J Clin Oncol 22: by American Society of Clinical Oncology INTRODUCTION

Transcription

1 VOLUME 22 NUMBER 16 AUGUST JOURNAL OF CLINICAL ONCOLOGY O R I G I N A L R E P O R T The Microanatomic Location of Metastatic Melanoma in Sentinel Lymph Nodes Predicts Nonsentinel Lymph Node Involvement D.J. Dewar, B. Newell, M.A. Green, A.P. Topping, B.W.E.M. Powell, and M.G. Cook From the Department of Plastic and Reconstructive Surgery, St George s Hospital, London; and Department of Histopathology, Royal Surrey County Hospital, Guildford, Surrey, United Kingdom. Submitted December 26, 2003; accepted May 27, Presented in part at the British Association of Plastic Surgeons Winter Meeting, London, United Kingdom, November 27-29, 2002; the EORTC Melanoma Group Spring Meeting, Paris, France, April 11-12, 2003; and the Third Research Meeting on Melanoma, Milan, Italy, May 25-27, Authors disclosures of potential conflicts of interest are found at the end of this article. Address reprint requests to Donald Dewar, MB, BS, Department of Histopathology, Royal Surrey County Hospital, Egerton Rd, Guildford, Surrey GU2 7XX, United Kingdom; DJDewar@aol.com by American Society of Clinical Oncology A B S T R A C T Purpose Sentinel node biopsy is now widely accepted as the most accurate prognostic indicator in melanoma, and is important in guiding management of patients with clinical stage I or II disease. Patients with a positive sentinel node have conventionally undergone completion lymphadenectomy (CLND) of the involved basin, but only 20% have involvement beyond the sentinel node, suggesting that CLND may be unnecessary for the other 80% of patients. This study seeks to identify criteria that might be used to be more restrictive in selecting those who should undergo CLND. Methods A total of 146 patients were identified who had had a positive sentinel node biopsy for malignant melanoma. Their sentinel nodes and lymphadenectomy specimens were re-evaluated pathologically. The metastatic melanoma in each sentinel node was assessed according to its microanatomic location within the node (subcapsular, combined subcapsular and parenchymal, parenchymal, multifocal, or extensive), and this was correlated with the presence of involved nonsentinel nodes in the CLND. The depth of the metastases from the sentinel node capsule was also recorded. Results The metastatic deposits in the sentinel node were subcapsular in 26.0% of patients. None of these patients had any nonsentinel nodes involved on CLND. In the patients whose sentinel node metastases had a different microanatomic location, the rate of nonsentinel node involvement was 22.2% overall. Conclusion The microanatomic location of metastases within sentinel nodes predicts nonsentinel lymph node involvement. In patients with only subcapsular deposits in the sentinel node, it is possible that CLND could safely be avoided. J Clin Oncol 22: by American Society of Clinical Oncology X/04/ /$20.00 DOI: /JCO INTRODUCTION Sentinel node biopsy is now the standard of care for accurate staging of patients with clinical stage I or II malignant melanoma. 1 In published series, between 15% and 34% of patients undergoing sentinel node biopsy will have metastases within the sentinel node(s), depending on the inclusion criteria and the techniques used for preparation and histologic examination of the nodes. 2,3 These patients have been shown to have a greater risk of further disease and shorter survival than sentinel node negative patients. This is the strongest prognostic criterion for melanoma. 4 In addition, sentinel node status is used to determine subsequent therapy. It is usual for patients with a positive sentinel node to proceed to removal of the remaining lymph nodes within the affected basin, termed completion lymphadenectomy (CLND). The presence of melanoma metastases in nonsentinel nodes (NSN) from the CLND specimen occurs in only 10% to 30% of cases. 5 Given the morbidity associated with lymphadenectomy, various groups have tried to predict NSN involvement on the basis of histologic features of the primary melanoma. 5,6 Although the likelihood of NSN involvement 3345

2 Dewar et al increases with increasing Breslow thickness of the primary tumor, even melanomas less than 2 mm in thickness have an 11% risk of having NSN metastases. 5 Starz et al 7 have described a process for morphometric analysis of metastatic deposits within the sentinel node itself, and used this to predict NSN involvement much more accurately. An alternative classification, incorporating histologic features of both the primary melanoma and the sentinel node metastasis, has been developed by Reeves et al. 8 Our study describes a simpler, alternative method for categorizing sentinel node metastases, based on physiologic microanatomy, and correlates this with NSN involvement. METHODS Patients whose sentinel node(s) contained metastatic malignant melanoma were identified from the St George s Melanoma Unit (London, United Kingdom) database. Eligibility criteria and techniques for sentinel node biopsy in the St George s Melanoma Unit have been described elsewhere. 9 Briefly, patients presenting with clinical stage I or II malignant melanoma were offered sentinel node biopsy if the Breslow thickness of the primary tumor was greater than 1.0 mm. Patients with a Breslow thickness less than 1.0 mm were included if the melanoma was Clark s level III or more, showed evidence of regression, or was in vertical growth phase. Sentinel nodes were identified using a standard technique of preoperative dynamic lymphoscintigraphy using radiolabeled colloid, followed by intraoperative injection of vital blue dye and use of a hand-held gamma probe. The sentinel node specimens were processed as previously reported, 2,3 using a bivalving technique based on that of Cochran. They were examined by both conventional histology and immunohistochemistry. For the purposes of this study, some sentinel node specimens from the earlier part of the series (before the development of the current processing protocol) were re-evaluated by the senior author. If necessary further sections were cut and stained so that all patients were assessed using the same sectioning and processing protocol. The metastatic deposit within each sentinel node was classified as subcapsular, combined subcapsular and parenchymal, parenchymal, multifocal, or extensive (as described in Table 1 and Fig 1), and its maximum depth from the capsule was measured. When more than one positive sentinel node had been removed from a single lymph node basin, the most extensive and/or deepest metastatic deposit overall was used to categorize that basin. In the six patients in whom sentinel nodes from more than one lymph node Microanatomic Location Subcapsular Combined Parenchymal Multifocal Extensive Table 1. Definition of Microanatomic Locations Definition Metastasis confined to subcapsular sinus Subcapsular and parenchymal metastases Metastasis entirely within paracortical area of parenchyma Multiple discrete deposits (must include some parenchymal deposits) Any metastasis larger than 5 mm, any node with extracapsular spread basin were found to contain melanoma, only the first basin to be entered into the database was included in the final analysis, because of the inherent correlation between multiple basins from an individual patient. CLND specimens were retrieved and reassessed for the presence of metastases. The rate of NSN involvement was calculated for each sentinel node subtype. Statistical analysis was performed using the 2 test, the Mann-Whitney U test, the Kruskal-Wallis test, and Newcombe s method, as indicated in Tables 2 and 3. RESULTS Sentinel node and corresponding CLND specimens were retrieved for 146 lymph node basins in 146 patients. The mean age of the patients was 49.9 years, and 55.2% were male. Patient, primary tumor, and sentinel node characteristics are described in Table 2. Patient age and sex, and site of the CLND, were not related to either microanatomic location of the sentinel node metastasis or NSN involvement (Tables 2 and 3). The number of sentinel nodes classified into each microanatomic location is listed in Table 2, with 38 (26.0%) in the subcapsular category. Of the CLND specimens, 24 (16.4%) showed evidence of melanoma metastases in NSN. There was a significant correlation between the microanatomic location and NSN involvement (Table 2). No patient with only a subcapsular deposit had subsequent NSN metastases (95% CI, 0.0% to 11.4%). Depth of the metastatic deposit from the capsule of the sentinel node also was strongly correlated with NSN involvement (Tables 2 and 3). However, three of 24 patients (12.5%) with positive NSNs had deposits less than 1 mm in depth. The shallowest was only 0.6 mm (but in a multifocal pattern). Although there was a relationship between the Breslow thickness of the primary melanoma and the microanatomic location of the sentinel node metastasis (Table 2), Breslow thickness was not significantly correlated with NSN involvement (Table 3). DISCUSSION This study categorizes metastases according to their microanatomic location within the sentinel node. It demonstrates that, for patients with sentinel nodes in which metastases are confined to a subcapsular site, the probability of NSN involvement is extremely low (zero in this series). This raises the question of whether these patients could safely be spared CLND and its associated morbidity. It is not surprising that the extent of the sentinel node metastasis predicts nonsentinel lymph node involvement, given that microscopic and macroscopic nodal involvement are known to have different prognostic significance. 10,11 This is recognized in the recent revision of the American Joint Committee on Cancer Staging System for Cutaneous Melanoma, 12 which differentiates between micro- and macrometastases in the N classification JOURNAL OF CLINICAL ONCOLOGY

3 Microanatomic Location of Metastatic Melanoma Fig 1. Microanatomic locations of metastatic melanoma within sentinel nodes. (A) Subcapsular; (B) combined; (C) parenchymal; (D) multifocal; (E) extensive

4 Dewar et al Table 2. Patient, Primary Tumor, and Sentinel Node Characteristics, and NSN Positivity by Microanatomic Location Characteristic Total Subcaps Comb Parench Multifocal Extensive P No % Sex, % Male Female Age, years Mean , % , % Breslow of primary, mm 1.0, % , % , % , % Mean Range SD Basin, % Axilla Groin Neck Other Depth of SN metastasis, mm Mean Range SD Positive NSN No % % CI, % 11.0 to to to to to to 66.0 Abbreviations: NSN, nonsentinel node; subcaps, subcapsular; Comb, combined; Parench, parenchymal; SD, standard deviation; SN, sentinel node. 2 test. Kruskal-Wallis test. Newcombe s method. Previous studies have estimated the size of sentinel node metastases to predict NSN involvement. Starz et al 7 proposed a classification based on the depth of the metastasis from the capsule and the number of 1-mm slices containing melanoma. Although a correlation between the classification by Starz et al and NSN involvement exists, subsequent study on a larger group of patients has shown that it does not predict NSN metastasis with complete reliability, given that a small group of patients with metastases less than 1 mm in depth shows NSN involvement (H. Starz, personal communication, April, 2002). Reeves et al 8 have devised a scoring system based on the extent of the tumor within the sentinel node and whether the primary melanoma is ulcerated. For patients with nonulcerated primary tumors and less than 2 mm of tumor in the sentinel node, there is a low probability of NSN metastasis (zero in their published series). In the present study, however, six of 24 (25%) of the NSN-positive patients had a sentinel node metastasis of less than 2 mm in maximum dimension. This concurs with earlier work by Carlson et al, 12 suggesting that size of the sentinel node metastasis is not a reliable predictor of NSN involvement. Reeves et al 8 also found a correlation between Breslow thickness of the primary melanoma and NSN status. Our study supports the alternative view, that Breslow thickness and NSN status are not significantly related. 5,6 This question warrants further investigation. Ulceration of the primary tumor was not reported routinely in the early part of the St George s series and therefore has not been investigated as a factor in this study. The proposed microanatomic classification predicts NSN involvement more accurately than size and depth alone because, in several patients, small deposits (some consisting of only a few cells) were associated with NSN metastasis, but none was confined to a subcapsular location. In contrast, no patient with only a subcapsular deposit was found to have tumor in NSNs. Although this study confirms the correlation between depth of the sentinel node metastasis from the capsule and NSN involvement, 12.5% of patients with positive NSNs had metastases less than 1 mm 3348 JOURNAL OF CLINICAL ONCOLOGY

5 Microanatomic Location of Metastatic Melanoma Table 3. Patient, Tumor, and SN Characteristics by NSN Status Characteristic Negative NSN Positive NSN No % Sex, % Male Female Age, years Mean , % , % Breslow of primary, mm 1.0, % , % , % , % Mean Range SD Basin, % Axilla Groin Neck Other Depth of SN metastasis, mm Mean Range SD Abbreviations: SN, sentinel node; NSN, nonsentinel node; SD, standard deviation. 2 test. Mann-Whitney U test. P in depth. Depth alone, therefore, is not considered a sufficiently reliable predictor of NSN involvement to identify patients who might safely avoid CLND. Furthermore, the proposed microanatomic classification has a physiologic basis, with metastatic melanoma being presumed to enter a node via the afferent lymphatics into the subcapsular sinus, before passing into the paracortex, and ultimately replacing large parts of the node. This is reflected in the steadily increasing mean depth of metastatic deposits, and increased frequency of NSN metastases, on progression through the microanatomic categories (Table 2). The classification is simpler to learn and easier to use than the more complex micromorphometric techniques, and involves no additional cost. Given the controversy surrounding the influence of Breslow thickness on NSN status, it is also advantageous that the classification does not depend on histologic characteristics of the primary melanoma. This study was not designed to investigate differences in survival between different microanatomic locations of sentinel node metastases. Whether it is indeed safe not to perform CLND in patients with metastases confined to the subcapsular zone in sentinel nodes would need to be addressed in a properly designed and, ideally, randomized trial. It is recommended, in the meantime, that the microanatomic location of the metastases becomes a routine part of the histopathologic reporting of positive sentinel nodes. Authors Disclosures of Potential Conflicts of Interest The authors indicated no potential conflicts of interest. REFERENCES 1. WHO: WHO declares lymphatic mapping to be the standard of care for melanoma. Oncology 13:288, Cook MG, Green MA, Anderson B, et al: EORTC Melanoma Group: The development of optimal pathological assessment of sentinel lymph nodes for melanoma. J Pathol 200: , Ruiter DJ, Spatz A, van den Oord JJ, et al: Pathology Committee of the European Organization Research and Treatment of Cancer (EORTC) Melanoma Group: Pathologic staging of melanoma. Semin Oncol 29: , Gershenwald JE, Thompson W, Mansfield PF, et al: Multi-institutional melanoma lymphatic mapping experience: The prognostic value of sentinel lymph node status in 612 stage I or II melanoma patients. J Clin Oncol 17: , McMasters KM, Wong SL, Edwards MJ, et al: Frequency of nonsentinel lymph node metastasis in melanoma. Ann Surg Oncol 9: , Wagner JD, Gordon MS, Chuang TY, et al: Predicting sentinel and residual lymph node basin disease after sentinel lymph node biopsy for melanoma. Cancer 89: , Starz H, Balda BR, Kramer KU, et al: A micromorphometry-based concept for routine classification of sentinel lymph node metastases and its clinical relevance for patients with melanoma. Cancer 91: , Reeves ME, Delgado R, Busam KJ, et al: Prediction of nonsentinel lymph node status in melanoma. Ann Surg Oncol 10:27-31, Topping AP, Dewar DJ, Rose V, et al: Five years of sentinel node biopsy for melanoma: The St George s Melanoma Unit experience. Br J Plast Surg 57:97-104, Balch CM, Soong SJ, Gershenwald JE, et al: Prognostic factors analysis of 17,600 melanoma patients: Validation of the American Joint Committee on Cancer melanoma staging system. J Clin Oncol 19: , Carlson GW, Murray DR, Lyles RH, et al: The amount of metastatic melanoma in a sentinel node: Does it have prognostic significance? Ann Surg Oncol 10: , Balch CM, Buzaid AC, Soong SJ, et al: Final version of the American Joint Committee on Cancer staging system for cutaneous melanoma. J Clin Oncol 19: ,