Distribution of 21-Gene Recurrence Scores Among Breast Cancer Histologic Subtypes

Transcription

1 Distribution of 21-Gene Recurrence Scores Among Breast Cancer Histologic Subtypes Scott Kizy, MD; Jing Li Huang, MD; Schelomo Marmor, PhD, MPH; Anne Blaes, MD, MS; Jianling Yuan, MD, PhD; Heather Beckwith, MD; Todd M. Tuttle, MD, MS; Jane Yuet Ching Hui, MD, MS Context. The 21-gene recurrence score (RS) provides a probability of distant recurrence for estrogen receptor positive, human epidermal growth factor receptor 2 (HER2) negative breast cancers. The utility of RS for rarer histologic subtypes of breast cancer is uncertain. Objective. To determine the distribution of RS among various histologic subtypes using a population database. Design. Women between the ages of 18 and 75 with estrogen receptor-positive, HER2-negative breast cancer and known RS results were identified using the Surveillance, Epidemiology, and End Results database. Recurrence scores were categorized into risk groups using both traditional and Trial Assigning Individualized Options for Treatment cutoffs. Multivariable logistic regression was used to determine factors associated with high-risk RS. Results. We identified patients with stage I to III, estrogen receptor positive, HER2-negative breast cancer who had RS available. Overall, 3087 (7%) and 6337 (14%) of cancers were classified as high risk based on traditional and Trial Assigning Individualized Options for Treatment RS cutoffs, respectively. The proportion of highrisk RS ranged from 1% (tubular, 2 of 225) to 68% (medullary, 13 of 19) and 4% (tubular, 10 of 225) to 79% (medullary, 15 of 19) for traditional and Trial Assigning Individualized Options for Treatment cutoffs, respectively. Based on multivariable logistic regression (excluding medullary), subtypes other than invasive ductal carcinoma and papillary carcinoma were significantly associated with lower RS. The strongest predictors of a high-risk RS were higher tumor grade and negative progesterone receptor status. Conclusions. We identified distinct distributions of RS among different histologic subtypes of breast cancer. Excluding medullary carcinoma, histologic subtypes other than invasive ductal carcinoma and papillary carcinoma all predict lower RS. (Arch Pathol Lab Med. 2018;142: ; doi: / arpa oa) Breast cancer is a heterogeneous disease comprising several subtypes that differ in clinical and pathologic characteristics. 1,2 Invasive ductal carcinoma (IDC) is the most common histologic variant and accounts for 70% to 80% of cases. 3 Invasive lobular carcinoma (ILC) is the next most common subtype, affecting 15% of breast cancer diagnoses. 2 Several genomic and prognostic differences have been demonstrated between ILC and IDC. 4 Other histologic subtypes, including mixed ductal/lobular, tubular, mucinous, papillary, and medullary carcinoma, all have distinct clinical and biologic features. 1,2,5 7 Although breast Accepted for publication September 15, Published as an Early Online Release March 12, Supplemental digital content is available for this article at www. archivesofpathology.org in the June 2018 table of contents. From the Department of Surgery (Drs Kizy, Huang, Marmor, Tuttle, and Hui), the Division of Hematology, Oncology, and Transplantation (Drs Blaes and Beckwith), and the Department of Radiation Oncology (Dr Yuan), University of Minnesota, Minneapolis. Dr Tuttle holds an advisory board position for Genomic Health. The other authors have no relevant financial interest in the products or companies described in this article. Presented at the American College of Surgeons; October 26, 2017; San Diego, California. Corresponding author: Jane Yuet Ching Hui, MD, MS, Department of Surgery, University of Minnesota, 420 Delaware St SE, Mayo Mail Code 195, Minneapolis, MN ( jhui@umn.edu). cancer remains the most commonly diagnosed malignancy in women, 8 relatively little is understood about these less common subtypes. Underscoring the heterogeneity of breast cancer is the differential effectiveness of adjuvant chemotherapy. For patients with estrogen receptor (ER) positive disease, clinical trials have shown improvement in disease-free and overall survival with adjuvant chemotherapy. 9 However, approximately 85% of women with early-stage breast cancer treated with endocrine therapy alone remain disease free at 10 years. 9 Taken together, these results suggest that a majority of patients with ER-positive breast cancer may be overtreated, especially if chemotherapy is administered. To this end, genomic markers that help predict the risk of recurrence and benefit of adjuvant chemotherapy have been extensively researched. Oncotype DX (Genomic Health, Redwood City, California) is a 21- gene expression assay currently used in clinical practice to estimate the risk of distant recurrence. 10 This reverse transcriptase polymerase chain reaction-based test assesses the expression of several cancer-related genes associated with tumor invasion and proliferation. 11 Patients are categorized based on recurrence score (RS) into low-risk, intermediate-risk, and high-risk groups. 12 The RS has been validated for predicting benefit from chemotherapy using retrospective analyses of the prospective NSABP B-20 and SWOG-8814 trials for ER-positive patients with node- Arch Pathol Lab Med Vol 142, June 2018 Breast Cancer Subtypes and Oncotype DX RS Kizy et al 735

2 negative and node-positive disease, respectively. 12,13 Prospective trials, including Trial Assigning Individualized Options for Treatment (TAILORx) and Treatment for Positive Node, Endocrine Responsive Breast Cancer, are ongoing to further validate the use of RS in node-negative and node-positive disease. 14 The use of RS is now recommended by both the National Comprehensive Cancer Network and the American Society of Clinical Oncology in select patients when making treatment decisions. 15,16 The RS is currently being used to guide adjuvant systemic therapy recommendations for patients with ER-positive, human epidermal growth factor receptor 2 (HER2) negative breast cancer, most commonly for IDC. 17 Several groups 18,19 have observed correlations between pathologic characteristics and RS. Previously performed database analyses and several single-institution studies have found that histologic subtypes characterized by an indolent clinical course may routinely have lower RS Given the rarity of certain histologic subtypes of breast cancer, most of these studies are underpowered to assess these correlations. Identifying pathologic features that can predict RS risk categories may preclude the need to routinely obtain RS and help to curb health care costs. Our objective was to determine the distribution of RS among various histologic subtypes of breast cancer as well as to identify the patient and pathologic factors associated with high RS using a population-based approach. METHODS Data We used an augmented version of the Surveillance, Epidemiology, and End Results (SEER) program database to identify women diagnosed with breast cancer from 2010 to The SEER cancer registries provide population-based cancer surveillance for 18 geographic areas that represent approximately 28% of the United States. SEER collects patient demographic and tumor characteristics, including age at diagnosis; race; primary tumor site; tumor laterality; histologic subtype; tumor stage; tumor grade; ER, progesterone receptor (PR), and HER2 receptor status; diagnostic confirmation; type of surgery; use of radiation; vital status; and cause of death (per death certificate). Of note, HER2 status has been consistently recorded in SEER only since Beginning in 2004, SEER has required the collection of breast cancer multigene test results, including the 21-gene RS. Patients We limited our study to women aged 18 to 75 years diagnosed with invasive breast cancer from 2010 to 2013 who had RS available. Women with any of the 7 histologic subtypes of invasive breast cancer (IDC not otherwise specified, ILC not otherwise specified, mixed ductal/lobular, tubular, papillary, invasive medullary, and mucinous, identified by International Classification of Diseases 3 codes [Supplemental Table; see supplemental digital content at in the June 2018 table of contents]) were included. We excluded patients diagnosed with lobular carcinoma in situ, ductal carcinoma in situ, or stage IV breast cancer and tumors categorized as grade IV, ER negative, ER unknown, HER2 positive, or HER2 unknown. Patients diagnosed at autopsy, on death certificate, or at nursing homes were also excluded. Mixed ductal/lobular codes included infiltrating duct and lobular carcinoma, infiltrating ductal mixed with other types of carcinoma, and infiltrating lobular mixed with other types of carcinoma. Specific subtypes of papillary tumors are not recorded within the SEER data set. Our study was exempt from review by the Human Subjects Committee of the University of Minnesota s (Minneapolis) Institutional Review Board because it used a deidentified data source. Statistical Analysis We categorized the RSs into risk groups using the traditional (high [.30], intermediate [18 30], and low [,18]) and TAILORx cutoffs (high [.25], intermediate [11 25], and low [,11]) for the overall cohort and for each histologic subtype. We performed a multivariable logistic regression to determine the factors associated with high-risk RS (traditional and TAILORx cutoffs). All regression models included the patients age, race, histologic subtype (IDC, ILC, mixed ductal/lobular, tubular, papillary, or mucinous), tumor grade, tumor size, PR status (positive/negative or unknown), lymph node status, and year of diagnosis. In the SEER database, tumor grade is scored I through IV, with grades I through III referring to the Nottingham grading system. Forty-five tumors were classified as grade IV, and given the unclear pathologic distinction of this grade, they were excluded in this analysis. Year of diagnosis was included to evaluate for differences in RS distribution over time. We excluded the medullary subtype in our regression model as it represented a small subsample of our cohort (n ¼ 19). All statistical analyses were completed using SAS software, version 9.3 (SAS Institute, Cary, North Carolina). RESULTS Tumor and Patient Characteristics Tumor characteristics and basic demographics are presented in Table 1. We identified patients with stage I to III, ER-positive, HER2-negative breast cancer with RS available. The various histologic subtypes were IDC ( of ; 74%), ILC (5351 of ; 12%), mixed ductal/ lobular (5069 of ; 11%), mucinous (792 of ; 2%) tubular (225 of ; 0.5%), papillary (197 of ; 0.4%), and medullary (19 of ; 0.1%). Overall, 7% (3087 of ) and 14% (6337 of ) of cancers were classified as high risk based on traditional and TAILORx RS cutoffs, respectively. Intermediate-risk breast cancers accounted for 35% ( of ) and 60% ( of ) for traditional and TAILORx cutoffs, respectively. Low-risk cancers accounted for 58% ( of ) and 26% ( of ) for traditional and TAILORx cutoffs, respectively. Most patients in this analysis had no nodal metastases ( of ; 81%) and had tumors that were less than 2 cm in size ( of ; 69%), were grade I or II ( of ; 82%), and were positive for PR ( of ; 91%). Overall, 21% (9431 of ) of patients received chemotherapy. Using the traditional RS cutoffs, 72% (2238 of 3087) of high-risk tumors, 34% (5443 of ) of intermediate-risk tumors, and 7% (1750 of ) of lowrisk tumors were treated with chemotherapy. Using the TAILORx RS cutoffs, 63% (4018 of 6337) of high-risk tumors, 18% (4841 of ) of intermediate-risk tumors, and 5% (572 of ) of low-risk tumors were treated with chemotherapy. RS Distribution Based on Subtype The distribution of high-, intermediate-, and low-risk groups based on traditional RS cutoffs is presented in Table 2. Overall, 8% (2710 of ) of IDC tumors were high risk, 35% ( of ) were intermediate risk, and 57% ( of ) were low risk. High RS based on traditional cutoffs was found in 2% (110 of 5351) of ILCs and 4% (211 of 5069) of mixed ductal/lobular, 1% (2 of 225) of tubular, 7% (14 of 197) of papillary, 3% (27 of 792) of 736 Arch Pathol Lab Med Vol 142, June 2018 Breast Cancer Subtypes and Oncotype DX RS Kizy et al

3 Table 1. Clinicopathologic Characteristics of Stage I to III, Estrogen Receptor Positive, HER2-Negative Breast Cancer Patients (N ¼ ) No. % Histology group IDC ILC Mixed ductal/lobular Tubular Papillary Medullary Mucinous Traditional RS High (.30) Intermediate (18 30) Low (,18) TAILORx RS High (.25) Intermediate (11 25) Low (,11) Year of diagnosis Age, y Race Non-Hispanic white Black Asian Other Tumor size,2 cm cm cm Unknown 51 0 Tumor grade I or II III Unknown Tumor PR status Positive Negative Unknown 90 0 Lymph node status Negative Positive Unknown Chemotherapy No/unknown Yes Abbreviations: HER2, human epidermal growth factor receptor 2; IDC, invasive ductal carcinoma; ILC, invasive lobular carcinoma; PR, progesterone receptor; RS, recurrence score; TAILORx, Trial Assigning Individualized Options for Treatment. mucinous, and 68% (13 of 19) of medullary carcinomas. Of ILCs, 37% (1970 of 5351) were intermediate risk and 61% (3271 of 5351) were low risk. Of the mixed ductal/lobular carcinomas, 35% (1755 of 5069) were intermediate risk and 61% (3103 of 5069) were low risk. Low-risk RS was seen in 75% (147 of 197) of papillary, 70% (553 of 792) of mucinous, and only one medullary carcinoma. The distribution of risk stratification by TAILORx cutoffs is presented in Table 3. Overall, more tumors were found to be high or intermediate risk when TAILORx cutoffs were used as compared with the traditional cutoffs. Of IDCs, 16% (5306 of ) were high risk, 58% ( of ) were intermediate risk, and 26% (8990 of ) were low risk. Of ILCs, 7% (394 of 5351) were high risk, 70% (3741 of 5351) were intermediate risk, and 23% (1216 of 5351) were low risk. Of mixed ductal/lobular carcinomas, 10% (528 of 5069) were high risk, 62% (3162 of 5069) were intermediate risk, and 27% (1379 of 5069) were low risk. Four percent (10 of 225) of tubular carcinoma were high risk and 79% (15 of 19) of medullary carcinomas were high risk by TAILORx cutoffs. Ranges of the RS scores by subtypes are demonstrated in the Figure. IDC had the widest range of RS, and tubular and papillary had the narrowest range of scores. Median RS ranged from 14 to 16 among ILC, IDC, mixed, tubular, and mucinous tumors. Papillary subtype had the lowest median at 7, and medullary subtype had the highest median at 36. Papillary tumors had the lowest distribution of RS with patients in the 3 lowest quartiles ranging below 20. Compared with other subtypes, medullary had the highest RS, with the 3 highest quartiles of patients having a RS above 25. Factors Associated With High-Risk RS Multivariable logistic regression analysis was performed to evaluate for factors associated with high RS (Table 4). Because of the low number of medullary subtype tumors (n ¼ 19), these tumors were not included in this analysis. Compared with IDC, all histologic subtypes other than papillary were less likely to be associated with a high-risk RS based on TAILORx cutoffs. Overall, increasing age was associated with decreasing likelihood of being associated with a high-risk RS. Patients diagnosed with breast cancer in 2011 and 2012 were significantly more likely to have a high-risk RS than patients diagnosed in 2010 (traditional cutoffs: odds ratio [OR] 1.22; 95% CI, ; TAILORx cutoffs: OR, 1.15; 95% CI, ). Black race was associated with higher likelihood of high-risk RS as compared with non-hispanic white in the traditional cutoff analysis (OR, 1.23; 95% CI, ) but not in the TAILORx cutoff analysis (OR, 1.14; 95% CI, ). Asian patients were less likely to be classified as high risk than non-hispanic whites (OR, 0.85; 95% CI, ) in the traditional cutoff analysis but not in the TAILORx cutoff analysis (OR, 0.97; 95% CI, ). Tumor size of 2 to 5 cm as compared with less than 2 cm (traditional cutoffs: OR, 1.73; 95% CI, ; TAILORx cutoffs: OR, 1.51; 95% ), grade III as compared with grade I or II (traditional cutoffs: OR, 8.88; 95% CI, ; TAILORx cutoffs: OR, 6.84; 95% CI, ), and PR negativity as compared with PR positivity (traditional cutoffs: OR, 6.59; 95% CI, ; TAILORx cutoffs: OR, 6.44; 95% CI, ) were significantly associated with a high-risk RS (Table 4). DISCUSSION In this study, we used the SEER database to assess the distribution of RS among distinct histologic subtypes of breast cancer and the factors associated with high-risk RS. Arch Pathol Lab Med Vol 142, June 2018 Breast Cancer Subtypes and Oncotype DX RS Kizy et al 737

4 Table 2. Distribution of Recurrence Score by Histologic Subtype by Traditional Cutoffs Recurrence Score High (.30) Intermediate (18 30) Low (,18) Histologic Subtype No. % No. % No. % IDC (n ¼ ) ILC (n ¼ 5351) Mixed ductal/lobular (n ¼ 5069) Tubular (n ¼ 225) Papillary (n ¼ 197) Medullary (n ¼ 19) Mucinous (n ¼ 792) Abbreviations: IDC, invasive ductal carcinoma; ILC, invasive lobular carcinoma. Based on both traditional and TAILORx cutoffs for recurrence risk classification, we found several differences among subtypes. In our multivariable analysis, ILC, mixed ductal/lobular, tubular, and mucinous subtypes were associated with a lower likelihood of having a high-risk RS when compared with IDC regardless of the cutoff model used. Overall, increasing age was associated with lower likelihood of a high-risk RS, irrespective of histologic subtype. The strongest predictors of a high-risk RS were higher tumor grade and negative PR status. Genomic assays, including the 21-gene RS, have allowed for more tailored treatment of patients with ER-positive breast cancer. By identifying patients who are at high risk for distant recurrence based on genomic profiles, the RS allows selection of patients who would benefit most from adjuvant chemotherapy, while sparing others from treatment toxicities. Early results of the TAILORx trial found that patients with node-negative, ER-positive breast cancer and low-risk RS (,11) had very low recurrence rates with endocrine therapy alone. 14 For the intermediate-rs (11 25) group, patients have been randomized to either endocrine therapy alone or endocrine therapy plus adjuvant chemotherapy; the results of this component of TAILORx are pending. Currently, for patients with intermediate risk by RS, clinicopathologic findings are used to guide recommendations on adjuvant chemotherapy. 17 Several patterns of care studies have shown variable practices of adjuvant chemotherapy administration among medical oncologists for this group of patients. 24,25 Patients with high RS demonstrate the most benefit from adjuvant chemotherapy and thus are typically recommended this treatment. 12 Despite the rarity of subtypes other than IDC, singleinstitution studies have evaluated the distribution of RS among distinct subtypes In a study by Bomeisl and colleagues, 20 substantive conclusions could not be reached given low numbers of subtypes other than IDC within a single-institution database that contained only 4 mucinous carcinomas and 1 tubular carcinoma. This group did not observe any high-risk RS in any subgroups other than IDC. In other prior studies, 22,23 ILC has been observed to have lower proportions of high-risk tumors, although sample sizes were small. Our population-based analysis validates these findings, as only 2% and 7% of ILC breast cancer had high-risk RS based on traditional or TAILORx cutoffs, respectively, compared with 8% and 16% for IDC. Although single-institution studies could verify that most high-risk ILC were pleomorphic on histologic examination, we were unable to confirm this finding, as SEER does not provide data on the further characterization of ILC subtype into pleomorphic versus not. We found that, in addition to ILC, most other histologic subtypes also had a lower distribution of high-risk RS when compared with IDC. On the other hand, patients with medullary carcinoma had a high likelihood of having high-risk RS (68% and 79% for traditional and TAILORx cutoff, respectively). However, because most medullary carcinomas tend to be ER negative, 2,25 the sample size in this subgroup is small (n ¼ 19). Medullary carcinomas that are ER positive, however, are at high risk of recurrence based on RS. 26 Overall, personalized medicine is making a significant impact with regard to management of breast cancer. Although RS provides individualized risk estimates, testing costs approximately $4000 per patient. 27 In that context, our analysis of the distribution of RS within individual histologic subtypes, in conjunction with additional clinicopathologic features, may serve to further refine the utility of RS testing. For patients with subtypes of breast cancer other than IDC, given the cost, RS testing can reasonably be omitted in patients with tubular and medullary carcinomas, whereas for patients with mixed ductal/lobular, papillary, and Table 3. Distribution of Recurrence Score by Histologic Subtype by Trial Assigning Individualized Options for Treatment Cutoffs Recurrence Score High (.25) Intermediate (11 25) Low (,11) Histologic Subtype No. % No. % No. % IDC (n ¼ ) ILC (n ¼ 5351) Mixed ductal/lobular (n ¼ 5069) Tubular (n ¼ 225) Papillary (n ¼ 197) Medullary (n ¼ 19) Mucinous (n ¼ 792) Abbreviations: IDC, invasive ductal carcinoma; ILC, invasive lobular carcinoma. 738 Arch Pathol Lab Med Vol 142, June 2018 Breast Cancer Subtypes and Oncotype DX RS Kizy et al

5 Box plot of the 21-gene recurrence score by histologic subtype. Abbreviations: D/L, ductal/lobular; IDC, invasive ductal carcinoma; ILC, invasive lobular carcinoma. mucinous carcinomas, RS testing continues to play a valuable role in guiding adjuvant treatment. We found that the majority of patients with tubular carcinoma had either low- (67%) or intermediate-risk (32%) RS. Historically, tubular carcinomas are mostly ER positive and HER2 negative and confer a greater overall survival and greater disease-free survival than do IDCs. 28 Furthermore, small tubular carcinomas were found to have a decreased risk of nodal metastasis compared with IDCs. 29 Thus, tubular carcinoma may represent one subtype for which RS testing may be omitted, as the majority of these patients will not be recommended to undergo adjuvant chemotherapy. Medullary carcinoma, on the other hand, consistently had a higher RS (68%) when compared with other histologic subtypes in this study. Coupled with previous reports 30 noting that recurrence and survival are improved with the administration of chemotherapy, our analysis shows that 21-gene RS testing may not add decision-making value to ER-positive medullary tumors. Finally, although our study demonstrates that only a small proportion of papillary (7%), mucinous (3%), and mixed ductal/lobular (4%) carcinomas were categorized as high-risk RS, the potential for adverse biology of these subtypes may still warrant testing. Specifically, when compared with ILC, mixed ductal/lobular carcinomas are associated with a greater risk of recurrence, distant metastasis, and disease-specific mortality. 2,31 Also, prior SEER analyses 7,29,32 have shown similarities among tubular, papillary, and mucinous carcinoma, such as likelihood of ER positivity, decreased tendency for nodal metastasis, and smaller tumor size. However, the overall and breast cancer specific survivals of mucinous and papillary carcinoma were not significantly different when compared with IDC. 1,7 Altogether, further studies are needed to evaluate the impact of distinct breast cancer subtypes on the utility of RS testing. We recognize several important limitations to this study. Because the design of this study is retrospective, selection bias due to preferential ordering of the genomic assay is likely present. For example, in our multivariable analysis, age, lymph node status, and year of diagnosis all were associated with lower RS. In particular, one would not typically expect nodal metastases to be associated with lower RS. This observation is likely attributable to the selection bias of obtaining RS in node-positive patients with other low-risk clinicopathologic features, whereas RS is likely to be omitted altogether in node-positive patients, for whom adjuvant chemotherapy will be readily recommended based on other high-risk clinicopathologic features. This selection bias likely contributes to why the year of diagnosis correlates with lower likelihood of a high-risk RS, as, over time, this bias is likely to increase. Another limitation of this study is that although the histologic subtypes are based on World Health Organization classification, the reproducibility of the histology interpretation among pathologists across distinct institutions that participate in SEER is unknown. As such, we are unable to determine the exact specification of tumors labeled in different subtypes without central pathologic review, including the inability to identify more specific subtypes (eg, pleomorphic ILC). Finally, the SEER data set does not report the use of endocrine therapy and may underreport outpatient chemotherapy use, precluding meaningful analysis of the impact of RS on treatment recommendation and, ultimately, long-term survival in our cohorts. Finally, although SEER collects data on cancer-specific survival, no data are collected regarding disease recurrence. Therefore, rates of local, regional, or distant recurrence are not analyzable outcomes with SEER data. Despite these limitations, we believe our findings underscore the important link between histologic subtypes and RS. This study represents the largest descriptive analysis of RS within histologic subtypes of breast cancer. In patients with ER-positive, HER2-negative tumors, we found distinct differences in patterns of RS among histologic subtypes of tumors. Other than medullary carcinoma, histologic subtypes all predict lower RS based on multivariate analysis compared with IDC. The public health impact of this study is substantial, as finding predictors of RS may alleviate the economic burden of performing RS analysis routinely on all patients. Based on this and other analyses, RS may not be required for tubular and medullary subtypes. Further Arch Pathol Lab Med Vol 142, June 2018 Breast Cancer Subtypes and Oncotype DX RS Kizy et al 739

6 Table 4. Factors Associated With High-Risk Recurrence Score (Traditional and Trial Assigning Individualized Options for Treatment [TAILORx] Cutoffs) a Model 1: Traditional Cutoffs Model 2: TAILORx Cutoffs OR 95% CI OR 95% CI Age Reference Reference Race Non-Hispanic white Reference Reference Black Asian Other Histologic subtype IDC Reference Reference ILC Mixed ductal/lobular Tubular Papillary Mucinous Size,2 cm Reference Reference 2 5 cm cm Unknown Grade I or II Reference Reference III Unknown PR status Positive Reference Reference Negative Unknown Lymph node status Negative Reference Reference Positive Unknown Year of diagnosis 2010 Reference Reference Abbreviations: IDC, invasive ductal carcinoma; ILC, invasive lobular carcinoma; OR, odds ratio; PR, progesterone receptor. a Bold type indicates statistical significance. Reference refers to the comparative factor. research is necessary to identify specific subgroups of patients who may not benefit from RS testing. References 1. Li CI. Risk of mortality by histologic type of breast cancer in the United States. Horm Cancer. 2010;1(3): doi: /s Li CI, Uribe DJ, Daling JR. Clinical characteristics of different histologic types of breast cancer. Br J Cancer. 2005;93(9): doi: /sj.bjc Lakhani S, Ellis I, Schnitt SJ, Tan PH, van de Vijver MJ. WHO Classification of Tumours of the Breast. 4th ed. Lyon, France: IARC Press; Desmedt C, Zoppoli G, Gundem G, et al. Genomic characterization of primary invasive lobular breast cancer. J Clin Oncol. 2016;34(16): doi: /jco Nguyen CV, Falcón-Escobedo R, Hunt KK, et al. Pleomorphic ductal carcinoma of the breast: predictors of decreased overall survival. Am J Pathol. 2010;34(4): doi: /pas.0b013e3181ce97bf. 6. Huober J, Gelber S, Goldhirsch A, et al. Prognosis of medullary breast cancer: analysis of 13 International Breast Cancer Study Group (IBCSG) trials. Ann Oncol. 2012;23(11): doi: /annonc/mds Di Saverio S, Gutierrez J, Avisar E. A retrospective review with long term follow up of 11, 400 cases of pure mucinous breast carcinoma. Breast Cancer Res Treat. 2008;111(3): doi: /s z. 8. American Cancer Society. Cancer facts & figures cancer.org/content/dam/cancer-org/research/cancer-facts-and-statistics/annualcancer-facts-and-figures/2016/cancer-facts-and-figures-2016.pdf. Accessed January 3, Fisher B, Jeong J-H, Bryant J, et al. Treatment of lymph-node-negative, oestrogen-receptor-positive breast cancer: long-term findings from National Surgical Adjuvant Breast and Bowel Project randomised clinical trials. Lancet. 2004;364(9437): doi: /s (04)16981-x. 10. Paik S, Shak S, Tang G, et al. A multigene assay to predict recurrence of tamoxifen-treated, node-negative breast cancer. N Engl J Med. 2004;351(27): doi: /nejmoa Gy}orffy B, Hatzis C, Sanft T, Hofstatter E, Aktas B, Pusztai L. Multigene prognostic tests in breast cancer: past, present, future. Breast Cancer Res Treat. 2015;17:11. doi: /s Paik S, Tang G, Shak S, et al. Gene expression and benefit of chemotherapy in women with node-negative, estrogen receptor-positive breast cancer. J Clin Oncol. 2006;24(23): doi: /jco Mamounas EP, Tang G, Fisher B, et al. Association between the 21-gene recurrence score assay and risk of locoregional recurrence in node-negative, 740 Arch Pathol Lab Med Vol 142, June 2018 Breast Cancer Subtypes and Oncotype DX RS Kizy et al

7 estrogen receptor-positive breast cancer: results from NSABP B-14 and NSABP B- 20. J Clin Oncol. 2010;28(10): doi: /jco Sparano JA, Gray RJ, Makower DF, et al. Prospective validation of a 21- gene expression assay in breast cancer. N Engl J Med. 2015;373(21): doi: /nejmoa Gradishar W, Anderson B, Balassanian R, et al. Invasive breast cancer version , NCCN clinical practice guidelines in oncology. J Natl Compr Canc Netw. 2016;14(3): doi: /bmj Harris L, Fritsche H, Mennel R, et al. American Society of Clinical Oncology 2007 update of recommendations for the use of tumor markers in breast cancer. J Clin Oncol. 2007;25(33): doi: /jco Fried G, Moskovitz M. Treatment decisions in estrogen receptor-positive early breast cancer patients with intermediate onco type DX recurrence score results. Springerplus. 2014;3:1 8. doi: / Petkov VI, Miller DP, Howlader N, et al. Breast-cancer-specific mortality in patients treated based on the 21-gene assay: a SEER population-based study. NPJ Breast Cancer. 2016;2: doi: /npjbcancer Wen HY, Krystel-whittemore M, Patil S, Pareja F. Breast carcinoma with an Oncotype DX recurrence score,18: rate of distant metastases in a large series with clinical follow-up. Cancer. 2017;1(123): doi: /cncr Bomeisl PE, Thompson CL, Harris LN, Gilmore HL. Comparison of Oncotype DX recurrence score by histologic types of breast carcinoma. Arch Pathol Lab Med. 2015;139(12): doi: /arpa oa. 21. Baehner F, Watson D, Ballard J, Millward C, Palmer G, Shak S. Gene expression by standardized quantitative RT-PCR in the special histologic subtypes of estrogen receptor positive invasive breast cancer. Abstract presented at: San Antonio Breast Cancer Symposium; December 3, 2007; San Antonio, TX. Abstract Felts JL, Zhu J, Han B, Smith SJ, Truica CI. An analysis of Oncotype DX recurrence scores and clinicopathologic characteristics in invasive lobular breast cancer [published online ahead of print January 18, 2017]. Breast J. doi: / tbj Tsai ML, Lillemoe TJ, Finkelstein MJ, et al. Utility of Oncotype DX risk assessment in patients with invasive lobular carcinoma. Clin Breast Cancer. 2016; 16(1): doi: /j.clbc Loncaster J, Armstrong A, Howell S, et al. Impact of Oncotype DX breast recurrence score testing on adjuvant chemotherapy use in early breast cancer: real world experience in greater Manchester, UK. Eur J Surg Oncol. 2017;43(5) doi: /j.ejso Carlson J, Roth J. The impact of the Oncotype Dx breast cancer assay in clinical practice: a systematic review and meta-analysis. Breast Cancer Res Treat. 2013;141(1): doi: /jid Bertucci F, Finetti P, Cervera N, et al. Gene expression profiling shows medullary breast cancer is a subgroup of basal breast cancers. Cancer Res. 2006; 66(9): doi: / can Raldow AC, Sher D, Chen AB, Recht A, Punglia RS. Cost effectiveness of the Oncotype DX DCIS score for guiding treatment of patients with ductal carcinoma in situ. J Clin Oncol. 2016;34(33): doi: /jco Rakha EA, Lee AHS, Evans AJ, et al. Tubular carcinoma of the breast: further evidence to support its excellent prognosis. J Clin Oncol. 2010;28(1): doi: /jco Kader HA, Jackson J, Mates D, Andersen S. Tubular carcinoma of the breast: a population-based study of nodal metastases at presentation and of patterns of relapse. Breast J. 2001;7(1): Mateo AM, Pezzi TA, Sundermeyer M, Kelley CA, Klimberg VS, Pezzi CM. Chemotherapy significantly improves survival for patients with T1c-T2N0M0 medullary breast cancer: 3739 cases from the National Cancer Data Base. Ann Surg Oncol. 2017;24(4): doi: /s Rakha EA, Gill MS, El-Sayed ME, et al. The biological and clinical characteristics of breast carcinoma with mixed ductal and lobular morphology. Breast Cancer Res Treat. 2009;114(2): doi: /s Zheng Y, Hu X, Shao Z. Clinicopathological characteristics and survival outcomes in invasive papillary carcinoma of the breast: a SEER population-based study. Sci Rep. 2016;6: doi: /srep Arch Pathol Lab Med Vol 142, June 2018 Breast Cancer Subtypes and Oncotype DX RS Kizy et al 741