Use of the 21-Gene Recurrence Score to Predict Clinical Outcomes in Early Breast Cancer

To the Editor Sparano and colleagues¹ published the results of arm D (recurrence scores, 26-100) of the Trial Assigning Individualized Options for Treatment (TAILORx), a prospective validation of the clinical utility of a 21-gene recurrence score in women with hormone receptor–positive, ERBB2-negative, lymph node–negative, early breast cancer. Since the 2018 trial² that addressed the role of adjuvant chemotherapy added to endocrine therapy in patients with a recurrence score of 11 to 25, the authors reported patients’ clinical risk classes based on a modified version of the Adjuvant! criteria. These criteria were the same as those used in another large, prospective randomized trial,³ the Microarray in Node-Negative and 1 to 3 Positive Lymph Node Disease May Avoid Chemotherapy trial (MINDACT), which tested a 70-gene risk signature. In this study, the criteria were used with the 70-gene signature to prospectively define groups based on concordant or discordant clinical and genomic risk assessment. One interesting finding of the TAILORx trial¹ is that by using a very simple clinical risk classifier based essentially on tumor grade and diameter, 91% of the patients at low clinical risk were confirmed as having a recurrence score of 25 or lower. The remaining 9% of patients at low clinical risk were classified as high genomic risk (recurrence score, 26-100) and constituted 43% of the patients addressed in the current publication.¹ The corresponding figure in MINDACT³ was 18%, but a subset analysis of patients in this group revealed no significant benefit from chemotherapy. Based on this latter finding, the 2017 American Society of Clinical Oncology guidelines recommended use of MammaPrint (Agendia), which is the genomic test used in MINDACT, only in patients classified as at high clinical risk.⁴ However, the abstract⁵ of an unpublished analysis suggested poor prognosis and potential chemotherapy benefit in patients with small tumors at low clinical risk/high genomic risk. We believe that it would be of great importance if the authors of the TAILORx trial¹ could provide information on the influence of concordant or discordant clinical risk in patients with a recurrence score of 26 to 100.