Biopsy type does not influence sentinel lymph node status

Biopsy type does not influence sentinel lymph node status

The American Journal of Surgery 190 (2005) 551–556 Presentation Biopsy type does not influence sentinel lymph node status Anees B. Chagpar, M.D., M...

82KB Sizes 1 Downloads 29 Views

The American Journal of Surgery 190 (2005) 551–556

Presentation

Biopsy type does not influence sentinel lymph node status Anees B. Chagpar, M.D., M.Sc.a,*, Charles R. Scoggins, M.D.a, Sunati Sahoo, M.D.b, Robert C. Martin, II, M.D.a, David J. Carlson, M.D.c, Alison L. Laidley, M.D.d, Souzan E. El-Eid, M.D.e, Terre Q. McGlothin, M.D.d, Robert D. Noyes, M.D.f, Phillip B. Ley, M.D.g, Todd M. Tuttle, M.D.h, Kelly M. McMasters, M.D., Ph.D.a, for the University of Louisville Breast Sentinel Lymph Node Study a

Department of Surgery, University of Louisville, 315 E. Broadway, Suite 312, Louisville, KY 40202, USA b Department of Pathology, University of Louisville, Louisville, KY, USA c St. Mary’s Medical Center and Deaconess Hospital, Evansville, IN, USA d Breast Surgeons of North Texas, Dallas, TX, USA e Hudson Valley Surgical, Kingston, NY, USA f Latter Day Saints Hospital, Salt Lake City, UT, USA g Surgical Clinic Associates, Jackson, MI, USA h Park Nicollet Clinic, Minneapolis, MN, USA Manuscript received June 5, 2005; revised manuscript June 10, 2005

Presented at the Sixth Annual Meeting of the American Society of Breast Surgeons, Los Angeles, California, March 16 –20, 2005

Abstract Background: This study sought to determine whether the type of biopsy examination independently affects sentinel lymph node (SLN) status in breast cancer patients. Methods: A prospective multicenter study of patients who had SLN biopsy examination followed by axillary node dissection was analyzed to determine whether the type of biopsy examination influenced SLN status. Results: Of the 3853 patients studied, 32% had a positive SLN. Patients were diagnosed by fine-needle (N ⫽ 293), core-needle (N ⫽ 2154), excisional (N ⫽ 1386), or incisional (N ⫽ 20) biopsy procedures. The rates of SLN positivity for these groups were 45%, 32%, 29%, and 65%, respectively (P ⬍ .001). Other factors predictive of SLN status included: patient age (P ⬍ .001), tumor size (P ⬍ .001), tumor palpability (P ⬍ .001), number of SLN removed (P ⬍ .001), type of surgery (mastectomy vs. lumpectomy) (P ⬍ .001), histologic subtype (P ⫽ .048), and the use of immunohistochemistry (P ⬍ .001). All of these factors remained significant in the multivariate model except for histologic subtype and biopsy examination type. Conclusions: Biopsy examination type does not independently influence the risk for nodal metastasis. © 2005 Excerpta Medica Inc. All rights reserved. Keywords: Breast cancer; Biopsy type; Sentinel node status

It previously has been speculated that tumor manipulation may lead to dissemination of viable cancer cells [1–9]. Although many have found this concept controversial, a recent study found that needle biopsy examination of breast cancer was associated with a higher frequency of sentinel lymph node (SLN) metastasis [10]. In the current era of minimally invasive * Corresponding author. Tel.: ⫹1-502-629-6950; fax: ⫹1-502-6293183. E-mail address: [email protected]

techniques for diagnosing, staging, and treating breast cancer patients, the implications of this finding are significant. Currently, core-needle biopsy examination, including examination with vacuum-assisted devices, is the diagnostic biopsy method of choice for many clinicians [11]. This allows the clinician to obtain an accurate preoperative pathologic diagnosis and ancillary studies for hormone receptor status using a minimally invasive office procedure, without the need for an experienced cytopathologist or a more extensive open biopsy procedure. If

0002-9610/05/$ – see front matter © 2005 Excerpta Medica Inc. All rights reserved. doi:10.1016/j.amjsurg.2005.06.009

552

A.B. Chagpar et al. / The American Journal of Surgery 190 (2005) 551–556

such biopsy techniques truly are associated independently with biologically significant lymph node metastasis, practice patterns may need to change away from this technique. We therefore sought to determine if the type of biopsy has a significant impact on sentinel node status independent of other clinicopathologic variables in a large prospective multicenter study.

Methods The University of Louisville Breast Sentinel Lymph Node Study is a multi-institutional prospective study involving more than 300 surgeons and more than 4,000 patients. This study was approved by the institutional review board at each site, and all patients were required to sign an informed consent form before their participation. Patients with clinical stage T1-2, N0 breast cancer were included in the study and underwent an SLN biopsy examination followed by completion axillary node dissection. None of these patients received neoadjuvant chemotherapy. Biopsy examination type and SLN biopsy technique was left to the discretion of the individual surgeon or institution. From May 7, 1998, to August 2, 2004, there were 4,131 patients enrolled. An SLN was identified in 3,882 of these patients. The biopsy examination type used for preoperative diagnosis was noted in all but 29 patients. The remaining 3,853 patients therefore formed the cohort of interest for this study. Patient demographic information, clinicopathologic variables of the primary tumor, and biopsy examination type were collected in a prospective fashion, along with the SLN pathologic result. The method of evaluating the sentinel node (using either routine hematoxylin-eosin staining alone or in combination with immunohistochemical stains) was left to the discretion of the individual site, and was recorded prospectively. Statistical analyses were performed to determine factors influencing sentinel node positivity using SPSS software (version 11.0; Chicago, IL). Univariate analyses were performed by using likelihood ratio tests for discrete variables and Mann-Whitney U tests for continuous variables. Binary logistic regression was used for the multivariate analysis. Significance was set at a P value of .05.

Table 1 Clinicopathologic variables Characteristic

Number of cases (%)

Tumor size* ⱕ2 cm ⬎2 to 5 cm ⬎5 cm Palpable Yes No Histologic subtype† Ductal Lobular Surgery type‡ Lumpectomy Mastectomy

2293 (60) 1295 (34) 109 (3) 2056 (53) 1797 (47) 3175 (82) 323 (8) 2567 (67) 1244 (32)

* Tumor size not specified in 156 (4%) patients. † Histologic subtype not specified in 355 (9%) patients. ‡ Surgery type not specified in 42 (1%) patients.

Results The median age of the patients enrolled in this study was 60 years (range, 27–100 years) and the median tumor size was 1.5 cm (range, .1–11.0 cm). The clinicopathologic features of the primary tumors in this cohort of patients are shown in Table 1. The median number of sentinel nodes removed was 2 (range, 1–18). A positive sentinel node was found in 1,240 (32%) patients. In 2,215 (57.5%) cases, immunohistochemistry was used to analyze the SLNs. All patients underwent an axillary node dissection, with a median of 11 nonsentinel nodes being removed (range, 0 – 42). Of the 3,853 patients in this study, 293 (8%) were diagnosed by fine-needle aspiration biopsy examination, 2,154 (56%) were diagnosed by core-needle biopsy examination, 1,386 (36%) were diagnosed by excisional biopsy examination, and 20 (⬍1%) were diagnosed by incisional biopsy examination. Biopsy examination type was correlated significantly with overall positive lymph node status (Table 2). As one would expect, the correlation between biopsy examination type and sentinel node positivity closely mirrored that with overall nodal status. Incisional biopsy examination was correlated with the highest rate of lymph node positivity, followed by fine-needle and core-needle biopsy examination. Excisional biopsy examination was associated with the low-

Table 2 Biopsy procedure type and lymph node positivity Number of patients (%)

Overall lymph node SLN positive Non-SLN positive Only 1 lymph node positive FNA ⫽ fine-needle aspiration.

P value

FNA

Core

Excisional

Incisional

144 (49) 133 (45) 70 (24) 59 (41)

747 (35) 691 (32) 321 (15) 357 (48)

439 (32) 403 (29) 163 (12) 233 (53)

13 (65) 13 (65) 5 (1) 7 (54)

⬍.001 ⬍.001 ⬍.001 .066

A.B. Chagpar et al. / The American Journal of Surgery 190 (2005) 551–556

553

Table 3 Factors associated with choice of biopsy procedure type Number of patients (%)

Patient age ⬍60 y ⱖ60 y Tumor size ⱕ2 cm ⬎2 to 5 cm ⬎5 cm Palpable Yes No Location Upper outer Upper inner Lower inner Lower outer Central Histologic subtype Ductal Lobular Surgery type Lumpectomy Mastectomy

P value

FNA

Core

Excisional

Incisional

134 (53) 150 (47)

1111 (52) 1023 (48)

630 (46) 737 (54)

13 (65) 7 (35)

115 (40) 163 (57) 10 (3)

1248 (60) 771 (37) 66 (3)

924 (71) 349 (27) 31 (2)

6 (30) 12 (60) 2 (10)

259 (88) 34 (12)

1053 (49) 1101 (51)

726 (52) 660 (48)

18 (90) 2 (10)

154 (54) 40 (14) 16 (6) 29 (10) 48 (17)

1129 (53) 301 (14) 160 (8) 235 (11) 295 (15)

671 (49) 202 (15) 117 (9) 171 (13) 197 (15)

6 (32) 1 (5) 1 (5) 5 (26) 6 (32)

201 (91) 19 (9)

1809 (91) 186 (9)

1154 (91) 112 (9)

11 (65) 6 (35)

188 (65) 102 (35)

1437 (67) 701 (33)

940 (69) 424 (31)

2 (11) 17 (89)

.002

⬍.001

⬍.001

.098

.029

⬍.001

FNA ⫽ fine-needle aspiration.

est rate of lymph node positivity. If this trend was related to tumor disruption in the first 3 techniques, one might expect there to be a higher rate of non-SLN metastasis in patients who had an excisional biopsy procedure rather than a needle biopsy procedure because the latter may have displaced cancer cells to only the first draining lymph nodes and not beyond. Instead, we found that patients who underwent either fine-needle or core-needle biopsy procedures had a higher rate of non-SLN positivity than those who underwent an excisional biopsy procedure (Table 2). Furthermore, in node-positive cases, we analyzed how often only 1 lymph node was involved with metastasis, and compared this between each of the biopsy techniques. If needle biopsy procedures disrupted tumors and shed cells that were displaced to the lymph nodes, one would expect the number of cases with only 1 lymph node positive to be higher in the needle biopsy procedure groups. Instead, we found that there was no statistically significant difference between the groups (P ⫽ .066); in fact, the excisional biopsy procedure group had a higher rate of metastasis to a single lymph node than either the fine-needle aspiration biopsy procedure group or the core-needle biopsy procedure group (Table 2). The choice of which biopsy procedure type to use in any given patient is dependent on a number of clinicopathologic variables that also may impact SLN status. We therefore investigated which factors were associated with choice of biopsy examination type (Table 3). Biopsy examination type was associated with patient age, tumor size, palpability, and histologic subtype. The effect of various clinicopathologic variables on SLN status also was examined by using

univariate analysis (Table 4). This showed that patient age, tumor size, tumor palpability, histologic subtype, primary tumor location, type of surgery, number of SLNs removed, use of immunohistochemistry, and biopsy examination type all were associated significantly with sentinel node positivity. To determine whether biopsy examination type affected SLN status independent of these clinicopathologic variables, a multivariate analysis was performed. All of the variables found to be significant on univariate analysis remained significant in the multivariate model with the exception of histologic subtype and biopsy examination type (Table 5).

Comments The concept of tumor violation leading to potential seeding is not new. A number of studies have shown tumor cell displacement along biopsy needle tracts in breast cancer [1–9] and in a number of other malignancies [12–15]. However, the biologic significance of such displaced tumor cells has been debatable. Although some have attributed local recurrences to shed tumor cells from percutaneous biopsy procedures [16,17], others have found that the rate of local recurrence does not vary with biopsy procedure type [18,19]. In fact, removal of the biopsy procedure tract generally is felt to be unnecessary [4]. Diaz et al [6] found that the rate of finding displaced tumor cells decreased with increasing time intervals from core biopsy procedure to excision, suggesting that such cells do not survive displacement. We

554

A.B. Chagpar et al. / The American Journal of Surgery 190 (2005) 551–556

Table 4 Univariate analysis of factors associated with a positive SLN Characteristic

Biopsy procedure type FNA Core Excisional Incisional Patient age ⬍60 y ⱖ60 y Tumor size ⱕ2 cm ⬎2 to 5 cm ⬎5 cm Palpable Yes No Histologic subtype Ductal Lobular Location Upper outer quadrant Upper inner quadrant Lower inner quadrant Lower outer quadrant Central Surgery type Lumpectomy Mastectomy Number of SLN removed ⬍2 ⱖ2 Use of immunohistochemistry Yes No

Number of patients with a positive SLN (%)

P value

⬍.001 133 (45) 691 (32) 403 (29) 13 (65) ⬍.001 684 (36) 536 (28) ⬍.001 539 (24) 594 (46) 74 (68) ⬍.001 838 (41) 402 (22) .048 1035 (33) 123 (38) ⬍.001 631 (32) 126 (23) 87 (30) 162 (37) 207 (38) ⬍.001 728 (28) 497 (40) ⬍.001 357 (29) 883 (35) ⬍.001 536 (24) 681 (42)

previously showed that the theoretical risk for tumor seeding along a percutaneous biopsy procedure tract did not affect margin status in women choosing to undergo breastconserving surgery [20]; some even have found that it may improve margin negativity by allowing surgeons to plan their resection more appropriately [21]. As patients become increasingly involved as decision makers in their own care and surgical options for the management of breast cancer increase, the importance of accurate preoperative histologic diagnosis in surgical planning becomes increasingly evident. Core-needle biopsy examination is a minimally invasive office procedure that allows for appropriate histopathologic and hormone-receptor status information to be obtained without the need for an additional surgical procedure. This procedure is cost effective, simple to perform, well tolerated by patients, and can be performed under image guidance for nonpalpable tumors. For many clinicians, core-needle biopsy examination has become the procedure of choice in the diagnosis of breast cancer [11,12]. The recent study by Hansen et al [10], which found an association between needle biopsy examination and sentinel

node metastasis, raised the issue of whether tumor disruption by percutaneous biopsy procedures could in fact lead to biologically relevant nodal metastasis. In their study of 663 patients, biopsy procedure type was found to be associated with SLN positivity independent of patient age, tumor size, and tumor grade (P ⫽ .04) [10]. The purpose of the current larger study was to evaluate the validity of their results. We found that biopsy examination type was associated with SLN metastasis on univariate analysis. However, on closer analysis, it is clear that the choice of biopsy technique is correlated with a number of other clinicopathologic factors that influence nodal status. Similar to Morrow et al [22], we found that choice of biopsy procedure was correlated with patient age and ultimate surgery choice (breast conservation vs. mastectomy). Furthermore, we found that biopsy examination type correlated with tumor size, palpability, and histologic subtype—all of which were significant factors influencing SLN status. Interestingly, Hansen et al [10] found that the majority of metastases to a sentinel node were macrometastasis, even in patients who had a needle biopsy procedure. Although no explanation was offered for their results, we found similarly that the majority of sentinel node metastasis were found on hematoxylin-eosin staining and not with immunohistochemical techniques. This suggests that the metastases found were not related to the needle biopsy procedure (where one would expect isolated tumor cells found on immunohistochemistry alone), but rather true metastases related to the underlying tumor itself. This supports our finding on multivariate analysis that the type of biopsy procedure was not an independent predictor of sentinel node metastasis. Unlike Hansen et al [10], we found biopsy procedure type on univariate analysis to be correlated significantly with nonsentinel node positivity as well. This would seem to be related more to the differences in the clinicopathologic features of the underlying tumors associated with each type of biopsy procedure, rather than the biopsy examination type itself. In addition, we found that there were no significant differences between biopsy examination types in terms of how often only 1 SLN was positive. If SLN metastasis

Table 5 Multivariate analysis of factors influencing SLN positivity Factor

Odds ratio (95% confidence interval)

P value

Patient age Tumor size Tumor palpability Tumor location* Surgery type Number of SLN removed Use of immunohistochemistry Histologic subtype Biopsy type*

.984 (.977–.990) 1.505 (1.394–1.626) 1.548 (1.302–1.841)

⬍.001 ⬍.001 ⬍.001 ⬍.001 ⬍.001 .001 ⬍.001 .532 .139

1.388 (1.167–1.651) 1.075 (1.029–1.124) 2.403 (2.046–2.822) 1.093 (.827–1.445)

* Odds ratios not given for tumor location and biopsy procedure type because of multiple comparisons.

A.B. Chagpar et al. / The American Journal of Surgery 190 (2005) 551–556

were the result of tumor displacement, one would expect the rate of a single positive SLN to be higher in the groups who had percutaneous needle biopsy procedures. The difference found in the significance of biopsy procedure type on multivariate analyses between the current study and that of Hansen et al [10] can be explained in a number of ways. To begin with, the current study was larger, permitting an increased power to resolve any issues of association that may be present. In addition, although the study by Hansen et al [10] included patients treated with neoadjuvant therapy thereby allowing a significant amount of time between the biopsy procedure and sentinel node biopsy examination for lymph node metastasis to occur naturally, the current study did not. The multivariate analysis by Hansen et al [10] only took into consideration patient age, tumor size, grade, and biopsy examination type; the current analysis is much more exhaustive. We found that many other factors besides age, tumor size, and biopsy examination type were significant predictors of SLN positivity on univariate analysis. These included tumor location, histologic subtype, and surgery type, all of which also were correlated with biopsy examination type. Although our study is limited by the lack of histologic grade, if we were to perform a multivariate analysis based only on age, tumor size, and biopsy examination type, we too would find that biopsy examination type is statistically significant in this model (P ⫽ .004) because biopsy examination type would be functioning as a surrogate for these other factors. Therefore, the inclusion of multiple additional factors that were significant on univariate analysis makes our multivariate analysis more robust. From a biologic standpoint, it is questionable whether isolated displaced tumor cells from a biopsy procedure tract would harness heretofore unused metastatic potential, gain access to the lymphatic channels, and become macrometastatic tumor deposits of significant biologic importance in the relatively short period of time between a preoperative diagnostic biopsy procedure and a definitive surgical procedure. Carter et al [23] described a phenomenon of benign epithelial transport in which cells (either benign or malignant) were transported to regional lymph nodes after either core-needle (N ⫽ 3) or excisional biopsy (N ⫽ 12) procedures. This suggests that any manipulation (regardless of biopsy procedure technique) could be responsible for epithelial displacement into a lymph node. The fact that benign epithelial elements were transported to lymph nodes leads one to speculate that such displacement simply may be the result of mechanical forces without any biologic relevance [23,24]. In addition, massage of the breast during SLN mapping also has been implicated in benign epithelial displacement [25]. Studies of patients with ductal carcinoma in situ with no evidence of microinvasion have found cytokeratin-positive cells in SLNs in up to 23% of patients [26]. With a mean follow-up period of 5 years, all were alive and free of any evidence of recurrence. Clearly, these epithelial cells appearing in lymph nodes do not carry with them the

555

same prognostic implications as true metastatic lesions. Thus, although it may be possible for cells to be displaced, they are not likely to have any significant biologic ramifications. In fact, the revised American Joint Committee on Cancer Staging system has classified such cells as “isolated tumor cells of no known significance” and placed them into the node-negative category [27]. Given the tremendous advantage of percutaneous biopsy examination in terms of surgical planning and decision making, and our findings that biopsy procedure type is not associated independently with SLN positivity, clinicians should continue to offer their patients percutaneous biopsy examinations without undue concern of promoting metastatic disease. We look forward to the results of the recently closed American College of Surgeons Oncology Group Z0010 trial and the National Surgical Adjuvant Breast and Bowel Project B-32 trial, which will elucidate the true significance of isolated tumor cells in SLNs and bone marrow aspirates found on immunohistochemistry.

Acknowledgments The authors would like to thank the Center for Advanced Surgical Technology, Norton Hospital, and the University of Louisville for their support. A complete list of investigators in the University of Louisville Breast Sentinel Lymph Node Study is provided in Wong SL, Chao C, Edwards MJ, et al. Frequency of sentinel lymph node metastases in patients with favorable breast cancer hystologic subtypes. Am J Surg 2002;184:496 – 8.

References [1] Harter L, Curtis J, Ponto G, et al. Malignant seeding of the needle track during stereotaxic core needle breast biopsy. Radiology 1992; 185:713– 4. [2] Youngson B, Cranor M, Rosen P. Epithelial displacement in surgical breast specimens following needling procedures. Am J Surg Pathol 1994;18:896 –903. [3] Youngson B, Lieberman L, Rosen P. Displacement of carcinomatous epithelium in surgical breast specimens following stereotaxic core biopsy. Am J Surg Pathol 1995;103:598 – 602. [4] Hoorntje LE, Schipper ME, Kaya A, et al. Tumor cell displacement after 14G breast biopsy. Eur J Surg Oncol 2004;30:520 –5. [5] Stolier A, Skinner J, Levine EA. A prospective study of seeding of the skin after core biopsy of the breast. Am J Surg 2000;180:104 –7. [6] Diaz LK, Wiley EL, Venta LA. Are malignant cells displaced by large-gauge needle core biopsy of the breast? AJR Am J Roentgenol 1999;173:1303–13. [7] Bott S, Mohsen V, Wells C, et al. Needle tract metastases in breast cancer. Eur J Surg Oncol 1999;25:553. [8] Grabau Dam Andersen JA, Graversen HP, Dyreborg U. Needle biopsy of breast cancer. Appearance of tumor cells along the needle track. Eur J Surg Oncol 1993;19:192– 4. [9] Lieberman L, Vuolo M, Dershaw DD, et al. Epithelial displacement after stereotactic 11-gauge directional vacuum-assisted breast biopsy. AJR Am J Roentgenol 1999;172:677– 81.

556

A.B. Chagpar et al. / The American Journal of Surgery 190 (2005) 551–556

[10] Hansen NM, Ye X, Grube BJ, et al. Manipulation of the primary breast tumor and the incidence of sentinel metastases from invasive breast cancer. Arch Surg 2004;139:634 – 40. [11] Crowe JP, Rim A, Patrik R, et al. A prospective review of the decline of excisional breast biopsy. Am J Surg 2002;184:353–5. [12] Mighell AJ, High AS. Histological identification of carcinoma in 21 gauge needle tracks after fine needle aspiration biopsy of head and neck carcinoma. J Clin Pathol 1998;51:241–3. [13] John TG, Garden OJ. Needle track seeding of primary and secondary liver carcinoma after percutaneous liver biopsy. HPB Surg 1993;6:199 –203. [14] Karcioglu ZA, Gordon RA, Karcioglu GL. Tumor seeding in ocular fine needle aspiration biopsy. Ophthalmology 1985;92: 1763–7. [15] Ayar D, Golla B, Lee JY, et al. Needle-track metastasis after transthoracic needle biopsy. J Thorac Imaging 1998;13:2– 6. [16] Thurfjell MG, Jansson T, Nordgren H, et al. Local breast cancer recurrence caused by mammographically guided punctures. Acta Radiol 2000;41:435– 40. [17] Chao C, Torosian MH, Boraas MC, et al. Local recurrence of breast cancer in the stereotactic core needle biopsy site: case reports and review of the literature. Breast J 2001;7:124 –7. [18] Chen AM, Hafty BG, Lee CH. Local recurrence of breast cancer after breast conservation therapy in patients examined by means of stereotactic core-needle biopsy. Radiology 2002;225: 707–12.

[19] King TA, Hayes DH, Cederborn GJ, et al. Biopsy technique has no impact on local recurrence after breast conserving therapy. Breast J 2001;7:19 –24. [20] Chagpar AB, Martin RC, Hagendoorn LJ, et al. Lumpectomy margins are affected by tumor size and histologic subtype but not by biopsy technique. Am J Surg 2004;188:399 – 402. [21] King TA, Cederborn J, Champaign JL, et al. A core breast biopsy diagnosis of invasive carcinoma allows for definitive treatment planning. Am J Surg 1998;176:497–501. [22] Morrow M, Venta L, Stinson T, et al. Prospective comparison of stereotactic core biopsy and surgical excision as diagnostic procedures for breast cancer patients. Ann Surg 2001;233:537– 41. [23] Carter BA, Jensen RA, Simpson JF, et al. Benign transport of breast epithelium into axillary lymph nodes after biopsy. Am J Clin Pathol 2000;113:259 – 65. [24] Rogers LW, Ries SG. Another variable in lymph node biopsy. Am J Clin Pathol 2000;114:293. [25] Diaz NM, Vrcel V, Centeno BA, et al. Modes of benign mechanical transport of breast epithelial cells to axillary lymph nodes. Adv Anat Pathol 2005;12:7–9. [26] Tamhane R, Dahlstrom JE, McCallum DD, et al. The clinical significance of cytokeratin-positive cells in lymph nodes at the time of mastectomy from patients with ductal carcinoma in situ. Ann Surg Oncol 2002;9:999 –1003. [27] Singletary SE, Allred C, Ashley P, et al. Revision of the American Joint Committee on Cancer staging system for breast cancer. J Clin Oncol 2002;20:3628 –36.