DCIS Test Identifies Patients With Low Recurrence Risk

Neil Osterweil

December 16, 2014

SAN ANTONIO — A clinical validation study in a general population of patients who had undergone breast-conserving surgery for ductal carcinoma in situ (DCIS) has shown that the 12-gene Oncotype DX DCIS Score (Genomic Health) is good at predicting whether patients are at risk for a recurrence of DCIS or invasive cancer in the treated breast.

Women classified by the test as being at low risk were at significantly lower risk of developing recurrent disease than women who had an intermediate- or high-risk DCIS score.

Because DCIS is a precancerous lesion, there is concern about the overtreatment of patients, some of whom will never go on to develop invasive breast cancer.

The study results indicate that the test could help prevent the overtreatment of these patients, allowing some to skip radiation after surgery, said Kent Osborne, MD, director of the Dan L. Duncan Cancer Center and the Lester & Sue Smith Breast Center at the Baylor College of Medicine in Houston. He moderated a press conference here at the San Antonio Breast Cancer Symposium 2014 during which the results were highlighted.

This study "will provide us with more information to allow us to not give radiation to a certain subset of our DCIS patients," he said. Dr Osborne told Medscape Medical News that he uses the genetic panel as one of several tools to help predict an individual patient's risk, along with patient age, surgical margin width, and tumor size.

However, another expert present at the press conference, John Robertson, MD, professor of surgery at the University of Nottingham, Royal Derby Hospital, United Kingdom, said that for any treatment to be withheld from a patient population, the outcomes should be studied first in a randomized clinical trial.

Eileen Rakovitch, MD, associate professor and radiation oncologist at the Sunnybrook Health Sciences Centre in Toronto, who presented the clinical validation study at the meeting, emphasized that these results show that the test predicts the risk for recurrence. This study did not look at the effect of subsequent treatment, but her team plans to present analyses of patients treated with both surgery and radiation at a future meeting.

Risk for Recurrence Predicted

The DCIS score is based on a panel of 12 genes (7 DCIS-related genes and 5 reference genes), and is derived from the 21-gene Oncotype DX Recurrence Score, which is used to establish the management of patients with early-stage node-negative invasive breast cancer.

The 12-gene DCIS score was validated in 327 women enrolled in the Eastern Cooperative Oncology Group E5194 trial.

The E5194 study looked at a carefully selected patient sample; Dr Rakovitch and her colleagues wanted to see whether the DCIS score was similarly predictive in a general population.

In their retrospective study, the team identified 3335 women diagnosed with DCIS from 1994 to 2003, and tested tumor samples from 718 women treated with breast-conserving surgery alone and 846 treated with surgery and radiation. In all, 1260 patients had clear surgical margins (defined as no ink on the tumor).

Of the 718 patients treated with surgery only, 571 had negative resection margins; it was this population that was evaluated. The investigators tested the tumor samples with the genetic panel and compared the results with reported outcomes.

After a median follow-up of 9.6 years, 100 patients had a local recurrence, defined as DCIS or invasive breast cancer in the same breast 6 months or more after a diagnosis of DCIS. Of the 100 recurrences, 44 had DCIS and 57 had invasive disease (one patient had DCIS that progressed to invasive disease).

The overall 10-year Kaplan–Meier risk for local recurrence was 19.2%.

On univariate analysis, the risk for local recurrence was more than 2 times higher in women with estrogen-receptor-positive DCIS who had a high score than a low score (hazard ratio [HR], 2.6; < .001). The same was true when all patients were included in the analysis (HR, 2.15; P < .001).

Table 1. Kaplan–Meier 10-year Risk for Local Recurrence by DCIS Score

Risk Group DCIS Score 10-year Risk, % Log Rank P Value
Low <39 12.7 <.001
Intermediate 39–54 33.0 <.001
High >54 27.8 <.001


"We found that the DCIS score was associated both with invasive local recurrence and DCIS local recurrence, Dr Rakovitch reported.

Table 2: Risk for Local Recurrence

Local Recurrence Low Risk, % Intermediate Risk, % High Risk, % Log Rank Value
Invasive 8.0 20.9 15.5 .003
DCIS 5.4 14.1 13.7 .002


On multivariable analysis, factors significantly associated with local recurrence included every 50 points on the DCSI score (HR, 1.68; P = .02), diagnosis at age 50 or older (HR, 1.75; P = .03), tumor size (>10 mm vs ≤10 mm; HR, 2.07; P = .01), tumor subtype (solid vs cribriform; HR, 1.63; = .04), and the presence of multifocality (HR, 1.97; = .003).

This is the "first multigene biomarker assay in DCIS that can provide individualized estimates of the risk of recurrence in women who are treated with breast-conserving surgery alone," Dr Rakovitch explained.

"This can help clinicians and patients make more informed decisions about their own risks of local recurrence and better understand the potential benefits of treatment. Hopefully, this can lead to improvements in the management of DCIS by reducing overtreatment for those at low risk and reducing undertreatment for those at high risk of recurrence," she said at the press conference.

The study was supported by grants from the Canadian Cancer Society Research Institute and a research grant from Genomic Health. Dr Osborne and Dr Rakovitch have disclosed no relevant financial relationships.

San Antonio Breast Cancer Symposium (SABCS) 2014. Abstract S5-03. Presented December 12, 2014.


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.