Flow Cytometric and Quantitative Histological Parameters to Predict Occult Disease in Clinical Stage I Nonseminomatous Testicular Germ Cell Tumors

0022-534 7 /93/1503-0879$03.00/0 THE J OURNAL OF UROLOGY Copyright © 1993 by AMERICAN URO LOGICAL ASSOCIATION , INC.

Vol. 150, 879-883, September 1993 Printed in U. S.A.

FLOW CYTOMETRIC AND QUANTITATIVE HISTOLOGICAL PARAMETERS TO PREDICT OCCULT DISEASE IN CLINICAL STAGE I NONSEMINOMATOUS TESTICULAR GERM CELL TUMORS JUDD W. MOUL,* JOHN P . FOLEY, CHARLES L . HITCHCOCK, WILLIAM F . McCARTHY, ISABELL A. SESTERHENN, ROBERT L . BECKER AND JOE L. GRIFFIN From the Urology Service and Department of Clinical Investigation, Walter Reed Army Medical Center and Departments of Genitourinary Pathology and Cellular Pathology, A rmed Forces Institute of Pathology, Washington, D. C., and Department of Surgery, Uniformed Services University of the Health Sciences, Bethesda and Henry M. Jackson Foundation for the Advancement of Military Medicine, Rockville, Maryland

ABSTRACT

The goal of this study was to determine if deoxyribonucleic acid (DNA) flow cytometric and quantitative histological parameters could predict occult metastases in clinical stage I nonsemino­ matous testicular cancer. Archival paraffin primary tumor tissue was available from 36 clinical stage I nonseminomatous germ cell testicular cancer patients who all had retroperitoneal lymphadenec­ tomy and followup defining 2 groups: pathological stage I (23) and occult pathological stage II ( 1 3 ) . Archival blocks were microdissected and individual histological components were subjected t o flow cytometry. In addition, the primary histology was reevaluated for vascular invasion and per cent composition of histological components of embryonal carcinoma and other histologies. For flow cytometry parameters, no tumor was uniformly diploid, and the DNA index and per cent S phase cells were not useful in differentiating stages. Although mean per cent S phase for the aneuploid cell population and proliferative index were significantly greater for stage II cases by univariate logistic regression analysis, they are approximately 70% accurate in predicting occult disease as single tests and were not significant by multivariate analysis. The calculation of per cent embryonal carcinoma was also significantly greater in stage II cancer by univariate logistic regression testing and remained significant by multivariate analysis. Vascular invasion was marginally predictive of occult disease but was also not significant by multivariate analysis . Calculating the percentage of embryonal carcinoma of a primary testicular tumor may be a useful method to assess clinical stage I cancer patients for risk of occult disease. A larger study is needed to confirm the importance of per cent embryonal carcinoma and to clarify further if flow cytometry in combination is useful. KEY WORDS : flow cytometry, histology, testicular neoplasms, germ cells

Despite the tremendous advances that have occurred in the care and curability of most testicular cancer patients, 1 refine­ ments in staging and optimal therapy are necessary for certain subgroups. 2 One of the main problems in the care of patients with clinical stage I nonseminomatous germ cell testicular cancer is the possibility of occult retroperitoneal or distant metastases. Current clinical staging, which includes serum tu­ mor markers and computerized tomography (CT), fails to iden­ tify the 30 to 40% of men who have occult metastases. 3 The current treatment strategies for these patients include staging retroperitoneal lymphadenectomy, surveillance and primary chemotherapy. 4 Opponents of retroperitoneal lymphadenec­ tomy contend that more than half of the patients undergo an unnecessary operation, while detractors of surveillance contend that the burden of uncertainty and the rigors of followup are excessive. 2 • Although primary chemotherapy has been used, most are reluctant to subject all patients to the potential morbidity of combination chemotherapy. 5 The idea of identifying factors in the primary tumor that may further stratify patients as to the risk of occult disease and dictate more precise therapy is not new. The presence of 4

Accepted for publication January 29, 1993. Supported by Grants 2862 from the Department of Clinical Investi­ gation, Walter Reed Army Medical Center, Washington, D. C., G190AZ-01 and R090BJ-Ol from the Uniformed Services University and the Henry M. Jackson Foundation, Bethesda, Maryland. The opinions and assertions contained herein are the private views of the authors and are not to be construed as reflecting the views of the United States Army or the Department of Defense. * Requests for reprints: Department of Surgery, Uniformed Services University of the Health Sciences, 4301 Jones Bridge Rd., Bethesda, Maryland 20814-4 799. 879

vascular invasion and the element of embryonal carcinoma clearly are helpful factors in raising suspicion of occult dis­ ease. 4- 7 A recent multicenter study with central pathological review confirmed the importance of vascular invasion and added the concept of calculating the percentage of embryonal carcinoma in the primary tumor. 8 For retroperitoneal lymph­ adenectomy treated pathological stages I and II cancer patients not treated with adjuvant chemotherapy, the risk of relapse was low until the incidence of embryonal carcinoma exceeded 30 to 40%. 8 Aside from these traditional histological parameters, it is conceivable that the measurement of certain deoxyribonucleic acid (DNA) flow cytometry variables would also aid in patient stratification. Although flow cytometry has been investigated in nonseminomatous germ cell testicular cancer, 9-23 little work has addressed its use in clinical staging. One study found that the DNA index and size of the S phase fraction could not discern clinical stage I cancer patients with occult disease. 20 Conversely, another study using DNA image analysis f ound that hyperpentaploidy was statistically more prevalent in pa­ tients with clinical stage I relapses. 24 Another flow cytometry parameter, the proliferative index/activity, has been found to be useful in predicting prognosis in advanced nonseminomatous germ cell testicular cancer patients although it had not been used for low stage nonseminomatous germ cell testicular cancer patient stratification. 1 7 Our hypothesis was that flow cytometry markers of proliferation and quantitative per cent embryonal carcinoma would predict clinical stage I nonseminomatous germ cell testicular cancer patients with occult disease. There­ fore, we collected primary tumor archival blocks from a cohort

880

CLINICAL STAGE I NONSEMINOMATOUS TESTICULAR CANCER

of surgically staged (retroperitoneal lymph node dissection) clinical stage I nonseminomatous germ cell testicular cancer patients, and subjected the material to flow cytometry and to quantitative histological analysis for per cent embryonal car­ cinoma, teratoma, yolk sac and other elements, and vascular invasion. We then compared these parameters in the ability to predict occult disease. MATERIALS AND METHODS

Clinical material. Of 230 patients with nonseminomatous germ cell testicular cancer treated between 1980 and 1990, 81 (35%) had clinical stage I disease before retroperitoneal lymph node dissection and followup. Most of these patients had under­ gone radical orchiectomy at the referring hospital before trans­ fer to our center. Archival pathological material from these institutions was obtained through correspondence to assemble the clinical stage I cohort. Other cases had already been cata­ loged at the Armed Forces Institute of Pathology (AFIP), and blocks were identified and added to the collection, as well as archival tissue in the Walter Reed Pathology Department from patients who underwent surgery locally by the urology service. Blocks were available in 40 cases (50%), of which 4 were technically unacceptable due to predominant necrosis (2), sem­ inoma (1) and an exhauste d block (1). The remaining 36 cases (23 pathological stage I and 13 pathological stage II disease) formed the basis of the study. All patients underwent clinical staging, including serum tumor markers (a-fetoprotein and {3human chorionic gonadotropin) that normalized within appro­ priate half-lives after orchiectomy, CT of the abdomen and pelvis, and either chest CT or whole lung tomography. All patients then underwent r etroperitoneal lymph node dissection and sufficient followup to define the 2 stages. Stage II cancer cases were further characterized as stage Ila-less than 6 positive nodes, all less than 2 cm. in diameter and stage IIB­ more than 6 positive nodes or any node more than 2 cm. in diameter. Sample preparation. Hematoxylin and eosin stained sections from each block available from each case were reviewed (R. L. B., C. L. H. and I. A. S.). Specific areas of immature and mature teratoma, mixed germ cell tumor, yolk sac and embryonal carcinoma were identified and isolated by scoring the block face before sectioning. Care was taken to avoid sampling regions of seminoma as well as areas of necrosis, fibrosis and hemor­ rhage. Nuclei were isolated from 2 or 3, 80 µm. thick sections of each block. Sections were deparaffinized in 2 changes in His­ toclear, rehydrated in a sequence of ethanol solutions and stored overnight in 50% ethanol. The sections were washed in phosphate buffered saline (pH 7.5) and resuspended in 0.25% pepsin (P-6887) in normal saline (pH 7.5) and incubated for 90 minutes at 37C. Tissue disaggregation was enhanced by re­ peated aspiration through a 3-way stopcock connected to 2, 5 ml. syringes. Digestion was stopped with Pepstatin and the nuclei were counted with a hemocytometer. The isolated nuclei were stained with propidium iodide as previously described. 25 In brief, 1 to 2 x 106 nuclei were washed in a 0.1% Triton­ phosphate buffered saline solution (Triton X-100 in pH 7.2) and incubated in ribonuclease (180 units per ml./106 nuclei) for 20 minutes at 37C. Nuclei were stained in propidium iodide (50 µg./ml. phosphate buffered saline per 106 nuclei), stored at 4C in foil-covered tubes and analyzed within 24 hours. Nuclei were filtered through a 37 µm. nylon mesh before flow cytometric analysis. Flow cytometry. DNA content was measured on a Coulter EPICS 541 flow cytometer equipped with a 5 watt argon laser (488 nm.) at 150 mw. DNA check beads (Coulter) were used for instrument alignment and linearity. Nuclei isolated from a dog lymph node were run each day as a control for sample staining and instrument function. The effects of debris and doublets were minimized by gating on the peak and integrated

red signals. A minimum of 2 X 104 nuclei per sample were analyzed at a flow rate of no more than 100 events per second for each DNA histogram. Histogram analysis. All DNA histograms were analyzed and classified by 2 of us (C. L. H. and J. L. G.) without knowledge of the pathology or the clinical course. Tumors were classified as DNA diploid (a single GO/Gl peak), DNA aneuploid (2 distinct GO/Gl peaks), DNA multiploid (more than 1 DNA aneuploid peak) or uninterpretable (excess debris, no defined peaks, or a coefficient of variation of more than 10). Cell cycle analysis of the DNA diploid and DNA aneuploid tumors was done with the multicycle program* to provide a proliferative index, defined as the percentage of cells in the S plus G2/M phases of the cell cycle. The proliferative index value is a calculated number generated by the computer multicycle pro­ gram as previously described. 1 7 Histological prognostic factor analysis. All hematoxylin and eosin stained slides were reviewed by 1 of us (I. A. S.), who was blinded to the stage of disease, to assess for the presence of vascular invasion, and the percentages of embryonal carcinoma, mature teratoma, immature teratoma, yolk sac carcinoma and other elements as previously described. 8 Statistical analysis. All flow cytometry, histological and clin­ ical data were provided to 1 of us (W. F. M.) for statistical analysis. To determine independent statistically significant risk factors associated with stage II disease, logistic regression was used. 26 Initially univariate logistic models were used to deter­ mine potential risk factors, that is those histological or DNA flow cytometric measurements with a p value of �0.10. Once this step was completed, those potential prognostic factors with a p value of �0.10 were considered in a multivariate logistic model to determine which factors were independent risk fac­ tors. In the multivariate model any risk factor with a p value of �0.05 was considered to be statistically significant. The clinical use of the statistically significant risk factors determined by the univariate and multivariate logistic models was determined via the classification table associated with each model. A risk factor was considered to have clinical use if it had high sensitivity and high specificity. Each classification table provided the following information for each risk factor that was determined by either the univariate or multivariate model: sensitivity, specificity, number correct, false-positive rate and false-negative rate. Before using the univariate logistic models, preliminary sta­ tistical analyses via descriptive statistics (mean, median and percentages) indicated that per cent S of the aneuploid only cell population, the proliferative index, per cent embryonal cancer and vascular invasion were potential risk factors for stage II disease. Univariate logistic models validated that these histological and flow cytometric factors were, indeed, statisti­ cally significant predictors of stage II disease. However, the classification tables associated with the univariate logistic models indicated that only per cent embryonal cancer was meaningful from a clinical viewpoint. The multivariate model indicated that the only independent risk factor that was statis­ tically significant was per cent embryonal cancer. The associ­ ated classification table also indicated that this risk factor had clinical use as well. RESULTS

For the 23 pathological stage I and 13 pathological stage II cancer patients, all of whom initially had clinical stage I disease before retroperitoneal lymphadenectomy, table 1 summarizes the study data. For stage I disease 1, 2, 3 and 4 blocks were examined for 7, 8, 7 and 1 patients, respectively, while for stage II cancer all but 1 patient had 2 blocks examined. Table 2 reveals the discordance between the original pathology reports and the AFIP rereview. The original studies had an overre* Phoenix Flow Systems, San Diego, California.

C L I NI C A L STAGE I N O N S E M I N O M AT O U S TE S T I C U LAR CANCER TABLE 1. Flow cytometric and quantitative histological parameters

comparing pathological stages I and II cancer patiente

Ploidy : * Aneuploid a n d diploid Aneuploid Multiploid Aneuploid a n d multiploid Mean D NA index: Primary stemline Secondary stemline Mean flow cytometry proliferation: S phase, all cells S phase, aneuploid cells Proliferative index:j: Quantitative histology: Vascular invasion* Mean % embryonal Ca Mean % mature teratoma Mean % immature teratoma Mean % yolk sac Ca

2 11 4 6

Statistical Significance

Stage II

Stage I

0 7 4 2

(9) (48) (17) (26)

Not significant Not significant Not significant Not significant

(- ) (54) (31) (15)

Not significant Not significant

1.42 1.96

1 .45 1 .93 16.4 20.5 30.5

17.2 31.5 45.4

Not significant 0.0254t 0.0134t

3/21 (14) 38 35 11 15

6/13 (46) 63 15 4 8

0.066t 0.018t Not significant Not significant Not significant

* Number o f patients per total (per cent ) . t Univariate logistic regression analysis. :j: Proliferative index calculated for 19 evaluable stage I and 10 stage II cancer patients. TABLE 2. Comparison of histology on original pathology reports and

AFIP review Stage I

Stage II

Histology

Original Report

AFIP Review

Original Report

AFIP Review

Embryonal Teratoma Mixed germ cell tumor

5 (22) 8 (35) 10 (43)

0 (0) 1 (4) 22 (96)

7 (54) 2 (15) 4 (31)

4 (3 1) 1 (8) 8 (61)

Number of patients (per cent) . TABLE 3. Predictive value of flow cytometry and histological variables

in iden tifying occult disease in clinical stage I nonseminomatous germ cell testis tumor

Sensitivity Specificity Correct False-pas. rate False-neg. rate

Univariate Analyses

Multivariate Analysis

Vascular S Phase, Embryonal ProliferaInvasion Aneuploid Ca ( % ) tive Index (%) (%)

With % Embryonal Ca Remaining Significant (%)

35 89.2 70.2 36.4 28.3

20 93.9 70.8 40.0 27.9

45.5 85. 7 71.9 37.5 25.0

63.6 76.2 71.9 41.7 20.0

71.4 85.7 81.0 28.6 14.3

porting of pure embryonal carcinoma and an underreporting of mixed germ cell tumors due to decreased recognition of yolk sac elements. In regard to flow cytometry, only 3 of the 73 blocks (4%) contained diploid neoplasm (1 block with mature teratoma, 1 block with a mixture of embryonal and yolk sac tumor, and 1 block with intratubular malignant germ cells) but these cases had other areas of aneuploid tumor. No case was uniformly diploid and none of the 13 stage II tumors had diploid areas identified. The DNA index for the primary and secondary stemlines did not differentiate between stages I and II. Similarly, the mean percentage of S phase of the entire sample was alike in both stages (16.4 versus 17.2). However, for the per cent S phase of the aneuploid cells a difference was noted between the mean values for stages I and II (20.5 versus 31.5). This increasing trend was validated by a univariate logistic model (p = 0.0254). Likewise, for proliferative index the mean for stage I was 30.5 compared to 45.4 for stage II. This increasing trend was also validated by a univariate logistic model (p = 0.0134) . Table 3,

881

however, reveals the predictive value of these 2 flow cytometry variables as calculated from the data. The per cent S phase for aneuploid cell population or proliferative index values alone would have predicted occult stage II disease correctly approxi­ mately 70% of the time. Of the quantitative histological parameters (table 1), the percentage of embryonal carcinoma calculation was a signifi­ cant predictor of occult stage II disease by univariate logistic regression analysis (p = 0.018). Vascular invasion (14% in stage I versus 46% in stage II) was marginally significant (p = 0.066). These parameters individually would correctly predict occult disease approximately 72% of the time (table 3). A multivariate analysis was then performed with 4 variables: per cent S phase for aneuploid cell population, proliferative index, per cent em­ bryonal carcinoma and vascular invasion. In the statistical model, only the per cent embryonal carcinoma remained as a significant predictor of occult stage II disease. In this model the overall accuracy of predicting occult stage II disease was 81% (table 3). Individual histological components of the tumors were sam­ pled and subjected to flow cytometry. Table 4 examines flow cytometry variables for the different histological components. Progressing from teratoma through embryonal carcinoma areas, there are increasing trends in mean and median flow cytometry values that have been validated by univariate logistic regression. Because embryonal carcinoma was a powerful variable in multivariate analysis, flow cytometry data for pure embryonal samples from stages I and II cases were examined to determine if flow cytometry of pure embryonal carcinoma areas might be a useful staging procedure (tables 4 and 5). For 9 stage I (12 samples) and 9 stage II (15 samples) tumors the aneuploid/ multiploid rate, DNA index, mean and median per cent S phase (all cells), and mean and median proliferative indexes were not significantly different between the stages. However, percent S phase of aneuploid cells (mean and median 29.9 and 30.9 for stage I versus 40.0 and 45.5 for stage II, respectively) was significant (p = 0.05, t test). Performing an analysis similar to that of Sledge et al for advanced nonseminomatous germ cell testicular cancer, 1 7 5 of 7 evaluable stage I cancer patients (71.4%) had a proliferative index below the group mean of 61 versus only 3 of 9 (33%) who had occult stage II disease. DISCUSSION

The most important finding of our study was that a quanti­ tative histological parameter of the primary tumor, the per­ centage of embryonal carcinoma, was a stronger predictor of stage II occult disease in clinical stage I nonseminomatous germ cell testicular cancer patients than flow cytometry meas­ urements. Despite this finding, a number of new findings are reported with regard to flow cytometry in nonseminomatous germ cell testicular cancer, and especially the analysis of indi­ vidual histological components and the use of S phase prolif­ eration parameters in low stage disease. Furthermore, flow cytometry parameters were analyzed in conjunction with the histological parameters of vascular invasion and per cent em­ bryonal carcinoma such that comparisons may be made. Virtually all samples were aneuploid or multiploid and none of our patients had a uniformly diploid tumor. This finding is in complete agreement with other investigators, who noted the majority of nonseminomatous germ cell testicular cancers to have an abnormal DNA stemline. 9-23 Our data would suggest that if 1 sampled area of a nonseminomatous germ cell testic­ ular cancer was diploid then other areas should be examined, and aneuploidy and multiploidy will be found. Although a larger cohort is necessary, our study found that the 3 patients with diploid tumor areas all had pathological stage I cancer. Perhaps a diploid tumor area in a nonseminomatous germ cell testicular cancer is a favorable prognostic marker for stage I disease. With regard to DNA index, this flow cytometry parameter

882

CLINICAL STAGE I NONSEMINOMATOUS TESTICULAR CANCER TABLE 4. Flow cytometric variables for individual histological components of testicular nonseminomatous germ cell testis tumor

% S Phase, All Cells Histology

No. Samples

Mature teratoma Immature teratoma Yolk sac Mixed germ cell tumor Embryonal Ca: Stage I Stage II Logistic regression analysis

10 7 3 10

8.27 ± 10.54 ± 6.27 ± 15.89 ±

12 15

TABLE

% S Phase, Aneuploid Cells Median

No. Samples

Mean ± SE

Median

No. Samples

0.91 2.60 1.20 1.28

7.95 7.50 7.10 16.45

8 5 3 8

6.78 ± 1.61 9.94 ± 3.08 17.47 ± 6.89 24.40 ± 6.22

6.40 11.40 20.50 25.45

9 5 3 5

24.71 ± 2.30 21.73 ± 2.28 p = 0.0354

22.90 23.50

9 14

29.89 ± 3.93 40.01 ± 4.43 p = 0.0214

30.90 45.45

7 9

Mean ± SE

5. Subanalysis of pure embryonal carcinoma samples for stages I and II testicular nonseminomatous germ cell tumors

No. pts. No. blocks evaluable No. aneuploid blocks (% ) No. multiploid blocks (% ) Mean DNA index % S phase (mean/median): All cells Aneuploid cells Proliferative index (mean/median) No. with proliferative index below group mean (mean 61) * P = 0.05, t test.

Proliferative Index

Stage I

9 12 7 (58.3) 5 (41.6) 1.49

Stage II 9 15 9 (60) 6 (40) 1.45

24.71/22.90 29.89/30.90* 54.33/53.20 5/7 (71.4)

21.73/23.50 40.01/45.45* 66.30/71.40 3/9 (33)

was unable to discern stage I from stage II disease. Fossa et al similarly found that the DNA index was unable to differentiate clinical stage I cancer patients with occult disease or recur­ rence.20 Similarly, the size of the S phase fraction from the entire sample did not appear to be predictive of subclinical metastases. Our mean values of 16.4% for stage I and 17.2% for stage II were equivalent. However, they were generally lower than values obtained by Fossa et al.20 In their study the S phase fraction of 40 clinical stage I tumors was high (median 30%, range 7 to 62% ) but they were also not able to discern occult metastatic disease. Other flow cytometry parameters, such as the S phase for the aneuploid cells and the proliferative index, have not been used previously to stratify clinical stage I nonseminomatous germ cell testicular cancer. Our calculation of mean per cent S phase for aneuploid cell population was significantly greater in the stage II cancer patients. However, there was a wide range of values and this variable alone was approximately 70% ac­ curate in predicting occult disease. Similarly for proliferative index, there was a statistically significant higher mean value for stage II cases but the accuracy for this variable alone was also approximately 71 %. The predictive value for the histolog­ ical parameters, vascular invasion and per cent embryonal carcinoma were similar to the flow cytometry variable (72% range). Because of the importance of embryonal carcinoma, a flow cytometry subanalysis of pure embryonal areas was done for stages I and II cases. There was a significant difference in mean and median per cent S phase for aneuploid cell population values for stage I versus stage II cancer. Also, using an analysis similar to that of Sledge et al for advanced nonseminomatous germ cell testicular cancer,1 7 most stage I cases (71.4% ) had proliferative index values (of embryonal areas) below the group mean, whereas most stage II tumors (66%) had a proliferative index (embryonal) above the group mean. Recognizing that this subanalysis was based on a small subgroup of patients with a limited number of specimens, it is presently unclear if flow cytometry variables will be clinically useful on pure embryonal areas of primary nonseminomatous germ cell testicular cancer. A larger study is needed to determine if a flow cytometry parameter, such as per cent S phase for aneuploid cell popula­ tion or proliferative index, performed on embyonal carcinoma

Mean ± SE 14.61 ± 19.78 ± 22.90 ± 35.00 ±

Median

2.23 6.28 6.25 9.40

11.00 25.80 24.40 29.50

54.33 ± 4.98 66.30 ± 7.31 p = 0.0202

53.20 71.40

areas from the primary tumor will add predictive power to the discovered calculation of the percentage of embryonal carci­ noma alone. Aside from the flow cytometry information, the determina­ tion of per cent embryonal carcinoma emerged as a promising calculation. Although it has been known for some time that the element of embryonal carcinoma favors occult stage II disease, to our knowledge this is the first study that specifically uses an embryonal percentage to stratify clinical stage I cases. Sesterhenn et al calculated per cent embryonal carcinoma for pathological stages I and II nonseminomatous germ cell testic­ ular cancer patients who had entered the Testicular Cancer Intergroup Study and found that per cent embryonal carcinoma predicted relapse in nonchemotherapy treated patients.8 When this parameter exceeded 30 to 40% the relative hazard of relapse markedly increased. However, in that study percentage of em­ bryonal carcinoma was not significant after adjustment for vascular invasion or nodal stage. That study was different from ours in that not all patients had clinical stage I disease and it addressed a different issue of relapse in already surgically staged cancer patients. Our preliminary data suggest that per cent embryonal carcinoma may be useful to predict occult disease in clinical stage I cancer patients. A larger study of clinical stage I nonseminomatous germ cell testicular cancer must be done with regard to per cent embryonal carcinoma. Despite the fact that the various flow cytometry parameters alone were suboptimal to stratify clinical stage I nonsemino­ matous germ cell testicular cancer patients accurately, the flow cytometry data from individual histologies were interesting. The mean and median values for per cent S phase, per cent S phase for aneuploid cell population and proliferative index all trended higher (significant by logistic regression) from the teratomatous elements through embryonal carcinoma (table 4). We know from clinical experience that the teratomas behave less aggressively than the embryonal carcinomas and this was confirmed quantitatively by the flow cytometry proliferative indexes. Clearly, flow cytometry in this study provided a general gauge of clinical behavior. Although DNA index has been compared for various nonseminomatous germ cell testicular cancer histologies,18· 20-23 to our knowledge this is the first investigation to analyze per cent S phase, per cent S phase for aneuploid cell population and proliferative index for various nonseminomatous germ cell testicular cancer individual his­ tologies. Aside from the fact that DNA content and cell proliferation may not be sufficiently different to distinguish between clinical stage I nonseminomatous germ cell testicular cancer with and without occult disease, technical factors related to flow cytom­ etry, use of archival material and sampling could have affected our results. Regarding flow cytometric DNA ploidy determi­ nation, Fossa et al found a high correlation coefficient between results from paraffin embedded and frozen primary testicular tumors.2° Conversely, with regard to proliferation parameters, such as S phase fraction, investigators have cautioned about unreliable results from archival material.20· 2 1· 27· 28 The enzy­ matic digestion required to produce a nuclear suspension for archival specimens invariably results in some degree of subcel-

CUNICAL STAGE I N O N S E MI N O MATOUS TESTICULAR CANCER

lular debris that is superimposed on the DNA histogram. 27 If this debris is not taken into account, serious overestimation of tumor S phase fraction can result. 27 In our study sample prep­ aration was repeated or blocks were eliminated for use if there was excess debris or a coefficient variation of greater than 10% . Additionally, blocks were scored, and areas of hemorrhage, necrosis, benign tissue and seminoma were avoided. Because our study used microdissection of blocks, the usual concerns regarding normal cells interfering with S phase measurements were lessened. 28 Despite these precautions, our results must be confirmed by other investigators using their own extraction, flow cytometry and computer analyses. Even though individual flow cytometry and histological fac­ tors could not completely predict occult disease, we can devise some general guidelines for clinical decision making. All diploid samples were in stage I cases and, although diploidy is uncom­ mon, if it is found it may suggest true stage I disease. Also, a high per cent S phase for aneuploid cell population (more than 40) or proliferative index (more than 50) may suggest occult disease. Only 1 of the stage I cases had per cent S phase for aneuploid cell population greater than 40 and only 2 had a proliferative index of more than 50. A high percentage of embryonal carcinoma and a lower content of yolk sac or tera­ toma elements portends occult disease. None of the stage I cases was pure embryonal carcinoma, whereas 4 of the stage II cancers had this histology. All but 2 stage I cases had some element of yolk sac tumor and 7 had greater than 20% , whereas only 6 stage II tumors had yolk sac elements and only 1 was greater than 20% . Similarly, only 1 of the stage I tumors had no teratomatous elements (but that case did have 45% yolk sac elements), whereas 5 stage II tumors had no teratoma. Pure embryonal carcinoma probably signifies occult disease and greater than 50% embryonal cancer suggests a likelihood of stage II. Likewise, greater than 20% yolk sac tumor bodes well for stage I but lack of yolk sac or teratoma suggests stage IL Finally, the presence of vascular invasion, especially coupled with other factors, would help to predict stage II disease. A larger prospective multicenter study using central reference pathology determination of quantitative histological and flow cytometry parameters to stratify clinical stage I nonsemino­ matous germ cell testicular cancer into treatment arms is necessary. Perhaps an algorithm of factors can better define patients for optimal therapy. REFERENCES

1. Einhorn, L. H.: Treatment of testicular cancer: a new and improved model. J. Clin. Oncol., 8: 1777, 1990. 2. Moul, J. W.: Pitfalls in the management of testicular cancer patients and complications of therapy. In: Urology Annual. Ed­ ited by S. N. Rous. East Norwalk, Connecticut: Appleton & Lange, 1992. 3. Foley, J. P., Moul, J. W. and McLeod, D. G.: Computed tomography retroperitoneal staging shortcomings in clinical stage I nonsem­ inomatous germ cell tumors. Urology, in press. 4. Moul, J. W. and Paulson, D. F.: The enduring operation: RPLND for testicular nonseminoma. Contemp. Urol., pp. 31-40, Decem­ ber 1990. 5. Pont, J., Holt!, W., Kosak, D., Machacek, E., Kienzer, H., Julcher, H. and Honetz, N.: Risk-adapted treatment choice in stage I nonseminomatous testicular germ cell cancer by regarding vas­ cular invasion in the primary tumor: a prospective trial. J. Clin. Oncol., 8: 16, 1990. 6. Fung, C. Y., Kalish, L. A., Brodsky, G. L., Richie, J. P. and Garnick, M. B . : Stage I nonseminomatous germ cell testicular tumor: prediction of metastatic potential by primary histopathology. J. Clin. Oncol., 6: 1467, 1988. 7. Freedman, L. S., Parkinson, M. C., Jones, W. G., Oliver, R. T., Peckham, M. J., Read, G., Newlands, E. S. and Williams, C. J.:

8.

9. 10. 11. 12. 13.

14. 15. 16.

17.

18.

19.

20.

21. 22. 23.

24.

25. 26. 27. 28.

883

Histopathology in the prediction of relapse of patients with stage I testicular teratoma treated by orchidectomy alone. Lancet, 2 : 294, 1987. Sesterhenn, I. A., Weiss, R. B., Mostofi, F. K., Stablein, R. G., Rowland, R. G., Falkson, G., Rivkind, 8. E. and Vogelzang, N. J.: Prognosis and other clinical correlates of pathologic review in stage I and II testicular carcinoma. a report from the testicular Cancer Intergroup Study. J. Clin. Oncol. , 1 0: 69, 1992. Atkin, N. B. and Kay, R.: Prognostic significance of modal DNA value and other factors in malignant tumours, based on 1465 cases. Brit. J. Cancer, 40: 2 10, 1979. Zimmermann, A.: Aneuploidie bei malignen Hodentumoren und ihren Lymphknoten metastasen. Urologe A, 19: 391, 1980. Barlogie, B.: Abnormal cellular DNA content as a marker of neoplasia. Eur. J. Cancer Clin. Oncol., 2 0 : 1123, 1984. Fossa, 8. D., Pettersen, E. 0., Thorud, E., Melvik, T. E. and Ous, S.: DNA flow cytometry in human testicular cancer. Cancer Lett., 28: 55, 1985. Buchner, T., Hiddemann, W., Wormann, B., Kleinemeir, B., Schumann, J., Gohde, W., Ritter, J., Miller, K M., von Bassewitz, D. B. and Roessner, A.: Differential pattern of DNA-aneuploidy in human malignancies. Path. Res. Pract., 1 79 : 310, 1985 Quirke, P., Dyson, J. E. D., Sutton, J., Anderson, C. K., Joslin, C. A. F. and Bird, C. C . : Assessment of germ cell tumours of testis by flow cytometry and histopathology. Adv. Biosci., 55: 45, 1986. Kleinhans, G., Hacker-Klom, U., Gohde, W., Korner, F. and Schumann, J.: Studies of the cell kinetics of human malignant testicular tumors. Urologe A, 2 5 : 294, 1986. Sledge, G. W., Jr., Eble, J. N., Roth, B . J., Wuhrman, B. P. and Einhorn, L. H.: Flow cytometry derived DNA content of the primary lesions of advanced germ cell tumors. Int. J. Androl., 10: 115, 1987. Sledge, G. W., Jr., Eble, J. N., Roth, B . J., Wuhrman, B . P., Fineberg, N. and Einhorn, L. H.: Relation of proliferative activity to survival in patients with advanced germ cell cancer. Cancer Res., 48: 3864, 1988. Oosterhuis, J. W., Castedo, 8. M., deJong, B . , Cornelisse, C. J., Dam, A., Sleijfer, D . T. and Schraffordt Koops, H.: Ploidy of primary germ cell tumors of the testis. Pathogenetic and clinical relevance. Lab. Invest., 60: 14, 1989. Yoshihiro, S., Yamamoto, M., Shimabukuro, T., Matsuyama, H., Yamamoto, N. and Sakatoku, J.: Flow cytometric analysis of DNA ploidy using paraffin-embedded germ cell tumors. Acta Urol. Jap., 3 7 : 601, 1991. Fossa, S. D., Nesland, J. M., Waehre, H., Amellem, 0. and Pettersen, E. 0.: DNA ploidy in the primary tumor from patients with nonseminomatous testicular germ cell tumors clinical stage I. Cancer, 6 7 : 1874, 1991. Fossa, S. D., Nesland, J. M., Petterson, E. 0., Amellem, 0., Waehre, H. and Heimdal, K: DNA ploidy in primary testicular cancer. Brit. J. Cancer, 64: 948, 199 1 . Codesal, J . , Paniagua, R., Regadera, J., Fachal, C. and Nistal, M . : Significance o f D N A quantification i n testicular germ cell tu­ mors. Andrologia, 2 3 : 381, 1991. El-Naggar, A. K., Ro, J. Y., McLemore, D., Ayala, A. G. and Batsakis, J. G.: DNA ploidy in testicular germ cell neoplasms: histogenic and clinical implications. Amer. J. Surg. Path., 1 6 : 611, 1992. A!lhoff, E., Wittekind, C., Liedke, 8., Tanke, H. and Jonas, U.: The modular image analysis computer for assessment of prog­ nosis in testicular tumors. Verh. Dtsch. Ges. Path., 74: 196, 1990. Hitchcock, C. L., Norris, H. J., Khalifa, M. A. and Wargotz, E. S.: Flow cytometric analysis of granulosa tumors. Cancer, 64: 2127, 1989. Hosmer, D. W., Jr. and Lemeshow, S.: Applied logistic regression. New York: John Wiley & Sons, 1989. Merkel, D. E. and McGuire, W. L.: Ploidy, proliferative activity and prognosis: DNA flow cytometry of solid tumors. Cancer, 6 5 : 1194, 1990. Seckinger, D., Sugarbaker, E. and Frankfurt, 0.: DNA content in human cancer. Application in pathology and clinical medicine. Arch. Path. Lab. Med., 1 13: 619, 1989.