Lymphatic Invasion Identified by Monoclonal Antibody D2-40, Younger Age, and Ulceration: Predictors of Sentinel Lymph Node Involvement in Primary Cutaneous Melanoma

Objectives  To assess whether lymphatic invasion identified by immunostaining with monoclonal antibody (Mab) D2-40 in primary cutaneous melanomas correlates with other clinicopathologic factors and to assess whether lymphatic invasion is a potential predictor of sentinel lymph node (SLN) status.

Design  Retrospective case-series study.

Setting  Academic referral center.

Patients  Ninety-six consecutive patients with primary cutaneous melanomas 1 mm thick or greater with adequate pathologic material available for immunohistochemical studies and SLN biopsy.

Main Outcome Measures  Association between lymphatic invasion identified by immunostaining with Mab D2-40 in primary cutaneous melanoma and correlation with the clinicopathologic features and the association of all of the factors with SLN status.

Results  Lymphatic invasion identified by immunostaining with Mab D2-40 was significantly associated with deeper Clark level of invasion (P < .001), and greater Breslow tumor thickness (P = .01) SLN positivity was identified in 23 of 96 cases (24%). At univariate analysis, younger age (P = .03), ulceration (P < .006), lymphatic invasion (P < .02), deeper Clark level of invasion (P < .008), Breslow tumor thickness (P = .008), and tumor site on the trunk (P = .02) were significantly associated with SLN metastases. At multivariate analysis, only younger age (P = .04), ulceration (P = .03), and lymphatic invasion detected by immunostaining with Mab D2-40 (P = .01) were significantly associated with SLN positivity. The probability of SLN positivity was 13% when all 3 independent prognostic factors yielded negative findings and increased to 61% when all 3 variables yielded positive findings.

Conclusions  Breslow tumor thickness, Clark level of invasion, and tumor site on the trunk predicted SLN status at univariate analysis. Multivariate regression analysis showed that lymphatic invasion identified by immunostaining with Mab D2-40, younger age, and ulceration were the only independent prognostic factors. The most significant predictor of SLN metastasis was the positivity of all 3 independent prognostic factors (61%). Findings of this study suggest that assessment of lymphatic invasion by immunostaining with Mab D2-40 with other clinicopathologic factors can be used to identify patients who could be spared the need for SLN biopsy.

The most important predictive factor of survival in patients with early-stage cutaneous melanoma is the status of the regional lymph nodes.¹ Biopsy of the sentinel lymph node (SLN) in patients with clinically localized melanoma provides valuable prognostic information and has a role in treatment planning.^2-5 Controversy exists about the therapeutic role of performing an SLN biopsy in patients with melanoma.^6-11 Nevertheless, the most recent American Joint Committee for Cancer staging system for all patients with melanoma recommends the discussion of SLN biopsy and incorporation of the results of intraoperative lymphatic mapping and tumor status of the SLN for clinical and pathologic staging.¹²

Several studies have proposed different clinicopathologic characteristics that are predictors of metastasis and survival in patients with primary melanoma.^13-23 These include age, sex, tumor thickness, Clark level of invasion, ulceration, lymphovascular invasion, mitotic index, and regression of the primary melanomas. Breslow tumor thickness is the most consistently reported important pathologic prognostic factor.^14,16,18 The presence of ulceration and lymphovascular invasion identified with hematoxylin-eosin staining were also found to have predictive value for lymph node involvement.^14-19 Findings from 2 recent studies have shown that younger age and mitotic index are significantly associated with positive SLN status.^21,22

The recently developed monoclonal antibody (Mab) D2-40 reacts with endothelial cells of lymphatics but not with endothelial cells of blood vessels in normal tissues.^24,25 Furthermore, immunostaining with Mab D2-40 has enabled identification of lymphatic invasion in sections of a variety of primary tumors including melanomas.^24,26 Our previous study showed an increase of 16% in the frequency of identification of lymphatic invasion using Mab D2-40 in primary melanomas compared with conventional staining with hematoxylin-eosin.²⁶ The purpose of the present study was to assess whether lymphatic invasion identified by immunostaining with Mab D2-40 in primary cutaneous melanomas correlates with other clinicopathologic factors and to assess whether lymphatic invasion might predict the risk of melanoma metastasis to the SLN.

Between 1998 and 2004, melanoma with a Breslow thickness of 1.00 mm or greater was diagnosed at the dermatopathology division in 386 consecutive patients at Sunnybrook and Women's College Health Sciences Centre, Toronto, Ontario, Canada. During this period, SLN biopsy was offered as a staging option to patients younger than 65 years who had primary melanoma with a Breslow tumor thickness of 1.0 mm or greater and who had no clinical evidence of metastases. For tumors less than 1.0 mm thick, SLN biopsy was offered if Clark level of invasion was 4 or greater. Of 386 patients, 123 underwent SLN biopsy at Princess Margaret Hospital, Toronto, Ontario, Canada, and 96 of these patients had blocks of the primary tumor available for immunohistochemical studies. The clinical and pathologic features of age, sex, Clark level of invasion, Breslow tumor thickness, presence of ulceration, and tumor site were evaluated for association with SNL status. All primary tumors were evaluated without knowledge of the status of the subsequent SLN biopsy specimen. Approval of the study was granted by the ethics review boards of both hospitals.

Biopsy of the SLN was performed according to the method of McCready et al.²⁷ In brief, peritumoral intradermal injections of unfiltered technetium 99–sulfur colloid (20-40 mBq) with or without patent blue dye (0.5 mL) were used to identify SLNs intraoperatively with a gamma probe. All hot or blue-stained SLNs were excised through a small skin incision until the background count was less than 10% of the hottest node.

Pathologic assessment of the lymph nodes was performed according to the method of Cochran et al.²⁸ In brief, the lymph nodes were either bivalved through the hilum or, if the node was larger than 6 mm, a section through the hilum was obtained and then 3-mm parallel sections of the remainder of the node were submitted. Six serial sections were obtained from each block. Sections 1, 3, and 5 were stained with hematoxylin-eosin; section 2 was stained with polyclonal antibody (S100, 1:3000 dilution; Dako, Carpenteria, California), section 4 with Mab HMB45 (1:50 dilution; Dako), and section 6 with melan A (prediluted solution; Ventana Medical Systems, Inc, Tucson, Arizona). Patients were grouped by the presence or absence of metastatic melanoma identified by hematoxylin-eosin or by immunohistochemical staining in at least 1 SLN.

The formalin-fixed, paraffin-embedded block of the tumor with the thickest area of invasion was selected for immunohistochemical studies after review of the slides stained with hematoxylin-eosin. Immunohistochemical staining was performed on one 5-μm-thick section from each biopsy specimen as follows: The slides were sequentially incubated with 0.1 μg/mL of Mab D2-40 followed by biotinylated goat antimouse immunoglobulin antibody (Zymed Laboratories, Inc, San Francisco) at a 1:200 dilution and the streptavidin–biotin peroxidase conjugate (Dako) at a 1:500 dilution. For color development, the sections were incubated with 3,3′ diaminobenzidine or Vector Nova RED (Novocastro Laboratories Ltd, Newcastle upon Tyne, England) and counterstained with hematoxylin. Double immunostaining with antibody to S100 protein and Mab D2-40 was performed on sections from 5 cases to confirm the presence of melanoma cells within the Mab D2-40–positive lymphatics. Appropriate positive and negative controls were used in all cases.

Two pathologists (H.J.K. and L.F.) from Sunnybrook and Women's College Health Sciences Centre reviewed all slides stained with Mab D2-40 for lymphatic invasion. Circumferential staining of endothelial cells with Mab D2-40 around the tumor cells was interpreted as positive for lymphatic invasion.

The association between lymphatic invasion identified by immunostaining with Mab D2-40 and clinicopathologic features and the association of all of the factors with SLN positivity were analyzed using the χ² test. The variables were age (<50 vs ≥50 years), sex, Clark level (2-3 vs 4-5), ulceration (present vs absent), Breslow tumor thickness (1-2 mm vs >2 to ≤ 4 mm vs >4 mm), lymphatic invasion identified by immunostaining with Mab D2-40 (present or absent), and tumor site (extremity, head and neck, and trunk). A multivariate logistic regression analysis was used to determine which clinicopathologic features were predictive of SLN metastases. All P values were 2-tailed, and P < .05 was considered statistically significant. All analyses were performed using commercially available software (SAS version 9.1; SAS Institute, Cary, North Carolina). The probability of SLN positivity was also assessed for combinations of the parameters that were significant on multivariate logistic regression analysis.

Immunohistochemical staining showed that Mab D2-40 reacted with the endothelial cells of lymphatic vessels, whereas the endothelial cells of blood vessels were negative for D2-40 (Figure 1). Occasional D2-40–positive lymphatics were identified within the tumor, whereas most lymphatics were present in the peritumoral tissue (Figure 2). Tumor emboli in D2-40–positive lymphatics were identified in the tumor mass (Figure 3). In many cases, the tumor emboli completely obliterated the lumen of the vessel. Lymphatic invasion identified by immunostaining with Mab D2-40 was present in 32 of 96 cases (33%). In the adjacent or overlying skin, D2-40 was occasionally positive around the cell membrane of some of the basal cells. However, these could easily be differentiated morphologically from the lymphatic endothelial cells that were positive for D2-40.

Lymphatic invasion identified by immunostaining with Mab D2-40 was significantly associated with deeper Clark level of invasion (P < .01) and greater Breslow tumor thickness (P = .01). No significant correlation was found between lymphatic invasion identified by immunostaining with Mab D2-40 and other clinicopathologic variables including age, sex, and ulceration.

Sentinal lymph node positivity was identified in 23 of 96 cases (24%). The frequency and percentage distribution of the clinicopathologic characteristics of the patients and their association with the presence of SLN positivity are given in Table 1. Lymphatic invasion identified by immunostaining with Mab D2-40 was present in 15 of 23 SLN-positive cases (65%), whereas no lymphatic invasion was present in 56 of 73 SLN-negative cases (77%) (P < .001). Other factors associated with SLN positivity were younger age (P = .03), presence of ulceration (P < .006), deeper Clark level of invasion (P < .008), greater Breslow tumor thickness (P < .008), and tumor site on the trunk (P = .02). No significant correlation was found between sex and SLN status. The assessment of lymphatic invasion by immunostaining with Mab D2-40 had a positive predictive value of 46.9% and a negative predictive value of 87.5%.

The results of multivariable logistic regression analysis to determine which parameters predicted the presence of a positive SLN are given in Table 2. Lymphatic invasion detected by immunostaining with Mab D2-40 was significantly associated with SLN positivity (P = .01; odds ratio, 6.7; 95% confidence interval, 1.64-27.5). Younger age was significantly associated with SLN positivity (P = .04; odds ratio, 3.8; 95% confidence interval, 1.0-13.7). In addition, the presence of ulceration was also significantly associated with SLN positivity (P = .03; odds ratio, 5.21, 95% confidence interval, 1.18-23.0). Sex, Clark level of invasion, Breslow tumor thickness, and tumor site were not significantly associated with SLN positivity on multivariate analysis.

Table 3 gives the probability of a positive SLN for different values of the significant variables (age <50 years, presence of ulceration, and lymphatic invasion identified by immunostaining with Mab D2-40). The other nonsignificant variables in the model were kept fixed at their average values. As given in Table 3, the probability of a positive SLN is 0.13 (13%) when the patient is aged 50 years or older, no ulceration is present, and no lymphatic invasion is identified by immunostaining with Mab D2-40. The probability of a positive SLN increases to 0.22 to 0.45 (22%-45%) when 2 of the 3 key variables are significant and dramatically increases to 0.61 (61%) when all 3 variables are positive.

Identification of lymphovascular invasion on sections of melanomas stained with hematoxylin-eosin is difficult and, thus, has not been consistently reported by pathologists. However, lymphatic markers such as LYVE-1, D2-40, and podoplanin have recently become available, enabling evaluation of lymphatic invasion and lymphangiogenesis.^24-26,29-32 Recently, it has been shown that Mab D2-40 is identical to podoplanin and that they are selective markers of lymphatic endothelium and are not expressed in the endothelium of blood vessels.³⁰ Furthermore, Mab D2-40 and podoplanin are the most sensitive and specific markers of lymphatic endothelium.³⁰ Although D2-40 and podoplanin are not tissue-specific, inasmuch as they have been identified in mesothelial cells, mesotheliomas, germ cell tumors, and some squamous cell carcinomas,^30,33 the positivity for Mab D2-40 in endothelial cells of lymphatic vessels enables identification of tumor emboli in the lymphatic vessels.

Findings of the present study show the presence of lymphatic invasion identified by immunostaining with Mab D2-40 in 32 of 96 patients (33%) with primary cutaneous melanoma. In our previous study using the Mab D2-40 lymphatic marker, the presence of lymphatic invasion in 16% of primary cutaneous melanomas was demonstrated. However, the previous study included patients with melanomas with Breslow tumor thickness greater than 0.76 mm, whereas the current study excluded melanomas with Breslow tumor thickness less than 1 mm. The increased frequency of lymphatic invasion in the present study is probably related to increased depth of the melanoma because melanomas 0.76 to 1 mm were excluded. This study also showed that lymphatic invasion identified by immunostaining with D2-40 correlates with deeper Clark level of invasion (P < .001) and greater Breslow tumor thickness (P = .01), indicating that lymphatic invasion occurs more frequently in later stages of melanoma.

Although the therapeutic effects of SLN biopsy are controversial, the SLN status is an important prognostic factor.^4,7,8,14,16 In this series, 23 of 96 patients (24%) with cutaneous melanoma had 1 or more positive SLNs, which is similar to the results obtained in other studies.^14,23 In the current study, lymphatic invasion identified by immunostaining with Mab D2-40 at multivariable analysis was a significant independent predictor of metastases to the SLN (P = .01). In addition, the assessment of lymphatic invasion by immunostaining with Mab D2-40 as a test for SLN status had a negative predictive value of 87.5%, which indicates that a negative Mab D2-40 immunostain result is a fairly strong predictor of a negative SLN result. This factor, in addition to other clinicopathologic factors, could identify patients who might be spared the need for SLN biopsy.

Breslow thickness of the primary tumor has been reported as the most significant factor correlated with SNL status.¹⁴ In our study, Breslow tumor thickness predicted SLN status at univariate analysis but not multivariate analysis. Similarly, deeper Clark level of invasion predicted SLN status at univariate analysis but not multivariate analysis. A possible explanation for Breslow tumor thickness and Clark level of invasion losing their significance at multivariate analysis could be the small number of positive SLNs (23/96) in this study. A future study with a larger sample size could address this issue. In contrast, lymphatic invasion identified by immunostaining with Mab D2-40, which correlated with both deeper Clark level of invasion and greater tumor thickness, was a significant independent predictor of metastases to the SNL. This indicates that lymphatic invasion identified by immunostaining with Mab D2-40 is a stronger predictor of SNL positivity than either greater tumor thickness or increased Clark level of invasion.

We noted the presence of fewer lymphatics within the tumor mass, whereas more lymphatics were noted in the peritumoral tissue. However, most tumor emboli in our study were identified within the tumor mass. Massi et al³⁴ also showed the presence of more numerous peritumoral lymphatics than intratumoral lymphatics in cutaneous melanomas by immunostaining with Mab D2-40. They also showed that the presence of intratumoral lymphatics was a predictor of SLN positivity.³⁴

Data from this study on multivariate analysis demonstrated that age was significantly associated with SLN biopsy—with younger age (≤50 years) more likely to be associated with SLN positivity than older age. This finding agrees with findings from other studies that have also demonstrated that younger age is a predictor of SLN positivity.^21,22 Paradoxically, increasing age in patients with melanoma is considered a negative prognostic factor for survival.⁴ This study showed that lymphatic invasion identified by immunostaining with Mab D2-40 did not correlate with increasing age, indicating that both variables were independent predictors of SLN positivity.

Ulceration was also an independent predictor of SLN positivity. However, the presence of ulceration did not correlate with lymphatic invasion identified by immunostaining with Mab D2-40. This finding would suggest that the presence of tumor emboli in dermal lymphatics is not a cause of ulceration in cutaneous melanomas. Massi et al³⁴ found that ulcerated melanomas were associated with increased numbers of lymphatic vessels without tumor emboli, within the tumor and in the adjacent peritumoral tissue.

The results of our study also showed that the probability of SLN positivity was 13% when lymphatic invasion identified by immunostaining with Mab D2-40 was negative, no ulceration was present, and the patient was 50 years or older. If 2 of the 3 risk factors are positive, the chance of SLN positivity significantly increases (22%-45%), and when all 3 risk factors are present, the chance of SLN positivity increases even more (61%). Using Table 3, which assesses the significance of the combination of the 3 significant risk factors (age, presence of ulceration, and lymphatic invasion identified by immunostaining with Mab D2-40), the probability of SLN metastases can easily be assessed. The findings from this study suggest that identification of lymphatic invasion identified by immunostaining with Mab D2-40 along with other clinicopathologic factors could result in the identification of populations of patients with very low risk of nodal metastasis who might be spared the need for SLN biopsy. This study also suggests that assessment of lymphatic invasion by immunohistochemical studies should be included in the synoptic report of primary cutaneous melanomas.

In summary, our results show that lymphatic invasion identified by immunostaining with Mab D2-40 correlates with deeper Clark level of invasion and increased Breslow tumor thickness. Furthermore, lymphatic invasion identified with Mab D2-40, younger age, and the presence of ulceration were independent predictors of SLN positivity. In addition, when all of these independent predictors are negative, the probability of SLN positivity is low (13%), whereas the probability of SLN positivity increases dramatically to 61% when these predictors are significant. Future studies to assess the significance of lymphatic invasion on survival are indicated.

Correspondence: Harriette J. Kahn, MD, Department of Pathology, Sunnybrook Health Sciences Centre, University Health Network, Room E4-33, 2075 Bayview Ave, Toronto, ON M4N 3M5, Canada (harriette.kahn@sunnybrook.ca).

Accepted for Publication: October 12, 2007.

Author Contributions: Dr Kahn had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Niakosari, Kahn, Marks, and From. Acquisition of data: Niakosari, Kahn, McCready, Ghazarian, Rotstein, and From. Analysis and interpretation of data: Niakosari, Kahn, Kiss, and From. Drafting of the manuscript: Niakosari, Marks, and Kiss. Critical revision of the manuscript for important intellectual content: Kahn, McCready, Rotstein, and From. Statistical analysis: Kiss. Administrative, technical, and material support: Niakosari, Kahn, Ghazarian, and From. Study supervision: Kahn, Ghazarian, and From.

Financial Disclosure: None reported.

Funding/Support: Drs Kahn and Marks received royalties from the sale of the antibody D2-40 to Dako and to Covance Inc (formerly, Signet Laboratories, Inc).

Role of the Sponsor: Dako and Covance Inc had no role in the design and conduct of the study; in the collection, analysis, and interpretation of the data; or in the preparation, review, or approval of the manuscript.