Should a Routine Metastatic Workup Be Performed for all Patients with Pathologic N2/N3 Breast Cancer?

Should a Routine Metastatic Workup Be Performed for all Patients with Pathologic N2/N3 Breast Cancer? Quyen D Chu, MD, FACS, Amanda Henderson, MD, Roger H Kim, MD, J Karen Miller, MA, Gary Burton, MD, Fred Ampil, MD, Benjamin DL Li, MD, FACS BACKGROUND: STUDY DESIGN: Node-positive breast cancer patients are at risk for metastatic disease. A routine metastatic workup might or might not be necessary for all patients with N2 or N3 diseases. The National Comprehensive Cancer Network guidelines recommend a metastatic workup for patients with T3N1 disease, yet no definitive recommendations are made for N2/N3 diseases. We hypothesized that for patients with operable pathologic N2/N3 diseases, a metastatic workup should only be considered for patients with T3/T4 lesions. Two hundred and fifty-six patients with pathologic N2/N3 diseases were identified from a prospective breast cancer database of 1,329 patients with stage 0 to III breast cancer. A metastatic workup included chest x-rays, bone scans, CT scans, and PET scans. Primary end point was incidence of stage IV disease at the time of diagnosis or within 1 month of definitive surgery. Statistical analysis included chi-square test, independent t-test, Kaplan-Meier Survival method, log-rank test, and Cox proportional hazard model.apvalue 0.05 was considered statistically significant. RESULTS: There were 158 patients with N2 disease (62%) and 98 with N3 disease (38%). Overall, 16% had stage IV disease (N2 15%, N3 16%). There was no significant difference in age (p 0.37), tumor size (p 0.89), tumor grade (p 0.09), estrogen-receptor status (p 0.23), or progesterone-receptor status (p 0.35) between the N2 and N3 groups. Incidences of stage IV disease were T0/T1, 0%; T2, 6%; T3, 22%; and T4, 36%. Multivariate analysis demonstrated that only T stage (p 0.0006) and grade (p 0.026) were independent predictors of overall survival. CONCLUSIONS: A metastatic workup is only indicated for N2/N3 patients with T3 or T4 primary lesions. (J Am Coll Surg 2012;214:456 462. 2012 by the American College of Surgeons) Breast cancer patients with regional lymph node involvement are at risk for concurrent distant metastases. 1 Therefore, an extensive metastatic workup is often obtained. The workup often includes a combination of imaging modalities, such as chest x-rays, bone scans, liver ultrasounds, abdominal CT scans, PET/CT scans, and MRIs, with the Disclosure Information: Authors have nothing to disclose. Timothy J Eberlein, Editor-in-Chief, has nothing to disclose. Presented at the Southern Surgical Association 123 rd Annual Meeting, Hot Springs, VA, December 2011. Received December 8, 2011; Accepted December 15, 2011. From the Departments of Surgery (Chu, Henderson, Kim, Miller, Li), Medical Oncology (Burton), Radiology (Ampil), and the Feist-Weiller Cancer Center (Chu, Kim, Burton, Li), Louisiana State University Health Sciences Center in Shreveport, Shreveport, LA. Correspondence address: Quyen D Chu, MD, FACS, Department of Surgery, Louisiana State University Health Sciences Center-Shreveport and the Feist-Weiller Cancer Center, 1501 Kings Hwy, Shreveport, LA 71130. email: qchu@lsuhsc.edu goal of identifying patients who have incurable disease and would receive palliative therapy only. During the past several years, there has been an ongoing debate about the use of these workups, especially in the context of cost containment. 2-4 These radiologic studies have the potential to add considerable time and expense to a patient s care, with unclear benefit. In addition, the concern with falsepositive results, which can lead to additional unnecessary test and procedures as well as increased patient anxiety, is not trivial. The National Comprehensive Cancer Network (NCCN) guidelines on the management of breast cancer recommend a metastatic workup for patients with T3N1 disease. 5 There are no definitive recommendations for a metastatic workup for patients who have pathologic N2 or N3 disease, irrespective of T stage. Currently, there are no compelling data to support a routine metastatic workup in these patients. Intuitively, one might suspect that, at the time of diagnosis, a high proportion of N2/N3 patients will 2012 by the American College of Surgeons ISSN 1072-7515/12/$36.00 Published by Elsevier Inc. 456 doi:10.1016/j.jamcollsurg.2011.12.014

Vol. 214, No. 4, April 2012 Chu et al Workup for N2/N3 Breast Cancer 457 have distant disease and, therefore, a metastatic workup is warranted. However, this might not necessarily be the case. We hypothesized that for patients with N2/N3 disease, a metastatic workup should only be considered for those with T3 or T4 tumors and not for those with T0, T1, or T2 lesions. METHODS We performed a retrospective analysis using a prospectively maintained breast cancer database that was created in 1998. Patients with operable pathologic N2 and N3 breast cancer diagnosed between September 1998 and September 2010 were analyzed after Institutional Review Board approval. Approximately 90% of patients were treated at the Feist-Weiller Cancer Center under the auspices of the Louisiana State University Health Sciences Center-Shreveport, and the remaining patients were treated at EA Conway Hospital, a sister safety-net hospital in Monroe, Louisiana. Of the 1,329 patients in the database, 256 (20%) patients with N2/N3 diseases were identified. Patients were staged based on the American Joint Committee on Cancer 6 th Edition. 6 Chest x-ray was obtained for all patients. Bone scans were performed in 62 patients (24%), PET scans in 39 patients (15%), and CT scans in 78 patients (30%). Only CT scans performed within 1 month of diagnosis were included for analysis, as this modality was not routinely ordered in the earlier years. Treatment and surveillance protocols were standardized to ensure study homogeneity. Standard treatment protocols for adjuvant and neoadjuvant chemotherapy and radiation therapy were offered to all qualified patients. Adjuvant chemotherapy consisted of Doxorubicin 60 mg/m 2 and Cyclophosphamide 600 mg/m 2 on day 1 (AC regimen), both repeated every 21 days for 4 cycles, followed by paclitaxel 175 mg/m 2 for 4 cycles, or after 2004, every 14 days with granulocyte colony-stimulating factor for bone marrow support. Patients with T3 or T4 lesions generally received the AC phase as neoadjuvant therapy. After neoadjuvant AC regimen, each patient was restaged with a complete history, physical examination, repeat laboratory tests consisting of whole blood count and liver function tests, and additional imaging directed by any abnormal findings. Patients without progression to systemic disease underwent a definitive surgery 2 weeks after completion of preoperative chemotherapy. The surgery was either breastconservation therapy (consisting of breast-conservation therapy, lumpectomy with tumor-free margin, axillary lymph node dissection, and breast irradiation) or a modified radical mastectomy. The majority of patients underwent a modified radical mastectomy. All final specimens had negative margins. After completion of adjuvant chemotherapy, patients received adjuvant external beam radiation therapy, including the internal mammary chain (as indicated by imagingdetected adenopathy), the supraclavicular fossa, and ipsilateral breast or anterior chest wall. The standard radiation dosage was 50 Gy, with an additional 10 to 15 Gy boost to the tumor bed (as considered appropriate, ie, positive surgical margins). Patients whose tumors were HER2/neu-positive received Trastuzumab at 6 mg/kg, which was initiated with paclitaxel and continued every 3 weeks for 1 year. Patients with estrogen-receptor and/or progesterone-receptor positive tumors received Tamoxifen or an aromatase inhibitor for 5 years after completion of radiation therapy. The pathologic diagnosis was made by standardized hematoxylin and eosin staining. HER-2 status was determined by fluorescent in situ hybridization. Estrogen-receptor and progesterone-receptor status was determined using immunohistochemical methods. Slides were stained and evaluated using the Dako Autostriker and the automated cellular imaging system. Activity 5% was considered positive. Patients were all offered a similar surveillance regimen. The minimum surveillance protocol consisted of a history, physical examination, complete blood counts, and chemistry every 3 months for 3 years, every 6 months in years 4 and 5, and annually thereafter. Chest x-ray and mammogram were obtained annually. Subsequent radiologic and/or histologic evaluation was performed based on the patient s findings. Clinical data were accrued and recorded prospectively and included age at diagnosis, stage, nodal status, grade, tumor size, and receptor status. The primary end point was the incidence of stage IV disease, either at the time of diagnosis or within 1 month after definitive surgery. Stage IV disease was defined as any radiologic evidence of distant metastases, as determined by bone scan, CT scan, and/or PET scan. The cases were discussed in a multidisciplinary tumor board and a general consensus was reached on whether the radiologic findings were consistent with the presence of metastatic disease. Biopsy for tissue confirmation of metastasis was not a requirement for determination of stage IV disease, but was performed on the majority of patients. MedCalc software (Microsoft, Inc) was used to perform statistical analyses. The mean age at diagnosis, grade, node status, stage, and tumor size were analyzed using the independent sample t-tests, and categorical data were analyzed with chi-square test. Overall survival (OS) was calculated from the date of surgery to the date of death from any cause or date of last follow-up. Survival analysis was performed using Kaplan-Meier method, the log-rank test was used to compare the curves, and Cox proportional hazard regres-

458 Chu et al Workup for N2/N3 Breast Cancer J Am Coll Surg Table 1. Distribution of Patient and Clinicopathologic Characteristics of N2/N3 Breast Cancer Patients N2 (n 158; 62%) N3 (n 98; 38%) p Value Characteristics Mean age, y (range) 58.7 (27 86) 57.2 (31 84) 0.37 Tumor size, cm 0.89 Overall mean 4.6 Overall median 4.0 Mean 4.65 4.60 Median 3.5 4.50 Mean tumor grade 2.5 2.4 0.12 Tumor grade, n (%) 0.09 1 1/94 (1) 1/62 (2) 2 46/94 (49) 35/62 (56) 3 47/94 (50) 26/62 (42) T stage, n (%) 0.94 T0 (n 2; 0.8%) 1 (1) 1 (1) T1 (n 35; 14%) 21 (13) 14 (15) T2 (n 105; 41%) 68 (43) 37 (38) T3 (n 50; 20%) 30 (19) 20 (20) T4 (n 64; 25%) 38 (24) 26 (27) Receptor status, n (%) ER-positive 55/121 (45) 42/76 (55) 0.23 PR-positive 41/121 (34) 32/77 (42) 0.35 HER-2 positive 27/103 (26) 31/72 (43) 0.03 ER, estrogen receptor; PR, progesterone receptor. sion models were applied for multivariate analysis. There was no violation of the underlying assumption of the Cox proportional hazards. Risk ratios and 95% confidence intervals were calculated from the model. Apvalue 0.05 was considered statistically significant. Table 3. Likelihood of Having Stage IV Disease Based on T and N Stage Stage IV Group stage n n % T0N2 1 0 0 T0N3 1 0 0 T1N2 21 0 0 T1N3 14 0 0 T2N2 68 3 4 T2N3 37 3 8 T3N2 30 8 27 T3N3 20 3 15 T4N2 38 14 37 T4N3 26 9 35 Table 2. Distribution of Stage III, Stage IV Disease Based on T and N Stage Stage III Stage IV Characteristics n % n % T stage T0 (n 2) 2 100 0 0 T1 (n 35) 35 100 0 0 T2 (n 105) 99 94 6 6 T3 (n 50) 39 78 11 22 T4 (n 64) 41 64 23 36 N distribution N2 (n 158) 134 84 24 15 N3 (n 98) 82 83 16 16 RESULTS Two hundred and fifty-six patients with pathologic N2/N3 diseases were identified. The median follow-up time was 45.5 months. Thirty patients (15%) received neoadjuvant chemotherapy. Table 1 provides the distribution of patients and their clinicopathologic characteristics. There were 61 (30%) patients who underwent breast-conservation therapy and 141 (70%) patients who underwent mastectomy. Thirteen patients had a palliative mastectomy. Of the clinicopathologic parameters examined, mean age of diagnosis (p 0.37), tumor grade (p 0.12), T-stage (p 0.94), estrogenreceptor status (p 0.23), and progesterone-receptor status (p 0.35) were not statistically different between the N2 and N3 groups. There was a significantly higher proportion of N3 patients (43%) with HER-2 positive tumors than N2 patients (26%) (p 0.03). Table 2 demonstrates the distribution of stage III and stage IV diseases based on T and N stage. There were 2 patients with T0 lesions who presented with axillary lymphadenopathy. Both were deemed to have occult breast cancer and underwent modified radical mastectomy. No patient with T0 or T1 disease had metastatic disease. Patients with T2 lesions had a 6% chance of having distant metastases. However, this risk increased dramatically from 22% for T3 disease to 36% for T4 disease. Overall risk of having stage IV disease in patients with N2/N3 was 16%. Table 3 demonstrates the likelihood of N2/N3 patients having stage IV disease based on T and N stage. The likelihood of having stage IV disease for T2N2 and T2N3 was 4% and 8%, respectively. This increased to 15% for T3 lesions and 37% for T4 lesions (T3N2 27%, T3N3 15%, T4N2 37%, and T4N3 35%). The 5-year OS for stage III and IV disease was 61% and 36%, respectively (p 0.01) (Fig. 1). Median OS was 164 months for the stage III group and 46 months for the stage IV group. The 5-year OS between the N2 and N3 groups was not significantly different (Fig. 2); the 5-year OS for

Vol. 214, No. 4, April 2012 Chu et al Workup for N2/N3 Breast Cancer 459 Figure 1. Overall survival curve for stage III and IV breast cancer (256 patients with N2/N3 breast cancer). The 5-year overall survival for stage III and IV disease was 61% and 36%, respectively (p 0.01). The median survival was 164 months for the stage III group and 46 months for the stage IV group. the N2 and N3 groups was 61% and 55%, respectively (p 0.71); median survival was 163 months for N2 disease and 164 months for N3 disease. Cox proportional hazard model was used to compare T stage, grade, age at diagnosis, and N stage for risk of death (Table 4). Note that T stage (p 0.006) and grade (p 0.026) were the only independent predictors of OS, and N stage (p 0.25) and age at diagnosis (p 0.26) were not. DISCUSSION It is accepted that patients with early breast cancer do not require a routine metastatic workup; the likelihood of having a metastasis is 5%. 7-9 However, for patients who are considered to be at high risk for metastatic disease developing, there is a lack of consensus about whether a metastatic workup should be done at diagnosis. The high-risk group is generally thought to include those with nodepositive disease and/or large primary tumors (T3/T4). 1 The current NCCN practice guidelines recommend a metastatic workup for patients with T3N1 disease, yet for those with N2/N3 disease, a metastatic workup is considered optional. 5 The level of evidence for these recommendations was classified as either category 2A (the recommendation is based on lower-level evidence and there is uniform NCCN consensus) or category 2B (the recommendation is based on lower-level evidence and there is nonuniform NCCN consensus, but no major disagreement). Although there are some data to support the need for a routine metastatic workup for T3N1 disease, 10,11 to our knowledge there are no compelling data to support such a practice for patients with N2/N3 disease. Figure 2. Overall survival curve for N2 and N3 breast cancer (256 patients with N2/N3 breast cancer). The 5-year overall survival for N2 and N3 disease was 61% and 55%, respectively (p 0.71). The median survival was 163 months for N2 disease and 164 months for N3 disease. This is unfortunate because surgeons and their medical colleagues are commonly faced with the dilemma of whether to pursue a metastatic workup for patients with pathologic N2/N3 disease. Compounding this dilemma is that the majority of these patients are asymptomatic and, therefore, relying on symptoms to direct a metastatic workup runs the risk of understaging patients. Because of the perception that N2/N3 patients are at a high risk for having concomitant distant disease, most oncologists will opt for a routine metastatic workup. Whether such a workup is necessary for all patients with N2/N3 disease has not been evaluated and was therefore the focus of this study. We demonstrated that not all patients with N2/N3 disease have distant disease, and that T stage is a useful barometer to select patients who might require additional metastatic workup. For patients with T0, T1 and T2 diseases, the incidence of stage IV disease was 0%, 0%, and 6%, respectively. This incidence increases with higher T stage; 22% for T3 and 36% for T4 tumors. Given these intriguing data, one might argue that T3/T4 disease should be considered for a routine metastatic workup, and this consideration for those with T0 to T2 lesions should be based on other considerations (eg, grade, symptoms). Patients with T2 lesions represent an interesting subgroup. One might expect a higher incidence of patients with stage IV disease but, in fact, only 6 (6%) of 105 patients had stage IV disease. In reviewing the 6 patients who had metastatic disease, we could not discern what was uniquely different about them when compared with the other 99 patients who did not have metastatic disease. Perhaps other factors, such as tumor grade, play an important

460 Chu et al Workup for N2/N3 Breast Cancer J Am Coll Surg Table 4. Overall Survival for Patients with N2/N3 Breast Cancer (Cox Proportional Hazard Model) Adjusted hazard ratio 95% CI p Value T stage 1.64 1.24 2.18 0.0006 Grade 1.81 1.08 3.02 0.026 Age at diagnosis 1.01 0.99 1.03 0.26 N stage 1.37 0.80 2.32 0.25 role. However, our small population and low incidence of event does not allow confirmation. We believe that this might be a worthwhile issue to pursue in larger datasets. Detection of metastases is heavily dependent on radiologic studies. The American Society of Clinical Oncology and NCCN guidelines recommended CT, PET, PET/CT, bone scan, and/or MRI for high-risk patients. Morris and colleagues recently reported that PET/CT is the preferred imaging modality for detecting osseous metastases in patients suspected of having metastatic disease. 4 Whether PET/CT was also sufficient for detecting nonosseous metastases was not reported. Port and colleagues reported on the utility of using PET for high-risk patients with operable breast cancer. 3 In their study of 80 patients with T3 and/or clinically N1/2, they found that although PET and conventional imaging were equally sensitive in detecting metastatic disease, PET had fewer false-positive results. 3 The incidence of metastatic disease was approximately 12% for all patients, although the exact incidence for each T and N grouping was not defined. Despite these emerging data, which strongly support the use of the newer imaging modalities such as PET, none provided definitive data that will preclude the need for conventional modalities (ie, bone scan). Consequently, the optimal strategy for detecting metastases in these high-risk patients is not known. 3-5,12 The task of analyzing the role and contribution of each radiologic modality relative to their radiologic counterparts can be a formidable task. This task becomes more complicated because not all physicians order the same battery of radiologic tests and the pattern of practice changes with time as newer modalities are introduced and older modalities are abandoned. For example, at our institution, we abandoned obtaining a routine bone scan when reports of its low reliability to detect metastases surfaced, 11,13 and such results were confirmed in our practice. 14 Given the lack of an optimal strategy for detecting metastases and the difficulties encountered in clinical practice, the true incidence of stage IV disease in patients with N2/N3 is not known. Until a carefully designed clinical trial can be conducted to address this question, our retrospective data are useful because they provide a focal point for others to begin discussing its merits and validate its accuracy. It is important to note that the purpose of our study was not to provide a comparative analysis of the different radiologic modalities (ie, bone scan, CT scan, or PET scan) and their accuracy or utility in detecting occult metastases; rather, the objective was to determine if the incidence of stage IV disease in N2/N3 patients warrants any form of metastatic workup after definitive surgery. The lack of a standardized method of detecting distant metastases helps explain the low number of CT scans that were performed in our patient group; only 30% of patients had a CT scan ordered within 1 month of diagnosis. Although this lack is a limitation of our study, we believe that our data are likely to be reflective of the general practice patterns for this patient population and are also reflective of the lack of consensus on the optimal diagnostic approach for these patients. Despite this, we found that only 9% of patients (16 of 179 patients) who did not have a CT scan performed within 1 month of diagnosis were found to have stage IV disease within 6 months of diagnosis. We are planning to conduct a prospective study for all patients with N2/N3 disease with a standardized metastatic workup to determine the true incidence of distant disease in this patient population. The determination of who should or should not undergo a metastatic workup is necessary because of the potential downsides of unnecessary metastatic workups. These downsides include increased health care costs, delays in treatment, complications from intravenous contrast agents, increased patient anxiety, additional tests/procedures, and inappropriate upstaging. The inappropriate upstaging of a patient s disease based on a false-positive result is a particularly troubling possibility, as this could lead to treatment with palliative intent rather than curative intent. In this challenging economic environment, health care personnel should be cost-conscious to optimize care for their patients and lower cost. Although we did not perform a specific cost analysis in our study, we do recognize some of the issues. Given a cost estimate of US$1,500 for an abdominal CT scan and $1,500 to $4,000 for PET scans, 15 the potential cost savings in the workup of the approximately 20,000 stage III breast cancer patients diagnosed annually in the United States is not insignificant, especially considering that more than half of the patients in our study might have been spared a metastatic workup based on their T stage. CONCLUSIONS Our data suggest that only patients with T3/T4 and pathologic N2/N3 disease should undergo a metastatic workup. Based on our data and recent NCCN guidelines, we summarize the following: any patients with T3 N-positive disease or T4 N2/N3 disease should undergo a metastatic workup. To our knowledge, the T4N1 subgroup remains the only high-risk subgroup that requires needed data to determine whether a metastatic workup is warranted. Our group is currently performing such an analysis.

Vol. 214, No. 4, April 2012 Chu et al Discussion 461 Author Contributions Study conception and design: Chu, Henderson, Burton, Ampil Acquisition of data: Chu, Henderson, Miller Analysis and interpretation of data: Chu, Henderson, Kim, Burton, Ampil, Li Drafting of manuscript: Chu, Henderson, Kim, Burton, Ampil, Li Critical revision: Chu, Kim, Miller, Burton, Ampil, Li REFERENCES 1. Chu QD, McDonald JC, Li BDL. Adjuvant therapy for patients with node-positive breast cancer. In: Cameron JL, ed. Advances in surgery. New York: Mosby; 2006:77 98. 2. Aberle D, Chiles C, Gatsonis C, et al. Imaging and cancer: research strategy of the American College of Radiology Imaging Network. Radiology 2005;235:741 751. 3. Port E, Yeung H, Gonen M, et al. 18 F-2-fluoro-2-deoxy-dglucose positron emission tomography scanning affects surgical management in selected patients with high-risk, operable breast carcinoma. Ann Surg Oncol 2006;13:677 684. 4. Morris P, Lynch C, Feeney J, et al. Integrated positron emission tomography/computed tomography may render bone scintigraphy unnecessary to investigate suspected metastatic breast cancer. J Clin Oncol 2010;28:3154 3159. 5. National Comprehensive Cancer Network. Clinical Practice Guidelines in Oncology for Breast Cancer. V.1. Fort Washington, PA: National Comprehensive Cancer Network; 2010. 6. American Joint Committee on Cancer. AJCC Cancer Staging Manual. 6 th ed. Chicago, IL: American Joint Committee on Cancer; 2002. 7. Samant R, Ganguly P. Staging investigations in patients with breast cancer: the role of bone scans and liver imaging. Arch Surg 1999;134:551 553, discussion 554. 8. Weir L, Worsley D, Bernstein V. The value of FDG positron emission tomography in the management of patients with breast cancer. Breast J 2005;11:204 209. 9. Newman L. The search for occult metastatic disease in breast cancer patients: how far should we go? Ann Surg Oncol 2006; 13:604 606. 10. Ravaioli A, Pasini G, Polselli A, et al. Staging of breast cancer: new recommended standard procedure. Breast Cancer Res Treat 2002;72:53 60. 11. Puglisi F, Follador A, Minisini A, et al. Baseline staging tests after a new diagnosis of breast cancer: further evidence of their limited indications. Ann Oncol 2005;16:263 266. 12. Khatcheressian J, Wolff A, Smith T, et al. American Society of Clinical Oncology 2006 update of the breast cancer follow-up and management guidelines in the adjuvant setting. J Clin Oncol 2006;24:5091 5097. 13. Hamaoka T, Madewell J, Podoloff D, et al. Bone imaging in metastatic breast cancer. J Clin Oncol 2004;22:2942 2953. 14. Ampil F, Li B, Burton G, et al. Is baseline bone imaging in high-risk breast cancer patients justified? Breast J (in press). 15. Berger M, Gould M, Barnett P. The cost of positron emission tomography in six United States Veterans Affairs hospitals and two academic medical centers. AJR Am J Roentgenol 2003;181:359 365. Discussion DR WILLIAM C WOOD (Atlanta, GA): I congratulate Dr Chu on yet another original contribution to the field of breast cancer presented at this society. To my knowledge, the observation that he makes with these data is unique. Twenty years ago, when bone marrow transplantation for ablation of breast cancer with aggressive chemotherapy and then bone marrow transplant was in vogue, a lot of staging studies were undertaken for potential transplant patients, with the somewhat surprising finding that there was more metastatic disease that they could discover than they had expected. And it led to an enthusiasm for bone scans and PET/CT scans in people with breast cancer of multiple nodepositive stage disease. Today we are on the cusp of a genomic revolution that is redefining breast cancer totally. But still, we are using many of these same ways of looking at our breast cancer patients. Stage is a snapshot in time. The cancer may have grown very slowly to its present state of size and level of nodal involvement, or it may have galloped. The rate of progress is very important to prognosis and is suggested by the Ki-67 level or the grade of the tumor, but not at all by the stage. N2 or N3 disease usually tells us a great deal about tumor prognosis and the likelihood of additional metastases. For this level of spread to be achieved by a tumor that is very small suggests a very aggressive tumor in terms of its ability to spread and metastasize. On the other hand, the same level of nodal disease from a very large tumor may have progressed slowly over a long time. But if the primary is still quite small, it seems intuitive that any systemic metastases from this tumor may still be pretty small also. In N2 disease, are people with T3 or T4 primary tumors more likely to have systemic metastases, or simply more likely to have metastases that have grown large enough that we can find them radiologically? What confidence do you have in recommending that N2/N3 patients can be sorted based on T size, as your conclusions suggest, in an original observation, in a relatively small study, where you don t see a difference between N2, N3, N4, and N2/N3 prognosis, as we typically see in very large data sets? Are we risking an observation based on the play of chance? In this era of excellent treatment for breast cancer at any stage, is there any reason to detect asymptomatic metastatic breast cancer? Not to do so would save an enormous amount of money nationally with these staging studies. It would avoid the terror of a false-positive staging study for a person. And it would avoid a longer period of awareness of having stage IV asymptomatic disease in a person who could otherwise be blissfully unaware of that. DR STEPHEN GROBMYER (Gainesville, FL): The identification of low-value medical procedures or tests, such as the authors have done in this paper, should be a priority in our efforts to maintain sustainability of our health care system. In this manuscript, the authors investigate retrospectively the yield of a metastatic workup in patients with pathologic N2 or N3 disease. All patients in the study had a chest x-ray, with or without other studies, including a PET

462 Chu et al Discussion J Am Coll Surg scan, bone scan, or CT scan. They find that the incidence of finding metastatic disease is very low in the subset of patients with T1 or T2 tumors. I have several comments and questions for the authors. Do you think your results might differ if you used a standard metastatic imaging protocol for all patients on the study? Based on your results, do you routinely omit a chest x-ray for all patients with T1 or T2 breast cancers? Further, along these lines, do your results suggest that we might alter our long-term follow-up of asymptomatic patients who present initially with metastatic or locally advanced breast cancer? Do you feel that there is a preferred metastatic workup for patients in whom a metastatic workup should be performed? Certainly, it would appear there s no reason to perform a bone scan, CT scan, and PET scan on a single patient. How do you handle patients who present with palpable adenopathy at the time of presentation? Do you recommend a metastatic workup in these patients before multidisciplinary treatment planning? At our institution, the finding of metastatic disease prompts a change in systemic and local regional therapy for many patients. In fact, antiestrogen therapy is commonly the treatment that is first used in patients with metastatic estrogen receptor-positive breast cancer. Does the finding of metastasis guide therapy changes in your patients at your institution? Finally, in breast cancer and in many other solid tumors, we are moving from an anatomic basis to a molecular basis of staging. It would be interesting to perform your type of analysis using your patient set and a molecular scoring system such as the Oncotype DX test. Have you considered this approach? DR MARK KELLY (Nashville, TN): I have one brief question, which echoes one of Dr Grobmyer s questions. It appears from the presentation that you actually looked at the pathologic stage of the lymph nodes and used that as your end point, which, of course, is not known until after the surgical staging, which occurs after the scans. So have you looked at the clinical status of the lymph nodes as a determinant, and how useful are scans in patients with clinically palpable nodes? DR QUYEN CHU (Shreveport, LA): Dr Wood, if I understood your question correctly, I believe that you are striking at the very heart of the issue, which is the biology of the disease. For instance, if you have a patient who has a big, bulky primary tumor, you would expect this patient to be at a high risk of having distant disease. But then again, there are the occasional patients with very small tumors (ie, T0 or T1 tumor) who also have N2/N3 disease. Why is this? So I think that our knowledge of the biology of the disease is still not complete. I also think that, with time and with advanced understanding of genomics and other molecular platforms, we ll be able to answer your question with better precision. Your second question is what confidence do we have, given our small data set, to recommend sorting N2/N3 patients based on T-size? Your question is very poignant and you are absolutely right. This paper by no means suggests that we should change our current standard of care. However, what it does suggest is that we should begin questioning the behavior of N2/N3 disease. I think that a lot of us have a preconceived notion that patients with N2/N3 disease have such a high risk of having distant disease that we automatically perform a full metastatic workup. However, based on our data, this may not necessarily be the case. There is a subset of patients who actually are the true stage III rather than stage IV disease. And finally, you asked, so what if one detects metastatic disease in an asymptomatic patient? I think that s a very valid question. I think that as surgeons we like to look at things in a linear manner. However, I think that if we were to talk to our medical colleagues, they would say that it does matter whether patients have stage III or stage IV diseases because there are newer agents or medications that could potentially extend their survival maybe for months at a time. So I think it is important to be as precise with staging as possible. Also, recent data, albeit controversial, suggest that the behavior of stage IV disease may not be homogeneous, and that if you aggressively resect distant disease in selected patients with stage IV patients, you may actually improve overall survival. For Dr Grobmyer s question, we perform the chest x-ray routinely preoperatively to avoid anesthesia delay. The second question is, do we suggest standardizing the metastatic workup for all of these patients? That s a very good question. In the ideal world, the answer would be yes. But I think getting our medical colleagues on board with this concept would be difficult. Also, in our institution, we use CT of the chest, abdomen, and pelvis for all patients with N2 and N3 disease. Certainly, we would need to review this practice in light of our recent results. The exact role of the different imaging modalities in the workup of patients with N2/N3 disease would need to be carefully studied, perhaps in a clinical trial. As to the question about how we handle patients who have palpable adenopathy, do we recommend performing a metastatic workup in all these patients? It s an interesting question because these patients represent a heterogenous group. Certainly, for a patient with bulky lymphadenopathy along with a large tumor, we are more inclined to proceed with a metastatic workup. On the other hand, if a tumor is very small, we might preclude the metastatic workup until we get the final pathology after a definitive operation. Dr Grobmyer s final question was in regard to using molecular signature to prognosticate. Our laboratory has reported the prognostic and predictive roles of eif4e and CXCR4 and found that these factors were indeed independent predictors of outcomes for patients with locally advanced breast cancer. The person who got me started on translational research is Dr Benjamin Li, who is now chairman of our program. He has created an excellent breast cancer database and we are continually reaping the benefits. Finally, Dr Kelly asked, if I understood his question correctly, how we addressed patients a priori. In other words, can we determine based on clinical examination of the lymph nodes whether a metastatic workup should be used. I think, again, such a decision will need to be tailored. For instance, if you have a patient with a very big, bulky tumor and a palpable lymph node, that patient is likely going to have distant disease, up to 30% chance, according to our data. And that may be the person on whom you might want to proceed with a full metastatic workup to see if there is indeed distant disease. Why? Because if such a patient has stage IV disease, we might want to reconsider whether aggressive surgical extirpation should be the way to go.