Enhanced risk of multiple sentinel lymph node basins in truncal and head and neck melanoma

Introduction

The cutaneous lymphatic drainage of the trunk was initially divided by Sappey into four quadrants (right upper and lower trunk and left upper and lower trunk) based on a vertical line separating the trunk into two halves and a horizontal line joining the iliac crest.¹ Sappey thought that each quadrant would drain predictably to its respective lymph node basin (for example, the right upper trunk region would drain to the right axilla). However, the advent of lymphoscintigraphy demonstrated the potential for cutaneous drainage within the trunk to less predictable locations that do not always match Sappey’s predicted drainage patterns. The introduction of the sentinel node biopsy (SeNBx) provided a means to track the inter-individual lymphatic drainage of a primary cutaneous melanoma of the trunk.¹ From these studies, it was noted that truncal melanomas have a propensity to drain to multiple lymph node basins (MLNB).² Thompson and colleagues used lymphoscintigraphy studies in 1759 patients with primary cutaneous melanoma to demonstrate that truncal lesions drain to an average of 1.78 sentinel nodes, that head and neck lesions drain to an average of 2.1 sentinel nodes, while extremity melanomas only drain to an average of 1.1 sentinel nodes.³ The authors further illustrated that both truncal and head and neck melanoma are more likely to drain to sentinel nodes in interval or unexpected locations, compared to extremity melanoma. Head and neck melanomas present a different set of issues due to their higher rates of false negative SeNBx (20.4%)⁴ compared to non-head and neck cutaneous melanomas (4.8%).⁵ Puza and colleagues demonstrated that drainage to MLNB in head and neck melanoma is associated with lower rates of false negative SeNBx, perhaps due to wider sampling.⁶

To date, consensus on whether drainage to MLNB confers a worse overall prognosis remains elusive. Some studies have shown that MLNB from truncal melanoma is associated with poor survival compared to drainage to a single lymph node basin (SLNB)^2,7; other studies found MLNB does not imply a worse prognosis.^8–10 In terms of established prognostic factors, the three most significant clinicopathologic factors known to affect melanoma prognosis include Breslow thickness (BT), tumour ulceration and SeNBx status. In the last decade, the SeNBx status has assumed greater importance in melanoma staging and prognostication; it potentially also plays a therapeutic role in early metastatic disease.¹¹ Even more recently, the SeNBx status determines a patient’s eligibility for immunotherapy trials, which have displayed an improvement in patient survival.¹² The eighth edition of the American Joint Committee on Cancer (AJCC) on melanoma staging emphasised the prognostic importance of even microscopic (< 1 mm) tumour burden within the sentinel lymph node in guiding adjuvant treatment eligibility and surveillance frequency protocols.¹³ A 2015 Cochrane review found an improved disease-free survival and rate of local and regional recurrence in patients undergoing SeNBx.¹⁴ This review also found higher rates of surgical morbidity relating to postoperative complications in patients who had undergone SeNBx.¹⁴

Therefore, a more thorough understanding of lymphatic drainage for head and neck and truncal melanoma may enable more accurate surgical localisation of likely metastatic disease and clinical monitoring. Our aim was to compare the difference in drainage patterns to an SLNB versus MLNB between truncal and head and neck melanomas and to identify further, potentially significant, prognostic factors that predispose to MLNB drainage. In addition, we aimed to develop a better understanding of the clinical significance of MLNB in terms of the rate of SeNBx positivity, and to ascertain whether such an observation may assist in predicting which primary tumours are more likely to metastasise to multiple nodal basins.

Methods

Following study approval by the Alfred Health Research and Ethics Committee [approval number: 349/19], this retrospective study was conducted within a statewide melanoma referral service, at a tertiary referral hospital in Melbourne, Australia. Patients were identified based on the plastics and reconstructive surgery theatre booking system during a five-year period (January 2014–December 2018). Patients were included if they:

• had a truncal or head and neck melanoma and were scheduled for a wide local excision (WLE) and SeNBx

• had initial biopsy results that stated the melanoma type, cell type and BT.

Patients were excluded if they underwent nodal biopsy alone and therefore did not undergo lymphoscintogram to detect the sentinel node.

Patients were followed up from the date of their lymphoscintogram within the data collection period to 31 December 2018. Patients underwent a SeNBx if they met the criteria detailed in the AJCC guidelines.¹³

Medical records and investigation results of the identified patients were reviewed. The patient-related demographics were collected, included age and sex. The following melanoma-related variables were collected:

• date of initial biopsy, melanoma type (superficial spreading, nodular, lentigo maligna, desmoplastic)

• melanoma location (trunk or head and neck)

• melanoma cell type (epitheliod, spindled, mixed, naevoid)

• BT

• Clark level, mitotic rate (mm²)

• presence of lymphovascular or perivascular invasion

• regression, desmoplasia, neurotropism

• microsatellites

• date of lymphoscintography and date of WLE and SeNBx

• number and location of drainage basins identified on lymphoscintography presence and number of positive sentinel lymph nodes biopsied

• consequent basin clearance and histopathology results

• presence of postoperative complications (limited to three months postoperatively for SeNBx and basin clearance) and

• presence of recurrence (limited to the duration of the retrospective review).

Drainage basins were classified as axilla, cervical (including parotid), groin and other (for basins without a large cluster of lymph nodes).

Continuous variables were summarised using means and standard deviations (SD) or medians and interquartile ranges (IQR) according to data type and distribution. Categorical variables were presented as counts and percentages. Comparisons between groups were made using Student’s t-test for normally distributed continuous variables, Wilcoxon rank-sum test for non-normally distributed continuous variables and chi-square or Fisher’s exact test as appropriate for categorical variables. No adjustment was made for multiple outcome comparisons. All calculated p-values were two-tailed and a p <0.05 indicated statistical significance. Analyses were performed with SAS® software version 9.4 (100 SAS Campus Dr, Cary, NC 27513, United States).

Results

A total of 143 patients were identified as being eligible for inclusion in this study, including 95 patients with truncal melanoma and 48 patients with head and neck melanoma.

Patient and melanoma demographics

Male subjects were found to have a higher incidence of truncal (64.2%, n = 61) and head and neck (77.1%, n = 37) melanoma (Table 1). The overall mean patient age was 57.8 years, with head and neck melanoma patients presenting at a slightly older average age (59.9 years) than truncal melanoma patients (56.8 years). Superficial spreading melanoma (SSM) was the most commonly diagnosed type of melanoma in the truncal melanoma group (53.9%), while the incidence of nodular and SSM was equivalent for head and neck melanomas (40.0%). Melanomas in the head and neck region exhibited a more invasive profile with a higher BT (2.1(1.73–4.3) versus 1.9(1.0–3.0), p <0.01), Clark level (4.0(4.0–4.0) versus 4.0(3.0–4.0), p < 0.01) and mitotic rate (4.0 versus 2.0, p < 0.05) compared to truncal melanomas. The epitheliod cell type was the predominant cell type in truncal and head and neck melanoma (88.2% and 72.5%, respectively).

In terms of drainage patterns, truncal melanomas drained predominantly to the axilla (71.2%), followed by cervical (14.4%) and groin (14.4%) lymph node basins. As expected, head and neck melanomas drained principally to the cervical region (97.9%), with the remaining patients draining primarily to the axilla (2.1%). Overall, a total of 29 patients (20.3%) were found to have a positive SeNBx. Of these patients, truncal melanomas trended towards a higher proportion of positive SeNBx compared to head and neck melanoma (24.2% versus 12.5%); however, this trend did not reach significance. Of the 29 positive SeNBx, the majority (n = 21) underwent a further lymph node basin clearance. One patient developed a wound seroma after SeNBx of the groin, while the other complications were observed in patients who had undergone completion of lymph node basin clearance. Complications following lymph node basin clearance occurred at a rate of 8.4 per cent in truncal melanoma and 4.2 per cent in head and neck melanoma. Complications recorded during the data collection time frame ranged from seroma formation, wound breakdown, lymphoedema and nerve injury. Melanoma recurrence was lower in the truncal melanoma cohort (4.2%) compared with the head and neck melanoma cohort (10.4%).

Single versus multiple nodal basins

Drainage to MLNB occurred in 27.9 per cent (n = 40) of all patients. Patients with MLNB had a higher positive SeNBx rate (40.0%) compared to SLNB (12.6%, p < 0.001) (Table 2). Melanomas that drained to MLNB were not found to have a more invasive profile, with no significant increase in BT or mitotic rate, compared to melanomas that drained to SLNB. However, nearly half (43.2%) of melanomas that drained to MLNB were ulcerated compared to those that drained to SLNB (26.0%, p < 0.05). Melanomas that drained to MLNB showed nearly twice the risk of complications compared to those that drained to SLNB (10.0% versus 5.8%, p = 0.38). Overall, the risk of recurrence was similar between MLNB (5.0%) and SLNB (6.8%).

Truncal melanomas

Over a third (36.8%) of all truncal melanomas were found to distribute to MLNB, whereas only 10.4 per cent of head and neck melanomas did so (p < 0.001) (Table 1). Truncal melanomas that drained to MLNB were not associated with a slightly younger patient cohort (53.5 versus 58.7 years, p = 0.07) but were associated with a higher BT (2.1 versus 1.5, p < 0.05), ulceration rate (40.6% versus 20.7%, p < 0.05) and mitotic rate (3.0 versus 2.0, p < 0.05) compared to drainage to an SLNB (Table 3). Drainage to MLNB was also associated with a higher rate of positive SeNBx (42.9% versus 13.3%, p < 0.05) compared to the rate in SLNB. Four patients experienced recurrence during the data collection period: two with each of SLNB (3.3%) and MLNB (5.7%) on SeNBx. The two recurrences occurred in patients with drainage to MLNB (both with positive SeNBx): one at the site of the primary lesion and the other as a subcutaneous deposit distant from the primary lesion. The recurrences in patients with drainage to SLNB (both with negative SeNBx) were found within the expected nodal basin in one case, and in an unexpected location in the other patient.

Further analysis of the truncal melanomas that drained to MLNB did not reveal differences between primary melanoma characteristics that would aid in predicting SeNBx positivity (Table 2). However, the average age of positive SeNBx patients (47.2 ± 12.8 years) was significantly younger compared to those with a negative SeNBx (58.2 ± 14, p < 0.05). As expected, truncal melanoma patients with a positive SeNBx were more likely to have an ulcerated melanoma (60.0% versus 23.5%) compared those with a negative SeNBx (p < 0.05). There were no significant differences identified in average BT or Clark level in patients with a positive SeNBx (2.4 mm (1.5–4.5) and level IV (4.0–4.0) respectively) compared to patients with a negative SeNBx (1.95 mm (1.2–3.5) and level IV (4.0 (3.0–4.0)) (p = 0.28 and p = 0.31, respectively).

Head and neck melanomas

Only 10.4 per cent of head and neck melanoma patients drained to MLNB. These patients tended to have ulcerated primary melanomas with a higher BT and mitotic rate. None of the patients with drainage to MLNB experienced complications or recurrences within the study period. Of note, 11.6 per cent (n = 5) of the patients with drainage to SLNB had recurrence of their disease: four within the sentinel node basin and one with a subcutaneous deposit.

Discussion

Here, we present a cohort of truncal and head and neck melanoma, both with the propensity to drain to MLNB, compared with the published rates of drainage to MLNB in extremity melanomas. Our rates of drainage to MLNB are similar to the 23.0–39.0 per cent range reported within the literature.^15,16 Among these studies, only Porter and colleagues reported a statistically significant increase in positive SeNBx with drainage to MLNB compared to SLNB.² Our results similarly demonstrated a statistically significant association between positive SeNBx and drainage to MLNB compared to drainage to an SLNB. This pattern correlated with a worse prognosis based on a higher melanoma stage.¹³ Furthermore, our 30.0 per cent rate of positive SeNBx with drainage to MLNB is higher compared to the 17.5–23.⁹ per cent published within the literature.^8,16,17 The presence of ulceration was found to be associated with drainage to MLNB within both the truncal and head and neck cohorts. This study highlighted the importance of ulceration in light of the recent AJCC guidelines by demonstrating its association with drainage to MLNB in both cohorts.¹³ More than one-third of all truncal melanomas drained to MLNB within our cohort, consistent with the 29.0–39.0 per cent range within the literature.^8,15,18 This study demonstrated that a BT of 2.1 mm, mitotic rate of 3.0 mitoses and the presence of ulceration are risk factors for drainage to MLNB in truncal melanoma. Furthermore, we demonstrated that patients with truncal melanomas and positive SeNBx were younger and have a potentially poorer prognosis given their likely longer lifespan compared to older patients.

Surgical complications in truncal melanomas were found to have a two-fold increase with drainage to MLNB compared to SLNB. Similar to the MSLT-I trial, the complications within our patient cohort included wound dehiscence/breakdown, seroma and lymphoedema.¹¹ In contrast, drainage to MLNB for head and neck melanoma was associated with no complications. This result may reflect that healing post-SeNBx within the cervical region is more uneventful compared to the axilla or groin for truncal melanoma.

Head and neck melanoma has been shown to have a worse five-year disease-free survival rate compared to melanomas in the trunk or extremities.¹⁹ Our study similarly demonstrated a higher recurrence rate within the head and neck melanoma cohort compared to the truncal melanoma cohort, despite having a lower rate of positive SeNBx. Our finding may be explained by De Rosa and colleagues’ study, which demonstrated that head and neck melanomas are associated with a 20.⁴ per cent rate of false negative SeNBx.⁴ Furthermore, our review supports evidence within the literature by demonstrating that head and neck melanomas that drain to MLNB have a lower incidence of recurrence.⁶ Overall, drainage to MLNB may allow for a complete sampling of sentinel nodes that drain the primary tumour. This may make the analysis of SeNBx more sensitive at detecting metastatic disease, especially within the head and neck melanoma cohort. While findings of metastatic disease prognosticate a more aggressive disease, it may also result in a patient’s eligibility for immunotherapy trials, more rigorous surveillance and therefore improved management of their melanoma. A prospective study comparing SeNBx in head and neck melanoma and truncal melanoma to extremity melanoma would further aid comparison, given the different number of drainage sites.

Limitations

Our study is limited by its retrospective nature and small sample size, especially for the head and neck melanoma cohort. Follow-up for a minimum of five years for all patients would have provided further information on complications and recurrence rates. A prospective study comparing SenBx in head and neck melanoma and truncal melanoma to extremity melanoma would aid further comparison given the different number of drainage sites.

Conclusion

We have demonstrated that drainage to MLNB has a higher positive SeNBx rate compared to SLNB. The presence of ulceration is a risk factor for drainage to MLNB in both truncal and head and neck melanoma, while a BT of greater than 2.1 mm and a mitotic rate of 3.0 is a further risk factor for drainage to MLNB in truncal melanomas. One-third of the truncal melanomas drained to MLNB within our cohort. Furthermore, truncal melanoma patients with a positive SeNBx tended to be younger and may potentially have a poorer prognosis due to their expected longer lifespan. Finally, within our small cohort, we showed that drainage to MLNB in both truncal and head and neck melanoma has a lower recurrence rate, potentially demonstrating some therapeutic benefit of performing SeNBx.

Conflict of Interest

The authors have no conflicts of interest to disclose

Funding declaration

The authors received no financial support for the research, authorship, and/or publication of this article.

Revised: 2021 May 26; September 1 AEST