Algorithms for prediction of the Oncotype DX recurrence score using clinicopathologic data: a review and comparison using an independent dataset.

Published

Journal Article

PURPOSE: Given the potential savings in cost and resource utilization, several algorithms have been proposed to predict Oncotype DX recurrence score (ODX RS) using commonly acquired histopathologic variables. Although it is promising, additional independent validation of these surrogate markers is needed prior to guide the patient management. METHODS: In this retrospective study, we analyzed 305 patients with invasive breast cancer at our institution who had ODX RS available. We selected five equations that provide a surrogate measure of ODX as previously published by Klein et al. (Magee equations 1-3), Gage et al., and Tang et al. All equations used estrogen receptor status and progesterone receptor status along with different combinations of grade, proliferation indices (Ki-67, mitotic rate), HER2 status, and tumor size. RESULTS: Of all surrogate scores tested, the Magee equation 2 provided the highest correlation with ODX both with regard to raw score (Pearson's correlation coefficient = 0.66 95% CI 0.59-0.72) and categorical correlation (Cohen's kappa = 0.43, 95% CI 0.33-0.53). Although Magee equation 2 provided a way to reliably identify high-risk disease by assigning 95% of the patients with high ODX RS to either the intermediate- or high-risk group, it was unable to reliably identify the potential for patients to have intermediate- or high-risk disease by ODX (66% of such patients identified). CONCLUSIONS: Although commonly available surrogates for ODX appear to predict high-risk ODX RS, they are unable to reliably rule out the presence of patients with intermediate-risk disease by ODX. Given the potential benefit of adjuvant chemotherapy in women with intermediate-risk disease by ODX, current surrogates are unable to safely substitute for ODX. Characterizing the true recurrence risk in patients with intermediate-risk disease by ODX is critical to the clinical adoption of current surrogate markers and is an area of ongoing clinical trials.

Full Text

Duke Authors

Cited Authors

  • Harowicz, MR; Robinson, TJ; Dinan, MA; Saha, A; Marks, JR; Marcom, PK; Mazurowski, MA

Published Date

  • February 2017

Published In

Volume / Issue

  • 162 / 1

Start / End Page

  • 1 - 10

PubMed ID

  • 28064383

Pubmed Central ID

  • 28064383

Electronic International Standard Serial Number (EISSN)

  • 1573-7217

Digital Object Identifier (DOI)

  • 10.1007/s10549-016-4093-4

Language

  • eng

Conference Location

  • Netherlands