ORIGINAL ARTICLE PROGNOSTIC IMPLICATION OF SENTINEL LYMPH NODE BIOPSY IN CUTANEOUS HEAD AND NECK MELANOMA

Transcription

1 ORIGINAL ARTICLE PROGNOSTIC IMPLICATION OF SENTINEL LYMPH NODE BIOPSY IN CUTANEOUS HEAD AND NECK MELANOMA Benjamin E. Saltman, MD, 1 Ian Ganly, MD, 2 Snehal G. Patel, MD, 2 Daniel G. Coit, MD, 3 Mary Sue Brady, MD, 4 Richard J. Wong, MD, 2 Jay O. Boyle, MD, 2 Bhuvanesh Singh, MD, PhD, 2 Ashok R. Shaha, MD, 2 Jatin P. Shah, MD, 2 Dennis H. Kraus, MD 2 1 Department of Otolaryngology, Division of Head and Neck Surgery, Long Island Jewish Medical Center, New Hyde Park, NY. bsaltman@nshs.edu 2 Head and Neck Service, Department of Surgery, Memorial Sloan-Kettering Cancer Center, New York, NY 3 Department of Surgery, Memorial Sloan-Kettering Cancer Center, New York, NY 4 Division of Surgical Oncology, Department of Surgery, Memorial Sloan-Kettering Cancer Center, New York, NY Accepted 13 January 2010 Published online 9 April 2010 in Wiley Online Library (wileyonlinelibrary.com). DOI: /hed Abstract: Background. Current therapy for intermediate thickness melanoma involves wide local excision with sentinel lymph node biopsy (SLNB). SLNB provides important prognostic information and immediate regional lymphadenectomy for a positive sentinel lymph node (SLN) may improve survival and identifies patients who are candidates for adjuvant therapy and/or clinical trials. The head and neck site is unique because of its complex lymphatic drainage pattern to multiple nodal basins and because of the risk of site-specific morbidity associated with regional lymphadenectomy when compared to other body sites. The goal of this study is to report the results of SLNB for head and neck cutaneous melanoma in locating the sentinel node and to report on the prognostic implications of SLNB for this cohort of patients. Methods. A prospectively entered melanoma database was used to review consecutive patients with head and neck cutaneous melanoma undergoing SLNB at Memorial Sloan-Kettering Cancer Center between 1996 and The database, along with a retrospective chart review, was used to evaluate the success of SLNB at locating an SLN and the success rate of frozen section and permanent section analysis at diagnosing Correspondence to: B. E. Saltman VC 2010 Wiley Periodicals, Inc. metastatic disease. Recurrence at all sites including the nodal basin and status at last follow-up was recorded. Characteristics of the patients primary melanoma were included. Descriptive statistics along with univariate and multivariate survival analysis were performed. Results. Between 1996 and 2007, 234 patients with a diagnosis of head and neck cutaneous melanoma underwent SLNB and had at least 1 month of follow-up. At least 1 SLN was identified in 218 of these patients (93%) by lymphoscintigraphy. In 16 patients, no SLN was found. These patients had a much shorter time to recurrence (4.75 months) than either the SLN-positive group (10.7 months) or the SLN-negative group (26.0 months). They had a disease-specific survival (DSS) in between the SLN-positive and SLN-negative group. Of the patients in whom an SLN was identified, 28 patients (12%) had at least 1 positive SLN. Of these, the SLNs of 14 patients (50%) were identified on frozen section; 14 (50%) could only be identified after further sectioning or immunohistochemical analysis postoperatively. Among 190 patients with a negative SLNB, 12 patients had recurrences in the nodal basin. This resulted in a sensitivity of 70%, a negative predictive value of 94%, and a false-negative rate of 30%. The 3-year diseasefree survival for SLN-negative and SLN-positive patients was 84% (p <.031) and 58% (p <.102), respectively. The 3-year melanoma-specific survival was 98% (p <.012) and 75% (p <.201), respectively SLN Biopsy in Cutaneous Head and Neck Melanoma HEAD & NECK DOI /hed December 2010

2 Conclusion. The SLN status is an important predictor of survival. The technique, performed in the head and neck is complex and associated with a high false-negative rate. VC 2010 Wiley Periodicals, Inc. Head Neck 32: , 2010 Keywords: head and neck; melanoma; sentinel lymph node; head and neck surgery; head and neck melanoma Since our first article was published in 2002 detailing our experience with sentinel lymph node biopsy (SLNB) for head and neck cutaneous melanoma, the incidence of cutaneous malignant melanoma has continued to grow faster than any other cancer worldwide. Evaluation of the Surveillance, Epidemiology, and End Results (SEER) data shows that in the United States, the incidence of melanoma is increasing at approximately 3% per year. One in 55 white people in the United States born today will develop melanoma. 1 In Australia, which has the highest reported incidence of melanoma in the world, the incidence is 1 in 35 men and 1 in 25 women. 2 Evaluation of the SEER data reveals that the majority of melanoma occurs on the trunk and extremities, although approximately 18% of the cutaneous melanoma is found in the head and neck. 3 Since the last article, Morton et al 4 reported the results of the prospective Multicenter Selective Lymphadenectomy Trial-I concluding that SLNB does identify positive regional lymph nodes; and that although it adds important prognostic information, it may not be therapeutic. This study also reported that patients with positive sentinel lymph nodes (SLNs) had an improved outcome compared with patients who developed clinically positive lymph nodes. Head and neck cutaneous melanoma may represent a unique entity versus cutaneous melanoma of the trunk and extremities not only in terms of prognosis but also with regard to treatment. A recent article using SEER data found a higher mortality for patients with head and neck cutaneous melanoma versus trunk and extremity cutaneous melanoma. 3 The application of SLNB in the head and neck has been well documented. 5,6 Lymphatic drainage from the head and neck is often both multiple and bilateral, making SLNB difficult compared to other sites. 7 The largest single report of SLNB for head and neck cutaneous melanoma comes from the Sentinel Node Working Group, derived from a pooled dataset from multiple institutions. 8 There are inherent problems with the reliability of such multi-institutional studies: institutions may perform this procedure differently; the pathology and processing, including immunohistochemistry may vary; and last, the diagnostics, including lymphoscintigraphy, may be performed differently. Our study is a retrospective review of SLNB in head and neck cutaneous melanoma from a single center. Our goals for this research were 3-fold. First, we wanted to evaluate the reliability of SLNB in head and neck cutaneous melanoma. Second, because of our observation that the frozen section has not been sensitive at identifying melanoma in the SLN, we wanted to evaluate its sensitivity versus permanent sectioning and immunohistochemistry. Last, we wanted to evaluate the prognostic value of SLNB in head and neck cutaneous melanoma by looking at the disease-free survival (DFS), neck recurrence-free survival (NRFS), and disease-specific survival (DSS). PATIENTS AND METHODS Patients. Between January 1996 and December 2007, all patients who underwent SLNB at Memorial Sloan-Kettering Cancer Center (MSKCC) were prospectively entered into a melanoma database. Multiple demographics and pathologic features were included in this database. All follow-up details are included in this database, especially date and site of first recurrence, site of recurrence, and melanoma-specific status at last follow-up. Methods. Our complete methodology has been published previously. 5 Preoperative lymphoscintigraphy was performed on all patients to identify drainage patterns and sites of the sentinel node(s). The majority of the patients (74.6%) also received intraoperative blue dye injection at 4 quadrants around the tumor (or excision site). A handheld gamma probe along with the lymphoscintigram was used to identify the sentinel nodes. All nodes were analyzed via frozen section. Hematoxylin-eosin stain, serial sectioning, and immunohistochemical analysis for HMB-45 and S-100 were performed during permanent pathologic determination as has been described. 5 Statistics were performed using a statistical software package (SPSS 15.0 for Windows, SLN Biopsy in Cutaneous Head and Neck Melanoma HEAD & NECK DOI /hed December

3 Chicago, IL). Follow-up information was calculated in months from the time of initial surgery to status at last follow-up. The DSS, neck recurrence-free survival, and disease-free survival were calculated using the Kaplan Meier method. Univariate comparisons between groups were performed using the log-rank test; multivariate analysis was performed using the Cox proportional hazards model. Variables for this model included those which have been found to be prognostic of survival in previous research. RESULTS Two hundred thirty-six consecutive patients underwent SLNB for head and neck cutaneous melanoma between 1996 and Two patients had no follow-up and were not included in this study. The follow-up of the study patients ranged from 1 to 124 months. The median follow-up was 37 months. Of these patients, an SLN was identified in 218 (92%); SLN could not be identified by either the lymphoscintigram or the intraoperative gamma probe in 16 cases (8%). In 5 of the 236 cases, the SLN mapped to the parotid gland. In 28 of 216 cases (13%), a positive SLN was identified; 14 (50%) of these were identified by frozen section, and 14 (50%) required immunohistochemistry or further sectioning on permanent histology. The descriptive statistics for the melanoma pathology is shown in Table 1. The median Breslow depth of melanoma undergoing SLNB was 2.8 mm, the median Clark level was IV, and the most common site of occurrence was the face. The majority of SLNBs were performed on patients with intermediate thickness melanoma between 1 and 4 mm in Breslow depth. There was a nodal recurrence in 12 of the 190 patients who had a negative SLNB. This resulted in a sensitivity of 70%, a negative predictive value of 94%, and a false-negative rate of 30% (in which the false-negative rate is defined as: the number of patients who recurred in the SLNB negative nodal basin divided by the sum of these patients and the positive SLNB). In the majority of patients (66.7%), identification of positive lymphadenopathy resulted in lymphadenectomy. In almost half of the patients (47.8%), immediate lymphadenectomy was performed, whereas in the rest of the patients a delayed lymph node dissection was performed. Recurrence. Fifty-five patients (24%) had a recurrence at a median time of 10 months. The site of initial recurrence was systemic recurrence in 36.4%, nodal in 32.7%, local in 23.6%, and in-transit in 3 patients (7.3%). One patient presented with both local and nodal recurrence. Both the mean time to nodal recurrence and the mean time to recurrence at any site were 20.7 months. The 3-year NRFS rate for SLN-negative and SLN-positive patients was 93% and 89%, respectively (p ¼.451). The Kaplan Meier curve for NRFS is shown in Figure 1 and clearly shows there to be no difference between the SLN-negative and SLN-positive groups. If we evaluate recurrence by SLN status, SLN-negative patients were most likely to recur regionally and systemically (31.6% and 36.8%), whereas SLN-positive patients were more likely Table 1. Descriptive statistics for the melanoma pathology. No. of patients (%) Depth*, mm <1 29 (12) (67) >4 47 (21) Clark level II 4 (2) III 24 (11) IV 139 (65) V 38 (22) Site Lip 2 (1) External ear 36 (15) Neck 40 (17) Scalp 52 (22) Face 104 (45) *Three patients had no depth measurement and were not included. FIGURE 1. Nodal recurrence free survival stratified by sentinel lymph node status. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.] 1688 SLN Biopsy in Cutaneous Head and Neck Melanoma HEAD & NECK DOI /hed December 2010

4 58%, and 75%, respectively. Figure 2B shows the Kaplan Meier survival curve for DSS. SLNnegative, SLN-positive, and not identified resulted in 3-year survival rates of 97%, 75%, and 90%, respectively. We then evaluated DSS data using the Kaplan Meier method for variables which have been shown to influence survival. Figure 3A shows the Kaplan Meier curve for DSS stratified by the presence of ulceration. The survival curve shows no significant difference in the first 3 years. The 3-year DSS for those tumors with ulceration versus those without were 94.2% vs 93.5%, respectively. This difference was not significant (p ¼.49). Last, our research did not show any significant difference in DSS between scalp, face, or neck melanoma. There was no FIGURE 2. (A) Kaplan Meier survival curve showing diseasefree survival (DFS) stratified by sentinel lymph node (SLN) status. (B) Kaplan Meier survival curve showing disease-specific survival (DSS) stratified by SLN status. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.] to recur systemically than regionally (46% vs 23%). The recurrence rate of SLN-negative patients was 16.8% versus 46.4% for SLN-positive patients. The 16 patients in whom an SLN was not identified had a recurrence rate in between these 2 groups at 25%. The median time to recurrence varied significantly among the groups. The SLN-negative group had a median time to recurrence of 19 months, whereas the SLN-positive group had a median time to recurrence of 8 months. Last, the median time to recurrence for the 16 patients in whom no SLN was found was 5 months. Survival Analyses. Figure 2A shows the Kaplan Meier survival curve for DFS stratified by SLN status. The 3-year DFS for SLN-negative, SLN-positive, and not identified are 84%, FIGURE 3. (A) Kaplan Meier survival curve showing diseasespecific survival (DSS) stratified by the presence of ulceration. (B) Disease-specific survival stratified by mitotic rate. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.] SLN Biopsy in Cutaneous Head and Neck Melanoma HEAD & NECK DOI /hed December

5 Table 2. A summary of the re-calculated false negative rates for 5 large studies on SLNB for head and neck cutaneous melanoma. Study No. of patients Positive SLNB (%) Regional recurrence Total positive False-negative rate* Follow-up MSKCC (12.0) % 37 mo (median) Sydney Cancer Centre (10.3) % 34 mo (mean) M. D. Anderson Cancer Center (20.4) % 34 mo (median) University of Michigan (17.5) % 25 mo (median) Sunbelt Melanoma (16.1) % 15.5 mo Abbreviations: SLNB, sentinel lymph node biopsy; MSKCC, Memorial Sloan-Kettering Cancer Center. *The false-negative rate is equal to the regional recurrence divided by the true positives (positive SLN and regional recurrences). While 5 patients recurred in the regional lymphatics. Two other patients in a prospective trial had a negative SLNB but a contemporaneous lymphadenectomy revealed a positive lymph node. mortality, however, in any of our patients with an ear or lip melanoma. Figure 3B shows DSS as stratified by mitotic index. Interestingly, we see a 100% survival rate for those patients pathologically categorized as having a mitotic rate <1. When we made the mitotic index into a categorical variable by stratifying it into 4 groups, 0, 1 4, 5 10, and >11, we found a stepwise decrease in survival which was not significant (data not shown). This method has been described elsewhere and when compared to other categorizations, was found to have a high likelihood ratio. 9,10 There was no difference in overall survival between those SLN-negative patients who recurred in the nodal basin (n ¼ 15), versus the remaining SLN-negative cohort (n ¼ 175). Multivariate Analysis of Survival. Multivariate analysis by means of the Cox regression model found 2 variables to be significantly associated with DSS, depending on which variables were entered into the model. When SLN status, ulceration, and depth were studied, both depth (p ¼.005) and SLN status (p ¼.01) were independently predictive of outcome. When the mitotic rate was entered into this model, only depth (p ¼.022) was statistically significant, whereas the SLN status approached significance (p ¼.067). Ulceration was not significant in any of these 3 models. DISCUSSION This paper represents 1 of the largest single institution s collection of patients with head and neck cutaneous melanoma undergoing SLNB. The Sydney Melanoma Unit has also published a large collection of more than 300 cases representing 14 years of experience. 11 However, their SLNB methods have evolved during this process which makes the results difficult to interpret over this time period. The descriptive statistics we used were different from all of the major published studies in that the false-negative rate that we quote is significantly higher. We believe that the correct false-negative rate should include nodal recurrence in a region which was originally deemed negative based on SLN. This will equal the regional recurrence in a previously SLNB-negative patient divided by the sum of the true positives and this regional recurrence group. Our results are comparable to those of 4 large previous studies providing data on subsequent nodal recurrence (Table 2). The results from the Sydney Melanoma Units review of 136 patients, after recalculation resulted in a false-negative rate of 44% with a mean follow-up of 34 months. The recent M. D. Anderson Cancer Center contribution to this literature is of 113 patients over 12 years with a median follow-up of 34 months. They had a false-negative rate of 23%. Interestingly, 83 patients were part of a prospective protocol in which the SLN and its corresponding nodal basin were removed regardless of SLN positivity. For 83 patients in this protocol, 11 had a positive sentinel node and 2 other patients had positive nodes which were not the sentinel node. This results in a false-negative rate of 15.4%, or 2 of 13. However, after taking into account the subsequent nodal basin recurrence in patients with a negative SLN (5 in this study), we found a false-negative rate of 23% using the criteria identified above. 12 The University of Michigan review from Schmallbach et al 13 resulted in a false-negative rate of 17.6% and a median follow-up of 25 months. Lastly, the Sunbelt Melanoma series evaluated SLN Biopsy in Cutaneous Head and Neck Melanoma HEAD & NECK DOI /hed December 2010

6 patients from multiple institutions and resulted in a false-negative rate of 12.2% with a followup of 15.5 months. 14 It is clear that with longer follow-up, the rate of false-negative goes up as more nodal basin recurrences in the SLNB-negative groups are detected. Taken altogether, this article does highlight a key limitation of SLNB its high false-negative rate. Unfortunately, the 2 alternatives to SLNB-elective lymphadenectomy and the watch and wait approach are not attractive alternatives. Elective lymphadenectomy on all intermediate-thickness patients may be too aggressive. Forty patients (17.1%) of 234 had regional disease over the 37 months, 28 at initial SLNB. Therefore, based on our study, >80% of patients would have had an unnecessary dissection. The alternative of watchful waiting is also not attractive. Morton s 2006 article, 4 although not showing a survival advantage for the 2 patient populations, randomized to SLN versus watchful waiting did show progression of disease. Furthermore, their 5-year survival rate among patients with positive SLNB was statistically higher than among those patients who were watched and developed regional disease. All of these false-negative rates are much higher than what the original articles evaluating SLNB for melanoma quoted. In the sentinel article entitled Technical Details of Intraoperative Lymphatic Mapping for Early Stage Melanoma, Morton 15 sites a false-negative rate of less than 1%. In this study, each patient not only had SLNB but also had a neck dissection with hematoxylin-eosin sectioning and immunohistochemical staining of the lymph nodes. In only 2 patients in whom an SLNB was negative did a non-sln show micrometastases. This equates to a false-negative rate of 4.8% and a sensitivity of 95.2%. This is for all anatomic sites and was performed using only blue-v or isosulfan blue. When Morton 16 looked at only areas draining to the neck, including the upper chest and back, he found an even lower falsenegative rate. When he evaluated the SLNB results from 72 patients with early-stage melanoma who underwent SLNB and immediate lymphadenectomy only if the SLN was positive, he did not have 1 regional recurrence with a mean follow-up of 25 months. The SLNB procedure was successful in identifying a node in 90% of cases and in 22 cases a neck dissection was performed to ensure the specificity of the procedure. Therefore, 42 of 79 lymphatic basins underwent neck dissections. With such a high rate of neck dissection and such a short followup, it is again difficult to ascertain whether such a low false-negative rate is realistic. One of the newer statistics that we evaluated during this retrospective study was NRFS (Figure 1). The 3-year NRFS rates for SLN-negative and SLN-positive patients were 93% and 89%, respectively, which were not statistically significant (p ¼.451). Neither a negative SLN, nor a positive SLN with a neck dissection decreased the risk of regional nodal recurrence. Other studies have shown similar findings. Our data compare favorably to more recent studies when length of follow-up is taken into account if we assume that those studies with shorter follow-up will result in less regional recurrence when compared to studies with greater than 30 months of follow-up. Our DFS and DSS rates for patients with SLN-positive and SLN-negative patients show similar trends to previously reported results. 8 Both show the significant prognostic implications of a positive SLN. Although it currently remains unclear what impact regional dissection of those patients with positive SLN will have on their survival, their poor prognosis necessitates aggressive adjuvant therapy and/or clinical trials. We also found that the initial recurrence site for our 16 patients in whom a sentinel node was not found was the regional nodes in 3 of the 4 cases. Furthermore, their time to recurrence is shorter than either the SLN-positive or -negative cohorts, an observation we are unable to explain. Although its effect on overall survival has been challenged by the Multicenter Selective Lymphadenectomy Trial-I, it may improve quality of life in these head and neck patients by increasing their time to recurrence. 4 Multivariate analysis revealed further interesting findings. Although ulceration has been shown to be significantly correlated with DSS, we did not find this to be a significant predictor by univariate or multivariate analysis. 16 As expected, based on the multiple multivariate analyses, depth was always significantly correlated with DSS. Depending on the covariables entered, the mitotic index approached significance. The importance of the mitotic index has been previously reported as an independent predictor of MSS. However, closer inspection of our data reveals that there was little difference between the different steps of the mitotic index, SLN Biopsy in Cutaneous Head and Neck Melanoma HEAD & NECK DOI /hed December

7 ie, from 1 4, 5 10, and 11þ, but that there is a difference between those melanomas with a mitotic index of 0 and the rest. This was also seen in Azzola et al s 10 article evaluating the mitotic rate of over 3500 patients from the Sydney Melanoma Unit. CONCLUSION This analysis of the MSKCC experience with SLN for head and neck cutaneous melanoma shows the important prognostic information gained from SLN. It confirms that SLNB is a reliable technique; however, all patients must be carefully followed because of the possibility of a false-negative result and the high rate of recurrence at other sites. Patients in whom an SLN is not identified should be monitored even more closely, especially in the short term, because our data show them to have a high incidence of early regional recurrence. REFERENCES 1. Ries LAG, et al, editors. SEER Cancer Statistics Review , National Cancer Journal, Bethesda, Maryland. 2. AIHW and AACR, AIHW National Mortality Database, Australia s Health 2004, AIHW. 3. Lachiewicz AM, Berwick M, Wiggins CL, Thomas NE. Survival differences between patients with scalp or neck melanoma and those with melanoma of other sites in the Surveillance, Epidemiology, and End Results (SEER) program. Arch Dermatol 2008;144: Morton DL, Thompson JF, Cochran AJ, et al. Sentinelnode biopsy or nodal observation in melanoma. N Engl J Med 2006;355: Patel SG, Coit DG, Shaha AR, et al. Sentinel lymph node biopsy for cutaneous head and neck melanomas. Arch Otolaryngol Head Neck Surg 2002;128: Morton DL, Wen DR, Foshag LJ, Essner R, Cochran A. Intraoperative lymphatic mapping and selective cervical lymphadenectomy for early-stage melanomas of the head and neck. J Clin Oncol 1993;11: Shah JP, Kraus DH, Dubner S, Sarkar S. Patterns of regional lymph node metastases from cutaneous melanomas of the head and neck. Am J Surg 1991;162: Leong SP, Accortt NA, Essner R, et al. Impact of sentinel node status and other risk factors on the clinical outcome of head and neck melanoma patients. Arch Otolaryngol Head Neck Surg 2006;132: Attis MG, Vollmer RT. Mitotic rate in melanoma: a reexamination. Am J Clin Pathol 2007;127: Azzola MF, Shaw HM, Thompson JF, et al. Tumor mitotic rate is a more powerful prognostic indicator than ulceration in patients with primary cutaneous melanoma: an analysis of 3661 patients from a single center. Cancer 2003;97: De Wilt JH, Thompson JF, Uren RF, et al. Correlation between preoperative lymphoscintigraphy and metastatic nodal disease sites in 362 patients with cutaneous melanoma of the head and neck. Ann Surg 2004;239: Gomez-Rivera F, Santillan A, McMurphey AB, et al. Sentinel node biopsy in patients with cutaneous melanoma of the head and neck: recurrence and survival study. Head Neck 2008;30: Schmallback CE, Nussenbaum B, Rees RS, Schwartz J, Johnson TM, Bradford CR. Reliability of sentinel lymph node mapping with biopsy for head and neck cutaneous melanoma. Arch Otolaryngol Head Neck Surg 2003; 129: Chao C, Wong SL, Edwards MJ, et al. Sentinel lymph node biopsy for head and neck melanomas. Ann Surg Oncol 2003;10: Morton DL, Wen DR, Wong JH, et al. Technical details of intraoperative lymphatic mapping for early stage melanoma. Arch Surg 1992;127: 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 2001;19: SLN Biopsy in Cutaneous Head and Neck Melanoma HEAD & NECK DOI /hed December 2010