Do Clinical Data Improve Risk Tests for Breast Cancer?

Kate Johnson

December 11, 2015

SAN ANTONIO ― Predicting risk for distant recurrence of breast cancer in patients receiving endocrine therapy and especially zeroing in on who might be spared chemotherapy are best done by combining both genetic and clinical predictors, according to new research.

A combination of the EndoPredict molecular score (EP) with a clinical score that includes node status and tumor size (EPclin) provided more prognostic information than the OncotypeDX risk score (RS), a molecularly based test, for estimating distant recurrence in estrogen receptor–positive HER2-negative patients in the TransATAC (TRANSlational arm of the Anastrozole or Tamoxifen Alone or Combined) study.

"The data highlight the importance of the inclusion of clinicopathologic factors...for estimates of residual risk of distant recurrence," emphasized investigator Mitch Dowsett, PhD, from the Institute for Cancer Research, London.

He spoke here at San Antonio Breast Cancer Symposium (SABCS) 2015.

"Currently, several tests involving the measurement of multiple genes are available for predicting outcome in patients with newly diagnosed breast cancer," according to an independent expert, Michael Duffy, PhD, from University College Dublin, who wrote a recent review entitled, "Biomarkers in Breast Cancer: Where Are We and Where Are we Going?" ( Adv Clin Chem. 2015;71:1-23).

But how clinically meaningful the differences are between available prognostic tests remains an open question, he told Medscape Medical News.

"It is unclear which of these tests best predicts outcome. Indeed, there may not be a 'best test,' as available data suggest that most provide essentially similar information. These [new] findings now require independent validation using a different cohort of patients."

The new analysis used likelihood ratio tests (LR-chi sq) to assess the prognostic information provided by the EP, EPclin, and RS tests in 928 chemotherapy-naive breast cancer patients (mean age, 65 years).

The EPclin score is created by combining the molecular score, based on expression of eight genes, with nodal status and tumor size.

Over a 10-year follow-up period, 59 of 680 node-negative patients and 69 of 248 node-positive patients experienced distant recurrence, reported Dr Dowsett.

Overall, EPclin provided substantially more prognostic information than RS for both node-negative and node-positive patients across three periods (0-10, 0-5, and 5-10 years' follow-up), except for node-negative patients at 0-5 years.

Looking at 10-year distant recurrence rates with respect to tertiles of risk in node-negative patients, the EPclin and RS both significantly distinguished risk groups, but EPclin was better (Chi2 for trend, 43.4, P < .001, vs Chi2 for trend, 16.9, P < .001).

The difference between the tests was most pronounced among node-positive patients, for whom the EPclin significantly distinguished risk groups (Chi2 for trend 23.0, P < .001), whereas the RS provided no significant distinction (Chi2 for trend = 1.4, P = .23).

Dr Dowsett is a consultant for Genoptix, Radius, Gtx, and Roche and is a speaker for AstraZeneca, Pfizer, and Purna. He has also received ICR Rewards to Inventors Scheme.

San Antonio Breast Cancer Symposium (SABCS) 2015: Abstract S3-01. Presented December 10, 2015.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.