ulm, Ulm, Germany
    • 13Now with Department of Interdisciplinary Medical Services, Breast Center, Cantonal Hospital St Gallen, St Gallen, Switzerland.
    • 14Division of Breast Surgery, Department of Surgery, Brigham and Women’s Hospital, Boston, Massachusetts
    • 15Breast Oncology Program, Dana-Farber/Brigham and Women’s Cancer Center, Boston, Massachusetts
    • 16Now with Yale Cancer Center, New Haven, Connecticut
    • 17Women and Infants Hospital of Rhode Island, Providence
    • 18University of Pittsburgh Medical Center, Hillman Cancer Center, Pittsburgh, Pennsylvania
    • 19Allegheny General Hospital, Pittsburgh, Pennsylvania
    • 20Institute of Physiology and Pathophysiology, Department of Medicine, Philipps-University Marburg and University Hospital of Giessen and Marburg, Marburg, Germany
    • 21The University of Texas MD Anderson Cancer Center, Houston
    JAMA Oncol. 2021;7(4):603-608. doi:10.1001/jamaoncol.2020.7310
    Key Points

    Question  Does triple-negative breast cancer (TNBC) subtyping inform on chances of achieving a pathologic complete response (pCR) and on the potential benefits of adding carboplatin to standard neoadjuvant chemotherapy?

    Findings  This prespecified secondary analysis of a randomized clinical trial of 634 patients with stages II to III TNBC found that in 482 women with evaluable RNA sequencing, basal-like and immunomodulatory subtypes were associated with higher pCR rates. After analyzing important biological processes related to these phenotypes, tumor cell proliferation and immune scores were identified as independent factors informing on chances of achieving pCR; TNBC subtyping was not informative for treatment decision.

    Meaning  With further validation and long-term survival data, RNA-based proliferation and immune scores may inform the development of novel therapies for patients with TNBC.

    Abstract

    Importance  Adding carboplatin to standard neoadjuvant chemotherapy (NAC) in triple-negative breast cancer (TNBC) likely benefits a subset of patients; however, determinants of benefit are poorly understood.

    Objective  To define the association of molecular subtype, tumor proliferation, and immunophenotype with benefit of carboplatin added to NAC for patients with stages II to III TNBC.

    Design, Setting, and Participants  This was a prespecified secondary analysis of a phase 3, double-blind, randomized clinical trial (BrighTNess) that enrolled 634 women across 145 centers in 15 countries. Women with clinical stages II to III TNBC who had undergone pretreatment biopsy were eligible to participate. Whole transcriptome RNA sequencing was performed on the biopsy specimens. The prespecified end point was association of pathologic complete response (pCR) with gene expression–based molecular subtype, with secondary end points investigating established signatures (proliferation, immune) and exploratory analyses of immunophenotype. Data were collected from April 2014 to March 2016. The study analyses were performed from January 2018 to March 2019.

    Interventions  Neoadjuvant chemotherapy with paclitaxel followed by doxorubicin and cyclophosphamide, or this same regimen with carboplatin or carboplatin plus veliparib.

    Main Outcomes and Measures  Association of gene expression–based molecular subtype (PAM50 and TNBC subtypes) with pCR.

    Results  Of the 634 women (median age, 51 [range, 22-78] years) enrolled in BrighTNess, 482 (76%) patients had evaluable RNA sequencing data, with similar baseline characteristics relative to the overall intention-to-treat population. Pathologic complete response was significantly more frequent in PAM50 basal-like vs nonbasal-like cancers overall (202 of 386 [52.3%] vs 34 of 96 [35.4%]; P = .003). Carboplatin benefit was not significantly different in basal-like vs nonbasal-like subgroups (P = .80 for interaction). In multivariable analysis, proliferation (hazard ratio, 0.36; 95% CI, 0.21-0.61; P < .001) and immune (hazard ratio, 0.62; 95% CI, 0.49-0.79; P < .001) signatures were independently associated with pCR. Tumors above the median for proliferation and immune signatures had the highest pCR rate (84 of 125; 67%), while those below the median for both signatures had the lowest pCR rate (42 of 125; 34%). Exploratory gene expression immune analyses suggested that tumors with higher inferred CD8+ T-cell infiltration may receive greater benefit with addition of carboplatin.

    Conclusions and Relevance  In this secondary analysis of a randomized clinical trial, triple-negative breast cancer subtyping revealed high pCR rates in basal-like and immunomodulatory subsets. Analysis of biological processes related to basal-like and immunomodulatory phenotypes identified tumor cell proliferation and immune scores as independent factors associated with achieving pCR; the benefit of carboplatin on pCR was seen across all molecular subtypes. Further validation of immunophenotype with existing biomarkers may help to escalate or de-escalate therapy for patients with TNBC.

    Trial Registration  ClinicalTrials.gov Identifier: NCT02032277

    Introduction

    Triple-negative breast cancers (TNBCs) are heterogeneous and their 2 most widely used gene expression–based molecular classifications are PAM501,2 and TNBCtype subtypes,3,4 both of which have been associated with response to neoadjuvant chemotherapy (NAC) and long-term prognosis.2,4 There is great interest in understanding if adding other agents to NAC, such as platinum analogs,5,6 poly-ADP ribose polymerase inhibitors,7 or immune checkpoint inhibitors, can improve pathologic complete response (pCR) rates and long-term outcomes. In concert, multiple efforts1-4,8 have identified potential biomarkers of response and resistance to NAC, including tumor-infiltrating lymphocytes9 and expression-based biomarkers; however, none of these has been prospectively shown to predict a benefit of combining additional agents with standard NAC. The objective of this study was to evaluate the association of molecular subtype, tumor proliferation, and immunophenotype with response to standard NAC alone vs NAC with carboplatin in a prospective phase 3 study of women with stages II to III TNBC and to perform exploratory immunophenotype gene expression analyses.

    Methods
    Study Population

    BrighTNess5 was a phase 3, randomized, double-blind, placebo-controlled trial of 634 women with clinical stages II to III TNBC conducted across 145 centers in 15 countries from April 2014 to March 2016. All patients provided written informed consent, including risk of biopsy. The study was approved by institutional review boards and ethics committees at investigational sites.5

    This prespecified secondary analysis was conducted from January 2018 to March 2019. Patients were randomized 2:1:1 to paclitaxel plus carboplatin plus veliparib, paclitaxel plus carboplatin plus veliparib placebo, or paclitaxel plus carboplatin placebo plus veliparib placebo. All patients subsequently received doxorubicin and cyclophosphamide for 4 cycles prior to surgery. The complete clinical study protocol is available in Supplement 1. Consolidated Standards of Reporting Trials (CONSORT) reporting guideline

    Whole Transcriptome Gene Expression Analyses

    Pretreatment biopsies were collected in RNAlater stabilization solution (ThermoFisher Scientific), and total RNA was subjected to whole transcriptome RNA sequencing on an Illumina HiSeq 3000. The PAM502 and TNBCtype3 molecular subtypes, PAM50 proliferation signature,2 GeparSixto immune signature,10 and proportion of infiltrating immune cell subsets (CIBERSORT,11 TIMER12) were calculated. The full analytical approach is available in the eMethods in Supplement 2.

    Statistical Analyses

    Association of pCR with categorical variables (basal-like/nonbasal-like subtype) or continuous variables (proliferation and immune scores), overall and by arm, was assessed via univariable and multivariable logistic regression modeling, adjusted for planned doxorubicin and cyclophosphamide administration schedule, lymph node stage, baseline Eastern Cooperative Oncology Group performance status, and germline BRCA variant status. Logistic model assumptions were tested, and multivariable model performance was assessed with Pearson goodness-of-fit tests. Statistical analyses were performed with SAS, version 9.4 (SAS Institute, Inc) and R, version 3.5.0 (R Foundation for Statistical Computing). Two-sided P ≤ .05 indicated significance.

    Results
    Study Population

    Of the 634 patients enrolled in BrighTNess, 488 (76.9%) underwent a pretreatment research biopsy, and of these, 482 (98.8%) had results of high-quality RNA sequencing data (Figure 1). Relative to the overall intention-to-treat population, patients who underwent biopsy were balanced across arms and baseline characteristics (eTable in Supplement 2). The records of all 482 patients in the present analysis had complete data for each covariate assessed.

    Molecular Subtypes and Response to NAC

    The 482 patients with evaluable RNA sequencing were balanced across arms by PAM50 and TNBC subtypes (eTable in Supplement 2). Overall, the tumors of 386 (80.1%) patients were basal-like PAM50 subtype. BRCA1 and/or BRCA2 variant carriers were not significantly different among basal-like vs nonbasal-like groups (61 of 386 [15.8%] vs 14 of 96 [14.6%]; χ2 P = .77) or among TNBC subtypes. For PAM50, basal-like had the highest pCR rate (202 of 386 [52.3%]), while among TNBC subtypes, immunomodulatory had the highest pCR rate (70 of 109 [64.2%]) and luminal androgen receptor, the lowest (8 of 29 [27.6%]) (eFigure 1 in Supplement 2).

    The prespecified primary end point was association of pCR with molecular subtype. Basal-like vs nonbasal-like were evaluated because of the low numbers of nonbasal-like subtypes. Basal-like TNBCs had a higher pCR rate compared with nonbasal-like overall (202 of 386 [52.3%] vs 34 of 96 [35.4%]; P = .003) and in the combined carboplatin-containing arms (167 of 289 [58.1%] vs 28 of 70 [40.0%]; P = .007), but not with standard paclitaxel followed by doxorubicin and cyclophosphamide (T-AC) alone (34 of 97 [35.1%] vs 6 of 26 [23.1%]; P = .25) (Figure 2A2,10; Table). As in the primary study,5 there was no significant difference in pCR across the carboplatin-containing arms (Figure 2A), so the carboplatin arms were combined in subsequent analyses. Carboplatin benefit was not significantly different in basal-like vs nonbasal subgroups (P = .80 for interaction; Figure 2A). All TNBC subtypes had higher pCR rate with carboplatin, but small numbers precluded statistical comparisons of carboplatin benefit (Figure 2B).

    Immune and Proliferation Signatures and Response to NAC

    We hypothesized that tumor cell proliferation and the immune microenvironment may refine understanding of biological features associated with response to NAC. The association between the GeparSixto immune signature score and PAM50 proliferation signature score was assessed with pCR (Table). The multivariable logistic models sufficiently met model assumptions and were considered correctly specified according to the Pearson goodness-of-fit tests. In multivariable analysis, higher immune score was significantly associated with pCR (overall P < .001; see supporting data in the Table). The effect of immune and proliferation scores on pCR did not differ significantly by carboplatin treatment.

    To understand if immune and proliferation signatures identified the same or distinct groups of patients who responded to NAC, patient findings were stratified above and below the median for proliferation and immune signatures (eFigure 2 in Supplement 2)—a prespecified threshold to avoid on-the-fly threshold determination. Among all patients, those above the median for both signatures had the highest pCR rate (84 of 125 [67%]), while those below the median for both had the lowest pCR rate (42 of 125 [34%]) (eFigure 3 in Supplement 2). A similar pattern was observed in the carboplatin-containing arms and for T-AC alone (Figure 2C and D).

    Immunophenotype and Response to NAC

    Using 2 gene expression–based approaches to determine relative proportions of intercalating immune cellular subtypes11,12 (eFigure 4 in Supplement 2), the analysis found that higher inferred CD8+ T-cell infiltration was associated with greater odds of pCR from carboplatin—a finding validated with CIBERSORT.11 By contrast, higher inferred total macrophage infiltration using TIMER12 and higher M2 macrophages using CIBERSORT11 were associated with greater odds of pCR from T-AC. Immune genes were evaluated in an unbiased manner via Gene Set Enrichment Analysis (eFigure 5 in Supplement 2), which revealed that immune high GeparSixto score tumors were enriched for gene sets, including T-cell receptor/CD3 complex, T-cell receptor signaling, T-cell exhaustion, and interferon-inducible genes.

    Discussion

    This study evaluated the patterns of response to standard NAC using whole genome expression data (RNA sequencing) among patients diagnosed with TNBC. While novel agents are being added to the treatment of TNBC, identifying subsets most likely to respond to therapies is of great importance.13 Recent data from a KEYNOTE-522 study14 pointed to increased pCR rate with the addition of pembrolizumab to a carboplatin-containing chemotherapy backbone in patients with TNBC. In the present study, carboplatin was associated with higher pCR rates across all molecular subtypes (including basal-like) but did not identify the subgroup of patients most likely to benefit from the addition of carboplatin. The results of this study concur with data from CALGB 406036 and confirm a positive association between immune activation (GeparSixto signature) and pCR.10 High proliferation and/or immune scores were associated with higher pCR rates when compared with those with low proliferation and/or immune scores. Importantly, immune score was associated with pCR, independent of proliferative score. Hypothesis-generating immunophenotype analyses implicate CD8+ T cells as potential immune biomarkers associated with benefit from carboplatin.

    Limitations

    There were several study limitations. The lack of tumor-infiltrating lymphocyte evaluation limited the ability to establish associations between immune scores and lymphocytic infiltration. Exploratory results inferring the immunophenotype and treatment benefit require validation using pathology-based markers.

    Conclusions

    Taken together, these study findings suggest that RNA-based metrics should be informative in the identification of subsets of patients most likely to benefit from novel therapies. As the field of TNBC evolves, it will be important to understand if immune checkpoint inhibitors will improve pCR rates among those patients less likely to respond to standard NAC (ie, with low proliferation and/or low immune scores). More importantly, it will be critical to define the subset of patients likely to be cured with standard chemotherapy and to concentrate efforts on patients known to have an unfavorable long-term outcome.

    Back to top
    Article Information

    Accepted for Publication: October 7, 2020.

    Published Online: February 18, 2021. doi:10.1001/jamaoncol.2020.7310

    Corresponding Author: Otto Metzger Filho, MD, Department of Medical Oncology, Dana-Farber Cancer Institute, Harvard Medical School, 450 Brookline Ave, Boston, MA 02215 (otto_metzger@dfci.harvard.edu).

    Author Contributions: Dr Metzger Filho had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Drs Metzger Filho and Stover contributed equally.

    Concept and design: Metzger Filho, Stover, Ansell, Loibl, Geyer, Untch, Golshan, Sikov, von Minckwitz, Maag, Wolmark, Symmans.

    Acquisition, analysis, or interpretation of data: Metzger Filho, Stover, Asad, Ansell, Watson, Loibl, Geyer, Bae, Collier, Cherian, O'Shaughnessy, Untch, Rugo, Huober, Golshan, Sikov, Rastogi, Maag, Denkert, Symmans.

    Drafting of the manuscript: Metzger Filho, Stover, Asad, Bae, Untch, Golshan.

    Critical revision of the manuscript for important intellectual content: All authors.

    Statistical analysis: Metzger Filho, Stover, Asad, Bae, Golshan.

    Obtained funding: Ansell, Wolmark.

    Administrative, technical, or material support: Metzger Filho, Ansell, Watson, Loibl, Cherian, Untch, Rugo, Huober, Golshan, Maag, Wolmark, Denkert, Symmans.

    Supervision: Metzger Filho, Stover, Geyer, Cherian, Untch, Golshan, von Minckwitz, Maag, Symmans.

    Otherconcept and design of the main trial: Rugo.

    Conflict of Interest Disclosures: Dr Metzger Filho reported research funding from AbbVie and personal compensation for consulting to AbbVie-sponsored advisory board outside of the submitted work. Dr Watson reported grants from Alliance Foundation Trials during the conduct of the study. Dr Loibl reported grants from AbbVie during the conduct of the study; grants from Amgen, AstraZeneca, Celgene, Daiichi Sankyo, Immunomedics, Novartis, Pfizer, and Roche outside of the submitted work; and honorarium for lectures and advisory boards paid to the institution by AstraZeneca, Bristol Myers Squibb, Celgene, Chugai, Immunomedics, Lilly, Merck, Merck Sharp & Dohme, Novartis, Pfizer, prIME/Medscape, Pierre Fabre, Puma, Roche, Samsung, and Seagen outside the submitted work. In addition, Dr Loibl reported having a patent to EP14153692.0 pending. Dr Geyer reported grants from AbbVie and the NSABP Foundation during the conduct of the study and grants from Genentech/Roche, NSABP Foundation, and AstraZeneca for NRG Oncology; travel and personal fees from Daiichi Sankyo and Exact Sciences; and personal and consulting fees from Athenex outside the submitted work. In addition, Dr Geyer reported serving without compensation on the advisory boards of Daiichi Sankyo and Exact Sciences. Dr O'Shaughnessy reported personal fees from AstraZeneca, Immunomedics, Lilly, Merck, Novartis, Pfizer, Roche, and Seattle Genetics outside the submitted work. Dr Untch reported personal fees paid to the institution by AbbVie, Agendia, Amgen, Celgene, Clovis, Daiichi Sankyo, Eisai, Lilly, Merck, Merck Sharp & Dohme, Molecular Health, Mundipharma, Myriad Genetics, Novartis, Pfizer, Pierre Fabre, Roche Pharma, and Sanofi Aventis outside the submitted work. Dr Rugo reported grants for clinical trial support to the UC Regents from Daiichi Sankyo, Immunomedics, Lilly, Macrogenics, Merck, Novartis, Odonate, Pfizer, Polyphor, Roche, Seattle Genetics, and Sermonix; consulting fees from Samsung; and honoraria from Puma outside the submitted work. Dr Huober reported personal fees from AstraZeneca, Celgene, Eisai, Lilly, Merck Sharp & Dohme, Novartis, Pfizer, and Roche; grants from Celgene, Hexal, and Novartis; and travel costs from Daiichi Sankyo, Pfizer, and Roche outside the submitted work. Dr Golshan reported personal fees from AbbVie during the conduct of the study. Dr Sikov reported travel costs from AbbVie during the conduct of the study, and patient payments for study participation paid to the institution by AbbVie outside the submitted work. Dr von Minckwitz reported equity in Cara GmbH. Dr Rastogi reported travel and lodging from AstraZeneca, Genentech/Roche, and Lilly outside the submitted work. Dr Denkert reported being founder and shareholder of Sividon Diagnostics (now Myriad); personal fees from Daiichi Sankyo, Merck Sharp & Dohme Oncology, Molecular Health, Novartis, and Roche; and grants from Myriad outside the submitted work. In addition, Dr Denkert reported having a patent to EP18209672 issued and a patent to EP20150702464 pending; and is involved in the development of VMScope digital pathology software. Dr Symmans reported being founder of Delphi Diagnostics and being a scientific advisor, shareholder, and having intellectual property in the company, as well as equity in IONIS Pharmaceuticals outside the submitted work. In addition, Dr Symmans reported having a patent issued for a method to measure residual cancer burden. No other disclosures were reported.

    Funding/Support: Research reported in this article was supported by the Alliance Foundation Trials and, in part, by funds from AbbVie, Inc. Dr Stover is supported by a Young Investigator grant from Pelotonia.

    Role of the Funder/Sponsor: AbbVie, Inc, was provided a draft of the manuscript prior to publication; however, the funders were not involved in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

    Disclaimer: The content is solely the responsibility of the authors and does not necessarily represent the official views of the Alliance Foundation Trials, LLC.

    Meeting Presentation: This work was presented in part as a poster discussion at the 2019 Annual Meeting of the American Society of Clinical Oncology; May 31-June 4, 2019; Chicago, Illinois.

    Data Sharing Statement: See Supplement 3.

    References
    1.
    Perou  CM, Sørlie  T, Eisen  MB,  et al.  Molecular portraits of human breast tumours.   Nature. 2000;406(6797):747-752. doi:10.1038/35021093PubMedGoogle ScholarCrossref
    2.
    Parker  JS, Mullins  M, Cheang  MC,  et al.  Supervised risk predictor of breast cancer based on intrinsic subtypes.   J Clin Oncol. 2009;27(8):1160-1167. doi:10.1200/JCO.2008.18.1370PubMedGoogle ScholarCrossref
    3.
    Lehmann  BD, Bauer  JA, Chen  X,  et al.  Identification of human triple-negative breast cancer subtypes and preclinical models for selection of targeted therapies.   J Clin Invest. 2011;121(7):2750-2767. doi:10.1172/JCI45014PubMedGoogle ScholarCrossref
    4.
    Lehmann  BD, Jovanović  B, Chen  X,  et al.  Refinement of triple-negative breast cancer molecular subtypes: implications for neoadjuvant chemotherapy selection.   PLoS One. 2016;11(6):e0157368. doi:10.1371/journal.pone.0157368PubMedGoogle Scholar
    5.
    Loibl  S, O’Shaughnessy  J, Untch  M,  et al.  Addition of the PARP inhibitor veliparib plus carboplatin or carboplatin alone to standard neoadjuvant chemotherapy in triple-negative breast cancer (BrighTNess): a randomised, phase 3 trial.   Lancet Oncol. 2018;19(4):497-509. doi:10.1016/S1470-2045(18)30111-6PubMedGoogle ScholarCrossref
    6.
    Sikov  WM, Berry  DA, Perou  CM,  et al.  Impact of the addition of carboplatin and/or bevacizumab to neoadjuvant once-per-week paclitaxel followed by dose-dense doxorubicin and cyclophosphamide on pathologic complete response rates in stage II to III triple-negative breast cancer: CALGB 40603 (Alliance).   J Clin Oncol. 2015;33(1):13-21. doi:10.1200/JCO.2014.57.0572PubMedGoogle ScholarCrossref
    7.
    Rugo  HS, Olopade  OI, DeMichele  A,  et al; I-SPY 2 Investigators.  Adaptive randomization of veliparib-carboplatin treatment in breast cancer.   N Engl J Med. 2016;375(1):23-34. doi:10.1056/NEJMoa1513749PubMedGoogle ScholarCrossref
    8.
    Curtis  C, Shah  SP, Chin  SF,  et al; METABRIC Group.  The genomic and transcriptomic architecture of 2,000 breast tumours reveals novel subgroups.   Nature. 2012;486(7403):346-352. doi:10.1038/nature10983PubMedGoogle ScholarCrossref
    9.
    Denkert  C, Loibl  S, Noske  A,  et al.  Tumor-associated lymphocytes as an independent predictor of response to neoadjuvant chemotherapy in breast cancer.   J Clin Oncol. 2010;28(1):105-113. doi:10.1200/JCO.2009.23.7370PubMedGoogle ScholarCrossref
    10.
    Denkert  C, von Minckwitz  G, Brase  JC,  et al.  Tumor-infiltrating lymphocytes and response to neoadjuvant chemotherapy with or without carboplatin in human epidermal growth factor receptor 2-positive and triple-negative primary breast cancers.   J Clin Oncol. 2015;33(9):983-991. doi:10.1200/JCO.2014.58.1967PubMedGoogle ScholarCrossref
    11.
    Newman  AM, Liu  CL, Green  MR,  et al.  Robust enumeration of cell subsets from tissue expression profiles.   Nat Methods. 2015;12(5):453-457. doi:10.1038/nmeth.3337PubMedGoogle ScholarCrossref
    12.
    Li  B, Severson  E, Pignon  JC,  et al.  Comprehensive analyses of tumor immunity: implications for cancer immunotherapy.   Genome Biol. 2016;17(1):174. doi:10.1186/s13059-016-1028-7PubMedGoogle ScholarCrossref
    13.
    Cortazar  P, Zhang  L, Untch  M,  et al.  Pathological complete response and long-term clinical benefit in breast cancer: the CTNeoBC pooled analysis.   Lancet. 2014;384(9938):164-172. doi:10.1016/S0140-6736(13)62422-8PubMedGoogle ScholarCrossref
    14.
    Schmid  P, Cortes  J, Pusztai  L,  et al; KEYNOTE-522 Investigators.  Pembrolizumab for early triple-negative breast cancer.   N Engl J Med. 2020;382(9):810-821. doi:10.1056/NEJMoa1910549PubMedGoogle ScholarCrossref
    ×