Genetic test predicts 10-year risk of DCIS recurrence

Key clinical point: A genetic assay can accurately predict the 10-year risk of local recurrence in women with ductal carcinoma in situ.

Major finding: Women with the highest risk as graded by Oncotype DCIS Score were 68% more likely to develop recurrence than were those in the lowest risk group.

Data source: The retrospective validation study comprised 257 women with DCIS who were treated with breast-conserving surgery alone.

Disclosures:Funding was provided by the Canadian Cancer Society Research Institute and a research grant from Genomic Health, which makes the test. Dr. Rakovitch had no conflicts of interest.




SAN ANTONIO – A genetic assay specific for ductal cancer in situ accurately predicted 10-year recurrence risk in women who underwent breast-conserving surgery as treatment for the disease.

Women whom the test identified as having a high recurrence risk were 68% more likely to have an ipsilateral recurrence of either ductal cancer in situ (DCIS) or invasive breast cancer, Dr. Eileen Rakovitch said at the San Antonio Breast Cancer Symposium. These results suggest that the Oncotype DX DCIS Score could be used as a risk stratification biomarker, said Dr. Rakovitch, a radiation oncologist at Sunnybrook Health Sciences Centre, Toronto.

“We hope this will enable both clinicians and patients to understand their individual risk of progression, and to weigh the risks and benefits of available treatment. This might reduce the problems of overtreating women at low risk and undertreating women at high risk of recurrence.”

The Oncotype DX DCIS is a multigene expression assay that identifies 12 of the 21 genes included in the Oncotype DX Recurrence Score; seven of these are cancer specific and five are reference genes. It stratifies women into three 10-year risk groups according to increasing score (1-100): low-risk (less than 39), intermediate risk (39-54), and high-risk (55 and higher).

It was previously validated in a retrospective analysis of 327 selected cases, all of which were estrogen receptor–positive. In that study, all women had DCIS treatment by breast-conserving surgery alone. Tumors had to be 2.5 cm or less with nuclear grades 1 or 2; or at least 1 cm with a nuclear grade 3. Negative surgical margins of at least 3 cm were also required. In that population, after adjustment for tamoxifen use, the intermediate- and high-risk groups faced more than a doubling of the 10-year ipsilateral local recurrence risk (HR 2.31) compared to the low-risk group.

The study Dr. Rakovitch reported investigated the score’s validity in a cohort with mixed hormone receptor status, and in the subgroup of women with only ER-positive tumors.

The cohort was drawn from several Canadian health care databases that linked diagnosis, treatment, and outcome. Cases were verified by chart reviews of each stage of care. The women were treated from 1994 to 2003 and followed for a mean of 10 years. The final diagnosis was DCIS, and they were treated only with breast-conserving surgery.

Most (81%) were at least 50 years old. About half had an intermediate nuclear grade; it was low in 10% and high in the remainder. Necrosis was present in 61% of the tumors; 63% were solid and 20%, multifocal. Tumor size data were missing in about half; of those with data, 25% were 10 mm or smaller and the remainder larger than 10 mm.

Over the follow-up period, there were 100 recurrences (44 DCIS and 56 invasive cancers). The overall 10-year recurrence risk was 10%. The groups diverged significantly in risk. It was 28% in the high-risk and 33% in the intermediate-risk groups – not significantly different. But both those were significantly higher than the 13% risk in the low-risk group. These results were remarkably similar to those of the earlier validity study in women with ER-positive tumors.

Risk was further stratified by whether the recurrence was local or invasive and the results were similarly significantly different. For invasive local recurrence, the risks were 15.5%, 21%, and 8% for the high-, intermediate-, and low-risk groups. For local recurrence, the risks were 13.7%, 14% and 5.4%, respectively.

A multivariate analysis controlled for age at diagnosis, tumor subtype, and multifocality – none of these were significantly related to the risk of recurrence.

Funding for the study was provided by grants from the Canadian Cancer Society Research Institute and a research grant from Genomic Health, which makes the test. Dr. Rakovitch had no conflicts of interest.

On Twitter @alz_gal

Next Article:

   Comments ()