Axillary recurrence after a tumour-negative sentinel node biopsy in breast cancer patients: A systematic review and meta-analysis of the literature

Results Forty-eight studies concerning 14 959 sentinel node-negative breast cancer patients followed for a median of 34 months were selected. Sixty-seven patients developed an axillary recurrence, resulting in a recurrence rate of 0.3%. The sensitivity of the sentinel node biopsy was 100%. Uni- and multivariable variable analyses showed that the lowest recurrence rates were reported in studies performed in cancer centres, in studies that described the use of 99m Tc-sulphur colloid, and also when investigators used the superficial injection technique or evaluated the harvested sentinel nodes with haematoxylin–eosin and immunohistochemistry staining ( p < 0.01). Conclusions In this systematic literature review, the axillary recurrence rate in sentinel node-negative patients is 0.3%, which is well within the desired range. The median sensitivity of the procedure appears to be as high as 100%. The recurrence rate is influenced by the differences in the lymphatic mapping technique. Keywords Breast neoplasms False negative reactions Lymphatic metastasis Sentinel lymph node biopsy Recurrence Introduction Sentinel node biopsy is widely used as a minimally invasive staging alternative to routine axillary lymph node dissection in patients with breast cancer. The morbidity of sentinel node biopsy is less than that of axillary lymph node dissection and the procedure may improve staging of the axilla. 1 Sentinel node biopsy quickly became the standard of care in clinically node-negative breast cancer patients, before consensus on the technique was reached and before randomized studies showed a similar or decreased axillary recurrence rate. In 1999, two comprehensive reviews were published of studies on sentinel node biopsy followed by routine axillary lymph node dissection to assess the learning phase performance. 2,3 These reviews demonstrated median false-negative rates of 7% and 5% (range 0–40%), which is not within the 5% range that is generally considered acceptable. A critical look at the studies revealed that the pathologists never performed serial sectioning on the lymph nodes that were recovered from the remainder of the axilla and that the sensitive immunohistochemistry staining technique was used in only two studies. This suggests that published false-negative rates are too favourable. A substantial number of papers have since reported on patients in whom routine axillary lymph node dissection was omitted when the sentinel node was free of disease. 4 The duration of follow-up was limited in many of these studies, but a few long-term follow-up results have now been published which prompted this systematic literature review. The primary aims of this systematic review and meta-analysis were to determine the axillary recurrence rate when a completion dissection was omitted after a tumour-negative sentinel node biopsy, and to determine the sensitivity of the sentinel node procedure. The collected data were compared based on different lymphatic mapping techniques and different hospital settings in an attempt to identify factors associated with a favourable outcome. Methods Systematic review A broad search of the published literature based on the subject of this article was conducted using Pubmed, Embase, The Cochrane Library and DARE (Database of Abstract and Evidence-Based Medicine) ( Fig. 1 ). First, a systematic search was done with the following medical subject headings (MeSH) terms: breast neoplasms, sentinel lymph node biopsy, lymphatic metastasis, recurrence and false-negative reactions. We used the MeSH terms simultaneously by ‘OR’ and combined them with ‘AND’ in the syntax of the databases. The search in Pubmed and Embase was carried out in ‘all fields’ without any ‘limits’, and in The Cochrane Library in ‘title, abstract and keywords’. From Pubmed 364 articles were retrieved, from Embase 395, and in The Cochrane library and DARE 182 papers (articles and meeting abstracts) were found. The computer program Reference Manager (version 11, for Windows, 2004, Thomson ISI ResearchSoft) was used to exclude overlapping articles of the combined database searches, which resulted in a total of 376 papers. Second, a review was done of the cited references in the identified papers to make sure that no manuscripts were overlooked. Finally, an additional search was performed to identify relevant studies described in abstracts or in letters to the editor. Overall, 404 publications were retrieved by the systematic literature search. Sixty-two papers in a language other than Dutch or English and reviews without original data (41) were excluded. Based on the following criteria, 253 articles were excluded after analysis of their title (153) or abstract (100). The population had to be an original study group. This means that different papers based on the same group of patients were avoided and only the most recent study on identical patients was used. Selected papers were required to concern patients with breast cancer, with a clinically node-negative axilla, a tumour-negative sentinel lymph node biopsy and no subsequent axillary lymph node dissection. The papers had to report the presence or absence of a primary axillary recurrence after a specified time of follow-up. When tumour-positive sentinel node biopsies were also mentioned in the same article, the paper was not excluded from further analysis. In the end, 48 publications were selected that met the set inclusion criteria. Meta-analysis The following information was gathered from each paper: name of the first author, year of publication, type of institution, number of patients, type of radiopharmaceutical, its injection site, use of lymphoscintigraphy, use of blue dye and where it was injected, method of pathology analysis, number of patients with axillary recurrences, disease-free interval and mean or median follow-up duration. The overall and subgroup follow-up times were derived from studies that mentioned a median follow-up duration. The studies that only described mean numbers were documented as missing numbers. The axillary recurrence rate was calculated as a percentage of the patients with a tumour-negative sentinel node biopsy. The sensitivity was calculated by dividing the number of patients with a tumour-positive sentinel node by the sum of all patients with a negative sentinel node who developed a recurrence plus the number of patients with a positive sentinel node. This implies that studies that reported no recurrences have a sensitivity of 100%. A generalized linear model was applied to identify factors associated with the recurrence rate. Uni- and multivariable analyses of recurrence rates were conducted considering the type of hospital, the technique of lymphatic mapping, method of pathology analysis, and median follow-up time in months as independent factors. All analyses were weighted by the number of patients per study. In addition, subgroup analyses with studies with a median follow-up time of at least 3 years were performed to assess the effect of follow-up on the results. All p -values and 95% confidence intervals were two-sided. All analyses were performed in SPSS 15 (version 15, for Windows, SPSS Inc, Chicago, IL, USA). Results The 48 studies that met the selection criteria were published between 2001 and 2007, 11 publications were abstracts and one was a letter to the editor ( Tables 1 and 2 ). 5–52 The selected papers included 14 959 patients with a sentinel node biopsy that did not reveal metastasis and that was not followed by axillary node dissection. Sixty-seven patients developed a recurrence in their axilla during a follow-up time of 34 months. This results in an overall weighted recurrence rate of 0.3%. The published sensitivities ranged between 97% and 100%. The median weighted combined sensitivity of the sentinel node biopsy was calculated from the 20 publications that provided the required data and was 100%. 6,7,9,10,12,17,23,24,29,30,33–36,38,40,41,49,50,52 The median time-interval after which a recurrence occurred was 20 months with a range from 4 to 63 months ( Fig. 2 ). Lymphoscintigraphic imaging was carried out in 30 studies and a radiopharmaceutical injection and a gamma ray detection-probe were used in all studies except one. Blue dye was injected intraoperatively in 37 studies, mostly patent blue (Laboratoire Guerbet, Aulnay-Sous-Bois, France), or isosulfan blue (Lymphazurin 1%, Hirsch Industries, Inc., Richmond, VA). The blue dye was injected around the tumour in the majority of studies. A combination of a radiopharmaceutical injection, lymphoscintigraphy, blue dye injection and a gamma camera detection-probe was used in 24 studies. The influence of blue dye injection only versus the use of both blue dye and a radiopharmaceutical or the impact of the use of lymphoscintigraphy upon the recurrence rate could not be calculated in subgroup analyses due to a lack of required data. Table 3 shows the results of the uni- and multivariable analyses of the investigated independent factors that were statistically significant ( p < 0.01). The lowest recurrence rates were reported in studies performed in cancer centres, in studies that described the use of 99m Tc-sulphur colloid (CIS-US, Inc., Bedford, MA), and also when investigators used the superficial injection technique or evaluated the harvested sentinel nodes with haematoxylin–eosin and immunohistochemistry staining. Duration of follow-up was not statistically significant in both the uni- and multivariable analyses ( p = 0.70). All independent factors remained significant if the analyses were weighted for follow-up duration or included as a covariate in the model. In the subgroup analyses of studies with a median follow-up exceeding 3 years ( Table 3 ), the results showed similar results compared with the analyses of all studies. The only differences were that university hospitals had the lowest recurrence rate instead of cancer centres and that the use of 99m Tc-nanocolloid led to a lower recurrence rate than the use of 99m Tc-sulphur colloid. Discussion In this study, we analyzed 48 papers describing studies that concerned 14 959 breast cancer patients with a clinically node-negative axilla, a sentinel lymph node without established disease and no subsequent axillary lymph node dissection. The median overall weighted axillary recurrence rate is 0.3% after a median follow-up time of 34 months, which is a satisfying result. The median sensitivity of sentinel node biopsy to detect axillary metastasis is 100% (range 97–100%), an unexpected favourable percentage in view of the 60–100% range in the learning phase studies. 2 This high median sensitivity may be somewhat skewed because the 20 papers that provided the correct data to calculate the sensitivity happened to report relatively few recurrences or none. There were no signs of bias in the description of patient selection in these studies that could explain this favourable outcome. The low recurrence rate found in this review indicates that the procedure is accurate. Surgeons may be tempted to think that this is because of their expertise, but other explanations need to be considered as well. The radiotherapy that is part of the breast-conserving therapy is likely to clean up some of the tumour-positive nodes that surgeons may overlook and leave behind in the adjacent axilla. The same can be said for the adjuvant systemic treatment that many of these patients receive. The increasing use of preoperative ultrasound identifies patients with relatively large, although still not palpable, metastatic nodes resulting in direct axillary dissection. 53 Without ultrasound, such patients would have been subjected to sentinel node biopsy. The size of such metastases could then have prevented accumulation of tracer fluids, causing the node to pass unnoticed and leading to a recurrence later on. Moreover, several retrospective and prospective studies suggest that only about a third of the breast cancer metastases in the axilla go on to develop into clinically relevant disease. 13,54,55 The median overall follow-up time of almost 3 years calculated in the current study may be relatively short to draw definitive conclusions because of the protracted growth that typifies the disease. In one study of breast cancer patients who underwent routine axillary clearance, the median time to an axillary recurrence was 77 months (range 10–353 months), compared to 20 months in the present study. 56 A few remarks concerning the statistical methods used in the subgroup univariate analyses are in order. This is a review of studies with both prospective and retrospective data, making outcomes less reliable than when data were retrieved from prospective studies only. Other heterogeneous factors between the included studies are patient selection criteria, the definition of a sentinel node, the lymphoscintigraphy visualization rate and the number of retrieved sentinel nodes. These figures were not described in detail or not al all. 57 Since all calculations were made weighing the numbers of patients in each study, the large study of Naik et al. with 2340 patients may have had a dominating impact in the subgroup analyses. To test this, these 2340 patients were excluded from subgroup analyses, which led to the same statistically significant results as the original analyses except for two factors: deep injection became most favourable and university hospitals, instead of cancer centres, had the lowest recurrence rate. The way including or omitting the large study by Naik and coworkers swings the results of subgroup analysis one way or the other illustrates that the meaning of such subgroup calculations is questionable. 25 The same can be said for the different follow-up durations per study. The subgroup analysis of the data derived from the studies that mentioned a median follow-up time of at least 3 years differed from the subgroup analyses based on the data obtained from all studies. Although the differences were small, it does suggest that the results of subgroup analyses should be questioned and interpreted with restraint. Despite these limitations, several observations can be made regarding the subgroup analyses performed that demonstrate factors influencing the recurrence rate after a tumour-negative sentinel node biopsy. The observation that the results in community hospitals are good, but not as good as in university hospitals or dedicated cancer centres, seems solid. The fact that studies describing the injection of 99m Tc-sulphur colloid have a significantly lower recurrence rate than those in which 99m Tc-nanocolloid (Nanocoll ® , GE- Healthcare, Amersham, UK) is used might partly be explained because the more specialized high volume centres, like the Memorial Sloan Kettering Center, use the former radiopharmaceutical. Several papers describe the use of a filtered form of 99m Tc-sulphur colloid, which is similar to 99m Tc-nanocolloid in particle size. When adjusting for filtered or unfiltered 99m Tc-sulphur colloid compared to 99m Tc-nanocolloid in the subgroup analyses, the recurrence rates associated with these three instead of two types of radiopharmaceuticals were very similar. The particle size of the radiopharmaceuticals plays a role in depicting sentinel nodes as well. Unfortunately, most papers did not mention the particle size and therefore conclusions concerning its influence on the recurrence rate cannot be drawn. The fact that the additional use of frozen section analysis resulted in a significantly lower median recurrence rate than the use of routine haematoxylin–eosin and immunohistochemistry staining only was a remarkable result. Perhaps the pathologists who make the extra effort of doing frozen sections are the more ambitious ones and also make more slides afterwards. The results, however, do suggest that the former thought, that frozen sections lead to less tissue for postoperative pathologic observation and therefore a less accurate analysis, may be incorrect. Randomized trials comparing sentinel node staging with axillary node dissection are in progress and will shed more light on the recurrence rate. The 0.3% median recurrence rate that the current study established suggests that the results of these studies can be awaited with confidence. In the future, even more definitive conclusions might be drawn with a longer follow-up time, but it seems unlikely that the axillary recurrence rate after lymphatic mapping will surpass the 1–2.3% recurrence rate after routine axillary clearance in node-negative breast cancer patients. 58–60 Conflict of interest With this statement all authors mentioned in this review article disclose that there are no financial or personal relationships with other people or organisations that could inappropriately influence (bias) this work. Examples of such potential conflicts of interest include employment, consultancies, stock ownership, honoraria, paid expert testimony, patent applications/registrations, and grants or other funding. All these examples of possible conflicts are not present in the current manuscript. Acknowledgements The authors would like to thank Ninja Antonini of the Department of Biometrics for her support in the analysis of the collected data in SPSS 15. References 1 U. Veronesi G. Paganelli G. Viale A randomized comparison of sentinel-node biopsy with routine axillary dissection in breast cancer N Engl J Med 349 2003 546 553 2 O.E. Nieweg L. Jansen R.A. Valdes Olmos Lymphatic mapping and sentinel lymph node biopsy in breast cancer Eur J Nucl Med 26 1999 S11 S16 3 D.N. Krag The sentinel node for staging breast cancer: current review Breast Cancer 6 1999 233 236 4 O.E. Nieweg M.C. van Rijk R.A. Valdes Olmos C.A. Hoefnagel Sentinel node biopsy and selective lymph node clearance – impact on regional control and survival in breast cancer and melanoma Eur J Nucl Med Mol Imaging 32 2005 631 634 5 R.M. Roumen G.P. Kuijt I.H. Liem M.W. van Beek Treatment of 100 patients with sentinel node-negative breast cancer without further axillary dissection Br J Surg 88 2001 1639 1643 6 B.D. Badgwell S.P. Povoski S.F. Abdessalam Patterns of recurrence after sentinel lymph node biopsy for breast cancer Ann Surg Oncol 10 2003 376 380 7 I. Bedrosian B. Heimbach F. Spitz A prospective study of sentinel lymph node biopsy without axillary node dissection for staging patients with breast cancer Ann Surg Oncol 12 3 2002 S57 8 D.K. Blanchard J.H. Donohue C. Reynolds C.S. Grant Relapse and morbidity in patients undergoing sentinel lymph node biopsy alone or with axillary dissection for breast cancer Arch Surg 138 2003 482 487 9 P. Carcoforo D. Sortini G. Soliani E. Basaglia L. Feggi A. Liboni Accuracy and reliability of sentinel node biopsy in patients with breast cancer. Single centre study with long term follow-up Breast Cancer Res Treat 95 2006 111 116 10 M.A. Chung M.M. Steinhoff B. Cady Clinical axillary recurrence in breast cancer patients after a negative sentinel node biopsy Am J Surg 184 2002 310 314 11 A.Y. de Kanter M.M. Menke-Pluymers M.W. Wouters I. Burgmans A.N. van Geel A.M. Eggermont 5-Year follow-up of sentinel node negative breast cancer patients Eur J Surg Oncol 32 2006 282 286 12 A. Domenech A. Benitez M. Bajen Patients with breast cancer and negative sentinel node biopsy without additional axillary dissection: evolution after a 5-year follow-up Eur J Nucl Med Mol Imaging 33 S155 2006 69 80 13 J.M. Guenther N.M. Hansen L.A. DiFronzo Axillary dissection is not required for all patients with breast cancer and positive sentinel nodes Arch Surg 138 2003 52 56 14 A. Haid M. Knauer R. Koberle-Wuhrer Medium-term follow-up data after sentinel node biopsy alone for breast cancer Eur J Surg Oncol 32 2006 1180 1185 15 S. Imoto N. Wada K. Murakami T. Hasebe A. Ochiai S. Ebihara Prognosis of breast cancer patients treated with sentinel node biopsy in Japan Jpn J Clin Oncol 34 2004 452 456 16 J.S. Jeruss D.J. Winchester S.F. Sener Axillary recurrence after sentinel node biopsy Ann Surg Oncol 12 2005 34 40 17 N. Khakpour K.K. Hunt H.M. Kuerer Sentinel lymph node dissection provides axillary control equal to complete axillary node dissection in breast cancer patients with lobular histology and a negative sentinel node Am J Surg 190 2005 598 601 18 M.C. Kokke M.F. Ernst F. Eijckeler Axillary recurrence after negative sentinel node procedure in breast cancer patients Eur J Surg Oncol 32 Suppl 1 2006 S79 19 P. Kristen R. Lorenz M. Gross Axillary recurrences after sentinel node (SLN) biopsy without complete axillary dissection in breast cancer patients Br J Cancer 92 2005 981 20 I. Langer W.R. Marti U. Guller Axillary recurrence rate in breast cancer patients with negative sentinel lymph node (SLN) or SLN micrometastases: prospective analysis of 150 patients after SLN biopsy Ann Surg 241 2005 152 158 21 J. Leikola T. Saarto H. Joensuu Ultrasonography of the axilla in the follow-up of breast cancer patients who have a negative sentinel node biopsy and who avoid axillary clearance Acta Oncol 45 2006 571 575 22 J. Loza F. Colo J. Nadal M. Viniegra R. Chacon Axillary recurrence after sentinel node biopsy for operable breast cancer Eur J Surg Oncol 28 2002 897 898 23 A. Marrazzo P. Taormina M. David The role of sentinel lymph-node biopsy (SLNB) in the treatment of breast cancer Chir Ital 58 2006 299 304 24 T. Nagashima M. Sakakibara S. Nakano Sentinel node micrometastasis and distant failure in breast cancer patients Breast Cancer 13 2006 186 191 25 A.M. Naik J. Fey M. Gemignani The risk of axillary relapse after sentinel lymph node biopsy for breast cancer is comparable with that of axillary lymph node dissection: a follow-up study of 4008 procedures Ann Surg 240 2004 462 468 26 H. Paajanen V.M. Kilponen J. Heikkinen H. Sammalkorpi M. Helle Feasibility of axillary ultrasound in the quality assessment of sentinel node biopsy in breast cancer surgery Scand J Surg 95 2006 195 198 27 J.A. Palesty J.M. Foster T.C. Hurd N. Watroba H. Rezaishiraz S.B. Edge Axillary recurrence in women with a negative sentinel lymph node and no axillary dissection in breast cancer J Surg Oncol 93 2006 129 132 28 S. Pejavar L.D. Wilson B.G. Haffty Regional nodal recurrence in breast cancer patients treated with conservative surgery and radiation therapy (BCS + RT) Int J Radiat Oncol Biol Phys 66 2006 1320 1327 29 P. Pizzocaro F. Bertagna G. Bosio Long-term follow-up of breast cancer patients undergone to mastectomy without axillary dissection after negative sentinel lymph node (SLN) biopsy Eur J Nucl Med Mol Imaging 33 2006 S145 30 R. Reitsamer F. Peintinger E. Prokop L. Rettenbacher C. Menzel 200 Sentinel lymph node biopsies without axillary lymph node dissection – no axillary recurrences after a 3-year follow-up Br J Cancer 90 2004 1551 1554 31 S. Roka P. Konstantiniuk D. Heck Axillary recurrence after sentinel node biopsy Eur Surg Acta Chir Austriaca 34 5 2002 280 282 32 D.K. Rosing C.E. Dauphine M.P. Vargas Axillary regional recurrence after sentinel lymph node biopsy for breast cancer Am Surg 72 2006 939 942 33 A. Sanjuan S. Vidal-Sicart G. Zanon Clinical axillary recurrence after sentinel node biopsy in breast cancer: a follow-up study of 220 patients Eur J Nucl Med Mol Imaging 32 2005 932 936 34 P. Schrenk M. Hatzl-Griesenhofer A. Shamiyeh W. Waynad Follow-up of sentinel node negative breast cancer patients without axillary lymph node dissection J Surg Oncol 77 2001 165 170 35 T. Schulze J. Mucke J. Markwardt P.M. Schlag A. Bembenek Long-term morbidity of patients with early breast cancer after sentinel lymph node biopsy compared to axillary lymph node dissection J Surg Oncol 93 2006 109 119 36 S. Shivers C. Cox G. Leight Final results of the Department of Defense multicenter breast lymphatic mapping trial Ann Surg Oncol 9 2002 248 255 37 M. Sinko M. Rajtar G. Cserni G. Boross E. Tekle Wolde E. Ambrozay Axillary recurrence in 292 breast cancer patients with negative sentinel node biopsy and no axillary dissection Ann Surg Oncol 224 2006 S272 38 H.C. Snider Jr. E. Rubin R. Henson Axillary ultrasonography to detect recurrence after sentinel node biopsy in breast cancer Ann Surg Oncol 13 2006 501 507 39 M. Snoj M. Bracko I. Zagar Axillary recurrence rate in breast cancer patients with negative sentinel lymph node Croat Med J 46 2005 377 381 40 N.K. Soni A.J. Spillane Experience of sentinel node biopsy alone in early breast cancer without further axillary dissection in patients with negative sentinel node ANZ J Surg 75 2005 292 299 41 T. Susini J. Nori E. Vanzi Axillary ultrasound scanning in the follow-up of breast cancer patients undergoing sentinel node biopsy Breast 16 2 2006 190 196 42 K.K. Swenson A. Mahipal M.J. Nissen Axillary disease recurrence after sentinel lymph node dissection for breast carcinoma Cancer 104 2005 1834 1839 43 B. Taback N. Hansen K. Conway A. Giuliano Recurrence patterns following a negative sentinel lymph node biopsy in patients with early-stage breast cancer J Clin Oncol 22 S14 2004 560 44 H. Takei K. Suemasu M. Kurosumi Recurrence after sentinel lymph node biopsy with or without axillary lymph node dissection in patients with breast cancer Breast Cancer 14 2007 16 24 45 H. Torrenga H. Fabry J.R. van der Sijp P.J. van Diest R. Pijpers S. Meijer Omitting axillary lymph node dissection in sentinel node negative breast cancer patients is safe: a long term follow-up analysis J Surg Oncol 88 2004 4 7 46 B. van der Vegt M.H. Doting P.L. Jager J. Wesseling J. de Vries Axillary recurrence after sentinel lymph node biopsy Eur J Surg Oncol 30 2004 715 720 47 K.J. van Wessem W.S. Meijer Sentinel lymph node biopsy in breast cancer: results of intradermal periareolar tracer injection and follow-up of sentinel lymph node-negative patients Breast 13 2004 290 296 48 U. Veronesi V. Galimberti L. Mariani Sentinel node biopsy in breast cancer: early results in 953 patients with negative sentinel node biopsy and no axillary dissection Eur J Cancer 41 2005 231 237 49 G. Zavagno P. Carcoforo Z. Franchini Axillary recurrence after negative sentinel lymph node biopsy without axillary dissection: a study on 479 breast cancer patients Eur J Surg Oncol 31 2005 715 720 50 R. Ponzone N. Biglia F. Maggiorotto Sentinel node dissection as definitive treatment for node negative breast cancer patients Eur J Surg Oncol 29 2003 703 706 51 B.J. van Wely L.J. Strobbe M. Smidt I.M. de Kievit Fout negatieve sentinel lymph node biopsie; follow up van 391 sentinel lymph node negatieve patiënten met borstkanker Ned Tijdschr Geneeskd 15 2006 1 52 D. Grigolato A. Ballarin Z. Franchini Sentinel node biopsy: our clinical experience in 232 T1 breast cancers Eur J Nucl Med Mol Imaging 33 S2 2006 128 162 53 A.Y. de Kanter A.N. van Geel M.A. Paul Controlled introduction of the sentinel node biopsy in breast cancer in a multi-centre setting: the role of a coordinator for quality control Eur J Surg Oncol 26 2000 652 656 54 J.S. Fant M.D. Grant S.M. Knox Preliminary outcome analysis in patients with breast cancer and a positive sentinel lymph node who declined axillary dissection Ann Surg Oncol 10 2003 126 130 55 M. Greco R. Agresti N. Cascinelli Breast cancer patients treated without axillary surgery: clinical implications and biologic analysis Ann Surg 232 2000 1 7 56 F.C. Wright J. Walker C.H. Law D.R. McCready Outcomes after localized axillary node recurrence in breast cancer Ann Surg Oncol 10 2003 1054 1058 57 O.E. Nieweg P.J. Tanis B.B. Kroon The definition of a sentinel node Ann Surg Oncol 8 2001 538 541 58 R. de Boer H.F. Hillen R.M. Roumen H.J. Rutten M.J. van der Sangen A.C. Voogd Detection, treatment and outcome of axillary recurrence after axillary clearance for invasive breast cancer Br J Surg 88 2001 118 122 59 I. Fredriksson G. Liljegren L.G. Arnesson Consequences of axillary recurrence after conservative breast surgery Br J Surg 89 2002 902 908 60 A. Recht S.M. Pierce A. Abner Regional nodal failure after conservative surgery and radiotherapy for early-stage breast carcinoma J Clin Oncol 9 1991 988 996