Fax: (011) 0049-07071-295187
Diagnostic value and prognostic significance of protein S-100β, melanoma-inhibitory activity, and tyrosinase/MART-1 reverse transcription-polymerase chain reaction in the follow-up of high-risk melanoma patients
Article first published online: 18 MAR 2003
Copyright © 2003 American Cancer Society
Volume 97, Issue 7, pages 1737–1745, 1 April 2003
How to Cite
Garbe, C., Leiter, U., Ellwanger, U., Blaheta, H.-J., Meier, F., Rassner, G. and Schittek, B. (2003), Diagnostic value and prognostic significance of protein S-100β, melanoma-inhibitory activity, and tyrosinase/MART-1 reverse transcription-polymerase chain reaction in the follow-up of high-risk melanoma patients. Cancer, 97: 1737–1745. doi: 10.1002/cncr.11250
- Issue published online: 18 MAR 2003
- Article first published online: 18 MAR 2003
- Manuscript Accepted: 25 NOV 2002
- Manuscript Revised: 11 NOV 2002
- Manuscript Received: 18 APR 2002
- protein S-100β;
- melanoma-inhibitory activity (MIA);
- tyrosinase reverse transcriptase-polymerase chain reaction (RT-PCR);
- lactate dehydrogenase (LDH);
- alkaline phosphatase (AP)
Cutaneous melanoma is the most aggressive form of skin carcinoma in humans, frequently with a rapid progression of disease. To detect early developing metastasis, laboratory tests to determine levels of lactate dehydrogenase (LDH) and alkaline phosphatase (AP) form part of the regular follow-up, but often cannot discover recurrent disease at a sufficiently early stage.
To evaluate the diagnostic accuracy of protein S-100β (S-100β), melanoma-inhibitory activity (MIA), LDH, AP, and tyrosinase/MART-1 reverse transcription-polymerase chain reaction (RT-PCR), the authors included 296 consecutive AJCC Stage II or III clinically disease-free melanoma patients. Follow-up examinations were performed every 3 months and blood samples were drawn to determine the levels of these tumor markers.
Metastasis occurred in 41 of the 296 patients during a median follow-up period of 19 months (range, 1–33 months). The sensitivity to detect new metastases was 29% for protein S-100β, 22% for MIA, 2% for LDH, 17% for AP, and 24% for RT-PCR. The diagnostic accuracy was best for MIA (86%) and S-100β (84%), whereas AP (79%), LDH (77%), and RT-PCR (72%) demonstrated lower values. Elevated values of S-100β and MIA during follow-up examinations were associated with decreased survival rates in the further course of the disease, but pathologic findings of the other tumor markers showed no prognostic impact.
To the authors' knowledge, the current study is the first comparison of the diagnostic accuracy of currently available tumor markers in the follow-up of high-risk melanoma patients. Protein S-100β and MIA demonstrated a higher sensitivity, specificity, and diagnostic accuracy in the diagnosis of newly occurring metastasis compared with to the tumor markers AP, LDH, and RT-PCR diagnostics. Therefore, the tumor markers S-100β and MIA may be useful in the follow-up of disease-free Stage II and III melanoma patients. Cancer 2003;97:1737–45. © 2003 American Cancer Society.
Cutaneous melanoma is the most aggressive form of skin carcinoma in humans and it is increasing in incidence worldwide.1–3 Efforts have concentrated on the early detection of tumors to ensure adequate and effective treatment and to improve patient prognosis. Unfortunately, a rapid progression of disease with the development of local and distant metastasis is a frequent feature of malignant melanoma. To detect and treat metastasis early, routine imaging methods and laboratory tests form part of the regular follow-up protocol. However, recurrent disease often cannot be detected at a sufficiently early stage.4 Serum lactate dehydrogenase (LDH) has been found to be an independent prognostic factor in patients with Stage IV metastatic melanoma.4–6 Some studies have reported that LDH has the highest specificity but lowest sensitivity of all currently available tumor markers.6–8
Two of the more promising serologic tumor markers are protein S-100β (S-100β) and melanoma-inhibitory activity (MIA)9 S-100β is a 21-kilodalton dimeric acidic protein with intracellular calcium-binding capacity.10 There exist three isoforms and several structurally related proteins. S-100β can be isolated from glial cells, Schwann cells, and Langerhans cells10, 11 and is currently used as a routine marker in the immunohistochemistry for the diagnosis of melanoma and melanoma metastases.12 Elevated serum levels of S-100β in melanoma patients first were detected in a study on the relevance of S-100β in serum for patients with various neurologic disorders, in which some patients with melanoma metastasis were included.13 A number of reports confirm S-100β as a prognostic marker in patients with metastatic melanoma.14–20 It also was shown to be significantly associated with survival and to be an appropriate marker for monitoring therapy.21, 22 Melanoma-inhibitory activity was identified as a protein that functions as a melanoma growth-inhibitory factor in the culture supernatant of a metastatic melanoma cell line.23 It is translated as a 131-amino acid precursor and processed into a mature 107-amino acid protein after cleavage of a putative secretion signal. Some studies have shown that increased serum concentrations of MIA in melanoma patients indicate the degree of metastasis.23–25 This suggested that MIA is a novel serum marker for the progression of melanoma in the later stages of the disease with a higher sensitivity and specificity than other available markers. However, it is currently unknown whether the detection of MIA or S-100β in the serum of melanoma patients enables the early detection of recurrence in patients without evidence of metastasis after resection of the primary tumor or previous metastases.
A number of groups have used reverse transcription-polymerase chain reaction (RT-PCR) of tyrosinase, a key enzyme involved in melanin biosynthesis, to detect tumor cells in the peripheral blood of melanoma patients. Molecular evidence of circulating melanoma cells has been found in 0–100% of patients with disseminated malignant melanoma.26–30 Optimized protocols use two or more tumor markers, particularly the combination of tyrosinase and MelanA/MART-1.31–35 This method detects circulating tumor cells in 50–60% of Stage IV patients and in 15–30% of patients with Stage II–III disease. The question arises: Are these findings of any significance in the early detection of metastasis or in the prediction of the further course of the disease among patients with Stage II–III malignant melanoma with no known metastasis?
Because limited disease is associated with a more favorable response rate to surgical or chemotherapeutic therapy, a tumor marker detecting single, small metastases is important for the increasing number of melanoma patients. Published results revealing elevated S-100β levels in patients with melanoma metastases36 prompted us to determine the S-100β value in the regular follow-up of disease-free melanoma patients.37–40
The current study was performed to determine the diagnostic value and the prognostic significance of S-100β and MIA levels and tyrosinase/MART-1 RT-PCR, compared with LDH and AP levels, in patients with Stage II–III malignant melanoma. These patients are clinically disease free following resection of the primary tumor, in-transit, or lymph node metastases.
MATERIALS AND METHODS
Between September 1997 and December 1998, 296 consecutive patients were enrolled in this study. They attended the melanoma outpatient clinic through the Department of Dermatology, Eberhard-Karls-University of Tuebingen (Tuebingen, Germany) for regular follow-up examination. All patients provided informed consent. To be eligible for this study, each patient was required to have histologically confirmed TNM Stage II or III melanoma41 and to be free of any signs of metastasis after resection of the primary tumor, in-transit, or lymph node metastases at the time of study inclusion. Staging of the patients has been performed according to the American Joint Committee on Cancer classification of 1994. Sentinel lymph node biopsy was not used for staging. Primary melanomas were excised with safety margins of 1–3 cm according to tumor thickness. One hundred twenty patients (49 males and 71 females) with a median age of 43 years, who did not have melanoma or other known malignancies, also were included in the current study to establish an independent reference value for serum S-100β and MIA. Informed consent was obtained in all cases.
Patients were included in the study at different time points during their follow-up and blood values of the tumor markers were determined at the time of study inclusion. Peripheral blood samples for the determination of S-100β, MIA levels, LDH, AP, and tyrosinase/MART-1 RT-PCR were simultaneously drawn. Cutoff levels were defined according to the 95th percentile of the distribution of S-100β and MIA values in the control groups. All measured values of the five tumor markers were included in the analysis. For patients with elevated levels of any of these markers, clinical reexamination and computed tomographic scans of the chest, abdomen, and pelvis were performed to screen for developing metastases. In cases of uncertain findings, additional technical evaluations such as magnetic resonance imaging or bone scans were conducted. Thereafter, thorough follow-up examinations were performed every 3 months. Chest X-rays, abdominal and lymph node ultrasounds, and blood evaluations were scheduled once a year for patients with Stage II disease and twice a year for patients with Stage III disease. Recurrence-free patients were defined as patients without recognizable melanoma metastases (local, lymph node, and visceral) in the regular follow-up examinations. The median follow-up period was 19 months (range, 1–33 months).
Serum S-100β levels were measured using an immunoluminometric assay (LIA-mat Sangtec 100; Byk-Sangtec Diagnostics, Dietzenbach, Germany). This assay is a monoclonal two-site immunoluminometric test based on the sandwich principle. Antibody-coated polystyrene tubes discriminate between the α and β subunits. Only the β subunit, as defined by the monoclonal antibodies SMST 12, SMSK 25, and SMSK 28, is measured. Blood samples were frozen immediately and stored at −80°C to avoid false-positive results. The assay was performed in a two-step incubation procedure according to the manufacturer's instructions. The detection limit of S100β ranges from less than 0.02 to 20 μg/L.
Serum concentrations of MIA were measured by a quantitative one-step enzyme-linked immunosorbent assay using a commercially available kit (Boehringer Mannheim, Germany). The test was performed according to the manufacturer's instruction. Standardization was performed with a provided recombinant MIA-protein and linear signals were measured at MIA concentrations of 0.1–50 ng/mL. When MIA serum levels exceeded 50 ng/mL, the serum was diluted, retested, and recalculated. The standard curve was calculated in a linear fashion. All serum samples and standards were measured in duplicate and results did not vary by > 5%. Duplicate experiments were performed for S-100β and MIA according to the instructions recommended by the manufacturer.
Determination of tyrosinase and/or MelanA/MART-1 by RT-PCR was performed as previously described.35 A BLAST search was performed for all primers because this is a routine procedure in primer design. All primers demonstrated homology with the target sequences only. Briefly, 10 mL ethylenediaminetetraacetic acid (EDTA)-blood from each melanoma patient was purified over a Ficoll gradient with a density of 1.077 g/mL within 1 hour after blood collection. 4 × 8-mL prewarmed Ficoll with a density of 1.077 g/mL (Biochrom, Berlin, Germany) was overlaid with each 5-mL EDTA-blood sample that had been diluted 1:2 with phosphate-buffered saline (PBS). After 10 minutes of constant centrifugation at 1200 rounds per minute (rpm) at room temperature, the top layer of the Ficoll sample was transferred to another tube and washed twice with PBS. RNA isolation, the RT, and nested PCRs for tyrosinase, MelanA, and GAPDH gene products were performed blindly. The following sizes were obtained for the PCR products from the second amplification: tyrosinase, 207 base pairs (bp); MelanA/MART-1, 266 bp; and glyceraldehyde-3-phosphate dehydrogenase (GAPDH), 246 bp. Genomic DNA was not amplified using the primers described above because they were specific for different exons of their respective gene. RNA isolation, cDNA synthesis, and PCR were performed in separate areas to reduce the danger of contamination. Two negative controls and one cDNA negative control were always included in the PCR. Forty analyzed blood samples from nonmelanoma patients were negative for tyrosinase and MelanA/MART-1. Each RT-PCR reaction was repeated at least twice. A sample was regarded as tyrosinase or MelanA positive when in one PCR performance a 207-bp band (tyrosinase) or a 266-bp band (MelanA) was detected on an ethidium bromide-stained agarose gel. For all samples, GAPDH was amplified to determine the integrity of the RNA. Samples that failed to amplify products for GAPDH RNA were not considered.
LDH and AP were determined by conventional analysis. An LDH value > 240 IU/L and an AP value > 168 IU/L were considered to be pathologic and indicative of possible metastasis.
The data were evaluated using SPSS for Windows, version 8.0 (SPSS, Chicago, IL) and S-Plus 4, Release 3 for Windows NT (Mathsoft, Seattle, WA). Standard measures of test validity including sensitivity, specificity (both with 95% confidence intervals), and predictive values were calculated. Sensitivity is defined as (true positives/[true positives + false negatives]) and specificity as (true negatives/[true negatives + false positives]). Diagnostic accuracy is defined as the ratio of all true-positive and true-negative values to all measured values.
Discrimination ability of the serum parameters in detecting metastasis was assessed by the standardized Mann–Whitney U test or by Somer Dxy rank correlation. The receiver operating characteristics curve (ROC) was also calculated. The ROC curve displays the variability of sensitivity and specificity for cutoff points of different concentrations of the serum parameters which were measured as continuous variables. The Mann–Whitney U test, which measures accuracy, is equal to the area under the ROC.
In the case of elevated blood values for the specified tumor markers, diagnosis was aided by imaging techniques. The development of metastasis had to be confirmed within 3 months after the determination of elevated blood values to relate the metastasis to the pathologic tumor marker values.
In addition, Kaplan–Meier curves were calculated to show the probability of recurrence-free survival according to pathologic or normal values of the tumor markers in the time period after the respective follow-up examination. Differences between the estimated survival curves were evaluated using the log-rank test.
To establish a control level for S-100β and MIA, serum samples were obtained from 120 patients without a history of melanoma or any other malignancy. The median S-100β value was 0.00 μg/L (range, 0.00–0.13 μg/L). The median MIA level 5.31 ng/mL (range, 0.71–18.77 ng/mL). Cutoff levels were defined as the 95th percentiles, i.e., 0.12 μg/L for S-100β (as recommended by the manufacturer) and 10.49 ng/mL for MIA. Serum values above these cutoff levels were regarded as elevated.
Included in this study were 296 patients with TNM Stage II–III melanoma who were clinically free of any signs of metastasis. Tumor characteristics and stages of the melanoma patients are summarized in Table 1. Tumor stages are provided for the time before the first serum analysis. Metastasis occurred in 41 of the 296 patients within the median follow-up period of 19 months. Patient groups with and without development of metastasis did not significantly differ in their age and gender distribution.
|Tumor characteristics||Stage II A/B (n = 167)||Stage III A/B (n = 129)|
|Tumor thickness (mm)|
|Mean of age (yrs)||55||57|
The sensitivity and specificity of each parameter for the detection of disease recurrence were calculated and compared with the confirmation of metastases by imaging examinations and/or histopathology (Table 2). The sensitivity of S-100β (0.29) was slightly superior to that of PCR (0.24) and MIA (0.22) and arguably better than that of AP (0.17) and LDH (0.02). MIA was the parameter with the highest specificity (0.97). The specificity of S-100β and LDH was also high (0.93 and 0.90, respectively). AP had a specificity of 0.89 and PCR had the lowest value, 0.80. The diagnostic accuracy was best for MIA (0.86) and S-100β (0.84), whereas lower values were obtained for AP (0.79), LDH (0.77), and RT-PCR (0.72; Table 2).
|Marker||Cutoff value||Sensitivity/95% CI||Specificity/95% CI||Diagnostic accuracy|
|Protein S-100β||0.12 μg/L||0.29||0.93||0.84|
|AP||> 168 IU/L||0.17||0.89||0.79|
The sensitivity and specificity of the various parameters in separating patients likely/not likely to develop metastasis were estimated using ROC analysis (Fig. 1). This analysis revealed a better diagnostic accuracy for S-100β and MIA than for LDH and AP (Table 3). The ROC-AUC values for S-100β and MIA were 0.66 and 0.62, respectively, whereas LDH (0.53) and AP (0.51) showed lower values. The Mann–Whitney u test revealed significant P values for S-100β (0.001) and MIA (0.011), but not for LDH (0.571), AP (0.807), and PCR (0.519).
Somer's Dxy was used to discriminate among the serum markers to predict metastasis (Table 3). S-100β had the highest predictive value of the four parameters tested with a marginally better result than MIA. Somer's Dxy values for both LDH and AP were 0.055 and 0.024, respectively. In all of the three tests, S-100β was slightly superior to MIA and both S-100β and MIA were superior to LDH, AP, and PCR.
Kaplan–Meier survival curves were plotted for S-100β, MIA, and RT-PCR (Fig. 2). The difference in recurrence-free survival time in the time period after the follow-up examinations between patients with normal and pathologic values was highly significant for both MIA and S-100β (P < 0.0001). The difference in recurrence-free survival time between patients with and without pathologic LDH, AP, and RT-PCR results was not significant (P = 0.99, 0.24, and 0.4, respectively).
Two factors that may be important in improving the overall prognosis of this disease are the detection of primary tumors at an early stage and the detection of metastasis at an early stage of development. Early detection and treatment of locoregional and distant metastasis may improve patient outcome, especially if complete tumor resection can be achieved.42–47
Both serum proteins S-100β and MIA have been suggested to be markers for staging and for monitoring therapy of malignant melanoma.15, 19, 24 However, this is the first time, to the best of our knowledge, that the diagnostic and prognostic impact of the various blood parameters, simultaneously including RT-PCR, in Stage II and III patients without metastasis has been evaluated.
LDH is a well known and established prognostic factor in patients with disseminated melanoma (Stage IV).5, 48 In accordance with previously published data, a high LDH value is associated with reduced survival among Stage IV patients. It has also been shown that protein S-100β and MIA have a similar prognostic impact,9, 14, 22, 24, 49 but do not appear to be superior to LDH in determining disseminated disease.7
In our study, S-100β and MIA were sensitive markers in the diagnosis of newly occurring metastases, which were later confirmed by technical examinations. In Stage II and III patients with an increased risk of developing metastasis, S-100β and MIA were superior to AP and LDH in sensitivity, specificity, and diagnostic accuracy.
S-100β and MIA have a similar diagnostic accuracy (84% for S-100β vs. 86% for MIA). The sensitivity of S-100β and MIA for the detection of disease recurrence was moderate (29% and 22%, respectively), which is low compared with values reported in previous studies.16, 19, 21, 50 The rather low sensitivity may be caused by the relatively short follow-up duration of 19 months. False-positive results may be caused by long storage at room temperature, which leads to a significant increase in S-100β values.51 Another explanation for this finding might be that this study refers mainly to patients without clinical evidence of disease at the time of S-100β and MIA measurement, whereas previous studies quoted protein levels of patients who had metastatic disease.13, 14, 18, 19 However, with a specificity of 93% for protein S-100β and 97% for MIA, a positive test result is relevant to the clinician.
Using the ROC, the Mann–Whitney U test, and Somer'sdx for statistical analysis, S-100β was slightly superior to MIA and both S-100β and MIA were superior to LDH, AP, and PCR.
Kaplan–Meier survival curves showed that positive values for S-100β and MIA were associated with a decreased recurrence-free survival after the respective follow-up examination, whereas LDH, AP, and RT-PCR did not show significant differences between patients with normal and pathologic values.
The detection of circulating tumor cells by RT-PCR for tyrosinase and MelanA/MART-1 in the peripheral blood of melanoma patients did not have a significant impact on their prognosis, with a sensitivity of 0.24 and a specificity of just 80%, the lowest of all tested parameters. This contrasts with the results of past studies that reported that the use of RT-PCR to detect circulating melanoma cells in the peripheral blood of melanoma patients was of prognostic importance.31, 52, 53 Two groups found that the RT-PCR for tyrosinase may be a prognostic marker in the early and late stages of the disease.54, 55 Kunter et al.56 showed a correlation between tyrosinase detection in the blood of patients with disseminated metastases and disease progression. However, they found no tyrosinase transcripts in patients with Stage II–III melanoma. We also shared these views in a previous publication.35 These beliefs are based on casual observations or analysis of patient collectives comprising patients with and without metastasis. The higher rate of positive test results and disease progression in the group with metastasis is misleading and overestimates the prognostic significance of PT-PCR for metastasis-free patients.32, 33, 48, 56, 57–60 However, it is possible that the presence of circulating tumor cells in the lymph nodes is of greater prognostic significance than their presence in the peripheral blood. A number of studies have shown that the detection of melanoma cells in sentinel lymph nodes by RT-PCR for tyrosinase may be more effective than conventional methods currently in use and may be an important predictor of patient prognosis.61–63
Serum S-100β and MIA can be regarded as possible tumor markers for the screening of melanoma metastasis in the follow-up of high-risk melanoma patients. Elevated parameters for S-100β and MIA often preceded detection of developing metastases by imaging techniques and motivated these staging evaluations. Therefore, in cases of elevated serum S-100β and MIA, control determinations and clinical reevaluations should be performed. However, LDH, AP, and RT-PCR are not sensitive markers of tumor progression in Stage II and III patients and positive values were not related to decreased recurrence-free survival rates in the further course of the disease. However, elevated values of serum S-100β and MIA were found to be associated significantly with decreased recurrence-free survival rates in the further course. Because S-100β demonstrated the highest diagnostic sensitivity for recognition of developing metastasis in our study, it appears to be most suitable as a diagnostic tumor marker in the monitoring of patients with high-risk malignant melanoma.
- 32Evaluation of the use of tyrosinase-specific and melanA/MART-1-specific reverse transcriptase-coupled–polymerase chain reaction to detect melanoma cells in peripheral blood samples from 299 patients with malignant melanoma. Br J Dermatol. 2001; 144: 279–287., , , .
- 35Amplification of MelanA messenger RNA in addition to tyrosinase increases sensitivity of melanoma cell detection in peripheral blood and is associated with the clinical stage and prognosis of malignant melanoma. Br J Dermatol. 1999; 141: 30–36., , , , .
- 59Interlaboratory evaluation of a new reverse transcriptase polymerase chain reaction-based enzyme-linked immunosorbent assay for the detection of circulating melanoma cells: a multicenter study of the Dermatologic Cooperative Oncology Group. J Clin Oncol. 2001; 19: 1723–1727., , , et al.