The Impact of Genomic Testing on the Recommendation for Radiation Therapy in Patients With Ductal Carcinoma In Situ: A Prospective Clinical Utility Assessment of the 12-Gene DCIS Score TM Result MICHAEL ALVARADO, MD, 1 DENNIS L. CARTER, MD, 2 J. MICHAEL GUENTHER, MD, 3 JAMES HAGANS, MD, 4 RACHEL Y. LEI, BS, MA, 2 CHARLES E. LEONARD, MD, 5 JENNIFER MANDERS, MD, 6 AMY P. SING, MD, 7 * MICHAEL S. BRODER, MD, 8 DASHA CHEREPANOV, PhD, 8 * EUNICE CHANG, PhD, 8 MARIANNE EAGAN, MSc, 8 WENDY HSIAO, BS, 9 AND MICHAEL J. SCHULTZ, MD 10 1 University of California, San Francisco, California 2 Rocky Mountain Cancer Centers, Aurora, Colorado 3 St. Elizabeth Healthcare, Edgewood, Kentucky 4 The Surgical Center of Central Arkansas, Little Rock, Arkansas 5 Rocky Mountain Cancer Centers, Littleton, Colorado 6 The Christ Hospital, Cincinnati, Ohio 7 Genomic Health, Inc., Redwood City, California 8 Partnership for Health Analytic Research, LLC, Beverly Hills, California 9 University of Southern California, Los Angeles, California 10 University of Maryland St. Joseph Medical Center, Towson, Maryland Background and Objectives: Twenty percent of breast cancers are ductal carcinoma in situ (DCIS), with 15–60% having a local recurrence (LR) after surgery. Radiotherapy reduces LR by 50% but has not impacted survival. The validated Oncotype DX 1 12-gene assay (DCIS Score) provides individualized 10-year LR estimates. This is the first study to assess whether DCIS Score impacts physicians’ recommendations for radiation. Methods: Ten sites enrolled women (9/2012–2/2014) with DCIS eligible for breast-conserving therapy, excluding patients with invasive carcinoma and planned mastectomy. Prospective data collected included clinicopathologic factors, DCIS Score assay, and treatment recommendation before and after the assay result was known. Results: In 115 patients (median age: 61 years; 74.8% postmenopausal), median DCIS size was 8 mm; 20% were nuclear grade 1, 46.1% grade 2; 64.4% reported necrosis. 86.1% were ERþ, 79.1% PRþ (immunohistochemistry assay). Median DCIS Score: 29 (range: 0–85). Pre-assay, 73% (95%CI: 64.0–80.9%) had radiotherapy recommendations vs. 59.1% (95%CI: 49.6–68.2%) post-assay (P¼ 0.008). Physicians rated DCIS Score as the most impactful factor in planning treatment. Conclusions: The radiotherapy recommendation changed from pre-assay to post-assay 31.3% (95%CI: 23.0–40.6%) of the time—a clinically significant change. This study supports the clinical utility of the DCIS Score and indicates that the test provides additional, individualized information on LR risk. J. Surg. Oncol. 2015;111:935–940. ß 2015 Wiley Periodicals, Inc. KEY WORDS: ductal carcinoma in situ; breast cancer; clinical utility; genomics; recurrence risk; adjuvant radiotherapy INTRODUCTION Noninvasive breast cancers include both lobular carcinoma in situ (LCIS) and ductal carcinoma in situ (DCIS), referred to as stage 0 breast cancers [1]. The incidence of DCIS has increased over the past 30 years with routine mammography screening [2]. The rate of invasive breast cancer and related death has not increased in incidence over the same time period [2]. DCIS now comprises 20% of all breast cancers diagnosed by mammography in the United States [3,4]. Treatment of DCIS aims to prevent the occurrence of invasive disease and any local recurrence (LR) of DCIS. In most patients, breast- conserving therapy is an appropriate treatment option [1]. LR rates with surgery alone range from 15% to 60%; half of which are invasive [1]. Adding whole breast irradiation (XRT) following excision reduces the relative LR risk by approximately 50% but has not been shown to impact survival [1,5]. Since studies to date have not identified any sub- groups that did not receive some benefit from XRT [5], many patients— with low likelihood of recurrence—continue to be treated with XRT. LR estimates have historically been based on clinicopathologic factors. For instance, young age or higher tumor grade are considered to be associated with higher LR risk, and generally treated more aggressively. However, clinicopathologic factors can only serve as a proxy for the biologic aggressiveness of the disease; an optimal method for determining whether XRT is necessary in an individual patient has not been established. Grant sponsor: Genomic Health, Inc.. Presented at: the American Society of Clinical Oncology 50th Annual Meeting in Chicago, IL, May 30–June 3, 2014, and at the American Society for Therapeutic Radiology and Oncology 56th annual meeting in San Francisco, CA, September 14–16, 2014. *Correspondence to: Dasha Cherepanov, PhD, Partnership for Health Analytic Research, LLC, Beverly Hills, CA 90212. Fax: 310 858-9552. E-mail: dasha@PHARLLC.com; Amy P. Sing, MD, Genomic Health, Inc., Redwood City, CA. E-mail: asing@genomichealth.com Received 25 November 2014; Accepted 18 April 2015 DOI 10.1002/jso.23933 Published online 28 May 2015 in Wiley Online Library (wileyonlinelibrary.com). Journal of Surgical Oncology 2015;111:935–940 ß 2015 Wiley Periodicals, Inc.