47 A Biomarker Assay Predicts Women Diagnosed With DCIS Without Microinvasion Are at Increased Risk for Breast Cancer– Specific Death

Miami Breast Cancer Conference® Abstracts Supplement, 38th Annual Miami Breast Cancer Conference® - Abstracts, Volume 35, Issue suppl 1

Charlotta Wadsten, MD, PhD1; Rachel Rabinovitch, MD2; Frank Vicini, MD3; Chirag Shah, MD4; Steven C. Shivers, PhD5; Anna-Karin Wennstig, MD1; Malin Sund, MD1; Fredrik Warnberg, MD6; Troy Bremer, PhD5

1Umea University, Umea, Sweden.

2University of Colorado Cancer Center, Aurora, CO.

3GenesisCare, Royal Oak, MI.

4Cleveland Clinic, Cleveland, OH.

5PreludeDx, Laguna Hills, CA.

6Uppsala University Hospital, Uppsala, Sweden.


While rare, patients diagnosed with ductal carcinoma in situ (DCIS) can die from breast cancer. We investigated the association of DCISionRT test results with breast cancer mortality (BCM). DCISionRT is a validated test to assess 10-year breast cancer risk for patients with DCIS. Due to the very low incidence of death in DCIS, we used a prior nested case control study.

Materials and Methods

The case control study identified 96 women who died of breast cancer and 318 controls from a population of 6964 in Sweden diagnosed with DCIS without microinvasion (1992-2012). DCISionRT testing was performed on a subset of patients with formalin-fixed embedded tissue microarray while blinded to outcome. Conditional logistic regression was used to calculate odds ratios (ORs) for the risk of BCM accounting for clinicopathologic factors, treatment, and continuous and categorical Decision Score (DS).


DS results were available for 157 of the 414 women in the original case control study.1 Primary DCIS was treated with breast-conserving surgery (BCS) alone (34%), BCS plus radiotherapy (29%), or mastectomy (37%). Clinicopathologic factors and treatment distributions were consistent with those in the original case control study, except there were 15% fewer patients who had larger tumors (>2.5 cm). Continuous and categorial DS were independently associated with BCM in multivariate analyses, accounting for treatment differences and clinicopathology. Patients with increasing continuous DS had increasing BCM (OR, 10 per 5 DS units; P = .004), and patients with high categorical DS (>6) were at greater risk of BCM (OR, 19; P = .007). Treatment was not independently associated with decreased BCM risk in this study. Patients selected for treatment with mastectomy tended to have increased BCM compared with those who were treated with BCS. Young age (<50 years), tumor size (>1 cm), and grade 3 disease were not statistically associated with BCM.


In this case-control study, patients with higher DCISionRT scores had an increased risk of BCM. DCISionRT may help to identify patients with the potential to develop more aggressive subsequent disease that warrants more aggressive up-front treatment. Additional data are needed to validate these findings.


  1. Wadsten C, Garmo H, Fredriksson I, Sund M, Wärnberg F. Risk of death from breast cancer after treatment for ductal carcinoma in situ. Br J Surg. 2017;104(11):1506-1513. doi:10.1002/bjs.10589

“This nested-case control study indicates a tool, DCISionRT, may be used to identify patients with pure DCIS who are at higher risk of breast cancer mortality. While higher DCISionRT score is associated with higher BCM, the clinical implications of this are not clear as other clinical factors may account for higher risk (for example, DCIS >5 cm where microinvasion may have been missed by pathologist). Moreover, the rarity of BCM from pure DCIS is quite notable, making the clinical implications of these data less impactful.” -Sara Hurvitz, MD