CMS MolDX: EndoPredict Breast Cancer Gene Expression Test Form

Summary Of Evidence

In 2016, approximately 247,000 cases of breast cancer were diagnosed in the United States.¹ Approximately 75% of early-stage breast cancers are estrogen receptor (ER)-positive and HER2-negative, leading to treatment with adjuvant endocrine therapy (e.g., tamoxifen or aromatase inhibitors) that significantly improves prognosis.^2,3 Determining which patients with ER+/Her2- breast cancer will have a low enough risk of DR after 5 years of endocrine therapy to forgo adjuvant chemotherapy is a priority for physicians who manage these patients.

A 2012 meta-analysis by the Early Breast Cancer Trialist’s Collaborative Group demonstrated that all clinical risk groups of patients with early breast cancer experience a ~30% benefit from chemotherapy, in terms of decrease in DR rate.⁴ Therefore, patients with a low underlying risk of DR will have a lower absolute benefit from chemotherapy, compared to patients with a high underlying risk of DR. For each patient, the expected absolute benefit of chemotherapy needs to be weighed against the 2-3% chance of treatment-related toxicity and long-term side effects.

Tumor size, grade, and nodal status are currently used for assessment of a patient’s distant recurrence risk to make decisions about the addition of chemotherapy to endocrine therapy. However, molecular tests have been shown to improve prognostic accuracy compared to standard clinical features and have become increasingly important for patients with ER+/Her2- breast cancer. These assays have become standard of care in the treatment of early stage breast cancer, to identify patients who have a low risk of DR such that chemotherapy would not provide an overall benefit, to directly predict chemotherapy benefit, and to help curtail costly overtreatment.

In determining the cutoff to identify low risk patients, breast cancer prognostic tests commonly use a threshold of a 10% risk of DR at 10 years. Patients with a risk under 10% are categorized as low risk. This 10% cutoff is a well-accepted standard used by many currently available breast prognostic tests and accepted by the American Society of Clinical Oncology (ASCO).⁵ In addition, while hormone receptor-positive early-stage breast cancer patients have a favorable prognosis overall, there is an ongoing risk of distant recurrence (DR) and death. ^15,16 Women with node-negative disease who received 5 years of endocrine therapy have a 13% risk of recurrence 5-20 years after diagnosis.¹⁵ Studies have offered conflicting results on the efficacy of continuing endocrine therapy beyond five years. A series of randomized trials that compared the benefit of extended endocrine therapy (EET; 10 years total) versus the 5-year standard of care demonstrated that EET led to a modest clinical benefit in disease free survival, although many were confounded by the inclusion of patients who received adjuvant chemotherapy or by unknown hormone receptor status. ^17-23

Recently presented results from the large, randomized, placebo-controlled NSABP-B42 trial showed that survival was not significantly improved with EET. ²⁴ Finally, EET is associated with a risk of serious, potentially life-threatening toxicities and side effects that may impact quality of life, which must be balanced with the potential for increased survival. ^25-27

Patients with hormone receptor-positive, HER2-negative breast cancer and their physicians face a challenging second treatment decision regarding continuation of endocrine therapy beyond 5 years. Current professional society guidelines state that women may consider EET beyond 5 years.

Because of the modest benefit possible, ASCO Guidelines suggest using clinical criteria, including assessing tolerability, for determining which patients should be given EET in the low-risk population of node-negative and limited node-positive patients. ^28-29

 Test Description and Intended Use

EndoPredict was developed in a training cohort of 964ER-positive and HER2-negative breast tumor samples.¹⁰ The EndoPredict test is composed of a 12 gene molecular score as well as clinicopathological features (tumor size and nodal status). Eight genes were selected as relevant for therapeutic decision making; they include proliferation-associated genes as well as estrogen receptor signaling-associated genes. The signature also includes three RNA normalization genes and one DNA contamination control gene. The 12 gene molecular score is calculated by the weighted expression of the 8 target genes, as normalized by the 3 normalization genes, as measured in formalin-fixed paraffin-embedded (FFPE) breast tumor tissue. The 12 gene molecular score is then combined with clinicopathologic features including tumor size and lymph node status to produce the EndoPredict (EPclin) score. Patients with an EPclin score of ≤3.3 are classified as low risk and those with an EPClin score >3.3 are classified as high risk. In addition to a clear bimodal result (low- or high-risk), the test report includes the the patient’s 10-year risk of distant recurrence and the likelihood of distant recurrence 5-15 years after diagnosis.

EndoPredict is intended for use in FFPE breast tumor tissue from postmenopausal women diagnosed with early-stage (TNM stage T1-3, N0-1) ER-positive, Her2-negative breast cancer, who are either lymph node-negative or who have 1-3 positive nodes, and for whom treatment with adjuvant endocrine therapy (e.g., tamoxifen or aromatase inhibitors) is being considered. The test is used by physicians in the management of early-stage breast cancer by identifying those patients with a low-risk EPClin score, for whom the absolute benefit of adjuvant chemotherapy is unlikely to outweigh the risks.

Analytical Validation

This assays analytical validation is consistent with industry standards and MolDX criteria. 

Clinical Validation

The prognostic ability of EndoPredict has been validated by prospectively designed-retrospective studies in three different cohorts from phase III trials [The Austrian Breast & Colorectal Cancer Study Group (ABCSG)-6 and-8, and Arimidex, Tamoxifen, Alone or in Combination (ATAC)] involving more than 2,600 patients,^7,10 satisfying a 1B level of evidence according to the classification for prognostic biomarkers proposed by Simon et al.¹¹ These studies collectively demonstrate the ability of EndoPredict to predict the primary endpoint of distant metastases in both early and late time periods, to accurately classify patients into a low or high risk group, and to identify a low risk group with excellent 10 year outcomes after treatment with 5 years of endocrine therapy only.^7,10,12-13 EndoPredict was determined to provide independent prognostic information compared to clinical and pathologic features alone.⁹

Filipits et al. (2011) described the initial development and clinical validation of the 12 gene molecular score and EPclin scores.¹⁰ The 12 gene molecular score and EPclin scores were developed in a training cohort of 964 ER-positive, HER2-negative tumors from both node-positive and -negative patients treated with adjuvant endocrine therapy only. The design and calculation of the 12 gene molecular score and EPclin scores was prespecified. The threshold for EPclin to discriminate patients into low and high risk of distant recurrence was pre-defined at 3.3 which corresponded to a 10% DR risk at 10 years. Scores were then independently validated on 1702 patients from two large randomized phase III trials (ABCSG-6 and -8 trials), with ER+, HER2- breast cancer who received endocrine therapy only for 5 years and included both node-positive and node-negative patients. In Kaplan-Meier analysis, EPclin low risk patients had a 4% distant recurrence rate (both ABCSG cohorts), while EPclin high risk patients had a 28% (ABCSG-6) or 22% (ABCSG-8) rate of distant recurrence (HR= 7.97 (ABCSG-6) or 4.27 (ABCSG-8), both P<0.001), demonstrating the ability of EPclin to accurately classify patients into low and high risk groups.

Buus et al. (2016) published a clinical validation study of EndoPredict in a third cohort of patients using the same pre-defined EPclin threshold of 3.3 to discriminate low risk from high risk patients (as described in Filipits et al.).⁷ The study included 928 women from the ATAC trial who had ER-positive, HER2-negative breast cancer, both node-negative and node-positive, chemotherapy-naïve, treated with endocrine therapy. The EPclin score classified high and low risk patients with both node negative disease (5.9% DR for low risk vs. 20.0% DR for high risk, HR=3.9 (2.33-6.52) p<0.001) and node positive disease (5.0% vs. 36.9%, HR=9.49 (2.33-38.75) p<0.001). Overall, the EPclin score classified 58.8% of patients as low risk in this study. The authors concluded that the superior performance of EPclin compared with another widely used breast cancer prognostic test was partly due to the inclusion of clinical variables (nodal status and tumor size) in the EPclin score, but also due to an improved molecular signature that better predicts late events in years 5-10 (x2=59.3, p<0.001).

Dubsky et al. conducted two secondary analyses to evaluate the performance of EndoPredict in different subsets of the 1702 ER-positive, HER2-negative breast cancer patients from the ABCSG-6 and ABCSG-8 phase III trials.^12-13 The EPclin score improved the classification of breast cancer compared to the prognosis assigned by standard guidelines that use clinical and pathological features (i.e., NCCN, German S3 and St. Gallen).¹² Using clinical guidelines, 6-19% of patients were classified as low risk; however, EPclin classified 63% of all patients as low risk. Furthermore, 58-61% of patients classified as high- or intermediate-risk according to clinical guidelines were reclassified by EPclin as low risk; this group of patients had a 5% rate of distant metastasis at 10 years, confirming the accuracy of EndoPredict’s assessment of risk. Finally, EndoPredict was compared to standard clinical parameters for predicting distant metastases in both early (0-5y) and late (5-10y) time periods, which is an important consideration based on the fact that 50% of recurrences in women with ER+, Her2- breast cancer occur after 5 years.¹³ The EPclin low risk group had significantly improved clinical outcomes compared to the EPclin high risk group in both the early (0-5y; HR 4.82, p<.001)) and late (5-10y; HR 6.25, p<.001) timeframes. The EPclin score identified 64% of patients at risk after 5 years into a low-risk subgroup with an absolute 1.8% risk of late distant metastasis at 10 years.

A more recent, longer term study of the ABCSG6/8 cohorts evaluated the prognostic value of EndoPredict in women who were DR-free 5 years after diagnosis.³⁰ Filipits et al suggests that women with low EndoPredict scores had reduced rates of DR, and the hazard ratio remained statistically significant even in the longer follow-up period from 5 to 15 years compared to those with high EndoPredict scores. The overall data demonstrated that EndoPredict could identify a difference in recurrence rates between EPclin high and low risk patients. A detailed breakdown of the findings (fig.2 of the study) identifies specific information related to the patients within the intended use of EndoPredict. A subset of the data (including the most robust data) limited to 5-10 years after ET demonstrates that EPclin –low risk patients are unlikely to have recurrence. In the subset of node negative women who did not have a recurrence by 5 years, the DR for low risk EndoPredict was 2.1% at 10 years, with a narrow confidence interval (CI) that does not exceed 4%. In the subset of node positive women who did not have a recurrence by 5 years, the DR for low risk EndoPredict was 1.7% at 10 years, with a narrow confidence interval that does not exceed 5%. However, these data omit later time points (years 10-15 of the study).

Clinical Utility

A retrospective analysis of the prospective use and impact of the EndoPredict assay in a clinical setting was published in 2013 by Müller et al.¹⁴ Samples from 167 women with primary invasive ER+, HER2- breast cancer were analyzed by EndoPredict performed at the molecular pathology laboratory (Institute of Pathology) at the Charite´ University Hospital in Berlin. The impact of EndoPredict on changes in therapy decisions was evaluated for 130 of the 167 patients, of whom 47.7% had low EPclin scores and 52.3% had high EPclin scores. There was a change in pre-test versus post-test therapy for 37.7% of patients with most of the changes due to reduction from combination therapy (chemotherapy plus endocrine) to endocrine therapy alone. Before the EndoPredict assay, 47 (36.2%) of patients had been scheduled for endocrine therapy alone and 83 (63.8%) had been scheduled for a combination of endocrine therapy and chemotherapy. After the EndoPredict results were available, the number of patients with endocrine therapy alone increased to 67 (51.5%). Changes in therapy were directionally aligned with the EPclin result as low or high risk.

A 2011 meta-analysis of Oncotype DX decision data determined that the test changed adjuvant chemotherapy treatment decisions in 33.4 % of patients (8 studies, 1,437 patients).¹⁶ The observed change in treatment recommendations for 37.7% and 35.8% of patients after use of EndoPredict is therefore consistent with the expected clinical utility of similar established tests.

Summary of Analytical and Clinical Performance

General

Analytical Performance

The prognostic ability of EndoPredict for 10 year distant recurrence has been validated in three different cohorts from phase III trials (ABCSG-6 and-8, and ATAC), involving more than 2,600 patients with ER-positive, HER2-negative, node-positive and –negative invasive breast cancer treated with endocrine therapy only.

*Subset analysis from Filipits et al. 2011

A longer term study of the ABCSG6/8 cohorts evaluated the prognostic value of EndoPredict in women who were DR-free 5 years after diagnosis.

Table 1: DRFR (Distant Recurrence-Free Rate) at 10 years for patients who were distant recurrence free after 5 years, according to nodal status. Filipits et al. 2019, Supplementary Table S5.

Analysis of Evidence

The EndoPredict assay is reasonable and necessary to assist physicians in the management of early stage breast cancer by identifying those patients with a sufficiently low risk of 10-year distant recurrence who may safely forgo chemotherapy. While EndoPredict may help assess risk of late recurrence after ET is complete, it is not clear at this time if this test is useful for identifying which patients should receive EET.

This contractor believes the clinical utility of this test for determining which patients would benefit from EET is not clear based on these data. This test is indicated in node-negative and 1-3 node-positive patients. First, the data presented in Filipits et al. for 1-3 node-positive patients was not significant for the period studied (5-15 years), and the difference in DRFR was 84% vs 87% between the EPclin high and low groupWhile the difference between EPclin high and low groups are significant in node–negative patients for the duration of the study (96% vs 85% DRFR), a subset analysis (described above) of the best supported data (in the supplemental data, Table S5, shown above) shows that 5-10 years after ET therapy there is little difference in the performance of these patients; neither the EPclin high or low group would be defined as “high risk” (>10% risk of recurrence), and the difference in DRFR between these two groups is 97.9% vs. 94.1%. Based on this study, this contractor does not believe there is sufficient evidence that this test can clearly delineate patients who can benefit from EET (whose benefits are already known to be ‘modest’). Given that EET carries significant risk for adverse events, it is not clear how much, if any, weight an oncologist should place on EndoPredict results for determining the use of EET in patients who have not recurred after initial ET.

The assay is covered for women with T1-3, N0-1 breast cancer when the following criteria are met:

• Patient is post-menopausal, and
• Pathology (excisional or biopsy) reveals invasive carcinoma of the breast that is ER-positive, Her2-negative, and
• Patient is either lymph node-negative or has 1-3 positive lymph nodes, and
• Patient has no evidence of distant metastasis, and
• Test result will be used to determine treatment choice between endocrine therapy alone vs. endocrine therapy plus chemotherapy.

Note: The EndoPredict test should not be ordered if a physician does not intend to act upon the test result.

This Medicare contractor will provide limited coverage for the EndoPredict^® breast cancer gene expression test (Myriad Genetic Laboratories Inc., Salt Lake City, UT) for the management of post-menopausal women diagnosed with early-stage (TNM stage T1-3, N0-1) estrogen-receptor (ER) positive, Her2-negative breast cancer, who are either lymph node-negative or who have 1-3 positive nodes, and for whom treatment with adjuvant endocrine therapy (e.g., tamoxifen or aromatase inhibitors) is being considered. The test is used by physicians in the management of these patients by identifying those who have sufficiently low risk of distant recurrence (DR) at 10 years and may safely forego chemotherapy.