Predicting Risk of Recurrence After Adjuvant Therapies in ER-positive Breast Cancer

In a recent study, Paik and colleagues evaluated the magnitude of benefit from adjuvant chemotherapy as a function of the Oncotype DX 21-gene Recurrence Score (RS) assay in a subset of the NSABP B20 randomized trial.⁶⁷ The biomarker study retrieved tumor specimens from 430 patients randomized to tamoxifen alone or tamoxifen plus chemotherapy and examined recurrence rates and benefit from chemotherapy by RS category. Patients who were identified as high-risk derived significantly more benefit from chemotherapy than did those in the RS intermediate- and low-risk categories. The addition of chemotherapy to tamoxifen in the high-risk patient group resulted in an improvement in freedom from distant recurrences from 60% to 88%.⁶⁸ The study suggested that the Oncotype DX assay was successful in predicting the magnitude of benefit from chemotherapy in the various subgroups of ER-positive patients. However, this analysis included some of the same patients who were included in the development and optimization of the assay and this can bias the results toward better performance.

Results from an analysis of the prognostic and predictive value of the Oncotype DX 21-gene RS assay in the SWOG 8814 (TBCI 0100) trial were recently presented and suggested that the Recurrence Score assay may be useful in determining patients with node-positive postmenopausal disease who will benefit from adjuvant anthracycline therapy added to hormonal therapy.⁶⁹ The SWOG 8814 phase 3 trial evaluated adjuvant treatment in postmenopausal, node-positive, ER-positive breast cancer patients. The study randomized 1477 patients to one of three tamoxifen regimens: (1) adjuvant treatment with 5 years of tamoxifen only (n=361), (2) CAF cyclophosphamide/doxorubicin/fluorouracil) plus concurrent tamoxifen (n=550), (3) CAF followed by tamoxifen (n=566). Study endpoints were disease-free survival and overall survival. The study demonstrated superior disease-free survival and overall survival for the group receiving CAF followed by tamoxifen.⁷⁰ A retrospective biomarker subset analysis of this study was performed by applying Oncotype DX assay to 367 patients randomized to the tamoxifen alone or the sequential tamoxifen CAF treatment arms. The 21-gene RS was prognostic for both disease-free survival and overall survival in the tamoxifen arm, demonstrating significant differences in survival outcomes among high-, intermediate-, and low-risk groups. However, the absolute recurrence rate (including contralateral breast cancer and ipsilateral locoregional recurrences) was 40% in the low-risk group. This rate is higher than most patients and physicians feel comfortable with. Also, the rate varies from the disease-free survival rate of ER-positive, node-negative patients with similar low RS. The RS assay was also predictive of therapeutic benefit of adding CAF to tamoxifen, demonstrating significant benefit in only high-risk patients. The absolute improvement due to chemotherapy was 4% in the low RS category and this was not statistically significant. However, the total number of patients in the low RS category was 146, which makes this subset analysis underpowered to rule out small but real benefit from adjuvant chemotherapy in this group. The significance of the RS assay predictive effect was retained in Cox models adjusted for covariates including age, race, tumor size, PR status, grade, and HER2 status. However, the significance was lost when the ER Allred score was added to the model.⁶⁹

Other studies examined the prognostic utility of Oncotype DX in node-negative and node-positive (1-3 nodes) patients who received adjuvant chemotherapy and hormonal therapy. Goldstein and colleagues conducted an analysis of Intergroup trial E2197 to evaluate the prognostic utility of Oncotype DX compared with Adjuvant Online (www.adjuvantonline.com).⁷¹ Archived tissue samples were retrieved from 465 patients with hormone receptor (HR)-positive disease who participated in the E2197 study. This study (N=2885) compared adjuvant doxorubicin plus cyclophosphamide (AC) with docetaxel plus cyclophosphamide (AT), followed by tamoxifen in each treatment arm. Among these patients, 366 had experienced a recurrence and 99 had no recurrence, the cases were selected by the investigators following a case-control study design. Oncotype DX RS assay was a significant predictor of recurrence (local and distant), in both node-negative (P=.0007) and node-positive (P=.0004) disease. Low RS was associated with 3% and 8% recurrence rates in the 0-1 and 2-3 positive lymph node categories at a median follow-up of 76 months. The RS provided additional prognostic information to the clinical variables evaluated by Adjuvant Online.

A limitation of the study is that all patients received chemotherapy and therefore it remains uncertain to what extent chemotherapy contributed to the good survival of the node-positive low RS group. Similarly, longer follow-up may be needed to truly compare the genomic prediction results with that of Adjuvant Online that was optimized for predicting 10-year survival outcome. This is important because different clinical and molecular factors may predict for early (5 year) and late relapse (>5 years).

In aggregate, these results support the prognostic value of Oncotype DX in ER-positive cancers. The data also suggest that nodal status caries additional prognostic value. The true risk of relapse of patients with node-positive ER-positive low RS category treated with endocrine therapy alone remains uncertain. The SWOG 8814 study suggests a high rate of breast cancer related events (including local recurrence and second primary cancers as well as distant metastasis).