A DEEPER DIVE INTO DUCTAL CARCINOMA IN SITU: CLINICAL AND PATHOLOGIC CONSIDERATIONS IN 2015

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 A DEEPER DIVE INTO DUCTAL CARCINOMA IN SITU: CLINICAL AND PATHOLOGIC CONSIDERATIONS IN 2015 Monograph based on a satellite symposium held in conjunction with the 16 th Annual Meeting of the American Society of Breast Surgeons (ASBS) in Orlando, Florida on Thursday, April 30, 2015 The Complexity of Ductal Carcinoma In Situ Monica Morrow, MD, FACS Memorial Sloan Kettering Cancer Center, New York, New York, United States Ductal carcinoma in situ (DCIS) comprises a broad range of noninvasive breast tumors that range from relatively indolent, low-grade lesions to high-grade tumors containing areas of microinvasion. 1 Prior to the routine use of screening mammography, DCIS accounted for <1% of breast cancer diagnoses. 2 Detection of DCIS has become substantially more frequent as screening has increased in the United States, with the incidence rising from 5.8 per 100,000 in 1975 to 32.5 per 100,000 in 2005. 1 This has engendered considerable controversy regarding whether the increased detection of non lifethreatening breast cancer may lead to overtreatment. This idea is supported by breast cancer incidence models based on SEER (Surveillance, Epidemiology, and End Results) data from 1975 to 2005. 2 While the detection rate of DCIS substantially increased over this time period, the incidence of localized invasive carcinoma did not fall proportionally, as would be expected if identification and treatment of DCIS prevented subsequent invasive disease. On the contrary, the incidence of localized invasive carcinoma increased substantially over the same 30-year period, leading clinicians to question the benefit of treating DCIS and whether DCIS is an obligate precursor of invasive breast cancer. To complicate matters further, the diagnosis of DCIS within the range of abnormal breast lesions can be challenging. A recent study published by Dr Elmore and colleagues examined concordance in breast biopsy diagnosis ranging from benign disease without atypia to invasive carcinoma. 4 A panel of three expert pathologists agreed on the diagnoses of 240 total cases and the slide set was then sent to a large group of pathologists practicing in the United States to see how often they concurred with the experts. For the cases diagnosed as DCIS by the expert panel, overall concordance of the individual pathologists interpretations was 84%. The individual pathologists overdiagnosed 3% of the DCIS cases as invasive carcinoma and underdiagnosed 13% as benign disease or atypia. The difficulty of managing DCIS revolves around the fact that the natural history is poorly understood. In spite of multiple studies that have attempted to define which DCIS will progress to invasive cancer, it is currently not possible to predict this or to predict the timing of recurrence. Furthermore, the presence of invasive cancer cannot be reliably excluded after a core biopsy diagnosis of DCIS unless the lesion is

41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 completely excised. Lastly, it is difficult to judge the success of cancer therapy for DCIS, as the traditional endpoint of breast cancer mortality is not useful in this setting. Breast cancer-specific survival (BCSS) for patients with DCIS following mastectomy, lumpectomy alone, or lumpectomy with radiotherapy exceeds 96% in all cases. 5-7 Mastectomy is an effective therapy for DCIS and is associated with a very low local recurrence rate (LRR) (only 1.4% in a meta-analysis of over 1500 patients). 9 However, there is no doubt that mastectomy may also be overtreatment in many cases. Thus, it is important to ask how much risk reduction is necessary to recommend a therapy. Radiotherapy for DCIS was evaluated in a meta-analysis published by the Early Breast Cancer Trialists Collaborative Group (EBCTCG) of four randomized trials of breast-conserving surgery (BCS) with or without radiotherapy in over 3900 women with DCIS. 7 The use of radiotherapy reduced overall ipsilateral breast tumor recurrence (IBTR) by approximately 50% for both invasive and noninvasive recurrences (Table 1) 8. However, it is important to point out that there was no impact on mortality. Specifically in the National Surgical Adjuvant Breast and Bowel Project (NSABP) B-17 trial, the addition of radiotherapy to local excision halved the 15-year rate of IBTR for both invasive recurrence (52% reduction; P<.001) and DCIS recurrence (47% reduction; P<.001). Despite the significant impact on LRR, the breast cancer-specific mortality was similar with or without radiotherapy (4.7% vs 3.1%, respectively; HR 1.44). 6 Additionally, all-cause mortality rates (17.1% with radiotherapy vs 15.8%) dwarfed breast cancer-specific mortality, reinforcing that most women with DCIS have a substantially higher risk of dying from something other than breast cancer. 6 10-year outcomes BCS BCS + RT N = 1851 N = 1878 P value IBTR 7 28.1% 12.9% <.00001 Invasive IBTR 8 15.4% 6.9% <.001 DCIS IBTR 8 14.9% 6.5% <.001 All deaths 7 8.2% 8.4% >.1 Breast cancer mortality 7 3.7% 4.1% >.1 Mortality without a breast event 7 5.6% 5.4% >.1 Table 1. EBCTCG Overview: Impact of Radiotherapy on Local Recurrence and Survival in DCIS 7-8 Abbreviations: BCS, breast conserving surgery; DCIS, ductal carcinoma in situ; EBCTCG, Early Breast Cancer Trialists Collaborative Group; IBTR, ipsilateral breast tumor recurrence; RT, radiotherapy A number of surveys have been performed to evaluate factors that influence treatment decisions for DCIS. A survey in 2004 of academic physicians in the United States showed that DCIS margin width and nuclear grade strongly influenced recommendations for radiation, with clinicians far less likely to recommend radiotherapy for patients with low-grade, widely-excised DCIS. 10 In a population-based study of 659 patients in the SEER registry from Los Angeles, California, and Detroit, Michigan, patients were surveyed regarding the treatment recommendations they were given and the factors that influenced those decisions. 11 BCS was the most commonly recommended surgical approach irrespective of tumor size and grade. However, not surprisingly, surgeons were more likely to recommend a mastectomy for those with larger and higher grade DCIS. When it came to surgeon recommendations for radiotherapy, approximately 90% of those with a high risk for recurrence according to tumor size and grade were recommended for radiotherapy.

77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 It is very important to involve patients in treatment decisions for DCIS given the favorable overall prognosis and uncertainty regarding the optimal extent of therapy. Unfortunately, surveys of patients with DCIS have demonstrated that they significantly overestimate the risk of DCIS recurrence or development of invasive breast cancer. In fact, one survey of 228 patients with DCIS or early breast cancer showed that women with DCIS thought their risk of dying of breast cancer was precisely the same as women with T1 and T2, node-negative breast cancers. 12 In the population-based study from the Los Angeles and Detroit SEER registry, the single most important factor influencing patient decisions was the risk of recurrence, followed by concerns about radiation therapy side effects and body image. 11 Not surprisingly, patients who were substantially concerned about recurrence risk and radiation side effects were far more likely to opt for mastectomy. Ultimately, clinicians and patients are left with the dilemma of how to approach DCIS. One option is to approach it like a precursor lesion and treat it like cancer with a mastectomy or lumpectomy and radiotherapy. Another option is to approach it as a risk marker and nonobligate precursor, excising it to exclude the presence of invasion and considering chemoprevention with hormonal therapy. Very conservative physicians may simply make a diagnosis of DCIS and observe the patient or put them on chemoprevention. To examine controversial issues regarding the management of DCIS and shed light on best practices for patient care, three clinical scenarios and accompanying questions were posed to the symposium audience by Dr Morrow and discussed by an expert faculty panel. Following the audience voting, each member of the panel provided insight on important considerations for therapeutic decision making, including pathologic and diagnostic issues, surgical management, and the role for radiotherapy. The multidisciplinary faculty panel consisted of Melissa Murray, DO, a pathologist from Memorial Sloan Kettering Cancer Center; Terry Mamounas, MD, MPH, FACS, a breast surgeon from the University of Florida Health Cancer Center; and Julia White, MD, a radiation oncologist from The Ohio State University Comprehensive Cancer Center. Highlights from the engaging case-based discussions are presented below. Multifocal DCIS Clinical Case #1: Multifocal DCIS A 51-year-old with no family history of breast cancer has a screening mammogram demonstrating two clusters of calcifications in the upper outer quadrant (UOQ) of a C cup breast. Calcifications measure 1.2 cm and 0.7 cm and are separated by a distance of 1.5 cm. Core biopsy of both demonstrates grade 3 DCIS cribriform and solid type, estrogen receptor (ER)-positive. Question #1: For a patient with DCIS presenting as two separate clusters of calcifications that can be removed through a single incision, which of the following would you do?

118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 Audience response: 1. 91.2% answered Recommend BCS 2. 2.3% answered Recommend mastectomy 3. 6.5% answered Advise mastectomy, but do BCS if the patient really wants it Question #2: For a patient with DCIS presenting as two separate clusters of calcifications that require two incisions to remove, which of the following would you do? Audience response: 1. 44.3% answered Recommend BCS 2. 21.7% answered Recommend mastectomy 3. 34.0% answered Advise mastectomy, but do BCS if the patient really wants it Pathology Considerations for Multifocal/Multicentric DCIS Melissa Murray, DO Multifocal disease is defined as 2 foci of tumor that are in the same quadrant, while multicentric carcinoma consists of 2 foci that are in different quadrants of the same breast. 13 Multicentric disease appears to be relatively rare in DCIS, as supported by a study evaluating the extent of DCIS within 119 mastectomy specimens using serial sectioning in pathology along with mammography. 14 Ninety-nine percent of the tumors were unicentric and only one case in this series had a multicentric distribution that was defined as greater than 4 cm apart. A second study using a stereomicroscopic 3D technique to examine the growth pattern of DCIS in 60 mastectomy specimens showed that approximately half of the specimens possessed a continuous growth pattern and the other half had a discontinuous or multifocal pattern. 15 These growth patterns related to the histology of the DCIS; 90% of high-grade comedo DCIS with necrosis had a continuous growth pattern, whereas low-grade DCIS with the cribriform or micropapillary architecture were most often seen in a multifocal growth pattern. Only 30% and 45% of low-grade and intermediate-grade DCIS, respectively, had a continuous growth pattern. The distance between tumor foci varied greatly, but 82% were less than 5 mm apart. In this clinical scenario, in which a core needle biopsy was performed, the size of DCIS cannot be accurately measured and the anatomic distribution of the tumor cannot be assessed to determine whether it is multifocal or a single lesion because a core biopsy does not provide a single intact specimen. Thus, the term multifocal in a pathology report from a core needle biopsy is not meaningful. Surgical Considerations for Multifocal/Multicentric DCIS Terry Mamounas, MD, MPH, FACS Factors for selection of BCS versus mastectomy include the distribution and extent of tumor in the breast by imaging; the breast size and tumor location; candidacy for radiotherapy with BCS; and factors that are associated with local recurrence, such as patient age, grade, presence of comedonecrosis, and

160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 ER status. 16 Patient preference for BCS or mastectomy is also very important, including consideration of the patient s tolerance for risk of local recurrence, concern regarding cosmesis, and acceptance of radiotherapy if BCS is chosen. Absolute indications for mastectomy include 1) multicentric DCIS or malignant-appearing microcalcifications in an area that is too large for BCS with acceptable cosmetic result and 2) persistent positive margins after several attempts of resection. This is particularly true if the extent of disease was not anticipated by imaging studies. Unfortunately, imaging studies may underestimate or overestimate the extent of DCIS. Mammography can typically identify the extent of DCIS presenting with microcalcifications. 17 The correlation between mammographic and pathologic assessment of the extent of DCIS is more accurate in high-grade tumors with comedonecrosis than in low-grade tumors, although assessment of low-grade tumors can be improved by employing magnification views. The role of magnetic resonance imaging (MRI) for DCIS is less well defined than in invasive breast cancer. 1,18 Historically, the sensitivity of MRI was less for DCIS than for invasive breast cancer, although improvements in MRI techniques have increased the sensitivity considerably. 18 Studies have now shown MRI to be more sensitive than mammography in detecting DCIS, although additional studies are needed to determine the precise role for MRI in assessment of DCIS. 18-19 MRI can be useful in patients with multicentric DCIS who wish to receive BCS in order to rule out more extensive disease involvement. 1 However, it is important to keep in mind that MRI can overestimate the extent of disease, thus disease that was detected only by MRI should be confirmed by biopsy if a mastectomy is being considered. The choice of surgical approach, whether to make one incision or two incisions, depends on the extent and location of each DCIS lesion, the distance between the lesions, the size of the breast, and the ability to perform some form of tissue rearrangement in order to minimize the lumpectomy defect. Radiotherapy Considerations for Multifocal/Multicentric DCIS Julia White, MD It is important to consider the goals of radiotherapy following lumpectomy for DCIS, as they differ from radiotherapy post lumpectomy in patients with invasive breast cancer. For invasive breast cancer, radiotherapy seeks to maximize local control and prevent distant metastases, making BCS equivalent to mastectomy in terms of reducing recurrence risk and optimizing BCSS and overall survival (OS). In patients with DCIS, post-lumpectomy radiotherapy also seeks to maximize local control, but with the goal of preventing development of first invasive breast cancer and maintaining freedom from mastectomy. In both DCIS and invasive disease, other important goals of postsurgical radiotherapy are to maintain physical sensation and acceptable cosmetic breast appearance. As discussed previously, the EBCTCG overview clearly showed that radiotherapy following BCS reduced the rate of IBTR by a relative 50% to 60% (Table 1). 7-8 Importantly, this benefit was durable 10 years from randomization, achieving the goal of long-term local control of DCIS. The question is: can those same goals be achieved in the setting of multifocal or multicentric DCIS that undergoes BCS? While data on radiotherapy for multifocal or multicentric DCIS is lacking, data from patients with invasive breast cancer can shed some light on this issue. A retrospective analysis of over 3700 patients from the MD

202 203 204 205 206 207 208 209 210 211 212 Anderson Cancer Center registry with unifocal or multifocal invasive breast cancer showed that when a patient could undergo a successful BCS with negative margins followed by radiotherapy, there was no difference in LRR at 5 years (1.02% for unifocal, n = 1757 vs 1.95% for multifocal, n = 256). 13 Similarly, a German multicenter, retrospective cohort study of over 8900 patients with breast cancer demonstrated that when patients received guideline-adherent therapy, their breast cancer recurrencefree survival (RFS) was the same whether the disease was unifocal, multifocal, or multicentric. 20 BCS with radiotherapy showed similar RFS benefit to mastectomy in patients with multifocal or multicentric disease (Figure 1). This data, together with the favorable outcome generally associated with DCIS, suggests that the addition of radiotherapy following successful BCS can achieve the desired goals for many patients with DCIS. Multifocal Multicentric 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 Figure 1. Recurrence-Free Survival Benefit by Treatment in Patients With Multifocal or Multicentric Invasive Breast Cancer 20 Abbreviations: BCT, breast-conserving therapy; HR, hazard ratio; RT, radiotherapy Another question pertinent to this case is how multifocality affects the radiotherapy method utilized for this patient with DCIS. In nearly all the published randomized trials, whole breast irradiation of 50 Gy administered over 25 treatments is the standard. 21-22 Though not studied well, hypofractionated whole breast irradiation is also currently being used for DCIS. While the available prospective data is lacking, it seems like a reasonable option. Accelerated partial breast irradiation was developed primarily for unifocal cancer and thus, is not the treatment of choice in this setting. Lastly, is a boost recommended for patients with high-risk DCIS? The rationale for a radiation boost in DCIS is borrowed from the experience in invasive breast cancer and is based on the fact that failure patterns following lumpectomy for DCIS are similar to those of invasive breast cancer. Thus, a higher dose of radiation to the highest-risk breast tissue around the surgical cavity should improve local control, preventing first invasive breast cancer and sustaining freedom from mastectomy. There is limited data to address the role of boost in DCIS. The NSABP B-24 trial comparing breast radiotherapy with or without tamoxifen in patients with DCIS had an optional boost that was delivered to 38.5% of patients. 23 An unplanned subset analysis of 1569 patients with a follow-up of 14.5 years revealed that

233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 there was no difference in rates of in-breast recurrence regardless of whether patients received boost or no boost. This included invasive and noninvasive recurrences and was not influenced by the presence of positive margins. In addition, administration of boost increases the risk of reducing good cosmetic outcomes. In summary, BCS is reasonable for this patient provided a lumpectomy can yield negative margins and a cosmetically acceptable surgical outcome. Whole breast radiotherapy is recommended and there is no evidence that a boost adds benefit in local control. Panel Discussion Dr Morrow: Dr White, you mentioned one relatively small subset analysis from NSABP B-24 suggesting radiation boost had no benefit. 23 However, if DCIS is sensitive to radiation and randomized trials of boost in invasive cancer show benefit, 24 why would it not also show benefit in DCIS? I see a lot of DCIS patients that have received radiation boost. Do you deliver boost in your practice? Dr White: In grade 1 or grade 2 DCIS, I do not typically boost because there is no randomized data to support its benefit in preventing a first invasive breast cancer and I know it can deteriorate the cosmetic appearance. For instance, 7-year outcomes from the RTOG 9804 randomized clinical trial for grade 1 or grade 2 DCIS, <25 mm in size, reported an in-breast recurrence of 0.9% in those on the treatment arm that received whole breast irradiation post lumpectomy. 25 None of these patients received a supplemental boost dose after whole breast irradiation. I may boost a young woman with a grade 3 DCIS because her risk of in-breast recurrence is on par to that of invasive cancer. However, if more than 30% of the entire breast volume is receiving boost, there is a significant risk for poor cosmetic outcome. In that case, I will have a conversation with the patient to discuss the potential risks and benefits. Dr Mamounas: The problem with a lot of these retrospective studies is the issue of patient selection and why certain patients received boost while others did not. One might expect that the patients with a higher risk of recurrence were more likely to receive boost. Although the statistical analysis tried to adjust for this selection bias, it may still remain a factor. Because of that caveat, I still believe that the boost could be helpful in selected patients with DCIS. Dr Morrow: So there is an agreement that selective boosting is an appropriate practice in DCIS, recognizing that supporting clinical trial evidence is lacking at this time. Dr Morrow: Dr Mamounas, you identified that mastectomy was indicated for patients with multicentric DCIS. However, you also stated you would perform a lumpectomy through two separate incisions on this patient, which usually means two separate quadrants or multicentric disease. Dr Mamounas: Mastectomy is indicated in multicentric DCIS with extensive calcifications that prevent a BCS with a reasonable cosmetic result. I feel it is completely appropriate to perform two lumpectomies in patients with multicentric DCIS as long as a mammography, and perhaps MRI, confirms there is no disease between the two lesions. Dr Morrow: Dr Murray, we did not touch on the importance or lack there of histologic subtypes of DCIS. Pathology reports always include descriptions of papillary, micropapillary, solid, or cribriform DCIS. Is there any reason that clinicians need to consider histology when making treatment decisions? Dr Murray: There is no universally accepted grading system for DCIS and the one that is most often recommended is to look at the nuclear grade. However, the World Health Organization (WHO)

275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 recommends that pathologists mention all the different architectural patterns, presence of necrosis, and calcifications in the report. This can be informative for surgeons and radiation oncologists. For instance, low-grade micropapillary DCIS is known to have a growth pattern that skips through the ducts, 15 making it difficult to remove with negative surgical margins. Dr Morrow: Do any of our other panelists pay any particular attention to the descriptors of the DCIS subtype? Dr Murray: Yes, I do. When I see a solid or comedo type DCIS that is nuclear grade 3, it suggests a histologically more aggressive DCIS. Dr Mamounas: I consider it as well. Low-Risk DCIS Clinical Case #2: Low-risk DCIS managed by excision alone, without radiotherapy A 48-year-old with no family history of breast cancer has calcifications in the left breast identified on screening mammogram. Core biopsy shows low-grade DCIS. Surgical excision shows a 0.8 cm grade 2 DCIS, ER-positive. The closest margin is 2 mm. The patient will not take tamoxifen. Question #1: In a patient like this, I: Audience response: 1. 61.7% answered Recommend whole breast radiotherapy 2. 11.2% answered Recommend partial breast radiotherapy 3. 22.7% answered Use DCIS Score for decision making 4. 4.3% answered Recommend no further treatment Question #2: What DCIS Score would prompt a change in your treatment plan? Audience response: 1. 8.4% answered 10 2. 13.1% answered 20 3. 16.8% answered 30 4. 61.6% answered DCIS score does not influence my choice of local or systemic treatments Radiotherapy Considerations: Who May Safely Avoid Radiotherapy? Julia White, MD To discuss the radiotherapy considerations in this patient, it is important to first define low-risk DCIS. While there are numerous definitions that exist based on clinical and pathologic factors, similar definitions were used for the low-risk cohort of the Eastern Collaborative Oncology Group (ECOG) E5194 trial and the Radiation Therapy Oncology Group (RTOG) 9804 trial. 25-26 These criteria included mammogram detected DCIS, nuclear grade 1 or 2, lesion size <25 mm, and a negative surgical margin by 3 mm.

316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 So, the next question is: what are the outcomes seen in clinical trials examining breast conservation with or without radiotherapy for this low-risk DCIS group? ECOG E5194 was a single-arm prospective trial of lumpectomy without radiotherapy in 665 women with DCIS divided into two cohorts: a low-risk cohort defined as above and a smaller higher-risk cohort comprised of high-grade DCIS with a tumor size <10 mm. 26 Dr Solin and colleagues recently reported updated data with a median follow-up of 12.3 years. Compared to the patient in this clinical scenario, the low-risk patient population of ECOG E5194 was considerably older (median age 60 years) and 76% were postmenopausal. The median tumor size was relatively small (6 mm) and only approximately one-fourth of the patients received tamoxifen. At 12 years, in-breast recurrence rates were nearly doubled from 14.4% in the low-risk cohort to 24.6% in the higher-risk cohort (P =.003). As expected, the distribution of invasive to noninvasive breast recurrence in each group was evenly divided. Importantly, the Kaplan Meier curves show there was no plateau in the rates of in-breast recurrence or invasive in-breast recurrence; they both continued to increase with longer follow-up (Figure 2). Interestingly, one of the important predictors of in-breast recurrence was tumor size, increasing from 11.3% for tumors 5 mm, 17.8% for tumors 6 mm to 10 mm, and 22.9% for tumors >10 mm. 333 334 335 336 337 338 339 340 341 342 343 344 345 Figure 2. ECOG E5194: In-Breast Recurrence for Low-Risk and High-Risk DCIS According to Tumor Grade and Size 26 Abbreviations: IBE, ipsilateral breast event RTOG 9804 enrolled only patients with good-risk DCIS, randomizing women post lumpectomy to either observation or whole breast irradiation. 25 The trial closed in 2006 with a total of 636 patients. In comparison to the patient in the clinical case, only approximately 20% of the patients in this study were less than 50 years of age and median tumor size was 5 mm. Close to 70% of the patients in the observation and radiotherapy arms received tamoxifen. With a median follow-up of 7 years, the observation arm had a relatively low rate of in-breast recurrence of 7.2% (Figure 3). However, the addition of radiotherapy reduced this risk significantly to 0.8% (HR 0.10; P <.001).

346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 Figure 3. RTOG 9804: In-Breast Recurrence With Observation Versus Radiotherapy 25 Abbreviations: HR, hazard ratio; RT, radiotherapy It is also important to consider how patient age affects the radiotherapy recommendation. In selected trials examining lumpectomy with or without radiotherapy for DCIS, younger women generally had higher rates of in-breast recurrence compared to their older counterparts (Table 2). 27-33 Thus, using standard clinical and pathologic criteria, a cohort of patients with DCIS at low risk for recurrence can be identified. Data from ECOG E5194 and RTOG 9804 demonstrate that observation after lumpectomy results in a 14% in-breast recurrence rate at 12 years and 7% at 7 years, respectively. 25-26 The addition of breast radiotherapy will reduce the risk of in-breast recurrence significantly if that is meaningful to the patient. However, there are limited data in women less than 50 years of age and patient-reported outcomes are lacking. Thus, decisions about the use of radiotherapy in this case should be patient driven. First author Year Study Time point Treatment In-breast recurrence (%) Age <50 Age 50 Fisher 27 8 year L 26.3 25.6 1998 NSABP B-17 L + RT 15.3 9.5 Holmberg 28 2008 SweDCIS Trial 8 year L 26.7 27.3 L + RT 20.5 9.4 Cuzick 29 10 year L 28.6 25.0 2011 UK/ANZ DCIS trial L + RT 28.9 9.9 Solin 30 2001 Multi-institution, USA 10 year L + RT 16.4 6.9 Age <40 Age 40 Bijker 31 10 year L 54.2 2006 EORTC 10853 26.3 a L + RT 23.5 a 16.4 a Alvarado 32 2012 Retrospective series 5 year L 22.9 5.6 from MDACC L + RT 6.4 2.3 Cutuli 33 2002 Multi-institution, 7 year L 36.4 30.7 France L + RT 23.5 11.2

361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 Table 2. In-Breast Recurrence in DCIS by Patient Age 27-33 Abbreviations: DCIS, ductal carcinoma in situ; EORTC, European Organisation for Research and Treatment of Cancer; L, lumpectomy; MDACC, MD Anderson Cancer Center; NSABP, National Surgical Adjuvant Breast and Bowel Project; RT, radiotherapy; UK/ANZ, United Kingdom/Australia/New Zealand; USA, United States of America a In EORTC 10853, age categories were 40 and >40 Pathology Considerations: The DCIS Score Melissa Murray, DO The 12-gene DCIS Score for DCIS is a modification of the 21-gene Recurrence Score (RS) used to predict the risk of recurrence for invasive breast cancer. 34 The DCIS Score is designed to predict the likelihood of local recurrence within 10 years of the initial DCIS diagnosis in patients who received surgical excision without radiotherapy. The 12-gene panel includes five reference genes and seven cancer-related genes associated with proliferation, susceptibility to carcinogens, and progesterone receptor (PR) signaling. (Figure 4). This RT-PCR-based panel predicts the likelihood of local recurrence with or without treatment with tamoxifen in patients with DCIS. 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 Figure 4. 12-Gene DCIS Score 34 Abbreviations: DCIS, ductal carcinoma in situ; PR, progesterone receptor The DCIS Score is a continuous variable from 0 to 100 similar to the 21-gene RS for invasive breast cancer. 34 A DCIS Score of <39 is defined as low risk, 39 to 54 is considered intermediate risk, and 55 is high risk. The output from this assay includes two graphs; one outlines the risk of any local event (DCIS or invasive breast cancer) in the ipsilateral breast within 10 years and one specifies the risk of an invasive local event occurring within 10 years. ER and PR are also quantified by RT-PCR. The intent of the test is to stratify risk and provide an individualized estimate of the 10 year local recurrence risk. Validation of the DCIS Score and Potential Clinical Use Terry Mamounas, MD, MPH, FACS The 12-gene DCIS Score was first evaluated by Dr Solin and colleagues in a prospective-retrospective study of tissue samples from a subset of patients enrolled in the previously mentioned ECOG E5194 trial evaluating observation in patients with either low/intermediate-grade DCIS or high-grade DCIS. 34 The

393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 minimum margin width was 3 mm and adjuvant tamoxifen was optional and not randomized. The prespecified primary objective was to determine whether there was a significant association between the DCIS Score and the risk of an ipsilateral breast event (IBE). The secondary objective aimed to identify whether the DCIS Score provided value beyond standard clinical and pathological factors. The third conditional objective, if the DCIS Score was validated, was to evaluate if the 21-gene RS for invasive breast cancer was a better predictor of risk of IBE than the DCIS Score. The patient and tumor characteristics were similar to the parent trial except tumor size, with a median of 7 mm. 34 Approximately 80% of patients had a tumor size of 10 mm and 65% had negative margins of 5 mm. The median age was 61 years and 76% of patients were postmenopausal. Almost all patients (97%) were ER-positive and 29% received tamoxifen. In the primary Cox proportional hazards regression analysis adjusted for tamoxifen use, the DCIS Score was an independent predictor of risk for an IBE with a HR of 2.31 (P =.02). In the conditional analysis, the 21-gene RS was not statistically significantly associated with development of an IBE, in contrast to the DCIS Score. As the Kaplan Meier curves demonstrate in Figure 5, the 10-year risk of developing an IBE or invasive IBE increased continuously as the DCIS Score increased. 34 Patients categorized as low risk by the DCIS Score had a 10-year risk of IBE or invasive IBE of 10.6% and 3.7%, respectively. In a multivariable analysis excluding the DCIS Score, tumor size and postmenopausal status were independent predictors for the risk of an IBE. Those with a larger tumor size had a 54% increased risk for IBE (HR 1.54; P =.006) and postmenopausal patients had approximately half the risk of an IBE compared to premenopausal patients (HR 0.49; P =.02). Importantly, when the DCIS Score was included, it remained an independent predictor of risk for IBE (HR 2.37; P =.02), along with tumor size and menopausal status. 417 418 419 420 421 422 423 Figure 5. Probability of Developing an IBE or Invasive IBE Based on the DCIS Score 34 Abbreviations: CI, confidence interval; DCIS, ductal carcinoma in situ; IBE, ipsilateral breast event A second validation study of the DCIS Score was recently presented by Dr Rakovitch and colleagues. 35 This prospectively-designed study evaluated a population-based cohort of 571 patients diagnosed with DCIS in Ontario, Canada, who received BCS alone with negative resection margins. The median follow-up

424 425 426 427 428 429 430 431 432 was 9.6 years and the 10-year Kaplan Meier risk of local recurrence for the entire patient cohort was 19.2%. Ten-year risk of local recurrence according to the DCIS Score risk group was similar to the results from ECOG E5194: 12.7% for low risk, 33.0% for intermediate risk, and 27.8% for high risk (Figure 6). DCIS Score was predictive of both invasive and DCIS local recurrence (P =.03 and P =.002, respectively) (Figure 7). When invasive and noninvasive local recurrence was examined separately, patients with a low-risk DCIS Score had a 10-year risk of only 8% for invasive local recurrence and 5.4% for DCIS recurrence. In the multivariable analysis, the DCIS Score was an independent predictor of local recurrence. Other independent predictors included age, tumor size, solid or cribriform DCIS subtype, and multifocality defined as 2 different foci of DCIS in the same quadrant at least 5 mm apart. 433 434 435 436 Figure 6. Ontario Group Study: 10-Year Risk of Local Recurrence by DCIS Score Risk Group 35 Abbreviations: CI, confidence interval; DCIS, ductal carcinoma in situ 437 438 439 440 441 442 Figure 7. Ontario Group Study: 10-Year Risk of Invasive and DCIS Local Recurrence by DCIS Score Risk Group 35 Abbreviations: CI, confidence interval; DCIS, ductal carcinoma in situ; LR, local recurrence Panel Discussion

443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 Dr Morrow: The 12-year follow-up data from ECOG E5194 troubles me because it shows there is a remarkably high failure rate in postmenopausal patients with small, low-grade DCIS and half of these local failures are invasive. 34 I think there was a prevailing myth that one of the reasons it was safe to treat low-grade DCIS without radiation was that it was more likely to recur as DCIS rather than invasive carcinoma. Unfortunately, every single randomized trial that has looked at that question shows that not to be true. Similarly, patients with a low-risk DCIS Score still have a local failure rate of 10% to 12%. 34-35 As patients continue to become more risk averse and are opting for not only unilateral mastectomy, but bilateral mastectomy for their DCIS, how do you use the DCIS Score in your discussions with patients regarding treatment decisions? Dr Mamounas: I think treatment decisions for low-risk DCIS depend on patient preferences and the level of risk tolerance. The DCIS Score is a useful prognostic assay, but its ability to predict who will specifically benefit from radiotherapy will require further clinical investigation. The results from RTOG 9804 indicate post-lumpectomy radiotherapy can significantly reduce the risk of an IBE, but how this translates specifically to patients with a low-risk DCIS Score remains unclear. 25 I use the DCIS Score primarily in patients who I consider low risk based on their age and tumor size. If a postmenopausal patient with a small tumor also has a low DCIS Score, I will feel more comfortable not recommending radiotherapy. I would not likely use the DCIS Score in a young patient with a large tumor, because even if the DCIS Score was low I would most likely administer radiotherapy. Dr White: When a patient with DCIS comes to me for a radiotherapy recommendation, the first thing I say is, Whether you undergo mastectomy, lumpectomy alone, lumpectomy with radiotherapy, or lumpectomy with radiotherapy and tamoxifen for 5 years, the good news is that none of these options will impact you being on earth one day longer. This reframes the conversation and allows us to focus on the patient s preferences and goals. The young age of the patient in this particular case would increase my tendency to recommend radiotherapy. However, I would be comfortable omitting radiotherapy if she understands she has a higher risk of recurrence and of losing her breast but that this will not shorten her lifespan. The DCIS Score provides personalized risk assessment for each patient, which can better focus the discussion on her case than trying to estimate a patient s risk based on a clinical trial patient population. Dr Morrow: In the modern era of computed tomography-based treatment planning, what do you tell patients about significant toxicities of radiation? Dr White: I discuss that their DCIS is not life threatening so I do not want to irradiate vital organs unnecessarily. If I need to irradiate a patient that has a left-sided DCIS or a patient who is a regular smoker, I will generally treat these patients prone. If treating supine, a breath hold or other respiratory motion control technique should be used as necessary to maximize cardiac avoidance. Dr Morrow: Dr Murray, how should we interpret the finding from the Elmore et al study I mentioned earlier regarding pathologic discordance in the diagnosis of DCIS? 4 Dr Murray: It is important to understand that this study was not done in a real world setting where pathologists study multiple sections from the same sample and can consult other pathologists. There is always going to be a gray area regarding the diagnosis of atypia versus low-grade DCIS and some lesions engender borderline diagnoses.

484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 DCIS With Microinvasion Clinical Case #3: DCIS with microinvasion A 58-year-old presents with a cluster of calcifications on screening mammogram. Core biopsy shows grade 3 DCIS. Patient undergoes lumpectomy, which demonstrates a 1.6 cm DCIS with 2 foci of microinvasion, each <1 mm in size. The closest margin is 1 mm. There are no residual calcifications on mammogram. The DCIS is ER-negative; the microinvasive foci are too small to assess receptor expression. Question #1: Would you return the patient to the operating room for a sentinel lymph node biopsy (SLNB)? Audience response: 1. 68.2% answered Yes 2. 31.8% answered No Question #2: Would you re-excise the 1 mm margin on this patient? Audience response: 1. 70.8% answered No 2. 24.5% answered Yes, margins should be >2 mm 3. 3.6% answered Yes, margins should be >5 mm 4. 1.1% answered Yes, margins should be >1 cm DCIS With Microinvasion: Role of Sentinel Lymph Node Biopsy Terry Mamounas, MD, MPH, FACS Is there a role for SLNB in patients with pure DCIS? Is there a role in patients with microinvasive breast cancer? When a core needle biopsy detects DCIS, there is a possibility that it could underestimate the presence of microinvasive or invasive breast cancer. In a retrospective study of 289 cases of DCIS diagnosed by core needle biopsy, the frequency of underestimation depended on the type of needle used. 36 An 11-gauge directional vacuum-assisted biopsy has a lower incidence of underestimation of invasive ductal carcinoma compared to an automated 14-gauge needle (10% vs 21%; P<.05). Likewise, a larger study of 1226 nonpalpable breast lesions with DCIS detected by core needle biopsy demonstrated an underestimation rate of 20.4% with a 14-gauge large-core biopsy versus 11.2% with 14-gauge or 11- gauge vacuum-assisted biopsy (P<.001). 37 Core needle biopsy was also more likely to underestimate the presence of invasive breast cancer when the lesions were masses compared to microcalcifications (24.3% vs 12.5%; P<.001). An increase in the number of specimens per lesion reduced the underestimation rate from 17.5% for 10 specimens to 11.5% for >10 specimens removed. Based on the limitations of core needle biopsy, what is the role for SLNB in patients with high-risk DCIS? Patients with high-risk DCIS who undergo mastectomy will lose the opportunity to undergo a SLNB later if microinvasive or invasive breast cancer is found on final pathology. In a study from Memorial Sloan-

525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 Kettering Cancer Center of patients with DCIS who underwent SLNB, 10% of the 31 patients with DCIS with microinvasion had a positive sentinel lymph node (SLN). 38 In comparison, 12% of patients with highrisk DCIS (defined as a palpable mass, mass on mammogram, extensive high-grade lesions, or multicentric disease) had a positive SLN. MD Anderson Cancer Center also reported a series of 398 patients with an initial diagnosis of DCIS in which 20% had invasive carcinoma on final pathology. 39 The predictors for invasive carcinoma were young age, diagnosis by core biopsy, high-grade tumors, and large tumors 4 cm on mammography. Ten percent of the patients who received a SLNB had positive SLNs in this highly selected group of patients. In summary, patients with pure DCIS on core needle biopsy who undergo BCS do not require SLNB unless there is a high risk for microinvasion or true invasion on the final pathology report as estimated by palpability of the tumor and tumor grade and size. Patients with DCIS on core needle biopsy who will undergo mastectomy should undergo SLNB since there will not be a chance to do it after the mastectomy. And finally, patients like the one in this case scenario with microinvasion on core needle biopsy or on final excision should undergo SLNB because most of the studies show approximately 10% to 12% have positive lymph nodes. DCIS With Microinvasion: Pathology Considerations Melissa Murray, DO Microinvasion is defined as invasive carcinoma with no focus measuring >1 mm. 40 If there are multiple foci, these foci should not be added together; however, an attempt to quantify them should be included in the pathology report. Microinvasive carcinoma, which is often associated with high-grade DCIS, is an infrequent finding and is commonly overdiagnosed. Several studies examining outcome for patients with DCIS with microinvasion have demonstrated a relatively low rate of local and distant failure (Table 3). 41-46 Moreover, the local and distant RFS rates at 5 years or 10 years in these analyses were greater than 90%. In addition, two studies with long-term follow-up beyond 8 years showed there is no difference in OS between pure DCIS and DCIS with microinvasion. 43,47 First author Year N Median follow-up, years BCT, % Local recurrence, % Distant metastasis, % OS, % Lyons 41 2012 112 6 54 4.5 0 100 Margalit 42 2013 83 6.4 63 7.2 2.4 100 Parikh 43 2012 72 8.9 100 8.3 2.8 95.7 Vieira 44 2010 21 3 66.7 0 0 100 Kwon 45 2010 120 5 52.5 2.5 0.8 NR Colleoni 46 2004 24 3.6 58 0 0 100 Table 3. Outcomes for Microinvasive DCIS 41-46 Abbreviations: BCT, breast conserving therapy; NR, not reported; OS, overall survival There is limited data on the clinical behavior of DCIS with microinvasion due to small patient numbers, varying definitions of microinvasion, and differing degrees of tissue sampling. For microinvasion defined as 1 mm, patient prognosis is excellent and not clearly different from the patients with pure DCIS of

560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 equivalent size and grade. No predictive factors have been identified for these lesions to further inform treatment decisions. DCIS With Microinvasion: Radiotherapy Considerations Julia White, MD The questions raised by this clinical scenario are: does microinvasive DCIS always require postlumpectomy radiotherapy and how do you decide if resection margins are adequate? Looking at the series of retrospective data on DCIS with microinvasion shown in Table 3, varying amounts of BCS are observed. 41-46 When breast conservation was performed, it almost always included breast radiotherapy. The rate of local recurrence is relatively low, though it remains higher than would be expected from a similar population of patients with pure DCIS. Distant metastases were rare and OS was very high, as mentioned previously. Thus, DCIS with microinvasion demonstrates similar outcomes to pure DCIS, although the risk of local recurrence may actually lie somewhere between that of pure DCIS and T1A node-negative breast cancer. Treatment decisions regarding radiotherapy should be individualized based on clinical and pathologic features and patient preferences. For the patient in the case scenario who is 58 years of age with grade 3 DCIS with microinvasion, I would recommend a course of postlumpectomy breast radiation. There is limited data regarding optimal margin width for microinvasive DCIS. A study from Dana- Farber/Brigham and Women s Cancer Center examined 83 cases of microinvasive breast cancer with a median follow-up of 77 months. 42 A total of 63% of patients had BCS. Compared to negative margins of >2 mm, having close ( 2 mm) or positive margins with either BCS or mastectomy was significantly associated with an increased risk of local recurrence (HR 8.8; P =.003). Therefore, margin status for microinvasive disease should be approached similarly to the approach used for pure DCIS. Panel Discussion Dr Morrow: Dr Murray, you mentioned the fact that microinvasive carcinoma is frequently overdiagnosed. What pathologic tools could help distinguish microinvasion from other diagnoses to prevent uncertainty? Dr Murray: We have good immunohistochemical stains to determine whether the breast ducts have two cell layers, the epithelial cells and the myoepithelial cells, which would be lost in a case of microinvasion. In order to perform immunohistochemical staining, additional tumor sections must be cut. If the area of microinvasion is very small, it may be gone as these additional sections are cut and prevent a definitive diagnosis. Dr Morrow: Dr Mamounas, you stated you would perform a sentinel node biopsy if a patient turned out to have microinvasion on final pathology after BCS. This had always been my practice as well. However, we recently published a study looking at nodal metastases in 414 patients with microinvasive DCIS in which 57% had a single focus and 43% has 2 foci. 48 The rate of macrometastases in the SLN was only 1.4% and only 6.3% had micrometastases. Given the low rate of macrometastases, which is more likely to change subsequent treatment approaches than micrometastases, is it really necessary to perform a SLNB in all patients with evidence of microinvasion? If a patient had ER-positive disease, we would most likely give endocrine therapy and a SLNB may not provide any information to alter the treatment course.

602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 Moreover, the New England Journal of Medicine paper published by the NSABP assessing the prognostic significance of micrometastases concluded that they were clinically insignificant. 49 Dr Mamounas: In that study, patients with occult metastases, such as micrometastases, had statistically significantly worse disease-free survival and OS compared to those with no occult metastases, although the actual difference in OS was only 1.2%. 49 However, I think the presence of micrometastases in the SLN makes medical oncologists uncomfortable because it is considered N1 disease. Nodal positivity is a spectrum ranging from micrometastases to macrometastases to multiple nodes involved. Outcomes worsen as you move along that spectrum, but the impact on treatment decisions depends on multiple factors. Patients with triple-negative breast cancer or HER2-positive breast cancer who have micrometastases in the SLN may warrant a change in treatment, while a patient with ER-positive disease may not. The morbidity of SLNB is low and I feel comfortable to go back and perform one if microinvasion is detected at final pathology. However, I do not routinely do it upfront since it will not be necessary for most patients with DCIS diagnosis on core needle biopsy. It is also important to remember we are talking about patients who received BCS with radiotherapy. Therefore, even if there are micrometastases in the SLN, local control would be good following radiotherapy. Dr Morrow: I too would perform a SLNB in a patient with triple-negative or HER-positive disease, but would potentially consider not doing one in a postmenopausal patient with ER-positive DCIS with microinvasion, particularly if it was not going to change my treatment approach. Dr Morrow: Dr White, you stated that you regard microinvasive DCIS the same as pure DCIS in terms of approaches to margin status. Is that related to the fact that patients will most likely receive radiotherapy if microinvasion is present? Dr White: If the pathology report showed microinvasion, especially in more than one area, I feel that the DCIS is biologically declaring itself to be more aggressive and I would recommend breast radiotherapy. Although the additional risk of recurrence is fairly low as a result of the presence of microinvasion, it remains an increased risk. The presence of microinvasion needs to be placed into the context of all of the clinical pathologic features. For instance, if an older postmenopausal patient had strongly ERpositive DCIS with only one area of microinvasion, I would consider observation following surgery. Especially, when I consider that observation could be a reasonable option for a similar patient with outright invasive breast cancer. Dr Morrow: If a patient received lumpectomy alone for DCIS and then recurred with more DCIS, would you consider that an indication for radiotherapy? Dr White: Yes, I would recommend radiotherapy following surgery. There is a significant lack of data on observation alone in the salvage setting, so my practice is to radiate this patient population. Questions From the Audience Audience member: I recently performed a mastectomy on a 35-year-old patient with multifocal, ER and PR-positive, invasive cancer with a T2 lesion and a histological score of 5. One SLN out of 14 showed micrometastasis. She also had multifocal DCIS with a nuclear grade of 2 to 3. The histology report showed solid, cribriform, and micropapillary subtypes, with central necrosis. The surgical margins were negative for invasive cancer, but one area was positive for DCIS. Would you give radiotherapy?