The study was performed in the dermatology department of Consorcio Hospital General Universitario de Valencia (CHGUV), Spain, a tertiary level hospital that serves a population of 378 138 inhabitants.

We performed an analytic cohort study with a longitudinal observational design in which data were collected retrospectively and prospectively.

We retrospectively selected patients diagnosed with melanoma who had undergone SLN biopsy between May 2001 (the year in which this technique was introduced in the CHGUV) and May 2009 (total of 150 patients) from the melanoma database in our hospital's dermatology department. In all cases, the clinical and histologic information was collected prospectively from biopsy samples taken at the time of diagnosis before the SLN biopsy. The inclusion criteria were a) a tumor thickness of 1 mm or more or of between 0.75 mm and 1 mm in the case of tumors with histologic signs of regression or ulceration; b) mitotic rate measured directly using the primary tumor biopsy specimen; c) an SLN biopsy at our hospital; and d) follow-up at our hospital. We excluded patients with multiple melanoma or extracutaneous melanoma, patients under 16 years of age, and patients for whom it was not possible to re-examine the original biopsy specimen (Fig. 1).

Figure 1.

Study population. SLN indicates sentinel lymph node.

(0,3MB).

The original histology report was reviewed in 2011 by a dermatopathologist with experience in pigmented lesions; in accordance with the recommendations of current clinical guidelines, a single dermatopathologist was used to eliminate interobserver variability.7 Regression was considered to be present when an inflammatory infiltrate was observed in over 50% of the lesion, with accompanying fibrosis (even if this was less prominent). The dermatopathologist systematically recorded the following information: diagnosis, tumor thickness (Breslow thickness in mm), presence of ulceration, mitotic rate, margin involvement, depth of invasion (Clark level),7 histologic subtype, presence of regression, T stage, and presence of vertical growth.13

SLN biopsy was performed following the preoperative intradermal injection of the radionuclide technetium 99 (Nancol, Molypharma, S.A.) in the peritumoral area on the day of the procedure, with detection by a gamma camera and subsequent gamma probe examination (Navigator GPS, Dynasil Corp.).

The SLN was excised along the vascular pedicle and fixed in formalin. The tissue was then cut into 3-μm serial slices using a microtome; 3 slices of each block were stained with hematoxylin-eosin and 1 with Melan-A.3,14 All SLNs with any number of hematoxylin-eosin- or Melan-A-positive cells (Menarini Diagnostics S.R.L) (no minimum threshold) were classified as node-positive.

All the patients included in the study were followed, regardless of subsequent treatment We made a note of all instances of relapse, including distant metastasis, local recurrence, and regional nodal metastasis. We also recorded deaths due to melanoma or other causes. The follow-up period lasted to February 2012.

Age, Breslow thickness, and mitotic rate were categorized for statistical analysis. The χ2 test and Fisher exact test were used to compare the clinical and histologic characteristics of node-positive and node-negative patients. To control for possible confounders, we performed stepwise logistic regression with an entry criterion of P<.20 for the initial model and a permanence criterion of P<.05 for the final model. Disease-free and overall survival were calculated using the Kaplan-Meier method and the log-rank test was used to detect differences between groups. Multivariate survival analysis of disease-free and overall survival using the Cox regression model, with the incorporation of statistically significant variables detected by the log-rank test (P<.05), was used to investigate independent predictors of survival in the melanoma patients in our series. Individual covariates were expressed using the hazard ratio (HR) and corresponding 95% CI. The level of statistical significance was set at a P value of less than .05. Statistical analysis was performed using the IBM SPSS statistics package 20 for MAC.

We included 139 patients for whom it was possible to re-examine the biopsy specimen from the original tumor. Table 1 summarizes the clinical and epidemiological characteristics of the patients. Their mean (SD) age at the time of SLN biopsy was 53.34 (16.05) years (range, 18-85 years), and there was a slight predominance of men over women (n=77, 55.4% vs n=62, 44.6%).

SLN biopsy was positive in 24.5% of the sample (34 patients) and negative in 75.5% (105 patients). Tables 2 and 3 summarize the clinical and histologic characteristics of the patients according to whether they had a positive or negative SLN biopsy.

In total, 97.1% of the node-positive patients (n=33) had a mitotic rate of ≥1 mitosis/mm,2 and just 1 patient (2.9%) had a lower rate. In the node-negative group, 15.2% of patients (n=16) had a mitotic rate of <1 mitosis/mm2 and 84.8% (89 patients) had a rate ≥1 mitosis/mm2 (Fig. 2). Univariate analysis with the Fisher exact test showed a nonsignificant association between mitotic rate and SLN status (P=.071). In the subgroup of patients with a Breslow thickness of 1 to 4mm, however, we detected a statistically significant association between mitotic rate and SLN positivity (P=.034, Fisher exact test). In this subgroup, all the node-positive patients (n=25) had a mitotic rate of ≥1 mitosis/mm2 and none of the patients with a lower rate had a positive SLN biopsy. In the node-negative group, 16.2% of patients (n=12) had a mitotic rate of <1 mitosis/mm2 and 83.8% (n=62) had a rate of at least 1 mitosis/mm.2 The likelihood (odds ratio) of having a mitotic rate of <1 mitosis/mm2 and a positive SLN biopsy in the subgroup of patients with a Breslow thickness of 1 to 4mm was 0.838 (95% CI, 0.758-0.926).

Figure 2.

Distribution of mitotic rate in our sample according to sentinel lymph node status. The percentages are shown in the corresponding bar.

(0,07MB).

The median overall survival in our series was 122 months and mean disease-free survival was 101.17 months (interquartile range, 93.59-108.75 months). The 5-year survival rate was 70.2% (95 CI, 54%-85%) in the node-positive group and 79% (95% CI, 70%-87%) in the node-negative group. Thirty-five patients died during follow-up and 3 experienced recurrence. As shown in Table 4, 12 of the patients who died were node-positive (35.3% of all node-positive patients) and 23 were node-negative (21.9% of all node-negative patients). The Kaplan-Meier curves show shorter overall survival in patients with a positive SLN biopsy, although the difference with node-negative patients was insignificant (P=.153, log-rank test) (Fig. 3).

Figure 3.

Survival curve according to sentinel lymph node status.

(0,11MB).

Univariate analysis of the association between SLN status and established prognostic factors in melanoma patients showed a significant association for Breslow thickness (P=.044), Clark level (P=.048), and ulceration (P<.001). Regression (P=.909), sex (P=.209), and age (P=.88), were not significantly associated with SLN positivity (Tables 2 and 3). The only factor that retained its statistical significance in the multivariate analysis was presence of ulceration (P<.001), which was associated with a 7.039-fold increased likelihood of a positive SLN result (Table 5).

Apart from studying the relationship between SLN biopsy results and mitotic rate, we also analyzed the relationship between these results and other established prognostic factors in melanoma.6,7 The univariate analysis showed a significant association with increased Breslow thickness, Clark level, and the presence of ulceration, and a nonsignificant association with age, sex, and the presence of regression.

Breslow thickness is considered to be the main predictor of SLN biopsy outcome in melanoma and is used as the major criterion for determining whether or not SLN biopsy is indicated.6,7 SLN biopsy is recommended for tumors with a Breslow thickness of over 1mm, as these are associated with a risk of nodal involvement of over 10%. The risk falls to 3% for thinner tumors1–3 and SLN is only considered in such cases when there are additional negative prognostic factors. In our study, patients with a Breslow thickness of over 2mm had a significantly higher rate of lymph node metastases than those with thinner tumors in the univariate analysis (P=.044). We chose 2mm for this analysis as, statistically, it was the best cutoff for creating 2 groups. Our findings are consistent with those of other studies that have shown a relationship between increasing primary tumor thickness and the presence of lymph node metastases,12,16,25 although the thicknesses used in the analyses differed among studies.

Breslow thickness, which is recognized as one of the most reproducible predictive factors in melanoma,11,19,21,24,25 was not significant in our multivariate analysis. This could be because many histologic variables, such as Breslow thickness, Clark level, and ulceration are interrelated and may act as confounders.11

The prognostic value of Clark level for SLN positivity has been a topic of debate in other studies.11,19,21,22,24 In our series, we found a significant association between increased Clark level and a positive SLN biopsy in the univariate analysis (P=.048), but the fact that all the patients had a Clark level of at least III might have influenced the result. The cutoff used for Clark level in other studies is highly variable and it is therefore difficult to compare results. An optimum cutoff for predicting the presence of a positive SLN result based on Clark level has not yet been established.

The presence of ulceration in our study was significantly associated with a positive SLN result in both the univariate and multivariate analyses (P<.001). This association has been reported in numerous series,18,22 and ulceration has been found to be negatively correlated with survival in melanoma patients for all Breslow thickness subgroups.18 Accordingly, ulceration has been used in the AJCC melanoma TNM staging system since its sixth edition.9 In one study in which ulceration was not found to be associated with positive SLN status, the authors, Sondak et al.,11 suggested that this might be because of possible differences in the causes of ulceration and the fact that the presence of ulceration may lead to erroneous measurement of Breslow thickness. In our study, patients with a Breslow thickness of over 0.75mm and ulceration were referred for SLN biopsy, as is recommended by the AJCC.6 Our findings support the indication of SLN biopsy in patients with ulceration, even for tumors with a Breslow thickness of less than 1mm.

Age, sex, and regression were not significantly associated with SLN status in the univariate analysis. The predictive value of age in this respect is a topic of debate.11,19,21,24 Several authors have reported lower rates of nodal involvement in older patients, despite their having shorter overall and disease-free survival.11,26 These differences could be explained by age-related lymphatic dysfunction, as shown by Conway et al.,27 who reported a decrease in radiocolloid transit and uptake with increasing age. The relationship between sex and SLN biopsy results is not well established. While some studies have found a higher proportion of positive results in male patients,17,22 others have not.11,19,21,24 In our study, more men than women were node-positive (28.6% vs 19.4%), but the difference was not statistically significant.

Histologic regression is observed in 10% to 35% of all melanomas, but there is a lack of consensus on how regression is measured and on what cutoffs should be used.28 Tumor regression areas are composed of variable quantities of fibrous tissue, lymphocytes, new vessels, and melanophages replacing the primary melanoma. Regression measurements vary according to the definition of regression and on the subjective assessment of the dermatopathologist.28 In our series we did not find a significant relationship between regression and positive SLN status.

A positive SLN biopsy result is considered to be the most specific and sensitive prognostic factor for overall and disease-free survival in melanoma.2 In our study, the 5-year disease-free survival rate was 67% in node-positive patients and 82% in node-negative patients (nonsignificant difference, P=.132). The difference for 5-year overall survival was also nonsignificant (P=.153), with rates of 70.2% and 79% for patients with a positive and negative SLN result, respectively. Although our results show longer survival in node-negative than in node-positive patients, the differences are not statistically significance, but this could have several explanations. First, survival in patients with a positive SLN biopsy may have been influenced by subsequent treatment. According to the AJCC, the presence of nodal metastasis indicates stage III disease, the standard treatment for which is lymph node dissection. In our series, node-positive patients were considered to have advanced regional disease and, in some cases, were treated with adjuvant therapy such as high-dose interferon alfa. This difference in treatment may have influenced the lack of significant differences in overall and disease-free survival between patients with positive and negative SLN status. A second possible explanation is that our sample might have been too small to detect significant differences; we cannot rule out the possibility that significantly better survival might have been detected in node-negative patients in a larger sample. Tumor thickness might also have had a bearing on our results. Morton et al.3 found significant differences in survival among melanoma patients, but they analyzed a subgroup of patients with a tumor thickness of 1.2 to 3.5mm. The authors chose this cutoff because results from preliminary studies had indicated that lymphadenectomy performed electively or after the clinical detection of nodal involvement would affect survival in this subgroup of patients.29 Our results support the use of SLN biopsy for prognosis and staging in melanoma but additional studies are needed to investigate its therapeutic value.

The authors declare that no tests were carried out in humans or animals for the purpose of this study.

The authors declare that they have followed their hospital's protocol on the publication of data concerning patients and that all patients included in the study have received sufficient information and have given their written informed consent to participate in the study.

The authors declare that no private patient data appear in this article.