Prognostic Value of S-100-β Serum Concentration in Patients With Uveal Melanoma

Background  In cutaneous melanoma, the S-100-β serum level is recognized as a marker of metastatic disease.

Objectives  To determine whether S-100-β is present in the serum of patients with uveal melanoma and to test whether the serum concentration of S-100-β is related to known clinical and histopathological prognostic factors in these patients.

Methods  The S-100-β concentration was measured in serum samples collected from 64 patients with uveal melanoma before enucleation and from 58 healthy control subjects. A 2-site immunoluminometric assay was used to quantify the S-100-β concentration in serum. S-100-β concentrations in the serum from patients were compared with clinicopathological tumor variables, sex, occurrence of metastasis, and survival.

Results  Thirty-seven (57.8%) of 64 patients with uveal melanoma showed detectable levels of serum S-100-β. There was, however, no significant difference between serum levels of patients and control subjects (P = .71). Statistical analysis showed no significant correlation between S-100-β concentration and any of the clinicopathological tumor variables, occurrence of metastases, or survival. Only sex was correlated with S-100-β serum levels, which was not observed in the control group.

Conclusions  In our study on patients with uveal melanoma, the S-100-β serum concentration was not correlated with any investigated prognostic factor and was not of prognostic value itself. Female patients appeared to have higher S-100-β concentrations than male patients.

S-100 IS A PROTEIN that is expressed in nerve tissue, normal melanocytes, and malignant melanoma.^1,2 In 1980, Gaynor et al³ reported that cutaneous melanoma cells are able to secrete a soluble form of S-100. In cutaneous malignant melanoma, the serum concentration of the S-100 protein and of its subunit S-100-β has been shown to be correlated with clinical stage, prognosis, and survival.^4-16

S-100 protein, an acidic protein of 21 k Da with calcium-binding properties, is a dimeric protein composed of 2 subunits, α and β, with the possible combinations of αα, αβ, and ββ. S-100-β is a homodimer ββ.^17,18 S-100 is present in the cytoplasm and nucleoplasm of cells; it may occur in a membrane-bound form or it may be secreted.^3,19 S-100 has been reported to play a role in cytoskeletal systems, such as microtubules and microfilaments.² It has been suggested that the presence of S-100 may play an important role in the cell cycle and the disassembly of cytoplasmic microtubules.^2,20,21

Cochran et al²² reported immunohistochemical expression of S-100 in ocular melanomas, in which 80% of the tumor cells contained S-100 in their cytoplasm. Later, Kan-Mitchell and colleagues²³ showed by polymerase chain reaction that S-100-β is expressed in uveal melanoma. Cochran and coworkers²⁴ investigated the level of the S-100 protein in ocular fluids and extracts of ocular tumors but found quantification in ocular fluids in uveal melanoma to be of limited use as a diagnostic tool.

To our knowledge, there are as yet no reports on serum levels of S-100-β in patients with ocular melanoma. Using a sensitive test for S-100-β, we determined serum concentrations of S-100-β in patients with uveal melanoma and correlated the data with clinicopathological tumor variables, sex, occurrence of metastases, and survival.

Serum samples were obtained from 64 consecutive patients (32 men and 32 women) with the diagnosis of intraocular uveal melanoma, seen at our ocular oncology service from November 9, 1988, to July 4, 1997. Serum was collected before the eye was enucleated and was frozen at −80°C until analysis. At the time of diagnosis, none of the patients had metastasis. The mean follow-up after enucleation was 60 months (range, 1-160 months).

The tumor tissues were classified histopathologically and showed that in 50 of 64 cases the uveal melanoma was located in the choroid, in 3 cases in the ciliary body, in 10 cases in the choroid and ciliary body, and in 1 case in the iris.

As a control, serum was obtained from 58 healthy individuals (41 men and 17 women), who voluntarily donated blood for this study.¹⁰ The study was performed in accordance with the Declaration of Helsinki, and informed consent was obtained from all patients and controls.

Serum levels of S-100-β were determined at the Netherlands Cancer Institute.¹⁰ S-100-β concentration was determined using a monoclonal 2-site immunoluminometric assay (LIA-mat Sangtec 100; AB Sangtec Medical, Bromma, Sweden). The detection limit of 0.02 µg/L was established. The measuring range was up to 20 µg/L.

S-100-β concentrations were compared with histopathological tumor variables, sex, and occurrence of metastasis. On the basis of the distribution in controls, S-100-β serum levels were divided into 3 categories for survival analysis: <0.02, 0.02-0.06, and >0.06 µg/L.

Mann-Whitney and Kruskal-Wallis nonparametric tests were used for the statistical analysis, and the survival curve was plotted according to the Kaplan-Meier method.

In 34 (58.6%) of the serum samples from 58 healthy volunteers, the S-100-β level was above the detection level, while 14 (24.1%) exceeded the 0.06 µg/L level. Among the patient samples, 37 (57.8%) of the 64 serum samples had detectable S-100-β levels, and 17 (26.6%) were higher than 0.06 µg/L. The difference between the patient and the control group was not statistically significant (P = .71).

The median value was 0.03 µg/L for the controls and 0.04 µg/L for the patients (P = .71). Because we observed high S-100-β serum values in some patients, we analyzed whether there was a correlation between S-100-β serum concentration and clinicopathological variables (Table 1). The variables shown in Table 1, including tumor diameter and prominence, showed no significant correlation with S-100-β except for the sex of patients. Female patients had a significantly higher S-100-β serum concentration than male patients; we did not find this difference in our controls (Table 2).

In addition, we compared S-100-β levels of the patients who had developed distant metastases vs those who had not. Table 3 shows no significant difference between these groups (P = .96). Moreover, 55% (11/20) of the patients with metastases vs 59.1% (26/44) of the patients without metastases showed detectable S-100-β levels (P = .95). Finally, we compared survival time between patients having S-100-β levels above and below 0.06 µg/L, because previous studies^5-16 showed that in cutaneous melanoma high expression levels were correlated with metastases and a worse prognosis. The mean survival time was 5 years. The survival curve(Figure 1) shows that the patients with higher S-100-β levels (>0.06 µg/L) did not have a lower survival rate compared with the patients with lower serum S-100-β levels (log rank P = .70).

Techniques such as ophthalmoscopy, fluorescein angiography, and ultrasonography are, in most cases, accurate enough to diagnose uveal melanoma. However, in some cases, such as nonexudative neoplastic detachment or rhegmatogenous retinal detachment with associated subretinal blood, identification of an intraocular melanoma may be difficult. Therefore, a serum marker would be desirable in distinguishing these melanomas from benign conditions with a similar appearance. Furthermore, it might be useful to have a serum marker that helps to detect metastases at an early stage.

In this study, we investigated the prognostic value of serum S-100-β in uveal melanoma. Because uveal melanoma, like cutaneous melanoma, is derived from the neural crest, the same variables used in immunohistochemical and serological analyses might be useful in the diagnosis and prognosis of the diseases. In earlier studies^5-16,25,26 on metastatic malignant cutaneous melanoma, serum levels of S-100-β were correlated with the course of the disease and the survival time. Only one study²⁷ found the diagnostic value of S-100-β serum levels insufficient to be considered a prognostic factor for cutaneous melanoma. S-100-β can be detected in serum using immunoradiometric assay and luminescence immunoassay–mat. The latter test has been available since 1997 and has a lower detection limit compared with the immunoradiometric assay (0.02 vs 0.20 µg/L). Using the luminescence S-100-β assay on malignant melanoma, Bonfrer et al¹⁰ confirmed the significant correlation between stage and disease in skin melanoma. Furthermore, they found that 85% of patients with metastatic disease had elevated S-100-β levels.

Although in general there is a positive significant correlation between serum levels of S-100-β and the presence of skin melanoma, we did not find such a correlation for ocular melanoma in this study. In cutaneous melanoma, a correlation between tumor thickness and S-100-β serum levels has been found.¹³ However, none of the histopathological data in our study showed a correlation with S-100-β. A surprising finding is a significant difference in S-100-β serum levels between female and male patients, for which we do not have an explanation. Our data show that determination of S-100-β levels does not help to determine the prognosis in patients with uveal melanoma.

Corresponding author: Guy S. O. A. Missotten, MD, Department of Ophthalmology, Leiden University Medical Center, PO Box 9600, 2300 RA Leiden, the Netherlands(e-mail: G.S.O.A.Missotten@lumc.nl).

Submitted for publication April 4, 2002; final revision received March 7, 2003; accepted March 26, 2003.

We thank the Rotterdamse Vereniging Blindenbelangen, Rotterdam, the Netherlands, the Gelderse Blinden Vereniging, Arnhem, the Netherlands, and the Stichting Blinden-Penning, Amsterdam, the Netherlands, for their support.