Randomized Trial of Standard Adjuvant Chemotherapy Regimens Versus Capecitabine in Older Women With Early Breast Cancer: 10-Year Update of the CALGB 49907 Trial.

PURPOSE: Older women with breast cancer remain under-represented in clinical trials. The Cancer and Leukemia Group B 49907 trial focused on women age 65 years and older. We previously reported the primary analysis after a median follow-up of 2.4 years. Standard adjuvant chemotherapy showed significant improvements in recurrence-free survival (RFS) and overall survival compared with capecitabine. We now update results at a median follow-up of 11.4 years. PATIENTS AND METHODS: Patients age 65 years or older with early breast cancer were randomly assigned to either standard adjuvant chemotherapy (physician's choice of either cyclophosphamide, methotrexate, and fluorouracil or cyclophosphamide and doxorubicin) or capecitabine. An adaptive Bayesian design was used to determine sample size and test noninferiority of capecitabine. The primary end point was RFS. RESULTS: The design stopped accrual with 633 patients at its first sample size assessment. RFS remains significantly longer for patients treated with standard chemotherapy. At 10 years, in patients treated with standard chemotherapy versus capecitabine, the RFS rates were 56% and 50%, respectively (hazard ratio [HR], 0.80; P = .03); breast cancer-specific survival rates were 88% and 82%, respectively (HR, 0.62; P = .03); and overall survival rates were 62% and 56%, respectively (HR, 0.84; P = .16). With longer follow-up, standard chemotherapy remains superior to capecitabine among hormone receptor-negative patients (HR, 0.66; P = .02), but not among hormone receptor-positive patients (HR, 0.89; P = .43). Overall, 43.9% of patients have died (13.1% from breast cancer, 16.4% from causes other than breast cancer, and 14.1% from unknown causes). Second nonbreast cancers occurred in 14.1% of patients. CONCLUSION: With longer follow-up, RFS remains superior for standard adjuvant chemotherapy versus capecitabine, especially in patients with hormone receptor-negative disease. Competing risks in this older population dilute overall survival benefits.

Implementing pharmacogenetic testing in rural primary care practices: a pilot feasibility study.

Aim: Assess feasibility and perspectives of pharmacogenetic testing/PGx in rural, primary care physician (PCP) practices when PCPs are trained to interpret/apply results and testing costs are covered. Methods: Participants included PCPs who agreed to training, surveys and interviews and eligible patients who agreed to surveys and testing. 51 patients from three practices participated. Results: Prestudy, no PCP had ever ordered a PGx test. Test results demonstrated gene variations in 30% of patients, related to current medications, with PCPs reporting changes to drug management. Poststudy, test cost was still a concern, but now PCPs reported practical barriers, including the utilization of PGx results over time. PCPs and patients had favorable responses to testing. Summary: PGx testing is feasible in rural PCP practices.

Multi-site investigation of strategies for the clinical implementation of CYP2D6 genotyping to guide drug prescribing.

PURPOSE: A number of institutions have clinically implemented CYP2D6 genotyping to guide drug prescribing. We compared implementation strategies of early adopters of CYP2D6 testing, barriers faced by both early adopters and institutions in the process of implementing CYP2D6 testing, and approaches taken to overcome these barriers. METHODS: We surveyed eight early adopters of CYP2D6 genotyping and eight institutions in the process of adoption. Data were collected on testing approaches, return of results procedures, applications of genotype results, challenges faced, and lessons learned. RESULTS: Among early adopters, CYP2D6 testing was most commonly ordered to assist with opioid and antidepressant prescribing. Key differences among programs included test ordering and genotyping approaches, result reporting, and clinical decision support. However, all sites tested for copy-number variation and nine common variants, and reported results in the medical record. Most sites provided automatic consultation and had designated personnel to assist with genotype-informed therapy recommendations. Primary challenges were related to stakeholder support, CYP2D6 gene complexity, phenotype assignment, and sustainability. CONCLUSION: There are specific challenges unique to CYP2D6 testing given the complexity of the gene and its relevance to multiple medications. Consensus lessons learned may guide those interested in pursuing similar clinical pharmacogenetic programs.

Proliferative potential and resistance to immune checkpoint blockade in lung cancer patients.

BACKGROUND: Resistance to immune checkpoint inhibitors (ICIs) has been linked to local immunosuppression independent of major ICI targets (e.g., PD-1). Clinical experience with response prediction based on PD-L1 expression suggests that other factors influence sensitivity to ICIs in non-small cell lung cancer (NSCLC) patients. METHODS: Tumor specimens from 120 NSCLC patients from 10 institutions were evaluated for PD-L1 expression by immunohistochemistry, and global proliferative profile by targeted RNA-seq. RESULTS: Cell proliferation, derived from the mean expression of 10 proliferation-associated genes (namely BUB1, CCNB2, CDK1, CDKN3, FOXM1, KIAA0101, MAD2L1, MELK, MKI67, and TOP2A), was identified as a marker of response to ICIs in NSCLC. Poorly, moderately, and highly proliferative tumors were somewhat equally represented in NSCLC, with tumors with the highest PD-L1 expression being more frequently moderately proliferative as compared to lesser levels of PD-L1 expression. Proliferation status had an impact on survival in patients with both PD-L1 positive and negative tumors. There was a significant survival advantage for moderately proliferative tumors compared to their combined highly/poorly counterparts (p = 0.021). Moderately proliferative PD-L1 positive tumors had a median survival of 14.6 months that was almost twice that of PD-L1 negative highly/poorly proliferative at 7.6 months (p = 0.028). Median survival in moderately proliferative PD-L1 negative tumors at 12.6 months was comparable to that of highly/poorly proliferative PD-L1 positive tumors at 11.5 months, but in both instances less than that of moderately proliferative PD-L1 positive tumors. Similar to survival, proliferation status has impact on disease control (DC) in patients with both PD-L1 positive and negative tumors. Patients with moderately versus those with poorly or highly proliferative tumors have a superior DC rate when combined with any classification schema used to score PD-L1 as a positive result (i.e., TPS ≥ 50% or ≥ 1%), and best displayed by a DC rate for moderately proliferative tumors of no less than 40% for any classification of PD-L1 as a negative result. While there is an over representation of moderately proliferative tumors as PD-L1 expression increases this does not account for the improved survival or higher disease control rates seen in PD-L1 negative tumors. CONCLUSIONS: Cell proliferation is potentially a new biomarker of response to ICIs in NSCLC and is applicable to PD-L1 negative tumors.

Implementing a personalized medicine program in a community health system.

The implementation of a de novo personalized medicine program in a rural community health system serving an underserved population is described. Focusing on the safe use of drugs impacted by genetic variations in the non-oncology setting, we first addressed drug-gene pairs designated by the US FDA in black-box warnings (codeine, clopidogrel, abacavir, carbamazepine). The program's first success was a policy change to remove codeine from the pediatric formulary, rather than a testing recommendation. Pilot studies were then conducted with primary care providers to get them familiar with pharmacogenetic testing, and a consultative outpatient clinic for patients was developed. The assessment, planning, implementation, challenges, successes and lessons learned are described.

Research Directions in the Clinical Implementation of Pharmacogenomics: An Overview of US Programs and Projects.

Response to a drug often differs widely among individual patients. This variability is frequently observed not only with respect to effective responses but also with adverse drug reactions. Matching patients to the drugs that are most likely to be effective and least likely to cause harm is the goal of effective therapeutics. Pharmacogenomics (PGx) holds the promise of precision medicine through elucidating the genetic determinants responsible for pharmacological outcomes and using them to guide drug selection and dosing. Here we survey the US landscape of research programs in PGx implementation, review current advances and clinical applications of PGx, summarize the obstacles that have hindered PGx implementation, and identify the critical knowledge gaps and possible studies needed to help to address them.

Patient perspectives on the barriers associated with medication adherence to oral chemotherapy.

Purpose Appropriate use of oral chemotherapy is a challenge for patients and clinicians. The purpose of this study was to analyze cancer patients' use of oral chemotherapies and identify opportunities to improve adherence. Methods We developed a 30-question survey to address frequency and reasons for reducing/skipping doses; sources of information for oral chemotherapy use; perceived importance of food-drug effects; and ease of understanding labeling directions. Results Ninety-three patients taking oral chemotherapies with chronic myeloid leukemia, renal cell carcinoma, breast cancer, and colorectal cancer completed the survey. This was a well-educated population with 69% (n = 62) having completed some college; 51% (n = 47) female and 59% (n = 54) older than 50 years of age. Thirty percent of patients reported forgetting to take their oral chemotherapy at least "sometimes". Younger patients (<50 vs. ≥50, p = 0.002), shorter treatment duration (<6 vs. ≥6 months p = 0.03), or with chronic myeloid leukemia (vs. other diagnoses, p = 0.015) forget to take their oral chemotherapy at higher rates. Twenty-three percent (n = 21) indicated they intentionally skipped their oral chemotherapies and 38% (n = 8) of those did not inform their physicians. Forty-one percent (n = 28) taking drugs with significant food-drug effects did not think about their last meal before taking their oral chemotherapy and 80% (n = 55) did not understand the potential interactions. Additionally, 39% (n = 36/92) never looked at labeling and 15% (n = 14/91) had difficulty understanding label directions. Conclusion There are three main barriers associated with appropriate use of oral chemotherapies: misunderstanding about the timing of drug with food; stopping drug without informing physicians; and difficulty understanding labeling directions. A multipronged approach is needed to optimize communication of directions for optimal oral chemotherapy use.

Participation in Cancer Pharmacogenomic Studies: A Study of 8456 Patients Registered to Clinical Trials in the Cancer and Leukemia Group B (Alliance).

BACKGROUND: Clinically annotated specimens from cancer clinical trial participants offer an opportunity for discovery and validation of pharmacogenomic findings. The purpose of this observational study is to better understand patient/institution factors that may contribute to participation in the pharmacogenomic component of prospective cancer clinical trials. METHODS: Patient demographic information (age, sex, self-reported race) and institutional characteristics (CALGB/CTSU site, "diversity," and accrual) were evaluated for 8456 patients enrolled in seven CALGB phase III studies with a pharmacogenomic component. All statistical tests were two-sided. RESULTS: The majority of patients (81%) consented to participate in the pharmacogenomic component. However, in a multivariable analysis, site (CALGB vs CTSU) and "institutional diversity" (percent minority cancer patients on national trials) were statistically significantly associated with participation. For both whites and nonwhites, patients from CALGB sites were more likely to participate compared with patients from CTSU sites (whites: odds ratio [OR] = 2.26, 95% confidence interval [CI] = 1.68 to 3.04, P < .001; nonwhites: OR = 1.79, 95% CI = 1.52 to 2.11, P < .001). However, as "institutional diversity" increased, the likelihood of participation in the pharmacogenomics component decreased for both white (OR = 0.94, 95% CI = 0.91 to 0.97, P < .001) and nonwhite patients (OR = 0.90, 95% CI = 0.81 to 1.00, P = .05). CONCLUSIONS: Most clinical trial cancer patients across geographical, racial, and practice settings are willing to participate in pharmacogenomic studies. However, to promote equitable benefit to the larger cancer community, optimization of both patient and institutional participation are needed. Institutional factors may be even more compelling than patient demographics. Prospective studies are needed to identify and address barriers/incentives to participation in pharmacogenomic research at the patient, clinician, and institutional levels.

Genomics education for the public: perspectives of genomic researchers and ELSI advisors.

AIMS: For more than two decades genomic education of the public has been a significant challenge. As genomic information becomes integrated into daily life and routine clinical care, the need for public education is even more critical. We conducted a pilot study to learn how genomic researchers and ethical, legal, and social implications advisors who were affiliated with large-scale genomic variation studies have approached the issue of educating the public about genomics. METHODS/RESULTS: Semi-structured telephone interviews were conducted with researchers and advisors associated with the SNP/HAPMAP studies and the Cancer Genome Atlas Study. Respondents described approach(es) associated with educating the public about their study. Interviews were audio-recorded, transcribed, coded, and analyzed by team review. Although few respondents described formal educational efforts, most provided recommendations for what should/could be done, emphasizing the need for an overarching entity(s) to take responsibility to lead the effort to educate the public. Opposing views were described related to: who this should be; the overall goal of the educational effort; and the educational approach. Four thematic areas emerged: What is the rationale for educating the public about genomics?; Who is the audience?; Who should be responsible for this effort?; and What should the content be? Policy issues associated with these themes included the need to agree on philosophical framework(s) to guide the rationale, content, and target audiences for education programs; coordinate previous/ongoing educational efforts; and develop a centralized knowledge base. Suggestions for next steps are presented. CONCLUSION: A complex interplay of philosophical, professional, and cultural issues can create impediments to genomic education of the public. Many challenges, however, can be addressed by agreement on a guiding philosophical framework(s) and identification of a responsible entity(s) to provide leadership for developing/overseeing an appropriate infrastructure to support the coordination/integration/sharing and evaluation of educational efforts, benefiting consumers and professionals.

Cancer pharmacogenomics, adoption by oncologists and patient benefit.

AIM: The study aim was to understand physician experience and factors influencing the adoption of cancer pharmacogenomic (caPGx) testing by oncologists practicing in academic and nonacademic settings. METHOD: Anonymous paper surveys were distributed to oncologists practicing in North Carolina (USA). RESULTS: Although 98% of oncologists see promise in utilizing PGx tests in their practice, few were comfortable with their knowledge (33%) or interpreting test results (37%). At one site, the survey was not distributed due to clinician unfamiliarity with the term 'pharmacogenomics'. Compared with oncologists in academia, community oncologists were more likely to order the new Oncotype Dx™ test for colon cancer (33% vs 0; p = 0.0071), more likely to indicate future use of caPGx tests (94 vs 75%; p = 0.012) and less likely to have never ordered a caPGx test (2 vs 35%%; p < 0.001). Nearly every oncologist was interested in additional PGx education. CONCLUSION: A critical need exists to disseminate accurate and updated caPGx information to oncologists practicing in both academic and nonacademic settings.

Integrating personalized genomic medicine into routine clinical care: addressing the social and policy issues of pharmacogenomic testing.

The provision of personalized genomic medicine presents significant policy challenges, such as ensuring equitable patient access to testing, preparing clinicians to manage genomic results, justifying test reimbursement, sharing genomic information for patient care, and protecting patients against misuse of genetic information.

Return of research results from pharmacogenomic versus disease susceptibility studies: what's drugs got to do with it?

One of the most controversial ethical issues in genomics research is the return of individual research results to research subjects. As new technologies, including whole-genome sequencing, provide an increased opportunity for researchers to find clinically relevant research results, the questions related to if, when and how individual research results should be returned become more central to the ethical conduct of genomic research. In the absence of federal guidance on this issue, many groups and individuals have developed recommendations and suggestions to address these questions. Most of these recommendations have focused on the return of individual results from disease susceptibility studies. However, in addition to predicting the development of disease, genomic research also includes predicting an individual's response to drugs, especially the risk of developing adverse events. This article evaluates and compares the return of individual research results from disease susceptibility studies versus pharmacogenomic studies.

Pathologic complete response predicts recurrence-free survival more effectively by cancer subset: results from the I-SPY 1 TRIAL--CALGB 150007/150012, ACRIN 6657.

PURPOSE: Neoadjuvant chemotherapy for breast cancer provides critical information about tumor response; how best to leverage this for predicting recurrence-free survival (RFS) is not established. The I-SPY 1 TRIAL (Investigation of Serial Studies to Predict Your Therapeutic Response With Imaging and Molecular Analysis) was a multicenter breast cancer study integrating clinical, imaging, and genomic data to evaluate pathologic response, RFS, and their relationship and predictability based on tumor biomarkers. PATIENTS AND METHODS: Eligible patients had tumors ≥ 3 cm and received neoadjuvant chemotherapy. We determined associations between pathologic complete response (pCR; defined as the absence of invasive cancer in breast and nodes) and RFS, overall and within receptor subsets. RESULTS: In 221 evaluable patients (median tumor size, 6.0 cm; median age, 49 years; 91% classified as poor risk on the basis of the 70-gene prognosis profile), 41% were hormone receptor (HR) negative, and 31% were human epidermal growth factor receptor 2 (HER2) positive. For 190 patients treated without neoadjuvant trastuzumab, pCR was highest for HR-negative/HER2-positive patients (45%) and lowest for HR-positive/HER2-negative patients (9%). Achieving pCR predicted favorable RFS. For 172 patients treated without trastuzumab, the hazard ratio for RFS of pCR versus no pCR was 0.29 (95% CI, 0.07 to 0.82). pCR was more predictive of RFS by multivariate analysis when subtype was taken into account, and point estimates of hazard ratios within the HR-positive/HER2-negative (hazard ratio, 0.00; 95% CI, 0.00 to 0.93), HR-negative/HER2-negative (hazard ratio, 0.25; 95% CI, 0.04 to 0.97), and HER2-positive (hazard ratio, 0.14; 95% CI, 0.01 to 1.0) subtypes are lower. Ki67 further improved the prediction of pCR within subsets. CONCLUSION: In this biologically high-risk group, pCR differs by receptor subset. pCR is more highly predictive of RFS within every established receptor subset than overall, demonstrating that the extent of outcome advantage conferred by pCR is specific to tumor biology.

Ethical aspects of participation in the database of genotypes and phenotypes of the National Center for Biotechnology Information: the Cancer and Leukemia Group B Experience.

BACKGROUND: The rapid pace of genetics research, coupled with evolving standards for informed consent, can create ethical challenges regarding future use of tissue or information from completed clinical trials. The Cancer and Leukemia Group B (CALGB) Oncology Cooperative Group was faced with an ethical dilemma regarding sharing genetic data from a completed genome-wide association study (GWAS) that was conducted as part of a large, multicenter breast cancer clinical trial with a national database: the Database of Genotypes and Phenotypes National Center for Biotechnology Information (dbGaP). METHODS: The CALGB Ethics Committee conducted a series of multidisciplinary meetings and teleconferences involving patient advocates, bioethicists, clinical researchers, and clinical oncologists to evaluate the ethical issues raised by this case and to identify lessons for improving informed consent to future genetics research in oncology trials. RESULTS: The Ethics Committee recommended that GWAS data be provided to dbGaP consistent with documented consent for future use of tissue among trial participants. Ethical issues, including adequacy of informed consent to future research, limitations of privacy in modern genetics research, the potential impact of population-based genetics research on health disparities, and recontact of research participants for clinical care or further research, were identified as major ethical considerations in this area. CONCLUSIONS: Although modern standards for informed consent should not prohibit research or sharing of data consistent with participant's intent and the public interest, there is an urgent need for national consensus on the appropriate use of archived tissue and standardized informed consent for future research among cancer clinical trial participants.

IRB perspectives on the return of individual results from genomic research.

PURPOSE: Return of individual research results from genomic studies is a hotly debated ethical issue in genomic research. However, the perspective of key stakeholders-institutional review board (IRB) professionals-has been missing from this dialogue. This study explores the positions and experiences of IRB members and staff regarding this issue. METHODS: In-depth interviews with 31 IRB professionals at six sites across the United States. RESULTS: IRB professionals agreed that research results should be returned to research participants when results are medically actionable but only if the participants want to know the results. Many respondents expected researchers to address the issue of return of results (ROR) in the IRB application and informed-consent document. Many respondents were not comfortable with their expertise in genomics research and only a few described actual experiences in addressing ROR. Although participants agreed that guidelines would be helpful, most were reticent to develop them in isolation. Even where IRB guidance exists (e.g., Clinical Laboratory Improvement Act (CLIA) lab certification required for return), in practice, the guidance has been overruled to allow ROR (e.g., no CLIA lab performs the assay). CONCLUSION: An IRB-researcher partnership is needed to help inform responsible and feasible institutional approaches to returning research results.

Chemotherapy response and recurrence-free survival in neoadjuvant breast cancer depends on biomarker profiles: results from the I-SPY 1 TRIAL (CALGB 150007/150012; ACRIN 6657).

Neoadjuvant chemotherapy for breast cancer allows individual tumor response to be assessed depending on molecular subtype, and to judge the impact of response to therapy on recurrence-free survival (RFS). The multicenter I-SPY 1 TRIAL evaluated patients with ≥ 3 cm tumors by using early imaging and molecular signatures, with outcomes of pathologic complete response (pCR) and RFS. The current analysis was performed using data from patients who had molecular profiles and did not receive trastuzumab. The various molecular classifiers tested were highly correlated. Categorization of breast cancer by molecular signatures enhanced the ability of pCR to predict improvement in RFS compared to the population as a whole. In multivariate analysis, the molecular signatures that added to the ability of HR and HER2 receptors, clinical stage, and pCR in predicting RFS included 70-gene signature, wound healing signature, p53 mutation signature, and PAM50 risk of recurrence. The low risk signatures were associated with significantly better prognosis, and also identified additional patients with a good prognosis within the no pCR group, primarily in the hormone receptor positive, HER-2 negative subgroup. The I-SPY 1 population is enriched for tumors with a poor prognosis but is still heterogeneous in terms of rates of pCR and RFS. The ability of pCR to predict RFS is better by subset than it is for the whole group. Molecular markers improve prediction of RFS by identifying additional patients with excellent prognosis within the no pCR group.

Ethical and practical challenges of sharing data from genome-wide association studies: the eMERGE Consortium experience.

In 2007, the National Human Genome Research Institute (NHGRI) established the Electronic MEdical Records and GEnomics (eMERGE) Consortium (www.gwas.net) to develop, disseminate, and apply approaches to research that combine DNA biorepositories with electronic medical record (EMR) systems for large-scale, high-throughput genetic research. One of the major ethical and administrative challenges for the eMERGE Consortium has been complying with existing data-sharing policies. This paper discusses the challenges of sharing genomic data linked to health information in the electronic medical record (EMR) and explores the issues as they relate to sharing both within a large consortium and in compliance with the National Institutes of Health (NIH) data-sharing policy. We use the eMERGE Consortium experience to explore data-sharing challenges from the perspective of multiple stakeholders (i.e., research participants, investigators, and research institutions), provide recommendations for researchers and institutions, and call for clearer guidance from the NIH regarding ethical implementation of its data-sharing policy.

p53 Expression in node-positive breast cancer patients: results from the Cancer and Leukemia Group B 9344 Trial (159905).

PURPOSE: p53 as a prognostic and predictive factor in early-stage breast cancer has had mixed results. We studied p53 protein expression, by immunohistochemistry, in a randomized clinical trial of stage II patients treated with adjuvant doxorubicin and cyclophosphamide with or without paclitaxel [Cancer and Leukemia Group B (CALGB) 9344, INT0148]. PATIENTS AND METHODS: Epithelial p53 expression was evaluated using two immunohistochemical antibodies (DO7 and 1801) in formalin-fixed, paraffin-embedded tissue from patients with node-positive breast cancer who were randomized to four cycles of cyclophosphamide and one of three doses of doxorubicin (60, 75, or 90 mg/m(2); AC) and to receive four subsequent cycles of paclitaxel (T) or not. Prognostic and predictive value of p53 protein expression was assessed, independent of treatment assignment, for escalating doses of doxorubicin or addition of T with endpoints of relapse-free (RFS) and overall survival (OS). RESULTS: Of 3,121 patients, 1,887 patient specimens treated on C9344 were obtained, passed quality control, and evaluated for p53 expression. Expression was 23% and 27% for mAbs 1801 and D07, respectively, with 92% concordance. In univariate analysis, p53 positivity was associated with worse OS with either antibody, but only p53 staining with monoclonal antibody 1801 had significantly worse RFS. In multivariate analysis, p53 was not predictive of RFS or OS from either doxorubicin dose escalation or addition of paclitaxel regardless of the antibody. CONCLUSION: Nuclear staining of p53 by immunohistochemistry is associated with worse prognosis in node-positive patients treated with adjuvant doxorubicin-based chemotherapy but is not a useful predictor of benefit from doxorubicin dose escalation or the addition of paclitaxel.

Intrinsic breast tumor subtypes, race, and long-term survival in the Carolina Breast Cancer Study.

PURPOSE: Previous research identified differences in breast cancer-specific mortality across 4 intrinsic tumor subtypes: luminal A, luminal B, basal-like, and human epidermal growth factor receptor 2 positive/estrogen receptor negative (HER2(+)/ER(-)). EXPERIMENTAL DESIGN: We used immunohistochemical markers to subtype 1,149 invasive breast cancer patients (518 African American, 631 white) in the Carolina Breast Cancer Study, a population-based study of women diagnosed with breast cancer. Vital status was determined through 2006 using the National Death Index, with median follow-up of 9 years. RESULTS: Cancer subtypes luminal A, luminal B, basal-like, and HER2(+)/ER(-) were distributed as 64%, 11%, 11%, and 5% for whites, and 48%, 8%, 22%, and 7% for African Americans, respectively. Breast cancer mortality was higher for participants with HER2(+)/ER(-) and basal-like breast cancer compared with luminal A and B. African Americans had higher breast cancer-specific mortality than whites, but the effect of race was statistically significant only among women with luminal A breast cancer. However, when compared with the luminal A subtype within racial categories, mortality for participants with basal-like breast cancer was higher among whites (HR = 2.0, 95% CI: 1.2-3.4) than African Americans (HR = 1.5, 95% CI: 1.0-2.4), with the strongest effect seen in postmenopausal white women (HR = 3.9, 95% CI: 1.5-10.0). CONCLUSIONS: Our results confirm the association of basal-like breast cancer with poor prognosis and suggest that basal-like breast cancer is not an inherently more aggressive disease in African American women compared with whites. Additional analyses are needed in populations with known treatment profiles to understand the role of tumor subtypes and race in breast cancer mortality, and in particular our finding that among women with luminal A breast cancer, African Americans have higher mortality than whites.