40 Characterization of Recurrence Risk After Lumpectomy and Radiotherapy in HER2-Positive Ductal Carcinoma in Situ of the Breast Using a 7-Gene Predictive Biosignature: Implications for the NSABP-B43 Trial Results

Background

HER2-positive versus HER2-negative ductal carcinoma in situ (DCIS) of the breast has been associated with an increased risk of local recurrence after breast-conserving surgery (BCS) and radiotherapy (RT). NASBP-B43 was designed to determine if 2 doses of trastuzumab would improve local control after BCS plus RT in women with HER2 (3+) DCIS. The trial demonstrated a statistically insignificant advantage to the drug in reducing ipsilateral breast recurrence (IBR). The predictive 7-gene DCIS biosignature, DCISionRT with Residual Risk Subtype (RRt) by PreludeDxTM, has been shown to classify DCIS patients into 2 distinct groups of patients with substantially different rates of IBR following BCS plus RT. Based upon these differences in outcome, we assessed the IBR rate in patients with HER2 (+) DCIS treated with BCS and RT who were or were not in the residual risk Subtype group defined using DCISionRT, while accounting for the varying clinicopathologic profiles of the patients.

Materials and Methods

DCISionRT with the integrated residual risk biosignature (DCISionRT with RRt) was evaluated on a subset of 178 women with HER2(3+) DCIS who were treated with BCS and RT in a multinational cohort of 926 patients from the United States of America, Sweden, and Australia who were used in the validation studies for DCISionRT. Central pathology review and biosignature testing were performed at a CLIA-certified lab. HER2(+) DCIS was defined as patients with a HER2 (3+) immunohistochemistry ≥10% (ASCO/CAP). The biosignature identified patients either classified into the Residual Risk group or not. The IBR rate was calculated for the overall group of HER2(3+) patients and those in the Residual Risk group. Individual patient outcome and biosignature results were analyzed using Kaplan-Meier and Cox proportional hazard analyses.

Results

The biosignature classified 113 of 178 HER2 (3+) women (63%) into the residual risk group (DS>2.8 with RRt). Grade 3 was more common in the residual risk group than the not residual risk group (87% vs. 63%). Patients in the residual risk group had a significantly higher 10-year total IBR rate of 16.2% (95% CI, 9.7%-26.5%) versus 1.6% (95% CI, 0.2%-10.9%; P = .012) for all other HER2 (3+) patients. In univariate analysis, younger patients tended to have higher IBR rates after BCS plus RT, but only residual risk was significantly associated with IBR rates after BCS plus RT. Multivariable analysis demonstrated that the residual risk group was eight times more likely to recur after BCS and RT, while clinicopathologic factors (age, grade, tumor size) were not associated with a higher IBR rate.

Conclusions

The new, integrated biosignature was predictive for 10-year IBR risk after BCS plus RT in women with HER2 (3+) DCIS. Approximately 40% of patients with HER2 (3+) DCIS would be expected to achieve low rates of recurrence with BCS and RT, while about 60% of these women (classified as residual risk) would have higher recurrence rates. Collectively, this suggests that there is a subpopulation of HER2 (3+) DCIS that may benefit from further therapy, such as HER2- directed therapies.

AFFILIATIONS:

Frank A. Vicini,¹ Chirag Shah,² Rachel Rabinovitch,³ Pat Whitworth,⁴ Julie A. Margenthaler,⁵ Brian Czerniecki,⁶ David J. Dabbs,⁷ Sheila Weinmann,⁸ Michael Leo,⁸ G. Bruce Mann,⁹ Fredrik Wärnberg,¹⁰ Jess Savala,⁷ Steven C. Shivers,⁷ Karuna Mittal,⁷ Troy Bremer⁷

¹GenesisCare, Farmington Hills, MI.

²Cleveland Clinic Taussig Cancer Institute, Cleveland, OH.

³University of Colorado Cancer Center, Aurora, CO.

⁴Nashville Breast Center, Nashville, TN.

⁵Washington University School of Medicine, St Louis, MO.

⁶Moffitt Cancer Center, Tampa, FL.

⁷PreludeDx, Laguna Hills, CA.

⁸Kaiser Permanente Center for Health Research, Portland, OR.

⁹University of Melbourne, Royal Women’s Hospital, Parkville, Australia.

¹⁰University of Gothenburg, Gothenburg, Sweden.