J Clin Oncol 24: by American Society of Clinical Oncology INTRODUCTION

Transcription

1 VOLUME 24 NUMBER 36 DECEMBER JOURNAL OF CLINICAL ONCOLOGY O R I G I N A L R E P O R T Locoregional Relapse and Distant Metastasis in Conservatively Managed Triple Negative Early-Stage Breast Cancer Bruce G. Haffty, Qifeng Yang, Michael Reiss, Thomas Kearney, Susan A. Higgins, Joanne Weidhaas, Lyndsay Harris, Willam Hait, and Deborah Toppmeyer From the Departments of Radiation Oncology, Medical Oncology, and Surgery, Cancer Institute of New Jersey, Robert Wood Johnson Medical School University of Medicine and Dentistry of New Jersey, New Brunswick, NJ; and Department of Therapeutic Radiology and Section of Medical Oncology, Yale University School of Medicine, New Haven, CT. Submitted March 14, 2006; accepted October 11, 2006; published online ahead of print at on November 20, Authors disclosures of potential conflicts of interest and author contributions are found at the end of this article. Address reprint requests to Bruce G. Haffty, MD, Cancer Institute of New Jersey, Robert Wood Johnson Medical School, 195 Little Albany St, New Brunswick, NJ ; hafftybg@umdnj.edu by American Society of Clinical Oncology X/06/ /$20.00 DOI: /JCO A B S T R A C T Purpose To determine the prognostic significance of triple negative breast cancers with respect to locoregional relapse and distant metastasis in conservatively managed breast cancer patients. Patients and Methods A database of conservative managed (conservative surgery followed by radiation) patients, in whom all three markers (estrogen receptor, progesterone receptor, and HER2/neu) were available, was reviewed. Patients were classified as triple negative if they tested negative for all three markers. Of 482 patients with all three markers available, 117 were classified as triple negative. Results As of September 2005, with a median follow-up time of 7.9 years, of the 482 patients in the study, there have been 53 in-breast relapses, 10 nodal relapses, 77 distant relapses, and 69 deaths. At 5 years, the triple negative cohort had a poorer distant metastasis-free rate compared with the other subtypes (67% v 82%, respectively; P.002). Triple negative subtype was an independent predictor of distant metastasis (hazard ratio 2.14; 95% CI, 1.31 to 3.53; P.002) and cause-specific survival (hazard ratio 1.79; 95% CI, 1.03 to 3.22; P.047). There was no significant difference in local control between the triple negative and other subtypes (83% v 83%, respectively). Of 99 BRCA-tested patients in this cohort, 10 had deleterious mutations in BRCA1, and seven had mutations in BRCA2. Of10BRCA1 patients, eight were triple negative, whereas only one of seven BRCA2 patients was triple negative (P.001). Conclusion Patients classified as triple negative have a poor prognosis. However, there was no evidence that these patients are at higher risk for local relapse after conservative surgery and radiation. Patients with BRCA1 mutations develop predominantly triple negative tumors. J Clin Oncol 24: by American Society of Clinical Oncology INTRODUCTION Breast cancers are represented by a heterogeneous group of tumors, characterized by a wide spectrum of clinical, pathologic, and molecular features. 1-3 This wide spectrum of factors accounts for variations in response to therapy and outcomes among women diagnosed with breast cancer. 4-6 Routine clinical variables have more recently been complemented by molecular profiling in an attempt to refine prognosis and response to therapy There have been numerous reports demonstrating that breast tumors will segregate into prognostic categories based on hierarchical cluster analysis of gene expression profiling, 1,4,9,11-14 which can help guide clinicians regarding prognosis and response to therapies. Recent attention has been devoted to a classification system that uses three common molecular markers, estrogen receptor (ER), progesterone receptor (PR), and HER2/neu, and classifies patients into subtypes. 2,4,6,15 Luminal subtypes make up the hormone receptor expressing tumors and generally carry a favorable prognosis. Human epidermal growth factor receptor 2 (HER2) subtypes refer to predominantly hormone receptor negative tumors with a specific gene expression pattern. Although not all tumors that are HER2/neu positive by clinical testing (immunohistochemistry and/or fluorescent in situ hybridization) strictly fall into this category, 5652

2 Triple Negative Breast Cancer most HER2/neu-positive tumors do fit this criteria. 2,4,6 Although these tumors generally carry a poorer prognosis than luminal types, the effectiveness of trastuzumab in HER2/neu-positive patients can significantly impact the outcome of these patients. 16 Basal-like tumors lack both hormone receptor and HER2/neu expression, and despite having a poor prognosis, these tumors have been shown in neoadjuvant studies to be responsive to chemotherapy, 5 are more commonly seen in African American women, and are commonly seen in women who are BRCA1 carriers. 2,4,12,17,18 These cancers are generally positive for HER1 expression, basal cytokeratins, and c-kit. 4,11,15 Although this classification system is based on extensive genetic profiling assays, a simplified method of classification, based on ER, PR, and HER2/neu assays, makes this system appealing and clinically useful. Accordingly, clinicians apply the combination of ER, PR, and HER2/neu to classify patients into these subtypes and are increasingly considering triple negative status (ER negative, PR negative, and HER2/neu negative) in clinical decision making and protocol design. Although triple negative tumors have been reported to be more aggressive, there are limited long-term clinical data evaluating outcome as a function of this classification. Despite the fact that clinicians routinely use conventional ER, PR, and HER2/neu status to classify patients as triple negative for clinical decision making, there are no large databases that have evaluated locoregional control in conservatively managed patients. The purpose of the current study was to provide additional validation of the prognostic significance of this classification scheme and to evaluate outcome in a cohort of conservatively managed breast cancer patients to determine the significance of triple negative classification for locoregional recurrence. PATIENTS AND METHODS Between 1980 and 2003, 1,990 patients were treated with breast-conserving surgery followed by radiation therapy to the intact breast, with or without systemic therapy, at the radiation oncology facility of Yale University School of Medicine. Review of the outcomes of these patients was approved by the institutional review board, with waiver of informed consent. Only those patients in whom ER, PR, and HER2/neu status was available were included in the current analysis. This limited the sample to a total of 482 patients in whom all three molecular markers were available. This introduces an unavoidable selection bias in that the group of 482 patients included in this study, compared with our larger conservatively managed database, had a higher percentage of patients younger than age 50 years (49% v 28%, respectively; P.01) and a higher percentage of patients with T2 tumors (26% v 17%, respectively; P.01). However, nodal status was not significantly different between the 482 patients included in this study and the overall population (25% node positive v 26% node positive, respectively). The data on ER, PR, and HER2/neu were obtained through standard clinical testing, using immunohistochemistry for ER and PR and the HercepTest (Dako, Glostrup, Denmark) for HER2/neu. For ER and PR, receptor positivity was based on more than 10% of cells testing positive, in accordance with our standard guidelines. HER2/neu data were obtained through routine clinical testing. Over the past 10 years, with the availability of HER2/neu testing through the HerceptTest, efforts were made to retrieve and process available paraffin-embedded specimens of patients from the database for HER2/neu testing. Scores of 0 and 1 were considered negative, and a score of 3 was considered positive. 19 Although there is general agreement that HER2/neu 0 and 1 are clinically HER2/neu negative and HER2/neu 3 is clinically HER2/neu positive, there is uncertainty about the appropriate classification of HER2/neu 2 patients. HercepTest scores of 2 were considered positive for the current study because this was the accepted classification scheme at the time of clinical treatment. More recently, it has become common to test HER2/neu 2 patients by fluorescent in situ hybridization (FISH) to determine the more appropriate classification. In the current study, we did not have FISH information available on the majority of HER2/neu 2 patients. To further validate the findings, two separate analyses were done, the first with HER2/neu 2 classified as positive and a second with HER2/neu 2 classified as negative. No attempt was made in this study to reclassify any of the patients in the database. However, all of the testing was performed in the same clinical pathology facility, using standard clinical protocols, by the same pathology staff. Patients were classified as triple negative if they were negative for all three receptors or as non triple negative if they were positive for any of the three markers. Although the triple negative classification is generally accepted based on negative status of all three assays (ER, PR, and HER2/neu), one must acknowledge some overlap between the luminal and HER2 subtypes based on immunohistochemical staining alone. 1-4 In the current analysis, more accurate subtyping of these categories could not be assessed without complete gene profiling studies, which are beyond the scope of the current study. Therefore, in comparing groups, we compared the triple negative group with the other two groups combined. All patients were treated with breast-conserving surgery followed by radiation therapy. Radiation was administered to the breast and/or regional lymphatics as clinically indicated and described elsewhere to a total median dose to the tumor bed of 64 Gy. Systemic therapy was administered in accordance with standard clinical practice during this time interval. All patient data were entered into a computerized database. Local or regional relapses were defined as clinically and histologically documented relapse in the ipsilateral breast or regional nodes. Distant metastasis was defined as clinical evidence of distant disease based on clinical and/or radiographic findings. All events were calculated from the time of initial diagnosis to the time of the event. Survival, distant metastasis, and local relapse free times were calculated using standard life-table methods. Differences between categoric variables were calculated using standard 2 methods, whereas differences in survival times were compared using standard life-table Cox regression models. RESULTS On the basis of ER, PR, and HER2/neu clinical testing, of the 482 patients in this analysis, 117 patients were classified as triple negative, and 365 were classified as non triple negative. Characteristics of the patient population as a function of whether they were triple negative are listed in Table 1. One of the more striking differences between the triple negative and the non triple negative groups was the higher percentage of younger women with triple negative tumors, with only 37% of the triple negative group over age 50 years at diagnosis compared with 55% of the non triple negative cohort. BRCA testing was performed on 99 patients in this cohort, of whom 10 had deleterious mutations in BRCA1 and seven had deleterious mutations in BRCA2. As expected, eight of the 10 BRCA1 patients were classified as triple negative. BRCA2 patients, who are more commonly receptor positive, were less likely to be in the triple negative category. Of note, the triple negative group had a higher proportion of patients with strong family histories (defined as at least one first-degree relative with breast cancer), and patients in the triple negative group were more likely to be African American. Although pathologic nodal status was evenly distributed among the three subtypes, triple negative subtype was more frequently associated with a higher pathologic T stage. Outcomes As of September 2005, with a median follow-up time of 7.9 years in this cohort, there have been 53 local (in-breast) relapses, 77 distant relapses, 10 nodal relapses, and 69 deaths. This resulted in a 10-year

3 Haffty et al Characteristic Total No. of Patients Table 1. Patient Characteristics of Different Tumor Subtypes Triple Negative Tumor Subtype No. % No. % Age, years Race.0172 White Black Other Family history.001 None Moderate Strong Histology.01 Ductal Lobular Other Tumor Size.0001 T T Nodal status NS Negative Positive p53 status NS Negative Positive BRCA1 mutation.001 No Yes BRCA2 mutation NS No Yes Adjuvant chemotherapy.0001 No Yes Tamoxifen treatment.0001 No Yes Abbreviation: NS, not significant. P overall survival rate of 73%, 10-year local relapse-free rate of 78%, distant relapse free rate of 71%, and 10-year cause-specific survival rate of 75%. In the initial analysis, HER2/neu 2 patients were classified as positive. When analyzed as a function of tumor subtype, the triple negative subtype had a significantly poorer prognosis compared with the non triple negative subtypes. As shown in Table 2, causespecific survival, nodal relapse-free, and distant metastasis-free rates were poorer among the triple negative group. Overall survival, although lower in the triple negative patients, was not statistically significant. Nodal control rate was significantly lower in triple negative patients, but given the small number of events, it is difficult to determine the clinical significance of this finding. The local in-breast relapse rate, however, was similar among the triple negative and non triple negative subtypes. In a multivariate Cox regression analysis taking into account patient age, margin status, nodal status, T stage, and use of adjuvant therapy, triple negative subtype was an independent predictor of distant metastasis (hazard ratio 2.15; 95% CI, 1.31 to 3.53; P.002) and cause-specific survival (hazard ratio 1.79; 95% CI, 1.03 to 3.22; P.047). Figures 1, 2, and 3 summarize the outcomes in distant metastasis-free survival, cause-specific survival, and ipsilateral breast relapse-free survival, respectively, as a function of triple negative classification. Because of the uncertainty in the appropriate classification of HER2/neu 2 patients by immunostaining as falling into a positive or negative category, a second analysis was performed with HER2/ neu 2 patients classified as negative. Using this strategy, similar results were found. Specifically, distant metastasis was poorer in those patients classified as triple negative, with the HER2/neu 2 classified as negative, although the differences were not as significant. Using this classification, the 5-year distant metastasis-free rate of the non triple negative patients was 88% compared with 79% in the triple negative cohort (P.05). Of note, there again 5654 JOURNAL OF CLINICAL ONCOLOGY

4 Probability of Survival (% ) Triple Negative Breast Cancer Table 2. Five-Year Outcomes As a Function of Tumor Subtype Outcome Tumor Subtype (%) Triple Negative Breast relapse-free rate NS Nodal relapse-free rate Distant metastasis-free rate Cause-specific survival rate Overall survival rate NS Abbreviation: NS, not significant. P Triple negative P = Cause-Specific Survival (years) Fig 2. Cause-specific survival as a function of subtype. was no difference in local relapse-free survival as a function of triple negative classification using these criteria. Using this classification, the nodal relapse rate was greater in the triple negative cohort compared with the other patients, but again, the difference was not statistically significant (99% v 96%, respectively). A subset analysis of overall survival, cause-specific survival, and distant metastasis-free survival in the 117 triple negative patients as a function of chemotherapy was conducted. As shown in Table 3, chemotherapy was administered to 77 of the 117 triple negative patients and was more frequently administered in patients with positive nodes and higher T stages. Chemotherapy regimens varied considerably over the time span of this study, reflecting common community practice patterns. Despite administration of chemotherapy in these patients, the distant metastasis-free survival rate was only 71% at 5 years in this group of triple negative breast cancer patients. In contrast, the distant metastasis-free survival rate in patients receiving chemotherapy who were not triple negative (n 151) was 83% at 5 years. Given the retrospective nature of this study and the relatively small numbers of patients in this subset analysis, clinical interpretation of this finding is limited. DISCUSSION Classification of breast cancers into basal type (triple negative), luminal, and HER2/neu has recently been proposed as a classification scheme based on gene expression profiles. 1-4 It has been demonstrated that this classification scheme has prognostic significance and implications with respect to response to therapy. 1-4,9 Although the more sophisticated hierarchical clustering analysis forms the basis of this classification scheme, currently available immunohistochemical markers can be used to closely replicate the more complex gene clustering analysis. Specifically, based on ER, PR, and HER2/neu analysis, tumors can similarly be classified into the categories of basal-like (negative for all three markers), luminal (ER/PR receptor positive), or HER2/neu. This classification scheme, using triple negative to characterize patients at high risk for systemic disease, is widely used. In the current study, we evaluated 482 patients with conservatively treated breast cancer in whom all three markers were available to validate the prognostic utility of this classification scheme and to determine whether triple negative breast cancers have a more aggressive locoregional relapse rate. One potential weakness of the current study is the lack of more detailed molecular profiling and hierarchical cluster analysis. Although there may be significant overlap and gene profiling analysis may modify the group that patients may fall into, there is general agreement that the majority of triple negative breast cancers based on simple analysis of ER, PR, and HER2/neu can be categorized as basal-like. 1,3,4,6,9,10,12,15 Another potential weakness of the study is unavoidable selection biases in a retrospective series such as this. For the current study, only patients who had available ER, PR, and HER2/neu data were included. Specifically, the 482 patients in this study included a higher percentage of patients younger than age 50 years than the overall population of 1,990 patients and a higher percentage of patients with T2 tumors. Although the overall survival of the 482 patients did not significantly differ from that of the overall patient population, the distant metastasis rate of these 482 patients was significantly higher than the distant metastasis rate of the overall population (16% v 10% at 5 years, respectively). This would indicate that patients included in the current study seemed to have biologically more aggressive disease than the Probability of Survival (%) Other Triple ve P =.002 Probability of Survival (%) Triple negative P = Distant Metastasis-Free Survival (years) Fig 1. Distant metastasis-free survival as a function of subtype Breast Relaspe-Free Survival (years) Fig 3. Breast relapse-free survival as a function of subtype

5 Haffty et al Variable Table 3. Triple Negative Group by Chemotherapy Received Adjuvant Chemotherapy (n 77) No. of Patients % Did Not Receive Adjuvant Chemotherapy (n 40) No. of Patients % T stage.002 T T Nodal status.003 Negative Positive Unknown Type of chemotherapy CMF Doxorubicin ATC 5 6 Other Outcomes at 5 years DMFS NS CSS NS Abbreviations: CMF, cyclophosphamide, methotrexate, fluorouracil; ATC, doxorubicin, docetaxel, and cyclophosphamide; DMFS, distant metastasis free survival; NS, not significant; CSS, cause-specific survival. overall population of 1,990 patients. Although we acknowledge this potential selection bias, all patients who had all three molecular markers available were included in this study, and no patients with invasive cancers who had all three markers available were excluded. Among all patients included, the triple negative cohort still had a poorer prognosis than the non triple negative cohort, but the two cohorts had a similar local relapse-free survival. Another weakness of the study is that the HER2/neu data was obtained through immunohistochemical testing and not by FISH. Although there is agreement that HER2/neu 3 patients are positive and 0 to 1 patients are negative, those patients who are 2 ideally should be tested by FISH. Our subset analysis, which demonstrates that the results were consistent whether HER2/neu 2 patients were classified as positive or negative, provides some reassurance, but ideally, these results should be reconfirmed with HER2/neu amplification data. It is notable that we were able to validate that triple negative breast cancer patients, even among this group of relatively early-stage cancer patients treated with breast-conserving therapy and radiation, have a poor prognosis, which was confirmed in multivariate regression analysis. It is reassuring that these patients do not seem to have a higher rate of local relapse and, therefore, should not be considered poor candidates for breast-conserving therapy. In fact, the local relapse rates among the triple negative cohort were nearly identical to those of the non triple negative cohort. This is of particular importance given the fact that triple negative patients are significantly P younger, and younger women have been shown to have a higher rate of local relapse compared with older women The fact that these women with triple negative tumors had similar local relapse rates indicates that these tumors are not radiation resistant. This is consistent with the fact that these basal-like tumors are theoretically responsive to DNA-damaging agents and should, therefore, be relatively radiation sensitive. Despite the fact that we included in this study nearly 500 patients with more than 5 years of follow-up, larger patient cohorts with even longer follow-up will be required to further validate these data. It is also notable that, despite the use of adjuvant systemic chemotherapy, triple negative patients seemed to have a poor prognosis, with a distant metastasis-free survival rate of only 71% at 5 years. The interpretation of these findings remains debatable given the relatively small number of patients and the retrospective nature of this study, and further studies evaluating the impact of adjuvant systemic therapy in triple negative patients is warranted. Although neoadjuvant studies have demonstrated that these cancers do respond well to chemotherapy, alternative chemotherapy programs have been considered in triple negative patients. 5,9,15 More extensive subtyping through molecular profiling may help to identify patients who may be more responsive to specific systemic regimens. By definition of their lack of receptors for ER, PR, and HER2/ neu, patients with triple negative tumors are not candidates for adjuvant hormonal therapy or trastuzumab. 16 Basal-like tumors have been shown to overexpress HER1 (epidermal growth factor), and patients with these tumors may be candidates for prospective studies evaluating targeted antibodies against epidermal growth factor that are already in clinical use. 11,15 Whether alternative forms of chemotherapy for these breast cancer patients, with or without biologic modifiers, will prove superior can only be addressed by well-designed prospective studies. Clearly, additional prospective and retrospective studies are warranted to further refine prognosis and optimize treatment in patients with triple negative breast cancers. We have demonstrated here, using the simple commonly available markers of ER, PR, and HER2/neu, that patients with triple negative breast cancers have a relatively poor prognosis, with a poorer distant metastasis-free, disease-free, and cause-specific survival. Overall survival was not significantly different, and larger data sets with even longer follow-up will be required to determine the impact of this classification on overall survival. Given the small number of nodal relapses, it is difficult to determine the clinical significance of the difference in nodal control. It is notable that, in these women treated with lumpectomy followed by radiation therapy, there was not a higher rate of local relapse, indicating that such patients are suitable candidates for breastconserving therapy. However, the high rate of distant metastasis, even among this cohort of women presenting with early-stage disease, reinforces the need to evaluate, in well-designed prospective studies, alternative forms of systemic therapy, including the use of targeted molecular therapies. REFERENCES 1. Perou CM, Sorlie T, Eisen MB, et al: Molecular portraits of human breast tumours. Nature 406: , Sorlie T: Molecular portraits of breast cancer: Tumour subtypes as distinct disease entities. Eur J Cancer 40: , Sorlie T, Perou CM, Tibshirani R, et al: Gene expression patterns of breast carcinomas distinguish tumor subclasses with clinical implications. Proc Natl Acad Sci U S A 98: , Brenton JD, Carey LA, Ahmed AA, et al: Molecular classification and molecular forecasting of breast cancer: Ready for clinical application? J Clin Oncol 23: , JOURNAL OF CLINICAL ONCOLOGY

6 Triple Negative Breast Cancer 5. Rouzier R, Perou CM, Symmans WF, et al: Breast cancer molecular subtypes respond differently to preoperative chemotherapy. Clin Cancer Res 11: , Sotiriou C, Neo SY, McShane LM, et al: Breast cancer classification and prognosis based on gene expression profiles from a population-based study. Proc Natl Acad Sci U S A 100: , Charafe-Jauffret E, Ginestier C, Monville F, et al: How to best classify breast cancer: Conventional and novel classifications. Int J Oncol 27: , Charafe-Jauffret E, Ginestier C, Monville F, et al: Gene expression profiling of breast cell lines identifies potential new basal markers. Oncogene 25: , Jacquemier J, Ginestier C, Rougemont J, et al: Protein expression profiling identifies subclasses of breast cancer and predicts prognosis. Cancer Res 65: , Jacquemier J, Padovani L, Rabayrol L, et al: Typical medullary breast carcinomas have a basal/myoepithelial phenotype. J Pathol 207: , Nielsen TO, Hsu FD, Jensen K, et al: Immunohistochemical and clinical characterization of the basal-like subtype of invasive breast carcinoma. Clin Cancer Res 10: , Sorlie T, Tibshirani R, Parker J, et al: Repeated observation of breast tumor subtypes in independent gene expression data sets. Proc Natl Acad Sci U S A 100: , Weigelt B, Hu Z, He X, et al: Molecular portraits and 70-gene prognosis signature are preserved throughout the metastatic process of breast cancer. Cancer Res 65: , Brenton JD, Aparicio SA, Caldas C: Molecular profiling of breast cancer: Portraits but not physiognomy. Breast Cancer Res 3:77-80, Matos I, Dufloth R, Alvarenga M, et al: P63, cytokeratin 5, and P-cadherin: Three molecular markers to distinguish basal phenotype in breast carcinomas. Virchows Arch 447: , Romond EH, Perez EA, Bryant J, et al: Trastuzumab plus adjuvant chemotherapy for operable HER2-positive breast cancer. N Engl J Med 353: , Turner N, Tutt A, Ashworth A: Hallmarks of BRCAness in sporadic cancers. Nat Rev Cancer 4: , Turner N, Tutt A, Ashworth A: Targeting the DNA repair defect of BRCA tumours. Curr Opin Pharmacol 5: , Perez EA, Roche PC, Jenkins RB, et al: HER2 testing in patients with breast cancer: Poor correlation between weak positivity by immunohistochemistry and gene amplification by fluorescence in situ hybridization. Mayo Clin Proc 77: , Bartelink H, Horiot JC, Poortmans P, et al: Recurrence rates after treatment of breast cancer with standard radiotherapy with or without additional radiation. N Engl J Med 345: , de la Rochefordiere A, Asselain B, Campana F, et al: Age as prognostic factor in premenopausal breast carcinoma. Lancet 341: , Fowble BL, Schultz DJ, Overmoyer B, et al: The influence of young age on outcome in early stage breast cancer. Int J Radiat Oncol Biol Phys 30:23-33, Harrold EV, Turner BC, Matloff ET, et al: Local recurrence in the conservatively treated breast cancer patient: A correlation with age and family history. Cancer J Sci Am 4: , Recht A, Connolly JL, Schnitt SJ, et al: The effect of young age on tumor recurrence in the treated breast after conservative surgery and radiotherapy. Int J Radiat Oncol Biol Phys 14:3-10, Vicini FA, Recht A: Age at diagnosis and outcome for women with ductal carcinoma-in-situ of the breast: A critical review of the literature. J Clin Oncol 20: , 2002 Authors Disclosures of Potential Conflicts of Interest The authors indicated no potential conflicts of interest. Author Contributions Conception and design: Bruce G. Haffty, Michael Reiss, Thomas Kearney, Qifeng Yang, Susan A. Higgins, William Hait, Deborah Toppmeyer Administrative support: Bruce G. Haffty Provision of study materials or patients: Bruce G. Haffty, Qifeng Yang, Susan A. Higgins Collection and assembly of data: Bruce G. Haffty, Qifeng Yang Data analysis and interpretation: Bruce G. Haffty, Thomas Kearney, Qifeng Yang, Deborah Toppmeyer Manuscript writing: Bruce G. Haffty, Thomas Kearney, Qifeng Yang, Susan A. Higgins, Joanne Weidhaas, Lyndsay Harris, Deborah Toppmeyer Final approval of manuscript: Bruce G. Haffty, Michael Reiss, Qifeng Yang, Susan A. Higgins, Joanne Weidhaas, Lyndsay Harris, William Hait, Deborah Toppmeyer