Key Points
Question
What is the association of complex lymphatic drainage patterns with sentinel lymph node biopsy outcomes for cutaneous melanomas of the head and neck?
Findings
In this cohort study of 1080 patients, including 147 with head and neck melanomas, a more complex lymphatic drainage pattern was associated with worse outcomes.
Meaning
False omission rate was 2 times higher, and false-negative rate was nearly 3 times higher in head and neck cutaneous melanomas compared with other body regions; lymphoscintigraphy with multiple hotspots was associated with poorer long-term outcomes, especially in the head and neck region.
This cohort observational study at a single UK University cancer center compares the accuracy, prognostic value, and long-term outcomes of sentinel lymph node biopsy in patients with head and neck melanoma with melanoma from the trunk and limb, focusing on the lymphatic drainage pattern.
Abstract
Importance
Although sentinel lymph node biopsy (SLNB) is a vital staging tool, its application in head and neck melanoma (HNM) is complicated by a higher false-negative rate (FNR) compared with other regions. This may be due to the complex lymphatic drainage in the head and neck.
Objective
To compare the accuracy, prognostic value, and long-term outcomes of SLNB in HNM with melanoma from the trunk and limb, focusing on the lymphatic drainage pattern.
Design, Setting, and Participants
This cohort observational study at a single UK University cancer center included all patients with primary cutaneous melanoma undergoing SLNB between 2010 to 2020. Data analysis was conducted during December 2022.
Exposures
Primary cutaneous melanoma undergoing SLNB between 2010 to 2020.
Main Outcomes and Measures
This cohort study compared the FNR (defined as the ratio between false-negative results and the sum of false-negative and true-positive results) and false omission rate (defined as the ratio between false-negative results and the sum of false-negative and true-negative results) for SLNB stratified by 3 body regions (HNM, limb, and trunk). Kaplan-Meier survival analysis was used to compare recurrence-free survival (RFS) and melanoma-specific survival (MSS). Comparative analysis of detected lymph nodes on lymphoscintigraphy (LSG) and SLNB was performed by quantifying lymphatic drainage patterns by number of nodes and lymph node basins. Multivariable Cox proportional hazards regression identified independent risk factors.
Results
Overall, 1080 patients were included (552 [51.1%] men, 528 [48.9%] women; median age at diagnosis 59.8 years), with a median (IQR) follow-up 4.8 (IQR, 2.7-7.2) years. Head and neck melanoma had a higher median age at diagnosis (66.2 years) and higher Breslow thickness (2.2 mm). The FNR was highest in HNM (34.5% vs 14.8% trunk or 10.4% limb, respectively). Similarly, the false omission rate was 7.8% in HNM compared with 5.7% trunk or 3.0% limbs. The MSS was no different (HR, 0.81; 95% CI, 0.43-1.53), but RFS was lower in HNM (HR, 0.55; 95% CI, 0.36-0.85). On LSG, patients with HNM had the highest proportion of multiple hotspots (28.6% with ≥3 hotspots vs 23.2% trunk and 7.2% limbs). The RFS was lower for patients with HNM with 3 or more affected lymph nodes found on LSG than those with fewer than 3 affected lymph nodes (HR, 0.37; 95% CI, 0.18-0.77). Cox regression analysis showed head and neck location to be an independent risk factor for RFS (HR, 1.60; 95% CI, 1.01-2.50), but not for MSS (HR, 0.80; 95% CI, 0.35-1.71).
Conclusions and Relevance
This cohort study found higher rates of complex lymphatic drainage, FNR, and regional recurrence in HNM compared with other body sites on long-term follow-up. We advocate considering surveillance imaging for HNM for high-risk melanomas irrespective of sentinel lymph node status.
Introduction
Cutaneous melanoma is the most lethal form of skin cancer, and its incidence is increasing faster than any other malignant disease worldwide.1 Sentinel lymph node biopsy (SLNB) is widely used as a staging procedure in patients with melanoma because it allows for improved identification of lymph node (LN) micrometastasis and directs adjuvant treatment options.2 Furthermore, the presence of micrometastasis within the SLNB sample has been demonstrated to be the most important prognostic indicator of disease-free survival.3
Around 13% to 18% of new melanoma cases head and neck melanomas (HNMs),4,5 with HMN incidence in pediatric, adolescent, and young adult patients being up by 51% in the past 2 decades.6 Historically, HNM is reported to have higher false-negative (FN) SLNB rates and worse prognosis compared with primary melanoma arising from the trunk or limb.4,7,8,9,10,11 Although more recent single-arm observational studies have suggested HNM may have lower FN rates and better survival than previously thought, these studies lack internal comparison groups, limiting conclusions we can draw from them.12,13,14,15
A commonly described cause for the high FN rates is the unpredictable and varied lymphatic drainage in the head and neck region.16 Several studies have attempted to capture the complex lymphatic drainage pattern, reporting high inter-patient variability of identified sentinel lymph nodes in HNM.17,18,19 However, whether this variability is an independent factor in the high FN rates observed in HNM remains to be fully elucidated.
The lack of clarity in the role SLNB plays in HNM is further affected by major clinical trials often excluding patients with HNM from their studies or, when they do include them, fail to separately analyze HNM.20,21,22,23 As a result, SLNB in patients with HNM remains less evidenced, with poorer long-term outcome data compared with melanoma in other body sites.
In this large single-institution population study, we sought to compare the accuracy and prognostic value of SLNB in patients with HNM compared with melanoma arising from other regions of the body by comparing their FN, recurrence, and survival rates. Furthermore, we hypothesize that HNM is associated with a more complex lymphatic drainage which predicts the higher FN and lymph node recurrence rates seen in patients with HNM compared with melanoma arising from other regions of the body.
Methods
Study Population
A retrospectively collected database was created for all patients with a histopathologically confirmed diagnosis of primary cutaneous melanoma referred to Addenbrooke’s Hospital for wide local excision with SLNB between the years 2010 and 2020.
Cambridge University Hospital NHS Trust is a large university cancer center that receives SLNB referrals from a population base of approximately 2.4 million people. All referrals are initially discussed by the Cambridge University Hospital melanoma multidisciplinary team. Data were collected centrally through the hospital medical record database (EPIC Systems Corporation) and, where possible, the most up-to-date follow-up details were retrieved through communicating with patients’ respective local hospitals. Data abstracted included age at primary diagnosis, sex, anatomical site, details of the histopathological reports, details of the sentinel LN surgery as well as long-term melanoma-specific outcomes (date and type of any recurrences and survival data). The follow-up period was defined as the time between melanoma diagnosis date to the last visit occurred or death.
Exclusion criteria included SLNBs undertaken for urogenital melanoma, nonmelanoma diagnosis, or SLNB abandoned due to technical failures or if cancelled after lymphoscintigraphy (LSG) (n = 40). This study did not include any patients with macroscopically enlarged lymph nodes either on clinical examination or imaging. Macroscopically enlarged lymph nodes were treated with surgical lymph node dissection, and further systemic therapy if appropriate as part of routine clinical care or therapeutic trials.
The study was registered with the Cambridge University Hospitals QSIS system (Project ID 4049). All patients were treated after appropriate informed consent. Patient records were accessed retrospectively for review after institutional review board registration with Cambridge University Hospitals.
Sentinel Lymph Node Biopsy
Patients underwent preoperative LSG using techniques previously described.22 Scanning with Tc 99m LSG was followed by intradermal patent blue V sodium dye injection (Guerbert Laboratories, Ltd) and handheld intraoperative γ probe SLN triple localization. Because this was a long study period, it is important to emphasize that the LSG protocol remained the same throughout, apart from the introduction of single-photon emission computed tomography in 2014 as an additional imaging modality after LSG for any drainage site including the neck for improved anatomical localization.24 The LSG was reported by a small group of experienced nuclear medicine physicians. All lymph nodes were evaluated histologically with the EORTC protocol for sentinel node processing.25,26
Head and neck LNs identified on the LSG findings were categorized into 5 groups modified from the American Academy of Otolaryngology–Head and Neck Surgery (AAO-HNS) description of head and neck lymph node levels27: facial LNs, upper neck LNs, lower neck LNs, posterior neck LNs, and posterior scalp LNs.
Follow-up and Identification of Disease Recurrence
Patients with no microscopic melanoma identified in the sentinel LN or wide local excision specimen were followed up in dedicated skin cancer outpatient clinics at 3 monthly intervals in the first 3 years, and 6 monthly intervals for an additional 2 years as per the UK national guidance current at the time.28
Patients with microscopic melanoma had baseline whole-body imaging, and were offered either completion lymph node dissection, observation, or ultrasound-guided LN basin imaging according to the Multicenter Selective Lymphadenectomy Trial (MSLT-II).5 In 2019, adjuvant systemic therapy was approved in the UK, and 29 patients had either adjuvant biological or immunotherapy. Any suspicion of metastatic disease was investigated with imaging, and tissue biopsy was sought where appropriate, with directed adjuvant medical or surgical therapy.
Statistical Analysis
Statistical analysis was conducted with GraphPad Prism (version 9.4.1; GraphPad Software, LLC). Patients’ details were summarized using descriptive statistics. All continuous variables were treated as skewed. Effect size metrics were calculated using RStudio (version 1.2.1335; RStudio, Inc). Cramer V was reported for contingency of the 3 body sites vs relevant categories with a weak relationship suggested if less than 0.1, a moderate relationship if between 0.10 to 0.25, and a strong relationship if greater than 0.25.29,30 For comparison of continuous variables, η2 was reported for comparison of continuous variables with strength of association interpreted as small (0.01), medium (0.06), or large (0.14).31 The analysis was conducted in December of 2022.
The SLNB positivity rate was calculated as number of patients with positive results over total number of patients undergoing SLNB. False-negative SLNB was defined by a negative SLNB result in a patient who developed regional recurrence during follow-up without previously or simultaneously diagnosed local or in-transit recurrence. The FN rate was defined as the ratio between FN and the sum of FN and true-positive results. False omission rate was defined by the ratio between FN and the sum of FN and true-negative results.16 Negative predictive value was defined by the ratio between true-negative and the sum of true-negative and FN results.
All time-to-event end points were calculated from the date of primary diagnosis. The melanoma-specific survival (MSS) was calculated from primary diagnosis until death due to melanoma or last follow-up date (treated as censored), and recurrence-free survival (RFS) was calculated from primary diagnosis until recurrence, death, or last follow-up (treated as censored). Recurrence was defined as local, in-transit, regional lymph node basin, or distant. Kaplan-Meier survival plots with log-rank tests were used to assess differences in MSS and RFS between different melanoma regions, number of positive lymph nodes and same day vs next day surgery. Cox multivariable regression analysis was employed to identify independent risk factors to RFS and MSS. Statistical tests were 2-sided and P < .05 was considered statistically significant.
Literature Review
We conducted a systematic literature review to find published cohorts of SLNB in the head and neck region in PubMed using the search terms melanoma, head neck, sentinel lymph node biopsy, and cutaneous restricted to the English language. We then performed a snowballing review of the references of individual articles of relevance. We arbitrarily excluded articles with less than 100 patients with SLNBs done for cutaneous head and neck melanoma.
Results
Population Characteristics
Overall, 1080 patients with cutaneous melanoma were included in the analysis. Median (IQR) follow-up was 4.8 (2.7-7.2) years. Patient demographic and clinical characteristics are sumarized in Table 1.
Table 1. Demographic, Tumor, and Lymph Node Characteristics of Patient Population According to Body Region.
| Characteristic | Patients, No. (%) | Effect size (95% CI)a | ||||
|---|---|---|---|---|---|---|
| Overall (n = 1080) | Head and neck (n = 147) | Trunk (n = 348) | Limb (n = 585) | Cramer V (95% CI) | η2 (95% CI) | |
| Sex | ||||||
| Female | 528 (48.9) | 47 (32.0) | 111 (31.9) | 370 (63.2) | 0.31 (0.25-0.37) | NA |
| Male | 552 (51.1) | 100 (68.0) | 237 (68.1) | 215 (36.8) | ||
| Age at diagnosis, median (IQR), y | 59.8 (48.7-69.0) | 66.2 (51.3-72.8) | 58.3 (47.3-68.5) | 59.7 (49.3-68.4) | NA | 0.01 (0.00-0.03) |
| Follow-up time, median (IQR), y | 4.8 (2.7-7.2) | 4.3 (2.3-6.6) | 4.5 (2.5-6.6) | 5.2 (2.9-7.7) | NA | 0.02 (0.01-0.05) |
| Breslow thickness, median (IQR), mmb | 1.6 (1.1-2.7) | 2.2 (1.3-4.2) | 1.6 (1.1-2.6) | 1.5 (1.1-2.5) | NA | 0.02 (0.01-0.04) |
| Breslow thickness, mmb | ||||||
| ≤1.0 | 219 (20.4) | 25 (17.0) | 69 (20.0) | 125 (21.4) | 0.12 (0.05-0.15) | NA |
| 1.1-2.0 | 453 (42.1) | 45 (30.6) | 150 (43.5) | 258 (44.2) | ||
| 2.1-4.0 | 267 (24.8) | 37 (25.2) | 84 (24.3) | 146 (25.0) | ||
| >4.0 | 137 (12.7) | 40 (27.2) | 42 (12.2) | 55 (9.4) | ||
| Ulcerationc | 0.04 (0.00-0.11) | NA | ||||
| Absent | 824 (79.7) | 109 (76.8) | 260 (77.2) | 455 (82.0) | ||
| Histological subtype | ||||||
| Superficial spreading | 703 (65.1) | 66 (44.9) | 233 (67.0) | 404 (69.1) | 0.21 (0.16-0.25) | NA |
| Nodular | 218 (20.2) | 35 (23.8) | 79 (22.7) | 104 (17.8) | ||
| Lentigo maligna | 29 (2.7) | 21 (14.3) | 3 (0.9) | 5 (0.9) | ||
| Spitzoid | 22 (2.0) | 4 (2.7) | 7 (2.0) | 11 (1.9) | ||
| Others/NOS | 108 (10.0) | 21 (14.3) | 26 (7.5) | 61 (10.4) | ||
| Hotspots on scan | ||||||
| <3 | 894 (82.8) | 105 (71.4) | 256 (73.6) | 533 (91.1) | 0.18 (0.13-0.22) | NA |
| ≥3 | 164 (15.3) | 42 (28.6) | 81 (23.3) | 41 (7.2) | ||
| Unknown | 21 (1.9) | 0 | 11 (3.2) | 10 (1.7) | ||
| Lymph node groups on scan | ||||||
| 1 | 823 (76.2) | 74 (50.3) | 206 (59.2) | 543 (92.8) | 0.31 (0.26-0.35) | NA |
| >1 | 236 (21.9) | 72 (49.0) | 131 (37.6) | 33 (5.6) | ||
| None identified | 21 (1.9) | 1 (0.7) | 11 (3.2) | 9 (1.5) | ||
| Lymph node basins at surgery | ||||||
| 1 | 881 (81.6) | 99 (67.3) | 222 (63.8) | 560 (95.7) | 0.28 (0.23-0.32) | NA |
| >1 | 193 (17.9) | 46 (31.3) | 123 (35.3) | 24 (4.1) | ||
| None identified | 6 (0.5) | 2 (1.4) | 3 (0.9) | 1 (0.2) | ||
Abbreviations: NA, not applicable; NOS, not otherwise specified.
Effect size metrics include Cramer V for contingency of the 3 body sites vs relevant categories and η2 for comparison of continuous variables.
Data available for 1077 patients.
Data available for 1034 patients.
When stratified based on melanoma arising from the 3 regions of the body (head and neck, trunk, or limb), the treatment groups had several differences from one another, summarized in Table 1. Median (IQR) age at diagnosis was 59.8 (48.7-69.0) years, with patients with HNM being the oldest (median [IQR], 66.2 [51.3-72.8] years), and patients with truncal melanoma the youngest (median [IQR], 58.3 [47.3-68.5] years). Men made up 51.1% (552/1080) of the study population and were represented most in HNM (68.0%, 100/147) and least in limb melanoma (36.8%, 215/585). Head and neck melanomas had a thicker Breslow depth compared with truncal or limb melanoma (2.2 mm vs 1.6 mm and 1.5 mm, respectively). Ulceration was present in 20.3% (210/1034) of the total study population. The most common histological variety was superficial spreading (65.1%, 703/1080) and nodular (20.2%, 218/1080).
Lymph Node Characteristics
Analysis of the LSG patterns showed a difference in number of LN hotspots and LN groups between the 3 different melanoma regions on the LSG (Table 1). Overall, HNM had the highest prevalence of patients, with 3 or more positive LN hotspots (42/147 [28.6%]) and multiple LN groups (72/147 [49.0%]) detected on LSG compared with melanoma arising from other regions. This was the lowest for limb melanoma, where 91.1% (533/585) of patients had 1 to 2 LN hotspots and 92.8% (543/585) had a single LN group on LSG.
Analysis of the SLNB data showed that trunk melanoma had the highest prevalence of cases with more than 1 LN basin, closely followed by HNM, and lowest in limb melanoma (123/348 [35.3%], 46/147 [31.3%], and 24/585 [4.1%], respectively).
SLNB Positivity, False-Negative Results, and Recurrence Patterns per Region
Overall, 20.9% (226/1080) of identified SLNs had microscopic melanoma (Table 2). This was lowest in HNM (19/147 [12.9%]) and highest in the trunk (86/348 [24.7%]).
Table 2. Sentinel Lymph Node Biopsy Test Characteristics and Melanoma Recurrence Pattern Stratified by Body Region.
| Characteristic | No. (%) | Effect size (95% CI)a | |||
|---|---|---|---|---|---|
| Overall (n = 1080) | Head and neck (n = 147) | Trunk (n = 348) | Limb (n = 585) | ||
| SLNB positivity rate (TP) | 226 (20.9) | 19 (12.9) | 86 (24.7) | 121 (20.7) | 0.08 (0.00-0.14) |
| False-negative results (FN) | 39 (3.6) | 10 (6.8) | 15 (4.3) | 14 (2.4) | 0.07 (0.00-0.13) |
| True-negative results (TN) | 815 (75.5) | 118 (80.3) | 247 (71.0) | 450 (76.9) | 0.06 (0.00-0.12) |
| Rate, % (95% CI) | |||||
| False negative (FN/FN+TP) | 14.7 | 34.5 (19.9-52.7) | 14.9 (9.2-23.1) | 10.4 (6.3-16.7) | NA |
| False omission (FN/FN+TN) | 4.6 | 7.8 (4.3-13.8) | 5.7 (3.5-9.2) | 3.0 (1.8-5.0) | |
| Negative predictive value (TN/FN+TN), % (95% CI) | 95.4 | 92.2 (86.2-95.7) | 94.3 (90.8-96.5) | 97.0 (95.0-98.0) | |
| Furthest first recurrence | |||||
| Local (including ITM) | 34 (16.6) | 2 (5.1) | 8 (12.3) | 24 (23.8) | 0.12 (0.00-0.21) |
| Regional (with or without local recurrence) | 51 (24.9) | 14 (35.9) | 16 (24.6) | 21 (20.8) | |
| Distant (with or without locoregional recurrence) | 120 (58.5) | 23 (59.0) | 41 (63.1) | 56 (55.4) | |
| Any time regional recurrence in SLNB negative | 81 (9.5) | 21 (16.0) | 25 (9.5) | 38 (8.2) | 0.08 (0.00-0.15) |
Abbreviations: FN, false-negative results; ITM, in transit metastasis; SLNB, sentinel lymph node biopsy; TN, true-negative results; TP, true-positive results.
Effect size Cramer V for contingency of the 3 body sites vs relevant categories.
The overall FN rate was 14.7% and was higher in HNM group (34.5%) and lowest in the limb group (10.4%). The false omission rate was highest in HNM (7.8%), whereas lowest in the limb group (3.0%). The negative predictive value was strongest in limb melanoma (97.0%) and weakest for HNM (92.2%).
There was a total of 205 (19.0%) patients who had at least 1 melanoma recurrence or progression. Overall, HNM had the highest rate of regional lymph node recurrences at any time in their follow-up (16.0% vs 9.5% trunk and 8.2% limb).
Time-to-Event Analysis
The region of the body that the melanoma originated from was associated with recurrence outcome. First, there was no difference in the MSS between HNM and other regions (Figure, A; HR, 0.81; 95% CI, 0.43-1.5), despite worse tumor characteristics in HNM. However, HNM had a lower RFS than other regions (Figure, B; HR, 0.55; 95% CI, 0.36-0.85).
Figure. Kaplan-Meier Survival Estimates According to Head and Neck Melanoma (HNM) vs Melanomas in Other Regions (Limbs and Trunk).
A, Melanoma-specific survival. B, Recurrence-free survival. C, Melanoma-specific survival. D, Recurrence-free survival. Kaplan-Meier survival estimates in HNM stratified by fewer than 3 hotspots and 3 or more hotspots on lymphoscintigraphic findings (B and D).
To further investigate the differences in outcomes observed, we substratified the entire cohort, and the HNM cases into those with 3 or more LN hotspots and less than 3 LN hotspots on LSG. The MSS and RFS were found to be lower in the overall cohort (eFigure, A and B in Supplement 1), with an HR of 0.38 (95% CI, 0.21-0.70) and 0.51 (95% CI, 0.34-0.77), respectively. This HR was pronounced in patients with HNM with 3 or more LN hotspots compared with those with fewer than 3 LN hotspots (Figure, C; MSS, 0.25; 95% CI, 0.07-0.93 and Figure, D; RFS, 0.37; 95% CI, 0.18-0.77).
Cox Multivariable Analysis
Cox multivariable analysis demonstrated age, Breslow thickness, and sentinel node status to be independently associated with both MSS and RFS (Table 3). Ulceration and melanoma region (HNM vs other regions) were independent predictors of RFS but not MSS. The number of positive LN groups detected on LSG were predictors for MSS and RFS.
Table 3. Multivariable Cox Proportional Hazards Analysis of Independent Risk Factors for Melanoma-Specific Survival and Recurrence-Free Survival.
| Variable | Hazard ratio (95% CI) | |
|---|---|---|
| Melanoma-specific survival | Recurrence-free survival | |
| Age | 1.03 (1.02-1.05) | 1.02 (1.01-1.03) |
| Sex (male) | 1.27 (0.79-2.06) | 1.20 (0.88-1.63) |
| Breslow thickness | 1.11 (1.05-1.16) | 1.11 (1.07-1.15) |
| Sentinel node (positive) | 6.51 (4.09-10.5) | 7.17 (3.05-18.54) |
| Ulceration (yes) | 1.07 (0.64-1.77) | 1.93 (1.39-2.68) |
| Melanoma region (head and neck) | 0.80 (0.35-1.71) | 1.60 (1.01-2.50) |
| LN hotspots on scan, No. | 1.06 (0.79-1.41) | 1.15 (0.95-1.39) |
| LN groups on scan, No. | 2.03 (0.92-3.61) | 1.37 (0.76-2.28) |
Abbreviation: LN, lymph node.
Discussion
This study reports long-term follow-up data of SLNB in melanoma subanalyzed by head and neck, limb, and trunk. We found that melanoma arising from the head and neck had a 2.8-times higher FN rate (34.5% vs 12.3% at any other body sites) and 2 times higher FOR (7.8% vs 3.9% at any other body sites) for SLNB compared to other regions. Head and neck melanoma also has higher regional recurrence compared with melanoma from the limb and trunk. Furthermore, we have demonstrated how HNM cases have significantly more LN hotspots and LN groups detected on LSG per patient and that cases with multiple positive LN hotspots were associated with lower RFS and MSS, suggesting that the complex lymphatic system of the head and neck is a contributing factor to the higher recurrence seen in HNM.
The literature review summary of previously published HNM SLNB series is outlined in Table 4. Of note, many had not reported their FNR, calculated FNR incorrectly, or used FNR and false omission rate interchangeably, highlighting the lack of consistency and quality in the current literature of SLNB for melanoma in the head and neck. Thus, where articles do not report their FNR or false omission rate by its standard definition, as outlined in our methods, or not reported at all, every effort has been made to calculate these based on the data set available in the published manuscript.
Table 4. Primary Patient Cohorts With Over 100 Head and Neck Melanomasa.
| Source | Total HNM | Median Breslow, mm | Positive SLNB, No. (%) | No. | % | Median follow-up, mo | ||
|---|---|---|---|---|---|---|---|---|
| False-negative results | True-negative results | False-negative rate | False omission rate | |||||
| Chao et al, 200310; multi-institution | 287 | 2.2b | 43 (15.0) | 6 | 238 | 12.2 | 2.5 | 15.5 |
| de Wilt et al, 200432; Sydney, Australia | 136 | NR | 14 (10.3) | 11 | 111 | 44.0 | 9.0 | 34b |
| Leong et al, 20063; multi-institution | 614 | 2.2b | 62 (10.1) | NR | NR | NR | NR | 39.6 |
| Teltztrow et al, 200733; Münster, Germany | 106 | NR | 17 (16.0) | 8 | 81 | 32.0 | 9.0 | 46.9b,c |
| Agnese et al, 200734; Columbus, US | 131 | 2.3b | 12 (9.2) | 7 | 112 | 36.8 | 5.9 | 38.4 |
| Carlson et al, 200811; Atlanta, US | 211 | 2.5d | 35 (16.6) | 14 | 162 | 28.6 | 8.0 | 44.3d |
| Gomez-Rivera et al, 200835; Houston, US | 113 | NR | 23 (20.4) | 7 | 83 | 23.3 | 7.8 | 34 |
| Saltman et al, 20109; New York, US | 218 | 2.8 | 28 (12.8) | 10 | 180 | 26.3 | 5.3 | 37 |
| Miller et al, 201136; Portland, US | 153 | 2.1 | 19 (12.4) | 9 | 125 | 32.1 | 6.7 | 28.8 |
| Fadaki et al, 20138; San Francisco, US | 360 | 2.4b | 39 (10.8) | 16 | 305 | 29.1 | 5.0 | 80.4d |
| Evrard et al, 201814; Paris, France | 124 | 2.5 | 23 (18.6) | 8 | 93 | 25.8 | 7.9 | 46.6 |
| Passmore-Webb et al, 201913; Southampton, UK | 143 | 1.8 | 29 (20.3) | 3 | 111 | 9.4 | 2.6 | 33 |
| Bachar et al, 202037; Tel Aviv, Israel | 141 | 2.1 | 18 (12.8) | 6 | 117 | 25.0 | 4.9 | 81.6 |
| Hanks et al, 202038; Ann Arbor, US | 356 | 2.5b | 75 (21.1) | 18 | 263 | 19.4 | 6.4 | 58.8 |
| Echanique et al, 202139; Los Angeles, US | 154 | 1.9b | 6 (3.9) | 7 | 141 | 53.8 | 4.7 | 15.8 |
| Current study, 2022; Cambridge, UK | 147 | 2.2 | 19 (12.9) | 10 | 118 | 34.5 | 7.8 | 51.6 |
| Overall | 3394 | NA | 462 (13.6) | 140 | 2240 | 25.9 | 5.9 | NA |
Abbreviations: HNM, head and neck melanoma; NA, not applicable; NR, not reported; SLNB, sentinel lymph node biopsy.
Where multiple reports exist from the same cohort, the most recent is included here to avoid duplication. The FNR and FOR have been calculated with the data set available if not already calculated according to a standardized definition.
Mean value only reported.
Only patients with negative sentinel lymph node biopsy results were followed up.
Follow-up period or mean reported for cohort inclusive of head and neck and other melanomas.
The positivity rate for SLNB in this study showed HNM to be the lowest at 12.9% compared with limb and trunk (20.7% and 24.7%, respectively) and was lower than most recent single-arm HNM studies as summarized in Table 4 (12%-20.3% for articles published in 2010 or after). However, it was comparable to figures from Chao et al10 for the Sunbelt Melanoma Trial who reported an SLNB positivity rate of 15% for HNM compared with 20% and 23% for the limbs and trunk, respectively (P < .001). In addition, our overall median positivity rate was 21.0%, consistent with the MSLT-1 trial21 (20.8% positivity for all melanoma sites).
We found a high SLNB FNR in HNM (34.5%) compared with the trunk (14.9%) and limb group (10.4%). Although there is large variability in FNRs reported in the literature, compounded by discrepancy on what defines a false-negative result in SLNB, it remains higher than most single-arm observational studies looking at HNM only (Table 4). However, these single-arm studies lack a comparison group, vary in follow-up time, and thus the comparison and conclusions we can draw from their results is limited. While Chao et al10 for the Sunbelt Melanoma Trial reported very low false-negative levels of 12.2% for HNM (median follow-up, 15.5 months), they acknowledge that a true FNR can only be established by longer follow-up,10 such as in the present study, which had a total median follow-up 4.8 years.
Numerous explanations have been attributed to the cause of FN results in SLNB. Sondak and Zager,16 outlined 4 possibilities for FN SLNBs: (1) the initial melanoma biopsy causing disruption in lymphatic drainage, (2) failure of correct LN or basin identification in nuclear medicine, (3) pathologist-associated error, and (4) surgeon-associated error. In line with UK melanoma guidelines, melanomas in the present study had a diagnostic 2-mm margin excision biopsy irrespective of site, and primary closure or secondary healing till biopsy results are available, with no initial flap-based reconstruction.28 Any flap-based reconstruction is done after wide local excision and SLNB, which then does not impede lymphatic drainage. As a single-center study with an experienced group of nuclear medicine physicians, surgeons, and pathologists, the difference in FNs observed between the different regions in this study is not explained by these factors.
However, 1 known difference between these 3 melanoma regions is their lymphatic drainage. Whereas the limbs have a largely linear lymphatic system, the head and neck region is known to be much more complex,18 likely resulting in more challenging sentinel lymph node identification during LSG and surgery. Indeed, in this study, we found that patients with HNM had significantly more positive LN hotspots and LN groups detected on LSG compared with the other regions and that those patients with a higher number of positive LN hotspots had significantly higher recurrence rates and lower MSS.
Limitations
There are some limitations in our study that must be kept in mind when applying to other populations. First, we did not have patients with flap-based reconstruction done prior to the SLNB procedure, which may alter the lymphatic flow compared with populations where a flap-based reconstruction has been performed at initial biopsy. Second, it is difficult to completely account for the group differences in melanoma features between body sites such a histologic subtype, Breslow thickness, and sex, all of which are known to be independent factors of long-term outcomes. Without true matched cohorts, Cox multivariable regression provided some indication of independent risk factors. Lastly, it is unclear what the effect of systemic adjuvant therapy in high-risk stage 2 melanoma, disproportionately represented in HNM, will have on the long-term outcomes in these groups as future practice evolves.40
Conclusions
What does this mean for clinical practice? Current international guidelines including those from the National Institute for Health and Care Excellence in the UK, and the National Comprehensive Cancer Network in the US, do not stratify surveillance methods based on primary site of melanoma.28,41 However, the findings of this cohort study suggest that for HNM, FNRs and regional recurrence rates are significantly higher compared with other body sites. Thus, we advocate for careful preoperative counselling, and consider enhanced regional surveillance in patients with HNM to enable timely detection of metastasis and guide management, leading to improved quality of life and outcomes for patients with HNM.
eFigure. Kaplan-Meier survival estimates for melanoma at all body sites (HNM + trunk + limb) sub-stratified by <3 hotspots and ≥3 hotspots on lymphoscintigraphy
Data Sharing Statement.
References
- 1.Melanoma skin cancer statistics. Cancer Research UK. Accessed September 10, 2022. https://www.cancerresearchuk.org/health-professional/cancer-statistics/statistics-by-cancer-type/melanoma-skin-cancer
- 2.Nakamura Y. The role and necessity of sentinel lymph node biopsy for invasive melanoma. Front Med (Lausanne). 2019;6:231. doi: 10.3389/fmed.2019.00231 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Leong SPL, Accortt NA, Essner R, et al. ; Sentinel Lymph Node Working Group . 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(4):370-373. doi: 10.1001/archotol.132.4.370 [DOI] [PubMed] [Google Scholar]
- 4.Sperry SM, Charlton ME, Pagedar NA. Association of sentinel lymph node biopsy with survival for head and neck melanoma: survival analysis using the SEER database. JAMA Otolaryngol Head Neck Surg. 2014;140(12):1101-1109. doi: 10.1001/jamaoto.2014.2530 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Faries MB, Thompson JF, Cochran AJ, et al. Completion dissection or observation for sentinel-node metastasis in melanoma. N Engl J Med. 2017;376(23):2211-2222. doi: 10.1056/NEJMoa1613210 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Bray HN, Simpson MC, Zahirsha ZS, et al. Head and neck melanoma incidence trends in the pediatric, adolescent, and young adult population of the United States and Canada, 1995-2014. JAMA Otolaryngol Head Neck Surg. 2019;145(11):1064-1072. doi: 10.1001/jamaoto.2019.2769 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Al Ghazal P, Gutzmer R, Satzger I, et al. Lower prevalence of lymphatic metastasis and poorer survival of the sentinel node-negative patients limit the prognostic value of sentinel node biopsy for head or neck melanomas. Melanoma Res. 2014;24(2):158-164. doi: 10.1097/CMR.0000000000000042 [DOI] [PubMed] [Google Scholar]
- 8.Fadaki N, Li R, Parrett B, et al. Is head and neck melanoma different from trunk and extremity melanomas with respect to sentinel lymph node status and clinical outcome? Ann Surg Oncol. 2013;20(9):3089-3097. doi: 10.1245/s10434-013-2977-7 [DOI] [PubMed] [Google Scholar]
- 9.Saltman BE, Ganly I, Patel SG, et al. Prognostic implication of sentinel lymph node biopsy in cutaneous head and neck melanoma. Head Neck. 2010;32(12):1686-1692. doi: 10.1002/hed.21390 [DOI] [PubMed] [Google Scholar]
- 10.Chao C, Wong SL, Edwards MJ, et al. ; Sunbelt Melanoma Trial Group . Sentinel lymph node biopsy for head and neck melanomas. Ann Surg Oncol. 2003;10(1):21-26. doi: 10.1245/ASO.2003.06.007 [DOI] [PubMed] [Google Scholar]
- 11.Carlson GW, Page AJ, Cohen C, et al. Regional recurrence after negative sentinel lymph node biopsy for melanoma. Ann Surg. 2008;248(3):378-386. doi: 10.1097/SLA.0b013e3181855718 [DOI] [PubMed] [Google Scholar]
- 12.Erman AB, Collar RM, Griffith KA, et al. Sentinel lymph node biopsy is accurate and prognostic in head and neck melanoma. Cancer. 2012;118(4):1040-1047. doi: 10.1002/cncr.26288 [DOI] [PubMed] [Google Scholar]
- 13.Passmore-Webb B, Gurney B, Yuen HM, et al. Sentinel lymph node biopsy for melanoma of the head and neck: a multicentre study to examine safety, efficacy, and prognostic value. Br J Oral Maxillofac Surg. 2019;57(9):891-897. doi: 10.1016/j.bjoms.2019.07.022 [DOI] [PubMed] [Google Scholar]
- 14.Evrard D, Routier E, Mateus C, et al. Sentinel lymph node biopsy in cutaneous head and neck melanoma. Eur Arch Otorhinolaryngol. 2018;275(5):1271-1279. doi: 10.1007/s00405-018-4934-3 [DOI] [PubMed] [Google Scholar]
- 15.McDonald K, Page AJ, Jordan SW, et al. Analysis of regional recurrence after negative sentinel lymph node biopsy for head and neck melanoma. Head Neck. 2013;35(5):667-671. doi: 10.1002/hed.23013 [DOI] [PubMed] [Google Scholar]
- 16.Sondak VK, Zager JS. Who is to blame for false-negative sentinel node biopsies in melanoma? Ann Surg Oncol. 2010;17(3):670-673. doi: 10.1245/s10434-009-0857-y [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Klop WMC, Veenstra HJ, Vermeeren L, Nieweg OE, Balm AJM, Lohuis PJFM. Assessment of lymphatic drainage patterns and implications for the extent of neck dissection in head and neck melanoma patients. J Surg Oncol. 2011;103(8):756-760. doi: 10.1002/jso.21865 [DOI] [PubMed] [Google Scholar]
- 18.Lavelli V, Ferrari C, Santo G, et al. The lymphoscintigraphic study of unpredictable head and neck cutaneous melanoma lymphatic drainage. Biomedicines. 2020;8(4):70. doi: 10.3390/biomedicines8040070 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Reynolds HM, Dunbar PR, Uren RF, Blackett SA, Thompson JF, Smith NP. Three-dimensional visualisation of lymphatic drainage patterns in patients with cutaneous melanoma. Lancet Oncol. 2007;8(9):806-812. doi: 10.1016/S1470-2045(07)70176-6 [DOI] [PubMed] [Google Scholar]
- 20.Leiter U, Stadler R, Mauch C, et al. ; German Dermatologic Cooperative Oncology Group . Final analysis of DeCOG-SLT trial: no survival benefit for complete lymph node dissection in patients with melanoma with positive sentinel node. J Clin Oncol. 2019;37(32):3000-3008. doi: 10.1200/JCO.18.02306 [DOI] [PubMed] [Google Scholar]
- 21.Morton DL, Thompson JF, Cochran AJ, et al. ; MSLT Group . Final trial report of sentinel-node biopsy versus nodal observation in melanoma. N Engl J Med. 2014;370(7):599-609. doi: 10.1056/NEJMoa1310460 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Morton DL, Thompson JF, Cochran AJ, et al. ; MSLT Group . Sentinel-node biopsy or nodal observation in melanoma. N Engl J Med. 2006;355(13):1307-1317. doi: 10.1056/NEJMoa060992 [DOI] [PubMed] [Google Scholar]
- 23.Leiter U, Stadler R, Mauch C, et al. ; German Dermatologic Cooperative Oncology Group (DeCOG) . Complete lymph node dissection versus no dissection in patients with sentinel lymph node biopsy positive melanoma (DeCOG-SLT): a multicentre, randomised, phase 3 trial. Lancet Oncol. 2016;17(6):757-767. doi: 10.1016/S1470-2045(16)00141-8 [DOI] [PubMed] [Google Scholar]
- 24.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. 2003;200(3):314-319. doi: 10.1002/path.1365 [DOI] [PubMed] [Google Scholar]
- 25.Moncrieff M, Pywell S, Snelling A, et al. Effectiveness of SPECT/CT imaging for sentinel node biopsy staging of primary cutaneous melanoma and patient outcomes. Ann Surg Oncol. 2022;29(2):767-775. doi: 10.1245/s10434-021-10911-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Cook MG, Massi D, Szumera-Ciećkiewicz A, et al. ; EORTC Melanoma Group . An updated European Organisation for Research and Treatment of Cancer (EORTC) protocol for pathological evaluation of sentinel lymph nodes for melanoma. Eur J Cancer. 2019;114:1-7. doi: 10.1016/j.ejca.2019.03.010 [DOI] [PubMed] [Google Scholar]
- 27.Robbins KT, Clayman G, Levine PA, et al. ; American Head and Neck Society; American Academy of Otolaryngology–Head and Neck Surgery . Neck dissection classification update: revisions proposed by the American Head and Neck Society and the American Academy of Otolaryngology-Head and Neck Surgery. Arch Otolaryngol Head Neck Surg. 2002;128(7):751-758. doi: 10.1001/archotol.128.7.751 [DOI] [PubMed] [Google Scholar]
- 28.National Institute for Health and Care Excellence NG14 . Melanoma: assessment and management. Accessed September 10, 2022. https://www.nice.org.uk/guidance/ng14 [PubMed]
- 29.Cramér H. Mathematical methods of statistics. Princeton University Press; 1946. [Google Scholar]
- 30.Cohen J. Statistical power analysis for the behavioral sciences. 2nd ed. L. Erlbaum Associates; 1988. [Google Scholar]
- 31.Cohen J, Cohen P. Applied multiple regression/correlation analysis for the behavioral sciences. 2nd ed. Lawrence Erlbaum Associates; 1983. [Google Scholar]
- 32.de Wilt JHW, Thompson JF, Uren RF, et al. Correlation between preoperative lymphoscintigraphy and metastatic nodal disease sites in 362 patients with cutaneous melanomas of the head and neck. Ann Surg. 2004;239(4):544-552. doi: 10.1097/01.sla.0000118570.26997.a1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Teltzrow T, Osinga J, Schwipper V. Reliability of sentinel lymph-node extirpation as a diagnostic method for malignant melanoma of the head and neck region. Int J Oral Maxillofac Surg. 2007;36(6):481-487. doi: 10.1016/j.ijom.2007.02.007 [DOI] [PubMed] [Google Scholar]
- 34.Agnese DM, Maupin R, Tillman B, Pozderac RD, Magro C, Walker MJ. Head and neck melanoma in the sentinel lymph node era. Arch Otolaryngol Head Neck Surg. 2007;133(11):1121-1124. doi: 10.1001/archotol.133.11.1121 [DOI] [PubMed] [Google Scholar]
- 35.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(10):1284-1294. doi: 10.1002/hed.20875 [DOI] [PubMed] [Google Scholar]
- 36.Miller MW, Vetto JT, Monroe MM, Weerasinghe R, Andersen PE, Gross ND. False-negative sentinel lymph node biopsy in head and neck melanoma. Otolaryngol Head Neck Surg. 2011;145(4):606-611. doi: 10.1177/0194599811411878 [DOI] [PubMed] [Google Scholar]
- 37.Bachar G, Tzelnick S, Amiti N, Gutman H. Patterns of failure in patients with cutaneous head and neck melanoma. Eur J Surg Oncol. 2020;46(5):914-917. doi: 10.1016/j.ejso.2019.12.021 [DOI] [PubMed] [Google Scholar]
- 38.Hanks JE, Kovatch KJ, Ali SA, et al. Sentinel lymph node biopsy in head and neck melanoma: Long-term outcomes, prognostic value, accuracy, and safety. Otolaryngol Head Neck Surg. 2020;162(4):520-529. doi: 10.1177/0194599819899934 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Echanique KA, Ghazizadeh S, Moon A, et al. Head & neck melanoma: a 22-year experience of recurrence following sentinel lymph node biopsy. Laryngoscope Investig Otolaryngol. 2021;6(4):738-746. doi: 10.1002/lio2.605 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Luke JJ, Rutkowski P, Queirolo P, et al. ; KEYNOTE-716 Investigators . Pembrolizumab versus placebo as adjuvant therapy in completely resected stage IIB or IIC melanoma (KEYNOTE-716): a randomised, double-blind, phase 3 trial. Lancet. 2022;399(10336):1718-1729. doi: 10.1016/S0140-6736(22)00562-1 [DOI] [PubMed] [Google Scholar]
- 41.NCCN Clinical Practice Guidelines in Oncology . (NCCN Guidelines) for Cutaneous Melanoma V.3.2022. Accessed September 10, 2022. https://www.nccn.org/guidelines/guidelines-detail?category=1&id=1492
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
eFigure. Kaplan-Meier survival estimates for melanoma at all body sites (HNM + trunk + limb) sub-stratified by <3 hotspots and ≥3 hotspots on lymphoscintigraphy
Data Sharing Statement.
