Effect of Multiple–Nodal Basin Drainage on Cutaneous Melanoma

Hypothesis  The number of nodal basins draining a primary cutaneous melanoma is not an independent predictor of outcome.

Design  Post hoc analysis of patients entered into a randomized, prospective study.

Setting  Multi-institutional academic and community medical centers.

Patients  Patients aged 18 to 70 years with melanoma 1.0 mm or greater Breslow thickness.

Interventions  Wide local excision and sentinel lymph node biopsy were performed on all patients; patients with sentinel lymph node metastases underwent completion lymphadenectomy. Patients with multiple–nodal basin drainage were compared with those with single–nodal basin drainage.

Main Outcome Measures  Sentinel lymph node status, locoregional recurrence-free survival, disease-free survival, and overall survival.

Results  A total of 2060 patients with single–nodal basin drainage (n = 1709 [83% of cohort]) were included in the analysis, with a median follow-up of 50 months. On univariate analysis, the group with multiple–nodal basin drainage (n = 351) was associated with female sex and primary tumor regression (P < .001). In addition, multiple–nodal basin drainage was associated with truncal primary tumor location (73.2%), while single–nodal basin drainage was more common for extremity tumors (50.9%; P < .001). On multivariate analysis, there were no differences in the rate of sentinel lymph node metastasis, disease-free survival, or overall survival between the groups. Interestingly, locoregional recurrence was significantly worse in the single–nodal basin drainage group (P = .003).

Conclusions  Multiple–nodal basin drainage does not confer a worse prognosis for patients with cutaneous melanoma. In fact, single–nodal basin drainage appears to be associated with a greater risk of locoregional recurrence.

Of the clinicopathologic factors known to impact outcome for early-stage melanoma, the status of the sentinel lymph node (SLN) appears to be the most important.^1,2 It has become increasingly clear that melanoma arising from the trunk and the head and neck regions have ambiguous drainage patterns, while extremity melanomas more reliably drain to either the inguinal or axillary basins, with occasional popliteal or epitrochlear drainage.^3-6 Interval nodes, or lymph nodes outside traditional nodal basins, are also found in a small proportion of cases.^7-12

Prior data demonstrate that nearly one-third of melanoma tumors arising on the trunk will drain to multiple nodal basins.^13,14 It has been suggested that cutaneous melanoma that has metastasized to more than 1 SLN basin carries a worse prognosis, independent of the actual tumor burden.¹⁵ However, there is conflicting evidence whether drainage to multiple nodal basins (regardless of the presence of nodal metastases) is associated with a worse outcome. Jimenez et al¹⁶ showed that multiple–nodal basin drainage for truncal melanoma was associated with a worse survival compared with patients who had drainage to only 1 basin, independent of SLN status. Porter et al¹⁴ also found multiple–nodal basin drainage to be independently associated with an increased risk of nodal metastasis. On the other hand, other studies have found that multiple–nodal basin drainage does not increase the risk of nodal metastasis or adversely affect prognosis compared with drainage to a single basin.^17,18 Because of the conflicting data regarding the impact of multiple basin drainage, we set out to determine the true effect of multiple–nodal basin drainage by analyzing the data from a large prospective, multi-institutional study.

The Sunbelt Melanoma Trial is a prospective, randomized, multi-institutional study that involved 79 centers in the United States and Canada. Each site's institutional review board approved the study, and all participants gave signed informed consent. Patients were eligible if they were aged from 18 to 70 years, had a cutaneous melanoma 1 mm or greater in Breslow thickness, and did not have clinical evidence of nodal or distant metastases. All patients underwent wide local excision of the primary melanoma and SLN biopsy (the procedural details of the Sunbelt Melanoma Trial have been previously described¹⁹). The Sunbelt Melanoma Trial was designed to evaluate the role of reverse transcriptase–polymerase chain reaction for ultrastaging as well as lymphadenectomy and adjuvant interferon alfa-2b for patients with early nodal metastasis.²⁰ Randomization was accomplished using a randomized permuted block design with stratification by tumor thickness and ulceration with reassignment to ensure that the number of patients in each arm in each center was roughly balanced.

For the current study, patients with complete data were included and divided into 2 groups: those with single–nodal basin drainage and those with multiple–nodal basin drainage. The 2 groups were compared in terms of SLN metastasis and survival. The survival analyses performed in the 2 groups included locoregional recurrence-free survival (LRRFS), disease-free survival, and overall survival. Locoregional recurrence-free survival included only those recurrences that were local, in transit, or in a regional lymph node basin, while disease-free survival included any type of recurrence of disease. There were 5 distinct lymph node basins used to categorize patients' sentinel nodes: the axilla, groin, epitrochlear fossa, popliteal fossa, and neck, which included supraclavicular and parotid lymph nodes. The neck was counted as a single nodal basin for the purposes of this trial, as the primary question of the Sunbelt Melanoma Trial was to determine the role of molecular staging as well as lymphadenectomy and adjuvant interferon.

Univariate analyses to determine clinical and pathologic factors correlating with the number of draining basins were performed using likelihood tests for discrete variables and Mann-Whitney tests for the continuous variables. Multivariate analyses were also performed using binary logistic regression for SLN status and the Cox proportional hazard model was used for the survival analyses. Survival curves were generated using the Kaplan-Meier method. P < .05 was considered significant. All analyses were performed with SPSS software, version 14.0 (SPSS Inc, Chicago, Illinois).

A total of 2060 patients were included in this analysis. There were 884 (42.9%) women and 1176 (57.1%) men, with a median age of 50.5 years. The median follow-up time was 50 months. Nine hundred twenty patients (44.7%) had extremity melanomas, 886 (43.0%) had truncal melanomas, and 253 (12.3%) had melanoma of the head and neck. There were 1709 (83%) patients with single–nodal basin drainage and 351 (17%) patients with multiple–nodal basin drainage. Of the patients with multiple-basin drainage, 267 (76.1%) did not have metastases in the SLN, while 84 (23.9%) had tumor-positive SLNs. Of the patients with tumor-positive SLNs, 53 (63.1%) had a single positive SLN, 26 (31.0%) had 2 positive SLNs, 4 (4.8%) had 3 positive SLNs, and 1 (1.2%) had 4 positive SLNs. Of the 31 patients with 2 or more positive SLNs, 21 (67.7%) had a positive SLN in 2 separate nodal basins, while the other 10 patients had nodal metastases in only 1 of the mapped basins. Only 3 patients (0.9%) in the multiple–nodal basin drainage group had melanomas that drained to more than 2 lymph node basins.

There were several interesting differences between the 2 groups with regards to clinicopathologic features (Table 1). Women were more likely to have single–nodal basin drainage than men (P < .001). The anatomic site of the primary tumor was also different between the 2 groups (P < .001). Extremity tumors were much more likely to drain to a single nodal basin (94.5% of cases), while the truncal melanomas were more likely to drain to multiple basins (29.0% of cases). Despite the fact that head and neck melanomas typically have ambiguous drainage and are believed to frequently exhibit drainage to multiple basins, head and neck tumors fell between the other 2 sites with 17.0% in the multiple–nodal basin drainage group.

The number of draining nodal basins was found to affect the rate of SLN metastasis by univariate analysis (P = .02); however, this difference did not remain significant on multivariate analysis (P = .14) (Table 2). The number of draining SLNs was not associated with the risk of nodal metastasis (P = .19). Interestingly, when considering only patients with a positive SLN, multiple–nodal basin drainage was associated with a higher risk of having more than 1 metastatic node (37% vs 21%; P = .004). Kaplan-Meier survival curves showed that LRRFS was significantly worse in the single–nodal basin drainage group (P = .003), but the number of draining nodal basins did not affect either disease-free survival (P = .78) or overall survival (P = .22) (Figure). The number of draining nodal basins remained a significant predictor of LRRFS on multivariate Cox regression analysis (Table 3). Additionally, analysis by the number of nodal basins containing metastatic disease again demonstrates that patients with single–nodal basin drainage have a worse outcome than patients with multiple–nodal basin drainage in terms of LRRFS (odds ratio, 0.344; 95% confidence interval, 0.187-0.633; P = .001), but there was no difference in disease-free survival or overall survival. Further stratified analysis to determine the effect modification between tumor location and the number of draining nodal basins was conducted using a proportional hazards model. This revealed that for patients with extremity melanoma, the number of nodal basins was not significant (P = .79). However, for patients with truncal melanoma, the number of draining nodal basins was significant (P = .01), indicating that anatomic site is an effect modifier. Likewise, sex is an effect modifier in the proportional hazards model (men, P = .02; women, P = .76).

There were 131 patients who had melanomas recur either locally or in transit; 39.7% of these patients had a positive SLN at the time of treatment for their primary melanoma, and they were more likely to have had single–nodal basin drainage (P = .02). There were 58 local/in-transit recurrences in patients with extremity melanomas compared with 51 local/in-transit recurrences in patients with trunk melanomas (P = .69). Given that only 5.5% of extremity melanomas drained to multiple basins, analysis to determine the impact of nodal basin drainage pattern on LRRFS for truncal and head/neck primary sites was conducted. Patients with nonextremity tumors that drained to a single nodal basin had a worse 5-year LRRFS (86.3%) compared with 95.7% in those whose tumors drained to multiple basins (P < .001).

There were 52 patients who had melanomas recur in a previously mapped nodal basin; 3.8% of these patients had a positive SLN. The number of draining nodal basins did not impact the risk of recurrence in a previously mapped basin (P = .19). Nineteen patients developed metastatic disease in a previously unmapped nodal basin; 47.4% of these patients had a positive SLN detected at the time of their melanoma treatment. Again, there was not a difference between single–nodal basin drainage and multiple–nodal basin drainage (P = .06).

Predicting the precise lymphatic drainage of melanomas remains elusive, despite efforts by investigators to map these lines with reasonable certainty.^21-23 A recent attempt to classify drainage patterns into distinct groups demonstrated that 6% of all primary melanoma tumors had drainage to multiple lymph node basins, and the tumor location with the highest rate of multiple–nodal basin drainage was the trunk.²⁴ Other studies have found that a greater percentage of melanomas drain to multiple basins, with a range of approximately 30% to 40%.^14,16 In the present study, 17% of melanomas exhibited multiple–nodal basin drainage.

Given the impact that lymph node biology has on the prognosis of melanoma, investigators have been interested in patients whose tumors drain to multiple basins. Several studies looked at patients with multiple–nodal basin drainage to determine if there was a survival difference compared with those with single–nodal basin drainage. These smaller studies were limited to patients with melanomas located on the trunk. The rationale for this approach is that the trunk is the most common location for multiple basin drainage.^25-27 We decided to expand our study cohort to include all anatomic sites. Our data demonstrate that drainage to a single nodal basin is a risk factor for local, in-transit, and regional recurrence. Additionally, those patients presenting with metastatic disease in the lymph nodes that drained to a single basin had a worse LRRFS than node-positive patients whose disease drained to multiple basins. These data did not translate into differences in overall survival, however. One hypothesis that potentially explains this phenomenon is that some patients might have significant endolymphatic tumor burden at the time of the lymphoscintigram and that these “clogged” lymphatic channels may (1) limit drainage to a single regional basin and (2) place them at a higher risk for local and in-transit recurrence. This theory is supported (at least partially) by the finding that patients with single–nodal basin drainage had a higher risk of local and in-transit recurrence. There was no difference in the risk of local or in-transit recurrence between extremity and trunk melanomas. However, when we considered only the nonextremity tumors (which comprised nearly 95% of the patients with tumors that drained to multiple basins), we found that these patients did have a higher risk of local and in-transit recurrence based on the number of draining nodal basins.

It makes sense intuitively that patients with metastatic melanoma draining to more than 1 lymph node basin would confer a worse prognosis, independent of the aggregate tumor burden, as shown by Dale et al.¹⁵ They postulated that this may indicate that the tumor is more aggressive in its inherent ability to spread, thus conferring a poorer prognosis.¹⁵ Perhaps the malignant cells seed other locations as they travel through the lymphatics, thus resulting in more metastatic deposits. However, debate remains about whether drainage to more than 1 lymph node basin alone, regardless of SLN status, affects survival. Jimenez et al¹⁶ found that multiple–nodal basin drainage was an independent risk factor for poorer survival, even when only patients with negative SLNs were analyzed. Porter et al¹⁴ found that multiple–nodal basin drainage was independently associated with an increased risk of nodal metastases. Positive SLN status in turn imposes an increased risk of melanoma-related death.² In contrast, other investigators have found no association between multiple–nodal basin drainage and melanoma outcome.^17,18 Our data do not support the notion that melanomas draining to multiple nodal basins are worse than those draining to a single basin. In fact, we found the opposite.

One notable difference between the single– and multiple–nodal basin drainage groups was that there was a greater proportion of men in the latter group, in keeping with results from other studies.^16-18 Furthermore, men had a greater proportion of truncal melanomas than women. Additionally, and not surprisingly, the multiple–nodal basin drainage group had the largest percentage of trunk melanomas. It is interesting that there was no difference in the number of patients with head and neck melanomas between the 2 groups.

We found no significant association between multiple–nodal basin drainage and SLN status, disease-free survival, or overall survival. The single–nodal basin drainage group was found to have worse LRRFS than the multiple–nodal basin drainage group, however, as seen on our multivariate analysis. This analysis included anatomic site as a covariate, making the possibility that this difference was due to the location of the tumor unlikely. From these data, we can conclude that the presence of multiple–nodal basin drainage does not confer a worse prognosis.

Correspondence: Charles R. Scoggins, MD, MBA, Division of Surgical Oncology, Department of Surgery, University of Louisville School of Medicine, Louisville, KY 40292 (charles.scoggins@louisville.edu).

Accepted for Publication: February 15, 2008.

Author Contributions:Study concept and design: McMasters and Scoggins. Acquisition of data: Federico, Ross, Noyes, Goydos, Beitsch, and Scoggins. Analysis and interpretation of data: Federico, Chagpar, Ross, Martin, Urist, Ariyan, Sussman, and Scoggins. Drafting of the manuscript: Federico and Scoggins. Critical revision of the manuscript for important intellectual content: Chagpar, Ross, Martin, Noyes, Goydos, Beitsch, Urist, Ariyan, Sussman, McMasters, and Scoggins. Statistical analysis: Chagpar. Administrative, technical, and material support: Federico and Sussman. Study supervision: Chagpar, Urist, McMasters, and Scoggins.

Financial Disclosure: None reported.

Previous Presentation: This paper was presented at the 2007 Annual Meeting of the Western Surgical Society; November 5, 2007; Colorado Springs, Colorado; and is published after peer review and revision. The discussions that follow this article are based on the originally submitted manuscript and not the revised manuscript.

Additional Contributions: We sincerely thank Deborah Hulsewede, CCRP, Sherrie Matthews, CCRP, Wanda Greenwell, RN, Pam Harlan, RN, and Ivan Deyahs for their dedication and hard work in keeping the Sunbelt Melanoma Trial up and running. We also express gratitude to Advertek Inc for the expert data management for the Sunbelt database.

Jeffrey E. Lee, MD, Houston, Texas: The data that Dr Scoggins and his coauthors have presented from the Sunbelt Melanoma Trial represent another example of how surgeons from academic and community centers who collaborate to answer 1 important clinical question can continue to use the same patient cohort to answer other important questions for years and even decades afterwards. The current study addresses a potentially significant issue not imagined in the original Sunbelt Trial: are patients whose primary melanomas drain to multiple regional nodal basins at increased risk for metastasis, recurrence, or death from their disease?¹⁴ The study is a large one, involving more than 2000 patients. The patients were followed prospectively, and the median follow-up of the group is now relatively mature at more than 4 years. Since the patients were treated at a variety of institutions, the data can probably be taken to represent an aggregate standard of care, rather than being biased by the approach of any single institution. The authors' primary conclusion, that patients whose melanomas drain to multiple regional nodal basins are not at elevated risk for nodal metastasis, recurrence, or death from their disease, appears sound. However, the intriguing and counterintuitive observation that patients with single–nodal basin drainage (particularly patients with trunk primaries) are more likely to develop local-regional recurrence than patients with multiple–nodal basin drainage raises several important questions.

First, what were the specific patterns of failure of the single– vs multiple–nodal basin drainage groups? If single-node basin patients with trunk primaries more commonly recurred in basins not originally mapped (eg, the opposite axilla) or recurred in a previously mapped basin, it would suggest that such patients may have undergone incomplete initial mapping procedures. If such a pattern were identified, it would suggest a follow-up question: Was single-basin drainage from a trunk primary more commonly seen at some institutions than others?

Second, can the authors provide a suggested biologic mechanism for the observation? Anecdotally, some patients with advanced primary melanomas will exhibit poor lymphatic drainage on lymphoscintigraphy or lymphatic mapping. Such patients may have multiple involved nodes, but a “choked” lymphatic system may prevent adequate sentinel node identification. Could the single-nodal basin patients be at increased risk for regional recurrence because of a less extreme example of a regional metastasis that “closes the door behind it” and thereby prevents complete identification of regional disease by lymphatic mapping?²⁸

Third and finally, how would the authors suggest the findings be incorporated into clinical practice? Should patients with trunk melanomas that drain to only a single-nodal basin be subjected to more aggressive evaluation or follow-up? Should such a patient have a preoperative US [ultrasound] of all potentially at-risk nodal basins, or is lymphoscintigraphy still sufficient? Should PET [positron-emission tomography] imaging be used? Should we follow such patients postoperatively with US or PET?^29,30 I congratulate the authors on a fine study.

Dr Scoggins: We were very interested to find that patients with single–nodal basin drainage actually had a shorter local and regional recurrence-free survival. You're right, it is counterintuitive, until you look a little further and see that the majority of those patients with single–nodal basin drainage indeed had extremity lesions. The nonextremity patients with single basin drainage do seem to have a higher incidence of local and in-transit recurrence as opposed to recurrence in a previously mapped basin.

Whether or not there is a biological mechanism that explains this phenomenon is a very interesting question. The hypothesis that you alluded to regarding choked lymphatic channels is an interesting one and perhaps carries some validity. Another possibility was raised recently by authors who demonstrated that patients with melanoma lesions that drain via multiple lymphatic channels, albeit single nodal basin, also have a worse prognosis. We did not capture number of draining lymphatic channels as part of the Sunbelt Trial, but those data would be intriguing to go back and look into.

Whether or not clinical practice should be changed or altered for patients with truncal melanomas draining to a solitary basin is a difficult one to know whether or not PET scan[s] done more frequently would pick up microscopic or very small disease in a patient who is ultimately deemed to in-transit or satellite failure is unknown and it certainly raises a question, and possibly one that could impact financially as well.

Financial Disclosure: None reported.