10.1245/ASO.2004.03.035
Annals of Surgical Oncology 11:85-93 (2004)
© 2004 Society of Surgical Oncology
Melanoma-Inhibiting Activity Assay Predicts Survival in Patients Receiving a Therapeutic Cancer Vaccine After Complete Resection of American Joint Committee on Cancer Stage III Melanoma
Mark B. Faries, MD,
Rishab K. Gupta, PhD,
Xing Ye, MS,
Eddy C. Hsueh, MD and
Donald L. Morton, MD
From the Sonya Valley Ghidossi Vaccine Laboratory of the Roy E. Coats Research Laboratories of the John Wayne Cancer Institute at Saint John’s Health Center, Santa Monica, California.
Correspondence: Address correspondence and reprint requests to: Donald L. Morton, MD, 2200 Santa Monica Blvd., Santa Monica, CA 90404; Fax: 310-582-7163; E-mail: mortond{at}jwci.org
 |
ABSTRACT
|
|---|
Background: More than half of melanoma patients rendered disease free by lymph node dissection will experience disease recurrence. We hypothesized that serum levels of melanoma-inhibiting activity (MIA) protein might be useful to stratify risk and identify subclinical recurrence in patients undergoing adjuvant immunotherapy. We examined MIA levels in the serum of stage III patients treated after surgery with a therapeutic cancer vaccine.
Methods: Three cohorts of 25 patients were randomly selected from our melanoma database on the basis of time to death (group 1, <1 year; group 2, 1–5 years; group 3, >5 years.) Prospectively collected serum samples were assayed in a blinded fashion for MIA by enzyme-linked immunosorbent assay.
Results: MIA was increased at any time in 19 (76%) of 25, 4 (16%) of 25, and 1 (4%) of 25 patients in groups 1, 2, and 3, respectively. The median survival was 11 months for the 25 patients with increased MIA and >75 months for the 50 patients with normal MIA. MIA increased above normal a median of 1 month (mean, 75 days) before clinical recurrence. All patients with increased MIA after 2 months of treatment subsequently died of melanoma. One patient in whom initially increased levels decreased to normal within 2 months is disease free.
Conclusions: Serum MIA levels provide important prognostic information early in the course of stage III melanoma and often detect melanoma recurrences before clinical evidence of disease.
Key Words: Melanoma-inhibiting activity • Stage III • Recurrence • Prognosis
|
INTRODUCTION
|
|---|
Patients with stage III melanoma are rendered clinically free of disease by complete lymphadenectomy. However, 50% to 80% of these patients will experience a disease recurrence, indicating the presence of occult disease that was clinically undetectable at the time of their initial treatment. Stage III patients are therefore attractive candidates for adjuvant therapy to address this microscopic disease. However, the only currently approved adjuvant agent, interferon alfa-2b, is associated with significant toxicity, and although it seems to improve disease-specific survival, beneficial effects on overall survival have not been clearly established.1 Identification of the patients at highest risk for recurrence would allow better selection of those who are most likely to benefit from this adjuvant therapy. In addition, several promising adjuvant therapies, including melanoma vaccines, are being studied. Accurate risk assessment of patients in these studies would allow for appropriate stratification of patients for randomization in such adjuvant trials.
Currently, risk of recurrence is assessed in melanoma patients on the basis of the degree of nodal involvement; characteristics of the primary lesion, such as location, thickness, or ulceration; and patient characteristics, such as age and sex. These factors can be used to develop a risk model, but actual detection of microscopic disease through laboratory assays would potentially be much more valuable. Several such laboratory assays have been assessed. These include serum markers and blood-borne molecular markers evaluated by polymerase chain reaction.
In 1989, autocrine growth-inhibiting activities of melanoma cell culture supernatants were reported.2 Subsequently, the melanoma-inhibiting activity (MIA) protein was isolated and its gene cloned. Because it is a small, soluble, secreted protein, it was believed to be a good candidate tumor marker. Subsequent studies have demonstrated the usefulness of MIA as a marker of advanced disease and response to treatment.3 Additional studies have demonstrated its sensitivity and specificity and have compared its usefulness with that of other melanoma tumor markers, such as S-100, lactate dehydrogenase, and soluble intercellular adhesion molecule-1.4–6 Our hypothesis was that MIA could serve as a useful indicator of the presence of subclinical disease in stage III melanoma patients. This might serve to stratify risk in patients and to detect recurrent disease. We examined serum that was prospectively collected from 75 patients while they were undergoing adjuvant immunotherapy with a therapeutic polyvalent cancer vaccine (Canvaxin; CancerVax Corp., Carlsbad, CA).
|
METHODS
|
|---|
Patient Population
Patients were selected from the population undergoing vaccine therapy for melanoma at the John Wayne Cancer Institute and were identified through a prospective melanoma database that included more than 11,000 patients. All selected patients had a history of nodal metastases and had undergone complete dissections of the affected lymph node basins to remove all clinically evident disease. Patients provided informed consent for participation in an institutional review board–approved protocol. Standard staging work-up included clinical examination, laboratory evaluation, and cross-sectional imaging, and those patients with evidence of systemic metastases were excluded from the study. Study cohorts consisted of three prognostic groups: patients who died of melanoma within 1 year (group 1), those who died of melanoma between 1 and 5 years (group 2), and those who were alive after 5 years (group 3). Patients in each category were selected at random by a computerized program for evaluation of serum markers. On the basis of prestudy power estimates, each group consisted of 25 patients.
Clinical Follow-Up
Patients underwent screening by clinical history and examination at each visit. Laboratory tests—including complete blood count, chemistry panel, liver function tests, and serum lactate dehydrogenase—were collected before treatment and at 2 weeks and 1, 4, and 6 months. A chest x-ray was performed at 3 and 6 months. Axial scanning of the trunk and brain was performed before treatment and at the 1-year follow-up. Additional follow-up studies were obtained when clinically indicated. Data regarding disease status, time and site of recurrence, and survival were collected prospectively during and after the study period.
Serum Collection
Patients underwent blood draws for serum collection before vaccination and at 2 weeks and 1, 2, 4, and 6 months after vaccination. Blood was collected in tiger-top Vacutainer (Becton Dickinson, Franklin Lakes, NJ) tubes and allowed to clot at room temperature for 1 hour. Serum was then separated by centrifugation at 1500 x g for 10 minutes, and serum was transferred to storage tubes. Specimens were frozen at -35°C until assayed. At the time of analysis, samples were thawed, assigned identification codes, and assayed in a blinded fashion. The blinded sample data were given to a biostatistics unit for decoding and data analysis.
MIA Enzyme-Linked Immunosorbent Assay
MIA was assayed by single-step enzyme-linked immunosorbent assay (Roche Molecular Biochemicals, Mannheim, Germany). Serum samples and standards were mixed with both biotin-conjugated capture antibody and peroxidase-conjugated detection antibody in streptavidin-coated microtiter plates. After incubation with shaking at room temperature for 90 minutes, plates were washed three times. A substrate solution was then added, and plates were again incubated with shaking for 15 minutes. Finally, the plates were read on a microtiter plate reader for absorbance at 405 nm. The concentration of MIA in test samples was calculated on the basis of a determined standard curve. On the basis of the results of earlier MIA evaluations and a desire to maximize the specificity of the assay, concentrations >8.5 ng/mL were considered increased. Any increase was sufficient to consider the patient as having had a positive test, and where specified in Results, some analyses considered results as positive on the basis of MIA values between 4 and 6 months.
Biostatistics
The 
2 test was used to determine differences between prognostic groups in terms of demographic characteristics and standard prognostic variables. Fisher’s exact test was used in comparing the presence of ulceration and Breslow thickness. Survival was estimated with the Kaplan-Meier method and compared between the patients with increased MIA between 4 and 6 months and those without by using the log-rank test. Cox regression was used to compare survival in univariate and multivariate analyses for MIA and other standard prognostic factors. Associations of tumor markers and disease-free and overall survival were analyzed at each time point individually and as an aggregate of any positive value between 4 and 6 months.
|
RESULTS
|
|---|
Demographics
Serum was available and assayed on all patients before and after vaccination. Five patients were unavailable for serum collection at the 6-month time point (because of progression of disease in four), but specimens were available for all patients at 4 months. A total of 517 serum samples were included in this analysis. There were no significant differences in age (
60 or <60 years), sex, Breslow depth, or number of positive lymph nodes (one to three vs. four or more) among the three groups (Table 1). The median follow-up was 50 months, and the median survival was 37 months for the group as a whole.
Predictive Value
Before vaccination, MIA was increased in five patients: three from group 1, one from group 2, and one from group 3. After vaccination, the level normalized in one patient, who continues to be disease free after 7 years. There was no clear correlation between baseline MIA level and either time since complete lymph node dissection or number of tumor-involved lymph nodes. There was no statistically significant correlation between MIA level and prognosis at baseline, although there were more patients with increased MIA in the worst prognostic group. This was true at all time points, but the relationship was most powerful at the later time points. We therefore analyzed differences in survival considering those with MIA >8.5 ng/mL between 4 and 6 months into the study.
Patients in the poor prognostic cohort, group 1, had increased MIA levels 4 to 6 months into therapy 76% of the time (19 of 25 patients). MIA was increased in 16% of the intermediate-prognosis group (4 of 25) and none (0 of 25) of the good-prognosis group during the same time period. There was a marked decrease in the disease-free survival of patients whose MIA level was increased between 4 and 6 months after therapy was initiated (Fig. 1). The same was true of overall survival (Fig. 2). The overall 5-year survival was 65% in those patients with normal MIA and 5% in those with increased levels. Differences in both disease-free and overall survival were highly significant by the log-rank test (P < .0001). In addition, all patients with increased MIA had recurrent disease within 2 years. In the group with normal MIA, 52% had no evidence of disease at 5 years.
View larger version (11K):
[in this window]
[in a new window]
|
FIG. 1. Kaplan-Meier plot of disease-free survival of stage III melanoma patients with or without increases in melanoma-inhibiting activity (MIA) between 4 and 6 months. Vertical tick marks are censoring points. P values are for the log-rank test.
|
|
View larger version (11K):
[in this window]
[in a new window]
|
FIG. 2. Kaplan-Meier plot of overall survival of stage III melanoma patients with or without increases in melanoma-inhibiting activity (MIA) between 4 and 6 months. Vertical tick marks are censoring points. P values are for the log-rank test.
|
|
Multivariate Analysis
Cox regression analysis of multiple prognostic variables was then performed (Table 2). By univariate analysis, sex, age, ulceration, Breslow thickness, and number of positive lymph nodes were not found to be significant variables in this relatively small group of patients. More patients in the poor prognostic groups had several of the poor prognostic variables, but the size of the cohorts did not allow sufficient power for these differences to reach statistical significance. Conversely, even with this limited sample size, a highly significant relationship between increases in MIA between 4 and 6 months and decreased disease-free and overall survival was evident. By multivariate analysis, an increase in MIA was the only significant indicator of prognosis (Table 3).
Recurrence Detection
Fifty patients have experienced disease recurrence during the postvaccination follow-up. Of these recurrences, 27 occurred during the initial 6-month study period. Seventeen (63%) of these recurrences also demonstrated increases of MIA at at least one time point. Eleven (41%) had an increase of MIA before the recurrence became clinically apparent through examination, laboratory tests, or radiographic studies; in six (22%), MIA levels increased only after clinical evidence of recurrence. Within the 6-month study period, nine patients developed recurrence without demonstrating increased MIA. Twenty-three patients had disease recurrence after the 6-month study period, and seven (30%) of those had had increased MIA during the study period. Sensitivity for recurrence during the initial 6-month period was 63%, and specificity was 85%. Sensitivity for recurrences beyond 6 months was 30%, and specificity was 100%.
MIA increased a mean of 75 days before clinical evidence of disease (median, 28 days). There was considerable variation in the timing of increases in MIA. Disease recurred clinically as many as 79 days before and as many as 628 days after MIA increases. Five of six patients who had disease recurrence before MIA increases had the recurrence within 2 months of starting vaccine therapy. Increases in MIA in the 4- to 6-month time period were uniformly associated with death from melanoma. MIA increases preceded death by a mean of 591 days (median, 262 days).
Pattern of MIA Increases
As a general rule, MIA levels in patients who had increases in MIA showed persistent or progressive elevation (Fig. 3). That is, most patients’ levels remained increased once the marker was demonstrated. When looked at individually, only five patients had transient increases in MIA, that is, increased at one time point with subsequent normalization. Two patients’ MIA levels decreased to normal after resection of clinically evident metastases. One patient in the good prognostic cohort had increased MIA at the baseline and 2-week blood draws, but it decreased to normal thereafter and has not increased again. Two patients had transient and low-level increases of MIA at intermediate time points, and both have experienced disease recurrence. In the last three patients, it is difficult to determine whether these increases reflect actual disease that was controlled for some period of time, loss of expression of MIA in evolving tumor clones, or false positives due to lack of specificity of the assay.
View larger version (41K):
[in this window]
[in a new window]
|
FIG. 3. Plots of melanoma-inhibiting activity (MIA) values for 25 individuals in the poor prognostic cohort. Disease recurrences are noted by a downward arrow. A horizontal arrow denotes recurrences that occurred after the 6-month study period. The timing of resection of one patient’s clinically evident disease is also noted.
|
|
|
DISCUSSION
|
|---|
Tumor markers have been used in several settings in cancer patients. These include screening measures, differentiating malignant from benign lesions, monitoring the response to treatment, and detecting recurrences. In melanoma, several tumor markers have been evaluated for use as prognostic variables, to monitor response to therapy, and to detect recurrence. These markers include MIA, S-100, soluble intercellular adhesion molecule-1, TA90, and 5-S-cysteinyldopa.7,8 In addition, other serum markers, including lactate dehydrogenase and liver function tests, have been used in follow-up for detection of recurrence. More recently, reverse transcriptase-polymerase chain reaction has been used to detect minute amounts of melanoma-related genetic material from peripheral blood or lymph nodes. This study evaluated the usefulness of MIA in the setting of patients with completely resected stage III melanoma who were undergoing adjuvant immunotherapy with a polyvalent cancer vaccine. The study evaluated MIA levels before therapy was initiated and during the first 6 months of follow-up. In that context, we were able to evaluate the marker’s utility in providing prognostic information for these patients, as well as its ability to detect recurrences earlier than standard clinical measures.
MIA was first detected in spent culture medium of melanoma cell lines, and, as a small (11-kDa) soluble protein, it was subsequently evaluated as a tumor marker. Its utility for providing prognostic information has been evaluated, as has its role in monitoring the response to therapy and its ability to detect recurrences. However, its use in the setting of adjuvant therapy for melanoma has not been thoroughly evaluated.
The prognostic information of MIA measured at baseline has been evaluated across all stages of melanoma. These results have demonstrated a correlation of MIA level with stage of disease9 and, similarly, have demonstrated that patients with visceral metastases have higher levels than those with disease limited to lymph nodes.10 However, MIA levels were believed not to add significant prognostic information beyond that provided by clinical stage. In this series, we limited our examination to stage III patients. A previous study of stage III patients found that median survival was significantly decreased in patients with increased MIA (14 vs. 28 months).11 Our study similarly found that increases in MIA after 4 to 6 months of treatment was a powerful predictor of poor outcome. The median survival of those with an increased level was 18 months, versus 63 months for those with normal levels. This was more powerful than the standard prognostic variables, such as tumor thickness or number of involved lymph nodes, which trended toward but did not reach significance in this limited cohort.
In contrast to the predictive value of MIA levels at the 4- to 6-month time point, baseline values were much less predictive of disease course. This was because so few of our patients had increases at baseline. Although a larger percentage of patients in the poor-prognosis group had increased levels, the number was too small to draw any conclusions. In addition, both of the treatment modalities used in these patients may have played a role in the importance of baseline values. In some cases, completion lymph node dissection was performed not long before vaccine therapy was initiated. In such cases, increased initial levels may reflect residual MIA released from tumor that was surgically removed. However, in four of five patients with increases at baseline, i.e., after they were rendered clinically disease free, persistently increased MIA levels were followed by recurrence of melanoma. This would tend to suggest that the increased levels reflected persistent, clinically occult disease.
Previously, MIA levels have been followed up in patients undergoing treatment for advanced melanoma with immunotherapy, surgery, or chemotherapy to determine whether it is a useful monitor of response to treatment. These previous studies demonstrated a correlation with disease burden and MIA level.5,6,9–14 However, because so few of our patients, treated in the adjuvant setting, demonstrated increased levels initially, the marker could not be used for monitoring responses in these patients. Also, because all patients were treated with vaccine, no treatment effect could be discerned. One patient who had a decrease in his MIA value (from increased at baseline to normal with treatment) has remained disease free.
This study was not designed to evaluate the relationship of MIA level with tumor burden. However, the progressive increase in MIA in most patients once an elevation was seen supports such a relationship. Indeed, relatively few patients had only a transient increase in MIA. In two of five of these patients, normalization of MIA corresponded to resection of clinically evident disease recurrences. Most patients with MIA increases had persistent or progressive elevations.
Perhaps the most useful context for monitoring MIA levels was demonstrated by the assay’s ability to detect recurrences. MIA was increased in most patients who experienced recurrence and became increased a mean of 75 days before the recurrence was detected by conventional means. The temporal relationship between the MIA increase and recurrence was highly variable. In some cases, the MIA increase was either significantly delayed or absent. It seems that, as is the case with a number of other tumor markers, MIA production varies from tumor to tumor. In some cases, the quantity of MIA produced is inadequate to be detected before disease recurrence or even in the setting of gross disease. However, in 18 of 50 cases with disease recurrence, MIA was significantly increased before disease detection by other means. More importantly, all cases of MIA increases after the first 2 months of therapy were followed by recurrence. Therefore, although the overall sensitivity of the assay was not great, the positive predictive value was excellent.
The optimal timing of MIA measurement has not been established. In our population, MIA was most predictive of outcome when measured several months into adjuvant treatment. However, an ideal tumor marker would identify patients with minimal, subclinical disease as early as possible to allow for the earliest institution of additional therapy. In the current group, patients whose clinical recurrence was preceded by a MIA increase had a mean lead time of 4 months (not shown). This 4-month interval might prove to be a reasonable one for monitoring over the early follow-up period in patients with stage III disease, but a firm recommendation for the optimal interval of MIA testing will require a study of a larger, less selected cohort. Currently, no therapy except surgical resection of isolated metastases has been shown to be effective in prolonging survival in the setting of recurrent disease. However, most of the likely candidate therapies for treatment of metastatic disease will entail significant morbidity. As such, it seems important to develop methods to identify patients who are most in need of such therapy and to allow adequate stratification of risk in studies of adjuvant therapies. Our data tend to support MIA as one clinical measurement that could aid in selecting patients likely to require additional treatment and stratify those in adjuvant therapy trials.
The most appropriate upper limit of normal for MIA is also not clear. Previous studies have used a range of potential normal values. The most frequently used value is 6.5 ng/mL. However, on the basis of normal population studies, 8.5, 8.8, and 18.7 ng/mL have been used.9–11,13 We chose 8.5 ng/mL in an effort to maximize the positive predictive value of the test. If the upper limit of normal is changed in our analysis to 6.5 ng/mL, our sensitivity for predicting any recurrence increases from 46% to 61%, but the specificity decreases as well, from 100% to 81%. As the upper limit of normal is decreased further, similar changes in sensitivity and specificity are seen (Fig. 4). If the test is to be used to change therapy or select patients for potentially toxic adjuvant treatments, we believe that it should be biased to avoid false positives. Our choice of normal limits allowed for this in our cohort. Additional studies in a less highly selected cohort will provide further data to help establish the most clinically important values for MIA.
View larger version (14K):
[in this window]
[in a new window]
|
FIG. 4. Plot of 1-specificity versus sensitivity for any recurrence at varying upper limits of normal for melanoma-inhibiting activity measured between 4 and 6 months of adjuvant therapy.
|
|
Overall, MIA seems to hold promise as a test in stage III melanoma patients. Values obtained in the first 6 months of patient follow-up provide excellent prognostic information that may be useful in selecting or stratifying patients for prospective trials. More importantly, surveillance levels may serve to detect recurrences early. Although there is still no proven systemic therapy in the setting of systemic recurrences of melanoma, a number of therapies currently in clinical trials, such as surgical resection and immunotherapy, may be more effective if instituted while disease burden is low. Although we believe that these observations, derived from a group of selected prognostic cohorts, are likely to be applicable to stage III melanoma patients in general, we are currently evaluating MIA in a prospective fashion in a larger patient population to validate these results. Comparison of this method of surveillance to other sensitive tests, such as the TA90 immune complex assay and polymerase chain reaction assays, is needed to determine the most useful method or combination of methods for patient follow-up.
|
ACKNOWLEDGMENTS
|
|---|
The acknowledgments are available online in the full-text version at www.annalssurgicaloncology.org. They are not available in the PDF version.
Supported by grant CA12582 from the National Cancer Institute and funding from the Wayne and Gladys Valley Foundation (Oakland, CA) and the Harold J. McAlister Charitable Foundation (Los Angeles, CA).
|
FOOTNOTES
|
|---|
Serial determination of melanoma-inhibiting activity in stage III melanoma patients undergoing adjuvant immunotherapy was found to be predictive of melanoma recurrence and overall survival.
Received for publication March 7, 2003.
Accepted for publication August 25, 2003.
|
REFERENCES
|
|---|
- Kirkwood J, Ibrahim J, Sondak V, et al. High and low dose interferon alpha-2b in high-risk melanoma: first analysis of intergroup trial E1690/S9111/C9190. J Clin Oncol 2000; 18: 2444–58.[Abstract/Free Full Text]
- Bogdahm U, Apfel R, Hahn M, et al. Autocrine tumor cell growth inhibiting activities from human malignant melanoma. Cancer Res 1989; 49: 5358–63.[Abstract/Free Full Text]
- Apfel R, Lottspeich F, Hoppe J, et al. Purification and analysis of growth regulating protein secreted by a human melanoma cell line. Melanoma Res 1993; 2: 327–36.
- Deichmann M, Benner A, Bock M, et al. S100-beta, melanoma-inhibiting activity, and lactate dehydrogenase discriminate progressive from nonprogressive American Joint Committee on Cancer stage IV melanoma. J Clin Oncol 1999; 17: 1891–6.[Abstract/Free Full Text]
- Juergensen A, Holzapfel U, Hein R, et al. Comparison of two prognostic markers for malignant melanoma: MIA and S100 beta. Tumour Biol 2001; 22: 54–8.[CrossRef][Medline]
- Djukanovic D, Hofmann U, Sucker A, et al. Comparison of S100 protein and MIA protein as serum marker for malignant melanoma. Anticancer Res 2000; 20: 2203–7.[Medline]
- Tsioulias G, Gupta R, Tisman G, et al. Serum TA90 antigen-antibody complex as a surrogate marker of the efficacy of a polyvalent allogeneic whole-cell vaccine (CancerVax) in melanoma. Ann Surg Oncol 2001; 8: 198–203.[Abstract/Free Full Text]
- Wakamatsu K, Kageshita T, Furue M, et al. Evaluation of 5-S-cysteinyldopa as a marker of melanoma progression: 10 years’ experience. Melanoma Res 2002; 12: 245–53.[CrossRef][Medline]
- Bosserhoff A, Kaufmann M, Kaluza B, et al. Melanoma-inhibiting activity, a novel serum marker for progression of malignant melanoma. Cancer Res 1997; 57: 3149–53.[Abstract/Free Full Text]
- Meral R, Duranyildiz D, Tas F, et al. Prognostic significance of melanoma inhibiting activity levels in malignant melanoma. Melanoma Res 2001; 11: 627–32.[Medline]
- Guba M, Steinbauer M, Ruhland V, et al. Elevated MIA serum levels are predictors of poor prognosis after surgical resection of metastatic malignant melanoma. Oncol Rep 2002; 9: 981–4.[Medline]
- Dreau D, Bosserhoff A, White R, et al. Melanoma-inhibiting activity protein concentrations in blood of melanoma patients treated with immunotherapy. Oncol Res 1999; 11: 55–61.[Medline]
- Schmitz C, Brenner W, Henze E, et al. Comparative study on the clinical use of protein S-100B and MIA (melanoma inhibitory activity) in melanoma patients. Anticancer Res 2000; 20: 5059–63.[Medline]
- Bosserhoff A, Dreau D, Hein R, et al. Melanoma inhibitory activity (MIA), a serological marker of malignant melanoma. Recent Results Cancer Res 2001; 158: 158–68.[Medline]
TOP
ABSTRACT
INTRODUCTION
