Gene Expression–Guided Adjuvant Chemotherapy in Breast Cancer


To the Editor

The Trial Assigning Individualized Options for Treatment (TAILORx) conducted by Sparano et al. (July 12 issue)1 showed that endocrine therapy alone was noninferior to chemoendocrine therapy in women with hormone-receptor–positive, human epidermal growth factor receptor 2–negative, axillary node–negative breast cancer who had a midrange 21-gene recurrence score. Recurrence scores based on the 21-gene breast cancer assay range from 0 to 100, with high scores indicating a chemotherapy benefit and low scores indicating a low risk of recurrence in the absence of chemotherapy. The Ki-67 index, a defined biomarker of the response to treatment and clinical outcome in patients with breast cancer,2 has been incorporated into the final calculation of the 21-gene recurrence score. However, in TAILORx, should the value of the Ki-67 index have been weighed alone or received special attention, as in a subgroup analysis?

Among women with node-negative, hormone-receptor–positive breast cancer, the level of Ki-67 determined luminal subtypes (luminal A or luminal B). Patients with luminal A breast tumors received endocrine therapy that antagonized hormone receptors, and patients with luminal B breast cancer received chemoendocrine therapy for suppression of tumor proliferation. For women with a midrange 21-gene recurrence score, it would be somehow arbitrary to ignore the fact that a high level of Ki-67 indicates a high proliferative burden of tumor cells. The chemotherapy-free strategy was not powerful enough to reduce the risk of recurrence in such a population. A stratified analysis that uses Ki-67 levels, especially in the prediction of recurrence among premenopausal women, may help to elucidate this issue.

Yanlei Cheng, M.D., Ph.D.
Yaogui Wu, M.D., Ph.D.
Long Wu, M.D., Ph.D.
Renmin Hospital of Wuhan University, Wuhan, China

No potential conflict of interest relevant to this letter was reported.

  1. 1. Sparano JA, Gray RJ, Makower DF, et al. Adjuvant chemotherapy guided by a 21-gene expression assay in breast cancer. N Engl J Med 2018;379:111121.

  2. 2. Dowsett M, Nielsen TO, A’Hern R, et al. Assessment of Ki67 in breast cancer: recommendations from the International Ki67 in Breast Cancer working group. J Natl Cancer Inst 2011;103:16561664.

To the Editor

Sparano and colleagues conducted an important trial assessing the noninferiority of endocrine therapy to chemoendocrine therapy for invasive disease–free survival. A hazard ratio for invasive disease recurrence, second primary cancer, or death (endocrine vs. chemoendocrine therapy) of 1.08 (95% confidence interval [CI], 0.94 to 1.24) was observed. On the basis of the upper bound of the confidence interval, a potential increase of 24% in the hazard ratio associated with endocrine therapy alone makes it difficult to claim the noninferiority of endocrine therapy to chemoendocrine therapy clinically. The concerns about the use of the hazard ratio in noninferiority studies have been discussed extensively.1-3

An alternative is to use the restricted mean survival time, which is defined as the average or expected duration of event-free survival over a specified follow-up period. The restricted mean survival time may be estimated with the use of the area under the empirical survival curve across the follow-up period. With the use of reconstructed data on invasive disease–free survival from Figure 2A of the article, at 108 months of follow-up, the restricted mean survival time would be 101.00 months among patients who received chemoendocrine therapy and 100.62 months among those who received endocrine therapy. That is, on average, patients who received chemoendocrine therapy had 101 months of invasive disease–free survival. The difference between the restricted mean survival times (chemoendocrine therapy vs. endocrine therapy) was 0.38 months (95% CI, −0.65 to 1.41). In the worst case, endocrine therapy might shorten invasive disease–free survival by 1.41 months, which is 1.31% of 108 months of follow-up. These quantifications provide more clinically interpretable summaries of the treatment effect than does the hazard ratio, and they should aid practitioners and patients during treatment selection.4

Zachary R. McCaw, M.A.
Lee-Jen Wei, Ph.D.
Harvard T.H. Chan School of Public Health, Boston, MA

Dae Hyun Kim, Sc.D., M.D.
Brigham and Women’s Hospital, Boston, MA

No potential conflict of interest relevant to this letter was reported.

  1. 1. Uno H, Wittes J, Fu H, et al. Alternatives to hazard ratios for comparing the efficacy or safety of therapies in noninferiority studies. Ann Intern Med 2015;163:127134.

  2. 2. Hasegawa T, Uno H, Wei LJ. Safety study of salmeterol in asthma in adults. N Engl J Med 2016;375:10971097.

  3. 3. Cheng D, Pak K, Wei LJ. Demonstrating noninferiority of accelerated radiotherapy with panitumumab vs standard radiotherapy with cisplatin in locoregionally advanced squamous cell head and neck carcinoma. JAMA Oncol 2017;3:14301431.

  4. 4. Merchant FM, Dickert NW Jr, Howard DH. Mandatory shared decision making by the Centers for Medicare & Medicaid Services for Cardiovascular Procedures and Other Tests. JAMA 2018 June 4 (Epub ahead of print).

To the Editor

Both the current article and previously reported data1 from TAILORx showed the value of the 21-gene recurrence score in women with hormone receptor–positive breast cancer. June 10, 2018, marked the 25th anniversary of the signing of the National Institutes of Health Revitalization Act, which promotes the inclusion of minorities in clinical trials. Viewed through the lens of this landmark event in medical research history, it is surprising that the TAILORx investigators did not report the distribution of the racial or ethnic backgrounds of the trial participants.

High rates of hormone-receptor–negative phenotypes among black patients may have resulted in accrual challenges. However, a subgroup analysis involving even a few hundred patients might yield valuable data regarding the performance of the assay in nonwhite patients. This is a relevant issue, since survival disadvantages have been reported among black patients with hormone-receptor–positive disease.2-4

There is an ethical obligation to request the outcomes of TAILORx according to race or ethnic group. The National Institutes of Health and the oncology community advocate for the accrual of diverse patients into clinical trials to strengthen the generalizability of results and to identify clinically significant outcome-related sociodemographic factors. Investigators should document and evaluate the results of these efforts in order to support sincere, trustworthy outreach to a diverse patient population.

Lisa A. Newman, M.D., M.P.H.
Weill Cornell Medicine, New York, NY

No potential conflict of interest relevant to this letter was reported.

  1. 1. Sparano JA, Gray RJ, Makower DF, et al. Prospective validation of a 21-gene expression assay in breast cancer. N Engl J Med 2015;373:20052014.

  2. 2. Warner ET, Tamimi RM, Hughes ME, et al. Racial and ethnic differences in breast cancer survival: mediating effect of tumor characteristics and sociodemographic and treatment factors. J Clin Oncol 2015;33:22542261.

  3. 3. Sparano JA, Wang M, Zhao F, et al. Race and hormone receptor-positive breast cancer outcomes in a randomized chemotherapy trial. J Natl Cancer Inst 2012;104:406414.

  4. 4. Iqbal J, Ginsburg O, Rochon PA, Sun P, Narod SA. Differences in breast cancer stage at diagnosis and cancer-specific survival by race and ethnicity in the United States. JAMA 2015;313:165173.


The authors reply: Cheng and colleagues recommend stratified analysis including Ki-67 RNA expression for the prediction of the benefit of chemotherapy. Ki-67 is one of 5 proliferation genes (MKI67, BIRC5, CCNB1, MYBL2, and AURKA) included in the proliferation group of the 21-gene recurrence score,1 which has previously been shown to predict chemotherapy benefit when the recurrence score is high, whether that score is defined as 31 or higher2 or 26 or higher.3 In our trial, the recurrence score was selected rather than the proliferation group or MKI67 alone, since the prognostic and predictive information provided by the recurrence score was more accurate than that provided by the group or individual gene.1,2 It seems unlikely that the proliferation score, or MKI67 alone, will predict chemotherapy benefit in the group of patients who underwent randomization and who had a recurrence score of 11 to 25, given the noninferiority of endocrine therapy alone within this range.4

McCaw et al. express concern that the use of the hazard ratio to establish the noninferiority of endocrine therapy alone did not involve exclusion of up to a 24% relative difference in invasive disease–free survival favoring chemoendocrine therapy. They confirmed the noninferiority using an alternative method — by comparing the restricted mean survival times, which they described as being “more clinically interpretable than the hazard ratio.” We think that the reported similar rates of invasive disease–free survival at 5 years (92.8% and 93.1%) and 9 years (83.3% and 84.3%) associated with the noninferiority hazard ratio margin described in our article provide an assessment that is as clinically meaningful as the difference in restricted median survival time provided by the alternative analysis. Nevertheless, we appreciate that their alternative analytic method confirms our major finding.

Newman advocates for reporting of the TAILORx outcomes according to race and ethnic group, a position that we fully agree with. In 2014, we reported that 7% of TAILORx patients were black and 9% were Hispanic and that black and Hispanic women were younger and more likely to have tumors that were larger than 2 cm in diameter, high-grade, and progesterone-receptor–negative than white or non-Hispanic women, respectively, although there were no differences in recurrence score according to race or ethnic group.4 When outcome data became available and were examined according to race and ethnic group, as Newman suggested, we observed overlapping confidence intervals for noninferiority of endocrine therapy alone in the cohort with a recurrence score of 11 to 25. These findings suggest that the TAILORx results are generalizable to black and Hispanic women with early breast cancer. An updated analysis of our previous findings that is focused on race and ethnic group is under way to provide the additional details that this subject deserves.

Joseph A. Sparano, M.D.
Montefiore Einstein Center for Cancer Care, Bronx, NY

Robert J. Gray, Ph.D.
Dana–Farber Cancer Institute, Boston, MA

Since publication of their article, the authors report no further potential conflict of interest.

  1. 1. Paik S, Shak S, Tang G, et al. A multigene assay to predict recurrence of tamoxifen-treated, node-negative breast cancer. N Engl J Med 2004;351:28172826.

  2. 2. Paik S, Tang G, Shak S, et al. Gene expression and benefit of chemotherapy in women with node-negative, estrogen receptor-positive breast cancer. J Clin Oncol 2006;24:37263734.

  3. 3. Sparano JA, Paik S. Development of the 21-gene assay and its application in clinical practice and clinical trials. J Clin Oncol 2008;26:721728.

  4. 4. Sparano JA, Gray RJ, Zujewski JA, et al. Recurrence score and clinicopathologic characteristics of TAILORx participants by race and ethnicity. J Clin Oncol 2014;32 (suppl 26).

Articles You May Like

Is this a standard file format?
Sloths are better at adapting than we ever gave them credit for
Our galaxy’s supermassive black hole could be pointing a relativistic jet right at us
Human ESC-Derived Chimeric Mouse Models of Huntington’s Disease Reveal Cell-Intrinsic Defects in Glial Progenitor Cell Differentiation
miCloud: a bioinformatics cloud for seamless execution of complex NGS data analysis pipelines

Leave a Reply

Your email address will not be published. Required fields are marked *