This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Save to my personal folders Download to citation manager Rights & Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Mazumdar, M. Articles by Bosl, G. J. Search for Related Content PubMed PubMed Citation Articles by Mazumdar, M. Articles by Bosl, G. J. Social Bookmarking                            What's this?

Cluster Analysis of p53 and Ki67 Expression, Apoptosis, Alpha-Fetoprotein, and Human Chorionic Gonadotrophin Indicates a Favorable Prognostic Subgroup Within the Embryonal Carcinoma Germ Cell Tumor

Madhu Mazumdar, Jennifer Bacik, Satish K. Tickoo, Deborah Dobrzynski, Alessia Donadio, Dean Bajorin, Robert Motzer, Victor Reuter, George J. Bosl

From the Department of Epidemiology and Biostatistics, the Genitourinary Oncology Service, Division of Solid Tumor Oncology, and Department of Pathology, Memorial Sloan-Kettering Cancer Center; and Departments of Medicine and Pathology, Weill Medical College, Cornell University, New York, NY.

Address reprint requests to Madhu Mazumdar, PhD, Memorial Sloan-Kettering Cancer Center, 307 E 63rd St, 3rd Floor, New York, NY 10021; email: mazumdar{at}biost.mskcc.org.

	ABSTRACT

TOP ABSTRACT INTRODUCTION PATIENTS AND METHODS RESULTS DISCUSSION REFERENCES

 
Purpose: The prognostic information provided by alpha-fetoprotein and human chorionic gonadotrophin in the management of germ cell tumor (GCT) patients is a biochemical reflection of tumor differentiation. Ki67, p53, and apoptosis have been found to be related to proliferation (Ki67), cell death (p53, apoptosis), and possibly differentiation chemoresistance (p53). We sought to determine whether simultaneous expression of one or more of these markers could identify clinically relevant subgroups of patients with nonseminomatous GCT (NSGCT).

Patients and Methods: These five marker values were obtained for 95 previously untreated patients with embryonal carcinoma with or without other germ cell components. A multivariate cluster analysis was performed to identify patients with similar marker patterns.

Results: One prominent cluster (n = 37; 36 testis retroperitoneum), consisting of 26 (70%) good-risk (GR), nine (24%) intermediate-risk (IR), and two (6%) poor-risk (PR) patients, as defined by the International Germ Cell Consensus Cancer Group (IGCCCG), was observed. The 5-year survival of the prominent cluster (with 30% IR/PR patients) was 94% (95% confidence interval [CI], 86% to 100%), which is comparable to the 91% (95% CI, 89% to 93%) 5-year survival of the IGCCCG GR patients. IGCCCG risk status (P = .005) and cluster affiliation (P = .04) were independent predictors of outcome with hazard ratios of 5.0 (95% CI, 1.6 to 15.4) and 4.6 (95% CI, 1.04 to 20.1), respectively.

Conclusion: These results suggest that there is a subgroup of NSGCT patients with embryonal carcinoma (with or without other histologies) with a specific tumor biology profile (high Ki67, low apoptosis, and low p53) whose survival is better than that of the overall patient group. The unexpectedly good outcome for the prominent cluster and independent-risk status suggest that subgroups of GCT reflecting different abilities to respond to treatment exist within IGCCCG prognostic categories.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION PATIENTS AND METHODS RESULTS DISCUSSION REFERENCES

 
GERM CELL tumors (GCT) in men are cured in more than 90% of newly diagnosed patients and in 70% to 80% of patients with advanced disease who require initial chemotherapy.¹ Clinical research during the last 15 years has focused on the identification of the 20% to 30% of patients with advanced GCT who do not achieve complete remission and ultimately die of disease. The recently validated International Germ Cell Consensus Cancer Group (IGCCCG) classification allocates patients who require initial chemotherapy into three risk groups on the basis of primary site, serum tumor marker levels, histology, and sites of metastasis.² Such risk stratification strategies have allowed the study of therapies directed to toxicity reduction without compromising efficacy in good-risk (GR) patients, and possible improved efficacy with tolerable toxicity in intermediate- and poor-risk patients.

In GCT, the importance of serum tumor markers alpha-fetoprotein (AFP), human chorionic gonadotropin (HCG), and lactate hydrogenase (LDH) in diagnosis, prognosis, and patient management is well established and embedded in the IGCCCG classification. These represent biochemical markers of differentiation (AFP, HCG) and perhaps proliferation (LDH). The histologic information used by IGCCCG characterizes the tumors as either seminoma or nonseminomatous GCT (NSGCT). However, there is considerable heterogeneity in serum tumor marker expression within these two histologies and the several NSGCT cell types. Hence, about 10% of IGCCCG GR patients, 20% of intermediate-risk patients, and 50% of poor-risk patients do not achieve a durable complete remission. Therefore, a better understanding of metastasis and resistance requires approaches that account for disparate serum tumor marker expression, the various histologic components (embryonal carcinoma [EC], choriocarcinoma, yolk sac, immature or mature teratoma, and seminoma), and new markers of tumor biology.

To investigate the biology of invasive and/or resistant GCT, the profile of proteins expressed in tumors that represent proliferation and resistance pathways needs to be established. Immunohistochemical characterizations of Ki67 and p53 expression and apoptotic index are indicators of tumor proliferative activity and cell death, respectively. Ki67 is a nuclear antigen expressed in the G₁, S, G₂, and M phases of the cell cycle, and its expression has been used to evaluate the proliferative activity of many tumors including GCT.³ Apoptosis is the end result of a series of genetic events that result in cell death. Homeostasis is maintained between proapoptotic and antiapoptotic proteins. Abrogation of proapoptotic signals, such as overexpression of bcl-2, has been shown to be associated with a poor outcome in mouse model systems.⁴ The p53 tumor-suppressor protein is a transcription factor known to regulate both apoptosis and cell cycle arrest. Mutations in TP53, the gene encoding p53, are one of the most common genetic defects in cancer and lead to a defective apoptotic response. The role of p53 expression in GCT is complex and not yet completely elucidated.⁵ However, we have recently described TP53 mutations in a subset of resistant GCTs. In vitro analysis of a cell line derived from a tumor with mutant TP53 showed an absent apoptotic response to cisplatin.⁶

In this paper, we examined these three indices of proliferation and cell death in combination as possible new markers in GCT. We chose to investigate EC because this NSGCT cell type is totipotential. It displays an aggressive phenotype and is associated with early metastatic disease in its pure form. It is more common than yolk sac or choriocarcinoma, and is the most common NSGCT component of mixed GCTs. To account for both serum and tumor marker expression without a priori assumption about risk strata, a multivariate cluster analysis was used.^7,8 Cluster analysis has been used in many investigations to identify subgroups with differing features and outcomes.^9–17 For example, Leoncini et al¹⁷ used it to separate non-Hodgkin’s lymphoma patients into three groups of varying proliferative activity (high, intermediate, and low) using markers such as Ki67, p34^cdc2, cyclin A2, cyclin B1, and mitotic indices. Gilks et al¹⁴ used this technique to cluster monoclonal antibodies into groups demonstrating similar patterns of reactivity in immunohistology experiments. In the present analysis, we used Ki67 and p53 expression levels and apoptotic rate to assess the role of these markers in the classification of GCT patients. Our goal was to identify a prognostic subgroup within EC combining Ki67 and p53 expression levels, apoptotic rate, and AFP and HCG expression.

	PATIENTS AND METHODS

TOP ABSTRACT INTRODUCTION PATIENTS AND METHODS RESULTS DISCUSSION REFERENCES

 
Specimens and Patients
We performed a retrospective review of all previously untreated patients who received first-line platinum-based chemotherapy for treatment of GCT between April 1975 and May 1996 at Memorial Sloan-Kettering Cancer Center (MSKCC; New York, NY) and for whom we had tumor specimens. The details of the therapy can be found in other publications.^18–22 We identified 95 chemotherapy-naïve NSGCT tumors for which histology contained EC; tumor blocks were obtained according to an institutional review board–approved protocol for tumor acquisition. Tumors were recut; histology of the specimen was determined; and immunohistochemical analyses for p53, Ki67, and terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate-biotin nick end-labeling (TUNEL) for apoptosis were performed. The recut tumors had EC alone; associated seminoma, syncytiotrophoblast, or intratubular germ cell neoplasm (ITGCN) components; or had EC in association with other nonseminomatous components (choriocarcinoma, yolk sac tumor, or immature or mature teratoma; Table 1). We selected all NSGCT specimens for which histology contained EC. We assayed them by using immunohistochemical techniques for p53 and Ki67 and by TUNEL method for apoptosis. This resulted in 95 specimens.

Immunohistochemical Staining for Ki67 and p53
Immunohistochemical staining was performed on 4-µm sections of representative formalin-fixed, paraffin-embedded archival tissue blocks. The immunostaining was performed by the streptavidin-biotin (horseradish peroxidase conjugated) technique using antibodies against Ki67 (diluted to 1:100; Beckman Coulter, Fullerton, CA) and p53 (clone) DO-7 (diluted to 1:500; Dako, Carpinteria, CA). Antigen retrieval was done by steaming the samples with 10 mmol/L citrate buffer before staining them.

Assessment of Apoptosis (TUNEL Method)
Apoptotic cells in situ were detected by the TUNEL method described by Gavrieli et al.²³ Briefly, the 4-µm sections were dewaxed and rehydrated, and the tissues were digested with 20 µg/mL proteinase K (Sigma, St Louis, MO) for 15 minutes at room temperature. The sections were then treated with a 2% hydrogen peroxide solution and preincubated with terminal deoxynucleotidyl transferase buffer. The sections were then incubated with 0.3 U/µL terminal deoxynucleotidyl transferase (GIBCO BRL, Gaithersburg, MD) and 0.04 nmol/µL biotinylated deoxyuracil triphosphate (Boehringer Mannheim, Indianapolis, IN) in a humid chamber at 37°C for 1 hour. The slides were rinsed in 30 mmol/L sodium citrate and 300 mmol/L sodium chloride for 30 minutes at room temperature, and then washed in phosphate-buffered saline (PBS). After sections were blocked with 2% human serum albumin for 10 minutes and rinsed briefly in PBS, they were incubated with the streptavidin–biotin peroxidase complex for 30 minutes at room temperature and again washed in PBS. Labeled cells were visualized with diaminobenzidine hydrogen peroxide solution. The sections were then counterstained with hematoxylin.

For all three stains, the distinct brown color of the nuclei was considered to be a positive result.

Serum Tumor Marker Assays
HCG was assayed using a modification of the double-antibody radioimmunoassay procedure.²⁴ The standard HCG for this assay was obtained from the National Institute of Arthritis, Musculoskeletal and Skin Disease (Bethesda, MD; standard CR117). The antiserum against the purified beta subunit was prepared in rabbits. AFP was assayed using a sandwich-enzyme immunoassay method (Hybritech, San Diego, CA). Both the capture and signal antibodies are monoclonal. For the purpose of patient description, an HCG value of more than 10 mU/mL and an AFP of more than 15 ng/mL were considered elevated.

Data Analysis
The values of Ki67 and p53 expression and apoptosis rate used in the analysis were assayed from the EC component of the tumor. The values of HCG and AFP used were the patients’ prechemotherapy baseline values. Each of the five variables was standardized by subtracting the mean and dividing by the standard deviation.

On the basis of the five variables mentioned, hierarchical agglomerative cluster analysis was used with the intention of discovering a subgroup of patients.^7,8 This is an exploratory multivariate technique that places individuals into unknown groups or clusters suggested by the data (not defined a priori). Individuals in a given cluster tend to be similar and individuals in different clusters tend to be dissimilar. The centroid method was used to compute closeness or similarity of two individuals. In this method, the distance between two clusters is defined as the distance between the group centroids, where the centroid is defined as the point at which the coordinates are the means of all the observations in the cluster. The cluster analysis begins by considering each observation as its own cluster. It then combines the two closest clusters according to the distance between their centroids, reducing the number of clusters by one. This step is then repeated until all observations are grouped into one cluster.

The output of a cluster analysis is a dendrogram (tree diagram), which illustrates the grouping. Clusters of substantial size were identified by observing the place in the dendrogram where the cluster structure remained stable for a long distance.⁸ The clusters of smaller size were not of interest. A profile plot was also used to illustrate the results of the cluster analysis.⁷ In a profile diagram, the variables are listed along the horizontal axis and the standardized value scale is listed along the vertical axis. Each point on the graph indicates the value of the corresponding variable for a patient. The points for each patient are connected to show the pattern of the values of the variables. Patients within a cluster should have similar patterns.

Survival curves were estimated by the Kaplan-Meier method,²⁵ with survival time defined from date of chemotherapy to date of death or last follow-up. The Cox proportional hazards model was used to test the association of cluster affiliation and IGCCCG risk grouping with overall survival.²⁶ All analyses were performed using SAS, Version 8.0 (SAS Institute, Cary, NC).

	RESULTS

TOP ABSTRACT INTRODUCTION PATIENTS AND METHODS RESULTS DISCUSSION REFERENCES

 
Patient Characteristics
Clinical characteristics for the 95 patients are listed in Table 1. The median age at time of chemotherapy was 26 years (range, 16 to 62 years). Tumor was obtained from the primary site in 55% of patients (53% testis, 2% mediastinum) and the remaining samples came from metastatic sites. Median time from the tumor acquisition to treatment was 19 days. All tumor specimens were obtained before treatment and 75% were acquired within 3 months of the start of chemotherapy.

The data set consisted of 60% GR, 33% intermediate-risk, and 7% poor-risk patients as defined by IGCCCG criteria.² The serum levels of AFP, HCG, and LDH were elevated in 47%, 38%, and 62% of the patients, respectively. The complete response rate to first-line therapy was 88% (95% confidence interval [CI], 80 to 94) and the 5-year survival rate was 83% (95% CI, 76 to 91). The distributions of p53, Ki67, apoptosis, HCG, and AFP are described in Table 1.

Cluster Analysis
The cluster analysis using p53, Ki67, apoptosis, HCG, and AFP revealed one prominent cluster at the 15th node, consisting of 37 patients. Figure 1 shows the dendrogram, with this cluster marked as cluster A. There is a possibility of a second cluster, but data are sparse (N = 20) and this cluster was not further investigated. The characteristics of the patients both in and not in cluster A are given in Table 2. The percentages of cluster A patients in each of the risk groups are 70% GR, 24% intermediate-risk, and 6% poor-risk, revealing a greater proportion of patients who have favorable characteristics than the overall group of patients (60% GR; Table 1).

View larger version (57K): [in this window] [in a new window]   Fig 1. Dendrogram for multivariate cluster analysis.

View larger version (48K): [in this window] [in a new window]   Fig 2. Distributions of apoptosis for patients in cluster A and patients not in cluster A.

View larger version (37K): [in this window] [in a new window]   Fig 3. Distributions of Ki67 for patients in cluster A and patients not in cluster A.

View larger version (39K): [in this window] [in a new window]   Fig 4. Distributions of p53 for patients in cluster A and patients not in cluster A.

View larger version (39K): [in this window] [in a new window]   Fig 5. Profile plot distinguishing between patients in cluster A and patients not in cluster A. Apopt, apoptosis; HCG, human chorionic gonadotropin. Each line represents one patient.

Survival Analysis
The survival curves for the overall group, for patients in cluster A, and for patients not in cluster A are given in Fig 6. The median time of follow-up for patients still alive is 9.3 years (range, 0.4 to 23 years). The 2- and 5-year survival rates for the overall group are 90% (95% CI, 84% to 96%) and 83% (95% CI, 76% to 91%). These survival rates for cluster A patients are 100% and 94% (95% CI, 86% to 100%) and are 84% (95% CI, 75% to 94%) and 77% (95% CI, 66% to 88%), respectively, for patients not in cluster A. These results suggest that there is a subgroup of NSGCT patients who have EC with or without other histologies with a specific immunohistochemistry and TUNEL profile (high Ki67, apoptosis, and low p53), and whose survival is better than that of the overall patient group.

View larger version (16K): [in this window] [in a new window]   Fig 6. Survival curves for all patients, patients in cluster A and patients not in cluster A. Tick marks indicate last follow-up.

View larger version (16K): [in this window] [in a new window]   Fig 7. Survival curve for patients in cluster A and patients not in cluster A for (A) GR patients and (B) intermediate- and poor-risk patients.

	DISCUSSION

TOP ABSTRACT INTRODUCTION PATIENTS AND METHODS RESULTS DISCUSSION REFERENCES

Both IGCCCG and cluster group were independent predictors of outcome. A comparison of GR patients from cluster A with those not in cluster A shows little separation in the survival curves (Fig 7). In contrast, the survival curves for the intermediate- and poor-risk patients combined are well separated. The survival curve for the subgroup of intermediate- and poor-risk patients belonging to cluster A is closer to the survival expected from IGCCCG GR patients. These data suggest that the underlying signaling pathways in which Ki67, p53, and apoptosis exist are relevant to clinical decision making. The subgroup of intermediate- and poor-risk patients with an extremely good prognosis could be a candidate for a less toxic regimen. Three cycles of bleomycin, etoposide, and cisplatin or four cycles of etoposide and cisplatin chemotherapy (the regimen used for GR patients) could be considered as an alternative to the standard four cycles of bleomycin, etoposide, and cisplatin currently used for poor-risk patients. Confirmatory studies on larger series of patients are warranted before any treatment alteration is attempted.

We reflect on each marker individually to seek the possible explanation for the differences in survival.

p53 expression is known to be present at baseline in GCT, possibly reflecting its origin in an early premeiotic germ cell precursor.²⁷ Univariate studies that examined p53 expression in GCT failed to show a difference in response or outcome on the basis of its expression.²⁸ We found that p53 expression was greater in those tumors not in cluster A compared with those in cluster A. Although TP53 mutations in GCT rarely have been reported, we previously have noted that 15% to 20% of the resistant GCTs harbored a TP53 mutation.²⁷ In one cell line derived from such a tumor, a mutant form of p53 was expressed, resulting in an absent apoptotic response to genotoxic (cisplatin) exposure.⁶ In both sensitive murine and human GCT cell lines, genotoxic damage results in a rapid increase in p53 expression followed by apoptosis.^6,29 In bladder cancer, p53 expression by immunohistochemistry in more than 20% of cells was nearly always associated with mutant p53.³⁰ Hence, the finding of high p53 expression in this group of patients could imply mutations in TP53, or possibly altered gene expression elsewhere in the p53 pathway. Detailed assays of untreated tumors, including high-throughput techniques, are warranted to test these hypotheses and to define the underlying mechanisms for elevated p53 levels in GCTs.

Lower Ki67 expression is seen in the tumors of patients not in cluster A. This can be seen in the histograms (Fig 3) and in the profile plot of standardized values. Our earlier data also show lower Ki67 with somatic (teratoma) and extraembryonic (choriocarcinoma and yolk sac tumors) tumors when compared with embryonal carcinoma.³¹ Clinically, teratoma clearly has the lowest proliferative capacity and displays resistance to genotoxic chemotherapy. This resistance is reflected by an in vitro model for GCT differentiation, in which all-trans retinoic acid can induce neuronal differentiation of the embryonal carcinoma cell line NT2D. Associated with this differentiation program is a reduced proliferative capacity and much attenuated p53 expression and apoptotic response to cisplatin treatment (unpublished observations). Thus, it is not surprising that EC not in cluster A with a poorer outcome exhibit lower Ki67 expression.

Similarly, in vitro, NTERA2d1 cells display neuronal differentiation when exposed to all-trans retinoic acid. This differentiation is associated with reduced proliferation and resistance to cisplatin.

Tumors in patients from cluster A show lower apoptotic indices compared with those not in cluster A. We have previously reported that GCTs with a lower apoptotic index were, in general, associated with lower p53 levels.³¹ This was particularly evident for seminoma, which had a lower apoptotic index, lower p53 levels, and phenotypically displayed a lower metastatic potential and a higher cure rate (stage for stage) than EC. Similarly, in this study, EC in cluster A with lower apoptotic index and lower p53 levels also exhibited a better outcome. The underlying biologic explanation for this association between histology, spontaneous apoptosis, p53 levels, and outcome remains unknown. It is possible that with EC and NSGCT, different apoptotic indices could reflect differentiation status, in which somatically differentiated tumors exhibit higher apoptotic indices.³¹

In addition, we found that the tumors not in cluster A have higher HCG values. HCG is a biochemical marker of trophoblastic differentiation and a well-known prognostic indicator. It is not surprising that cluster analysis would separate high and low HCG values. HCG production is a powerful predictive factor in logistic regression studies,³² and high values have been associated with an increased risk of chemotherapy resistance and/or death from tumor progression.

In contrast to HCG, tumors in cluster A had higher AFP values, a result that may seem counterintuitive at first because we expect patients with better survival to have lower AFP values. However, the percentage of patients with elevated AFP in cluster A and not in cluster A is similar. AFP had the least prognostic impact in the IGCCCG algorithm and failed to reach statistical significance in several MSKCC analyses using biochemical markers as continuous variables.^32,33 In a study evaluating surveillance in clinical stage I NSGCT, high AFP is considered a good prognostic marker.³⁴ Hence, AFP levels might have a marginal role in predicting survival that is abrogated by inclusion of better markers.

The evaluation of tumor proliferation by Ki67 and p53 expression and apoptotic rate has been found to be associated with outcome in many tumor types.^35–39 In GCT, high proliferation indices have generally been reported to predict a higher pathologic stage or to be a risk factor for relapse.^40,41 However, some others report lack of such an association.⁴² p53 expression by immunohistochemistry has been reported in a large percentage of GCTs in several prior studies, but its prognostic role remains controversial.^43,44 Apoptotic index in GCT has rarely been studied,⁴⁵ and no study investigating the prognostic implications of this index in a large sample of untreated tumors is available. It is worth noting that these studies primarily relied on univariate analyses and did not control for histologic subtype.

Finally, because cluster analysis is an exploratory technique and no assumptions regarding the nature of the relationship between the prognostic factors and survival time are required, it is an appropriate tool to use in the present setting. For example, in this analysis, it was not necessary to assume that high p53 is related to low survival. Some of the variables in this analysis are novel and their directionality in terms of their relationship to outcome is unknown. Thus, cluster analysis allowed us to conduct this study without a priori assumptions about the data and yet was able to provide us with useful information regarding patient classification.

Cluster analysis has been used only recently in the clinical literature for classification purposes. Lazzi et al¹¹ used cluster analysis of mitotic and apoptotic indices in tumors of the thyroid to reveal the existence of three groups of neoplasms with highly distinct growth characteristics. Querzoli et al¹⁵ used cluster analyses of estrogen receptor, progesterone receptor, proliferation, and c-erbB-1/neu status to identify two groups of breast cancer tumors with different pathologic features. Cluster analysis has been particularly useful in discovering classes from gene expression data or confirming already existing classes to generate new hypotheses.^46,47 These results plus ours in GCT suggest that this technique, in particular, is capable of recognizing relationships not predictable by observation alone.

This study shows a complex relationship between standard (AFP and HCG) and tumoral markers (p53, Ki67, and apoptotic index). Hence, an assessment of tumor biology may add prognostic information to known anatomic and serum markers and additional studies may identify pathways associated with clinical phenotype. However, the results also imply that immunohistochemistry alone is unlikely to solve the resistance riddle given the large number of genes and redundancy in resistance pathways. High-throughput analyses of GCT cell lines and primary and metastatic tumors of varying differentiation, resistance, and TP53 mutation status are ongoing to examine changes in gene expression and elucidate the relevant pathways.⁴⁸

	NOTES

 
Presented in part at the Thirty-Seventh Annual Meeting of the American Society of Clinical Oncology, May 2001, San Francisco, CA.

Supported in part by the National Institutes of Health grants CA05826 and CA60126, and the Byrne Foundation.

	REFERENCES

TOP ABSTRACT INTRODUCTION PATIENTS AND METHODS RESULTS DISCUSSION REFERENCES

 
1. Bosl GJ, Motzer RJ: Testicular germ-cell cancer. N Engl J Med 337:242–253, 1997[Free Full Text]

2. International Germ Cell Consensus Classification (IGCCC): A prognostic factor-based staging system for metastatic germ cell cancers. J Clin Oncol 15:594–603, 1997[Abstract/Free Full Text]

3. Albers P, Orazi A, Ulbright TM, et al: Prognostic significance of immunohistochemical proliferation markers (Ki-67/MIB-1 and proliferation-associated nuclear antigen), p53 protein accumulation, and neovascularization in clinical stage A nonseminomatous testicular germ cell tumors. Mod Pathol 8:492–497, 1995[Medline]

4. Knudson CM, Johnson GM, Lin Y, et al: Bax accelerates tumorigenesis in p53-deficient mice. Cancer Res 61:659–665, 2001[Abstract/Free Full Text]

5. Lutzker SG: P53 tumour suppressor gene and germ cell neoplasia. APMIS 106:85–89, 1998[Medline]

6. Houldsworth J, Xiao H, Murty VV, et al: Human male germ cell tumor resistance to cisplatin is linked to TP53 gene mutation. Oncogene 16:2345–2349, 1998[CrossRef][Medline]

7. Afifi AA, Clark V: Computer-Aided Multivariate Analysis. Belmont, CA, Wadsworth, Inc, 1984, pp 379–411

8. Aldenderfer MS, Blasenfield RK: Cluster Analysis. Newbury Park, CA, Sage Publications, 1984

9. Thykjaer T, Workman C, Kruhoffer M, et al: Identification of gene expression patterns in superficial and invasive human bladder cancer. Cancer Res 61:2492–2499, 2001[Abstract/Free Full Text]

10. Stemmelin J, Lazarus C, Cassel S, et al: Immunohistochemical and neurochemical correlates of learning deficits in aged rats. Neuroscience 96:275–289, 2000[CrossRef][Medline]

11. Lazzi S, Spina D, Als C, et al: Oncocytic tumors of the thyroid: Distinct growth patterns compared with clinicopathological features. Thyroid 9:97–103, 1999[Medline]

12. Megha T, Lazzi S, Ferrari F, et al: Expression of the G2-M checkpoint regulators cyclin B1 and P34CDC2 in breast cancer: A correlation with cellular kinetics. Anticancer Res 19:163–169, 1999[Medline]

13. Wetherton BM, Leonard NL, Renehan WE, et al: Structure and function of gustatory neurons in the nucleus of the solitary tract III: Classification of terminals using cluster analysis. Biotech Histochem 73:164–173, 1998[Medline]

14. Gilks WR, Oldfield L, Wild P: Data analysis of the Second International Workshop on Small Cell Lung Cancer Antigens. Br J Cancer 14:3–9, 1991 (suppl)[Medline]

15. Querzoli P, Ferretti S, Albonico G, et al: Application of quantitative analysis to biologic profile evaluation in breast cancer. Cancer 76:2510–2517, 1995[CrossRef][Medline]

16. Haroske G, Dimmer V, Steindorf D, et al: Cellular sociology of proliferating tumor cells in invasive ductal breast cancer. Anal Quant Cytol Histol 18:191–198, 1996[Medline]

17. Leoncini L, Cossu A, Megha T, et al: Expression of p34^cdc2 and cyclins A and B compared to other proliferative features of non-Hodgkin’s lymphomas: A multivariate cluster analysis. Int J Cancer 83:203–209, 1999[CrossRef][Medline]

18. Reynolds TF, Vugrin D, Cvitkovic E, et al: VAB-3 combination chemotherapy of metastatic testicular cancer. Cancer 48:888–898, 1981[CrossRef][Medline]

19. Vugrin D, Cvitkovic E, Whitmore WF Jr, et al: VAB-4 combination chemotherapy in the treatment of metastatic testis tumors. Cancer 47:833–839, 1981[CrossRef][Medline]

20. Vugrin D, Whitmore WF Jr, Golbey RB: VAB-5 combination chemotherapy in prognostically poor-risk patients with germ cell tumors. Cancer 51:1072–1075, 1983[CrossRef][Medline]

21. Bosl GJ, Gluckman R, Geller NL, et al: VAB-6: An effective chemotherapy regimen for patients with germ-cell tumors. J Clin Oncol 4:1493–1499, 1986[Abstract/Free Full Text]

22. Bosl GJ, Geller NL, Bajorin D, et al: A randomized trial of etoposide + cisplatin versus vinblastine + bleomycin + cisplatin + cyclophosphamide + dactinomycin in patients with good-prognosis germ cell tumors. J Clin Oncol 6:1231–1238, 1988[Abstract/Free Full Text]

23. Gavrieli Y, Sherman Y, Ben-Sasson S: Identification of programmed cell death in situ via specific labeling of nuclear DNA fragmentation. J Cell Biol 119:493–501, 1992[Abstract/Free Full Text]

24. Vaitukaitis JL, Braunstein GD, Ross GT: A radioimmunoassay, which specifically measures human chorionic gonadotropin in the presence of human luteinizing hormone. Am J Obstet Gynecol 113:751–758, 1972[Medline]

25. Kaplan EL, Meier P: Nonparametric estimation from incomplete observations. J Am Stat Assoc 53:457–481, 1958[CrossRef]

26. Collet D: Modelling Survival Data in Medical Research. London, United Kingdom, Chapman and Hall, 1994

27. Chaganti RSK, Houldsworth J: Genetics and biology of human germ cell tumors. Cancer Res 60:1475–1482, 2000[Abstract/Free Full Text]

28. Kersemaekers AM, Mayer F, Molier M, et al: Role of P53 and MDM2 in treatment response of human germ cell tumors. J Clin Oncol 20:1551–1561, 2002[Abstract/Free Full Text]

29. Lutzker SG, Mathew R, Taller DR: A p53 dose-response relationship for sensitivity to DNA damage in isogenic teratocarcinoma cells. Oncogene 20:2982–2986, 2001[CrossRef][Medline]

30. Cordon-Cardo C, Sheinfeld J, Dalbagni G: Genetic studies and molecular markers of bladder cancer. Semin Surg Oncol 13:319–327, 1997[CrossRef][Medline]

31. Bosl GJ, Reuter V, Mazumdar M, et al: Significant differences in Ki67, p53 expression and apoptosis characterize germ cell tumor (GCT) differentiation and progression. Proc Am Soc Clin Oncol 19:660a, 2000

32. Bajorin DF, Mazumdar M, Meyers M, et al: Metastatic germ cell tumors: Modeling for response to chemotherapy. J Clin Oncol 16:707–715, 1998[Abstract]

33. Bosl GJ, Geller NL, Cirrincione C, et al: Multivariate analysis of prognostic variables in patients with metastatic testicular cancer. Cancer Res 43:3403–3407, 1983[Abstract/Free Full Text]

34. Read G, Stenning SP, Cullen MH, et al: Medical Research Council prospective study of surveillance for stage I testicular teratoma: Medical Research Council Testicular Tumors Working Party. J Clin Oncol 10:1762–1768, 1992[Abstract/Free Full Text]

35. Kubicka S, Claas C, Staab S, et al: p53 mutation pattern and expression of c-erbB2 and c-met in gastric cancer: Relation to histological subtypes, Helicobacter pylori infection, and prognosis. Dig Dis Sci 47:114–121, 2002[CrossRef][Medline]

36. Skirnisdottir I, Sorbe B, Karlsson M, et al: Prognostic importance of DNA ploidy and p53 in early stages of epithelial ovarian carcinoma. Int J Oncol 19:1295–1302, 2001[Medline]

37. Baekelandt M, Holm R, Nesland JM, et al: Expression of apoptosis-related proteins is an independent determinant of patient prognosis in advanced ovarian cancer. J Clin Oncol 18:3775–3781, 2000[Abstract/Free Full Text]

38. Ohtani M, Isozaki H, Fujii K, et al: Impact of the expression of cyclin-dependent kinase inhibitor p27Kip1 and apoptosis in tumor cells on the overall survival of patients with non-early stage gastric carcinoma. Cancer 85:1711–1718, 1999[CrossRef][Medline]

39. Garzetti GG, Ciavattini A, Goteri G, et al: Ki67 antigen immunostaining (MIB 1 monoclonal antibody) in serous ovarian tumors: Index of proliferative activity with prognostic significance. Gynecol Oncol 56:169–174, 1995[CrossRef][Medline]

40. Leibovitch I, Foster RS, Kopecky KK, et al: Identification of clinical stage A nonseminomatous testis cancer patients at extremely low-risk for metastatic disease: A combined approach using quantitative immunohistochemical, histopathologic, and radiologic assessment. J Clin Oncol 16:261–268, 1998[Abstract/Free Full Text]

41. Albers P, Siener R, Hartmann M, et al: Risk factors for relapse in stage I non-seminomatous germ cell tumors: Preliminary results of the German Multicenter Trial—German Testicular Cancer Study Group. Int J Cancer 83:828–830, 1999[CrossRef][Medline]

42. Heidenreich A, Schenkmann NS, Sesterhenn IA, et al: Immunohistochemical expression of Ki-67 to predict lymph node involvement in clinical stage I nonseminomatous germ cell tumors. J Urol 158:620–625, 1997[CrossRef][Medline]

43. Ulbright TM, Orazi A, deRiese W, et al: The correlation of p53 protein expression with proliferative activity and occult metastases in clinical stage I non-seminomatous germ cell tumors of the testis. Mod Pathol 7:64–68, 1994[Medline]

44. Eid H, Geczi I, Magori A, et al: Drug resistance and sensitivity of germ cell testicular tumors: Evaluation of clinical relevance of MDR1/Pgp, p53, and metallothionein (MT) proteins. Anticancer Res 18:3059–3064, 1998[Medline]

45. Schmelz HU, Abend M, Kraft K, et al: Apoptosis in human embryonal cell carcinoma: Preliminary results. Urol Res 27:368–375, 1999[CrossRef][Medline]

46. Eisen, MB, Spellman PT, Brown PO, et al: Cluster analysis and display of genome-wide expression patterns. Proc Natl Acad Sci U S A 95:14863–14868, 1998[Abstract/Free Full Text]

47. Alizadeh AA, Eisen MB, Davis RE, et al: Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling. Nature 403:503–511, 2000[CrossRef][Medline]

48. Donadio AC, Houldsworth J, Bosl GJ, et al: TP53 mutation analysis in germ cell tumors (GCT): A mechanism of GCT resistance. Proc Am Soc Clin Oncol 2002

Submitted March 26, 2002; accepted April 21, 2003.

CiteULike    Complore    Connotea    Del.icio.us    Digg    Facebook    Reddit    Technorati    Twitter    What's this?

This article has been cited by other articles:

				J. E. Korkola, J. Houldsworth, D. R. Feldman, A. B. Olshen, L.-X. Qin, S. Patil, V. E. Reuter, G. J. Bosl, and R.S.K. Chaganti Identification and Validation of a Gene Expression Signature That Predicts Outcome in Adult Men With Germ Cell Tumors J. Clin. Oncol., November 1, 2009; 27(31): 5240 - 5247. [Abstract] [Full Text] [PDF]

				D. PECTASIDES, G. PAPAXOINIS, M. NIKOLAOU, C. VALAVANIS, G. ARAVANTINOS, G. FOUNTZILAS, N. TAMVAKIS, E. PECTASIDES, I. LEKKA, P. ARAPANTONI-DADIOTI, et al. Analysis of 7 Immunohistochemical Markers in Male Germ Cell Tumors Demonstrates the Prognostic Significance of p53 and MIB-1 Anticancer Res, February 1, 2009; 29(2): 737 - 744. [Abstract] [Full Text] [PDF]

				C. M. Sturgeon, M. J. Duffy, U.-H. Stenman, H. Lilja, N. Brunner, D. W. Chan, R. Babaian, R. C. Bast Jr., B. Dowell, F. J. Esteva, et al. National Academy of Clinical Biochemistry Laboratory Medicine Practice Guidelines for Use of Tumor Markers in Testicular, Prostate, Colorectal, Breast, and Ovarian Cancers Clin. Chem., December 1, 2008; 54(12): e11 - e79. [Abstract] [Full Text] [PDF]

				J. Houldsworth, J. E. Korkola, G. J. Bosl, and R. S. K. Chaganti Biology and Genetics of Adult Male Germ Cell Tumors J. Clin. Oncol., December 10, 2006; 24(35): 5512 - 5518. [Abstract] [Full Text] [PDF]

				G. J. Kelloff, R. C. Bast Jr., D. S. Coffey, A. V. D'Amico, R. S. Kerbel, J. W. Park, R. W. Ruddon, G. J. S. Rustin, R. L. Schilsky, C. C. Sigman, et al. Biomarkers, Surrogate End Points, and the Acceleration of Drug Development for Cancer Prevention and Treatment: An Update Prologue Clin. Cancer Res., June 1, 2004; 10(11): 3881 - 3884. [Full Text] [PDF]

				G. F. V. Woude, G. J. Kelloff, R. W. Ruddon, H.-M. Koo, C. C. Sigman, J. C. Barrett, R. W. Day, A. P. Dicker, R. S. Kerbel, D. R. Parkinson, et al. Reanalysis of Cancer Drugs: Old Drugs, New Tricks Clin. Cancer Res., June 1, 2004; 10(11): 3897 - 3907. [Full Text] [PDF]

This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Save to my personal folders Download to citation manager Rights & Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Mazumdar, M. Articles by Bosl, G. J. Search for Related Content PubMed PubMed Citation Articles by Mazumdar, M. Articles by Bosl, G. J. Social Bookmarking                            What's this?