Population-based study of hormone receptor status comparing OncotypeDX with standard immunohistochemistry

Objective

Accurate measurement of hormone receptor (HR) status is important as hormonal therapy reduces risk of recurrence by >50% in breast cancer patients with hormone-sensitive tumors. We compared estrogen receptor (ER) and progesterone receptor (PR) results obtained with Oncotype DX, a validated 21-gene recurrence score assay that utilizes quantitative RT-PCR, with results from established immunohistochemistry (IHC).

Methods

Breast cancer specimens from the Kaiser Oncotype DX study were evaluated by IHC for ER (SP1) and PR (636) using ≥ 1% staining for positivity. ER and PR status was also determined using Oncotype DX (RT-PCR) with pre-defined positivity cutoff values of 6.5 units and 5.5 units, respectively (each unit = twofold change in expression). HR positivity was defined as ER⁺ and/or PR⁺.

Results

The overall concordance in 607 evaluable patients (95% CI) between IHC and Oncotype DX was 96% for ER, 90% for PR, and 95% for HR (all P < 0.0001). The kappa (95% CI) between IHC and Oncotype DX was 83% for ER, 76% for PR, and 81% for HR (all P < 0.0001). For discordant pairs that were within cutoff limits, there were 20 that were IHC ER^- but ER⁺ by Oncotype DX, 17 (85%) of which were within cutoff limits; another 38 were IHC PR^- but PR⁺ by Oncotype DX, 23 (61%) of which were within cutoff limits; seven pairs were IHC ER⁺ but ER^- by Oncotype DX, six (86%) of which were within cutoff limits; and 22 pairs were IHC PR⁺ but Oncotype DX PR^-, 19 (86%) of which were within cutoff limits.

Conclusion

There was a high degree of concordance between Oncotype DX and standard IHC methods for ER, PR and HR status, indicating Oncotype DX may be a valid alternative to IHC. The notable incidence of IHC HR-negative patients who were Oncotype DX HR-positive, which could affect treatment decisions, deserves further study.

Disclosure

FLB is an employee and shareholder of Genomic Health, Inc.