Tumor Markers in Hydrocele Fluids of Patients With Benign and Malignant Scrotal Diseases

0022-5347/97/1583-0A51$03.00/0

THEJOIWNAI. OF uROl.OCU Copyright 0 1997 by

AMERIGV~ UIWl.Oc;lCAL

Vol. 158, 851-855. September 1997 Prrrrlerl rn U S A .

ASSOCIATIOK, INC

TUMOR MARKERS IN HYDROCELE FLUIDS O F PATIENTS WITH BENIGN AND MALIGNANT SCROTAL DISEASES K. DORFINGER, C. KRATZIK, S. MADERSBACHER, G. DORFINGER, P. BERGER M. MARBERGER

AND

From the Department of Urology, UniLersity of Vienna, Department of Laboratory Medicine of the Center of Pulmonary Diseases Baunigartner Hohe, Vienna and In5titute for Biomedical Aging Research of the Augtrian Academy of Sciences, Innshruck, Austria

ABSTRACT

Purpose: We evaluated the presence of human chorionic gonadotropin (hCG), a-fetoprotein (AFP) and a panel of other tumor markers in the compartment next to the tumor (that is, the malignant hydrocele fluid). Materials and Methods: We measured hCG, AFP, neuron-specific enolase, carcinoembryonic antigen and cytokeratin-19 fragments in cubital vein sera and in hydrocele fluids of patients with testicular cancer. Results were compared with those obtained from hydrocele fluids of patients with benign disease. Results: All tumor markers remained under the respective cutoff values in benign hydroceles. In patients with pure seminomas, hCG levels were elevated in 66% of hydroceles but only once in peripheral sera, whereas AFP remained low in both compartments. Furthermore, of 11 cases of nonseminomatous germ cell tumor hydrocele fluids, 3 with negative peripheral tumor marker values had to be reclassified marker positive, of which 2 showed elevated hCG levels and 1 had increased levels of AFP. Significant changes of neuron-specific enolase and carcinoembryonic antigen concentrations could not be observed. However, a cytokeratin-19 fragment measured by Cyfra 21-1 assay was elevated in 2 of 3 seminomatous and in 4 o f 8 nonseminomatous hydroceles. Conclusions: These data give a new insight into the in vivo secretion pattern of testicular germ cell neoplasms, which demonstrates t h a t the term “marker negative” should be restricted to selected cases of testicular cancer. Analysis of tumor markers in hydrocele fluids may be a helpful tool in patients with scrotal swelling if clinical and sonographic results remain uncertain. KEYWORDS:alpha fetoproteins, testicular neoplasms, hydrocele Patients with testicular cancer have a favorable prognosis compared with patients who have other solid malignancies, not only because of powerful therapeutic regimens but also because of early detection by high frequency ultrasonography and by measurement of specific tumor markers. Additionally, the preoperative levels of a-fetoprotein (AFP) or human chorionic gonadotropin (hCG) have been recognized as independent prognostic factors;’ no decrease after surgery indicates persistence of the disease. Therefore, quantification of hCG and AFP is routinely performed in patients with suspected or known testicular cancer for diagnosis and staging purposes as well as to monitor therapeutic effects.‘ Recently, the accuracy of peripheral tumor marker measurement has been debated; several authors reported higher hCG levels in spermatic vein blood compared with cubital vein blood.:+-” It has been suggested that the relative decrease of tumor marker concentrations in the periphery is caused by a diluting effect3 or by metabolic clearance.fi The development of highly sensitive and specific immunoassays for selective quantification of holo-hCG and its free 0-subunit was the next step to improve identification of marker positive patients? Consequently, the term “marker negativeseminoma” was questioned because P-hCG secretion could be detected in the spermatic vein blood of about 80% of patients with pure semin0mas.x Accordingly, we recently demonstrated that 92% of patients with testicular cancer had to be categorized within the marker positive group, regardless of histological classification, when a combined measurement of total hCG and

P-hCG-subunit was performed in another compartment, namely the respective hydroceles.6 In the present study, we evaluated for the first time the presence of other tumor markers in the hydroceles of testicular cancer patients. The questions t o be evaluated were 1) whether a similar increase could be observed for AFP, the second most important tumor marker in testicular cancer, 2) whether neuron-specific enolase can be detected, because this marker has been shown to be elevated in 11 of 16 testicular cancer patients with distant metastasis,”. 10 3) whether carcinoembryonic antigen, which should be negative in testicular cancer patients, might also be elevated and 4) whether a fragment of cytokeratin-19, which is known to be increased in malignant pleural effusions, represents a comparable nonserum compartment next to the tumor.” Therefore, we analyzed tumor marker concentrations in hydrocele fluids and in cubital vein blood of patients with benign and malignant diseases of the scrotal contents. PATIENTS AND METHODS

Patients. The hydrocele fluids of 35 male patients ( 5 to 74 years old), who underwent surgery either for benign diseases or for testicular cancer, were obtained between July 1994 and March 1995. The benign group initially consisted of 14 patients (mean age 31.6 i 23.2 years). The diagnostic study included a thorough exploration of patient history, a complete physical examination and scrotal sonography (SI 450*, 10 MHz.) if necessary. A color Doppler ultrasonography study was performed as well to exclude varicocele or inflammatory diseases. In 12 patients, clinical examinations did not

* Siemens AG, Germany

Accepted for publication January 17, 1997

85 1

852

TUMOR MARKERS IN HYDROCELES

were aspirated intraoperatively under sterile conditions, centrifuged and frozen immediately. Subsequently, all samples were assayed with the use of the same test kit to minimize No Age Type Stagmg interassay variation. Quantitative analysis of tumor marker 1 43 Seminoma TlNOMO levels was performed using t h e following standard immuno39 Seminoma TmnMo 2 assays: B-hCG and AFP were measured by a microparticle TlNOMO 0 3* Stminoma T2N2MO 4 36 Seminoma enzyme immunoassay.* The p-hCG assay measured the holoTlNOMO 5 36 Seminoma. multifocal hCG molecule as well as t h e p-subunit. Neuron-specific enoTlNOMO :I3 Seminoma 6 lase levels were determined by a solid-phase, 2-site fluoroimTlNOMO 7 30 Seminoma munometric assay (Delfia neuron-specific enolase k i t t j and TlNOMO 8 26 Seminoma. multifocal TlNlMO 9 26 Seminoma carcinoembryonic antigen as well as cytokeratin-19 frag.%minoma * intermediate malignant teratoma TlNOMO 28 10 ments (Cyfra 21-1$) were detected by a n enzyme-linked imSeminoma intermediaw malignant teratoma T4NlM1 32 11 munosorbent assay using streptavidin technology. Lactate Seminoma undifferentiated malignant teratoma TlNlMl 12 38 dehydrogenase, C-reactive protein, glucose and total protein 13 41 MTL' TlSOMO TlNOMO 14 30 MTL' levels were analyzed on a Hitachi 717 analyzer.$ 15 25 MTU TZNlMO As indicated by the distributors, the cut-off levels were 5 16 40 MTL' T3N2M1 ng./ml. for P-hCG, 8.6 ng./ml. for AFP, 10 ng./ml. for carcinoTIN 1510 17 2 3 MTU embryonic antigen, 12.5 ng./ml. for neuron-specific enolase 18 2:3 MTL' TlNOMO TZN2MO 19 29 MTC' and 3.3 ng./ml. for Cyfra 21-1. The expected values for pa20 29 MTI' TlNOMO rameters in normal serum ranged as follows: 120 to 240 TlNOMO 21 49 Leydig cell Ca unitsfl. for lactate dehydrogenase, 76 to 110 mg./100 ml. for glucose and 66 to 88 gmA.for total protein. Statistics. Data were compared using the Wilcoxon ranking reveal any primary cause for the hydrocele and were there- test. Results were estimated to be significant at p 50.05. fore classified as "idiopathic" hydroceles. Two patients were excluded from the study, 1 had a secondary hydrocele after RESULTS traumatic hematoma of the scrotum and 1 had a spermatoHydrocele fluid composition. As indicated in table 2, hydrocele, which was revealed when cytology was performed by a as the malignant group were cele fluids of the benign a s well remarkably dark-colored hydrocele fluid. analyzed for their composition compared with cubital vein The group with malignant diseases consisted of 21 patients blood serum levels. Similarly, mean glucose concentrations with a mean age of 32.1 2 6.1 years. Principal patient characteristics a r e listed in table 1. Clinical study included a were within normal range in cubital vein sera and hydrocele fluids of both groups. Although mean total protein content complete physical examination, high resolution sonography was slightly decreased in the hydroceles (mean plus standard of the scrota1 contents and the retroperitoneum, tumor marker analysis, abdominal computerized tomography, bi- deviation, benign 43 2 16 gA.,malignant, 4 1 -t 12) compared phase radiographs of the chest and routine preoperative with serum values (benign 70 % 12 gA., malignant, 69 2 blood parameters. All patients underwent inguinal semicas- 8 gm.fl.), no difference was observed between benign and tration. Histological examination revealed pure seminomas malignant diseases. In addition, the values for lactate dehyin 9 patients (seminomatous germ cell tumors) and nonsemi- drogenase were higher in patients with testicular cancer nomatous germ cell tumors in 12 patients. One patient was (277.6 2 256.8 units/l.) and lower in those with idiopathic found to have a rare Leydig cell tumor and was therefore not hydroceles (175.4 -t 64.3 unitsfl.) compared with t h e sera, included. Blood samples were obtained before surgery in 4 although markedly wide ranges were observed that lacked cases of the benign group and in all of the cancer patients. significance. Tumor marker concentrations of serum samples Germ cell tumor classification was performed according to were measured in 4 patients with idiopathic hydrocele before the 1992 TNM system.12 Further treatment was adapted to the operation. All of them remained under the respective the stage of the disease and additional risk factors.'" Seven of cutoff value (data not shown). Tumor markers in idiopathic benign hydroceles. The mean the seminoma patients who had stage 1 (or higher) disease received 2 cycles of carboplatin monochemotherapy and 2 concentrations of tumor markers in benign hydroceles ( 12) stage l a patients were observed. In the nonseminomatous are shown in figure 1. After we compared the data with the germ cell tumor group, 6 patients with stage 1 (or higher) expected values in normal serum, only 2 patients revealed were treated with 4 cycles of the bleomycin/etoposide/cispla- moderately elevated hCG concentrations in hydrocele fluids tin scheme and 2 of them had a retroperitoneal lymphade- (12.9 and 10.5 ng./ml.), both of whom were smokers. AFP as nectomy because of residual disease. Stage la was diagnosed well a s neuron-specific enolase were raised independently in the remaining patients without additional risk factors and only once, AFP 28.4 ng./ml., neuron-specific enolase, 25.8 they were observed without chemotherapy at close followup ng./ml.). In contrast, cytokeratin-19 fragments were slightly examinations. There was no case of residual disease or re- raised in more t h a n 70% of the benign hydrocele samples (12, lapse in any patient (mean followup seminoma, 39.1 % 14.8 mean 14.3 2 12.01 ng./ml.), which suggested that t h e serum months, nonseminomatous germ cell tumor, 38.2 -C 7.5 cutoff (3.3 ng./ml.) was of indefinite validity. As mentioned months ). Abbott Laboratories, Wiesbaden, Germany. Imniunonssays. All cubital vein blood samples were centriWallac Oy, Turku, Finland. fuged a t 500 gravity and frozen a t -7OC. Hydrocele fluids t Roehringer-Mannheim Immunodiagnostics, Mannheim, Germany.

TABLE1. Histopathological classification cancer

of

patients with testicular

7

r

t

TUMOR MARKERS IN HYDROCELES

853

cele 1067.9 2 1940.6 ng./ml., p = 0.0058, fig. 3, A ) . Consequently, 2 patients who seemed t o be marker negative revealed elevated hCG levels in malignant hydroceles (28.6 ng./ml. and 5.6 ng./ml.) and had to be reclassified within the marker positive group. In 1 patient, /3-hCG concentration was slightly higher in cubital vein serum (41.9 ng./ml.) than in the corresponding hydrocele (15.7 ng.lml.1. AFP concentrations were elevated in hydroceles of 9 patients with nonseminatous germ cell tumors, 8 of whom also exhibited a peripheral increase (fig. 3, B ) . Again, hydrocele levels were higher than in the corresponding cubital vein sera (mean plus or FIG. 1. Tumor marker concentrations (ng./ml.) in benign "idiopathic" hydroceles are shown. Values are mean of 12 patients (0)minus standard deviation sera 241.8 2 429.1 ng./ml., hydroplus or minus standard deviation (0). I shows cut-off value for cele 3369.2 ? 9137.1 ng./ml.), although they lacked signifinormal sera of respective tumor marker. CEA, carcinoembryonic cance ( p = 0.0738). No significant results were obtained antigen. N S E , neuron-specific enolase. regarding the various histological entities, because of the low number of samples. Neuron-specific enolase was higher in 2 above, the concentrations in cubital vein blood remained of the serum samples but in only 1of the hydrocele fluids (6, normal. Excessive levels of cytokeratin 19 fragments were data not shown). Cytokeratin-19 fragments in testicular cancer. Figure 4 observed in the hydrocele as a result of testicular trauma, which probably indicates a higher cytokeratin release caused shows the mean concentration of cytokeratin-19 fragments in by tissue damage (data were not included in the study). A sera and hydrocele fluids of patients with benign and maligtheoretical cutoff value for cytokeratin 19 fragments with nant diseases. Although cytokeratin-19 fragments were evalsensitivity of 9 6 8 was calculated to be 51.5 ng./ml. in benign uated in only 3 of the seminomatous hydroceles, surprisingly, hydrocele fluids. Carcinoembryonic antigen levels consis- 2 of them exhibited exceedingly elevated cytokeratin-19 fragtently remained negative in all cases. ment concentrations within hydrocele fluids. Moreover, Tumor markers in hydroceles of patients with seminoma- cytokeratin-19 fragment were increased in 50% of the nontous germ cell tumors. The distribution of hCG levels in seminatous germ cell tumor hydroceles (8) but remained hydrocele fluids from patients with pure seminomas (9) com- consistently negative in all corresponding sera of both pared with serum levels in cubital vein blood is indicated in groups, seminomatous as well as nonseminatous patients figure 2, A. Whereas raised serum hCG was observed only (fig. 4). once, 6 of 9 seminomatous hydrocele samples exhibited increased hCG levels. In contrast, AFP was slightly raised only in 1 of 9 patients (1 serum 18.4 ng./ml., hydrocele 10.5 ng./ ml.) in both compartments (fig. 2, B ) . Because of the limited hydrocele volume available, further tumor marker analysis could only be performed in 3 seminoma patients. Neuronspecific enolase concentration was increased only once, in the cubital vein blood (41.3 ng./ml.) as well as in the corresponding malignant hydrocele (1.733 ng.Iml.1. Carcinoembryonic antigen levels were consistently within the normal range (data not shown). Tumor markers in hydroceles of patients with nonseminomatous germ cell tumors. In the group of patients with nonseminomatous germ cell malignancies, 0-hCG was elevated in 7 of 11 sera and 9 of 11 hydrocele fluids, which always exhibited convincingly higher concentrations (mean plus or minus standard de&tionsera 201.8 2 352.8 ng./ml.,-hydro-

FIG. 2. A, hCG (ng./ml.)concentrations (lefl)and hydrocele fluids (right)of 9 patients with pure seminomas are demonstrated in logarithmic scale. B , AFP concentrations are shown in same sera and hydroceles. Thin horizontal line shows normal serum cutoff value.

FIG.3. hCG tA) values and AFP ( B )levels of 1 1 nonseminomatous germ cell tumors are shown on logarithmic scale (ngJm1.L represents serum values and I; corresponding concentrations within hydrocele fluids. Thin horizontal line shows normal serum cutoff.

TUMOR MARKERS IN HYDROCELES

854

Idiopathic

malignant

FK;.4. Mean values of cvtokeratin-19 fragment concentrations in idiopathic (.I a n d malignant IF111 probes are demonstrated on logarithmic scale I: ., standard deviation]. S.serum. H , hydrocele. Thin horizontal line shows theoretical cutoff value calculated out of hydrocelt. cytokeratin-19 fragment concentrations with sensitivity of 96';.

DIS('USSI0N

In the present study, we evaluated for the first time the presence of various tumor markers in secondary hydrocele fluids of patients with testicular cancer. Results were compared with those of the benign group of primary "idiopathic" hydroceles. Simultaneously, peripheral tumor marker concentrations were determined in cubital vein blood samples. Once a tumor is detected, the discrimination between benign and malignant lesions is frequently hampered by various difficulties in routine clinical examination. Searching for accurate methods, several investigators have analyzed tumor marker levels in compartments close to the tumor.i4-16 In testicular cancer tumor markers are not only helpful tools in diagnostic study but are also important prognostic factors and therefore must be considered for therapeutic management.l.2.17 In accordance with the review of Mann,' approximately 80% of the nonseminatous germ cell tumor patients reveal increased levels of either AFP or hCG in peripheral blood samples. However, the diagnostic sensitivity of AFP alone ranges between 50 and 80% and that of hCG is estimated to be around 60%.In a large series of 106 patients with pure seminomas, Ruther e t al noted elevated peripheral blood levels of hCG in only 30.24, although they used highly sensitive assays.'* In 1983, Fiet et al" gave the first report on higher P-hCG levels in spermatic cord blood in 2 with seminomatous carcinoma of the testis who were marker negative in peripheral blood samples. However, syncytiotrophoblastic p a n t cells were detected in only 1 of them. In a series of 47 seminoma patients, Mumperov and Hartmann observed elevated hCG levels in 80% of spermatic cord blood samples versus 22%in cubital vein blood." I t h a s been speculated that a low density of syncytiotrophoblastic giant cells is sometimes responsihle for this finding, particularly because the cells are not always detected, despite careful histological preparation. We demonstrated recently that u p to 92% of testicular carcinomas must be classified marker positive, regardless of the histology, when combined analysis of hCG and its free suhunits in hydrocele fluids was performed." These data therefore suggest a higher sensitivity of tumor marker assays when assessed directly a t the site of the tumor. The wide range of percentages of hCG positive peripheral hlnod samples of seminomatous patients indicated in several publications was a t least in part attributed to cross-reactive assays or different sensitivities.'. I!' Therefore, immunoass a y s with a higher sensitivity and specificity were developed.

The commercially available assay kits used in this study have a cross-reactivity below 1%. In t h e present study, it is demonstrated that tumor marker levels can be established in hydrocele fluids even when commercial assays a r e used. The data presented herein indicate for the first time that AFP levels, compared with hCG levels, a r e significantly higher in hydrocele fluids than in the periphery. However, whereas hCG levels in pure seminomatous hydroceles were positive in 2 of 3 patients, AFP levels were below the cutoff in all but 1 of the seminoma series (fig. 2, B ) . Because this was only a minor elevation, intensively increased AFP levels within the hydrocele fluid might be helpful parameters to distinguish seminomatous from nonseminomatous germ cell tumors, which is in accordance to the literature based on cubital vein blood analysis.' From the group of patients with negative serum tumor markers, 5 seminomatous (fig. 2, A ) and 2 nonseminomatous germ cell tumor hydroceles (fig. 3, €3) disclosed elevated levels either for hCG or for AFP and therefore had to be recategorized as marker positive tumors. As indicated in figure 1, from the whole panel of tumor markers (hCG, AFP, carcinoembryonic antigen, neuron-specific enolase and cytokeratin-19 fragments) only cytokeratin-19 fragments exhibited raised values in the benign control group of idiopathic hydroceles compared with normal serum cutoff values as measured by Cyffa 21-1 assay. This fragment of cytokeratin 19, a n intermediate filament protein, that is produced by epithelial cells of several tissues (for example, the lung), has been shown to be a useful tumor marker for lung cancer.20 If we calculate a hydrocele specific cutoff with 95% specificity for Cyfra 21-1, a theoretical value of 51.5 ng./ml. results, which is in good accordance with that of pleural effusions of lung cancer patients (234),where a value of 53.9 ngJml. has been calculated using the same type of assay (unpublished data). Several other tumor markers have been analyzed in the sera of testicular cancer patients, such as neuron-specific antigen?, 1" all of which have a low specificity. To address the question of whether the specificity was higher in the neighborhood of the tumor, we also measured neuron-specific antigen in the hydroceles, but no significant alterations could be observed in hydrocele fluids. Remarkably, this was not true for cytokeratin-19 fragments; this marker was elevated in 2 of 3 seminoma patients, even if we used the higher cutoff value calculated for benign hydroceles. In nonseminomatous germ cell tumor hydroceles, Cyfra 21-1 revealed elevated cytokeratin-19 fragment concentrations in 4 of 8 patients. To evaluate the feasibility of the latter marker, analysis of larger series has been started recently in our laboratory. CONCLUSIONS

Overall, 7 of 21 patients with testicular cancer had to be reclassified a s marker positive by analysis of the compartment near the tumor. The assessment of tumor markers in hydrocele fluids can be helpful in the discrimination between benign and malignant diseases of the scrota1 content. I n cases of uncertain clinical results, this evaluation may be a n argument for watchful waiting rather than immediate surgery. REFERENCES

1. Aass, N.. Klepp. 0..Cavallin-Stahl, E.. Dahl. O., Wicklund. H., Unsgaard, €3.. Baldetorp, L.. Ahlstrom. S., and Fossa. S. D.: Prognostic factors in unselected patients win nonseminomatous metastatic testicular cancer: A multicenter experience. J. Clin. Oncol.. 9 818. 1991. 2. Mann. K.: Tumor markers in testicular carcinoma. Urologe A, 2 9 77, 1990. 3. Fiet. .I.. .Jardin. A,, Guechot, J . and Vilette, ,J. M.: High levcls of beta-hCC in spermatic vein of patients with scminoma. I a n cet. 2: 1196. 1983.

855

TUMOR MARKERS IN HYDROCELES 4. Light, P. A. and Tyrell, C. J.: Testicular tumor markers in spermatic vein blood. Brit. J. Urol., 59: 74. 1987. 5. Mann, K. and Siddle, K.: Evidence for free beta-subunit secretion in so-called human chorionic gonadotropin-positive seminoma. Cancer, 62: 2378, 1988. 6. Madersbacher, S., Kratzik, C., Gerth, R., Dirnhofer, A. and Berger, P.: Human chorionic gonadotropin (hCG) and its free subunit in hydrocele fluids and neoplastic tissue of testicular cancer patients: Insights into the in vivo hCG-secretion pattern. Cancer Res., 54: 5096, 1994. 7. Marcillac, I., Troalen, F., Bidart, J. M., Ghillani, P., Ribrag, V., Escudier, B., Malassagne, B., Droz, J. P., LHomme, C., Rougier, P., Duvillard, P., Prade, M., Lugagne, P.-M., Richard, F., Poynard, T., Bohoun, C., Wands, J. and Bellet, D.: Free human chorionic gonadotropin beta-subunit in gonadal and nongonadal neoplasms. Cancer Res., 52: 3901, 1992. 8. Mumperow, E. and Hartmann, M.: Spermatic cord p-human chorionic gonadotropin levels in seminoma and their clinical implications. J. Urol., 147: 1041, 1992. 9. Fossa, S. D., Klepp, 0. and Paus, E.: Neuron-specific enolase-a serum tumor marker in seminoma? Brit. J. Cancer, 65: 297, 1992. 10. Gross, A. J. and Dieckman, K. P.: Neuron-specific enolase: enolase-a serum tumor marker in malignant -germ-cell tumors? Eur. Urol., 24: 277, 1993. 11. Satoh. H.. Sumi, M.. Y a r n , H., Suyama, T., Naitoh, T., Saitoh, T. and Hasegawa, S.-Clinical evaluation of CYFRA 21-1 in malignant pleural fluids. Oncology, 52: 211, 1995. 12. TNM Classification of Malignant Tumours, 4th ed. Edited by P. Hermanek and L. H. Sobin. New York: Springer-Verlag, pp. 145-147, 1992. 13. Krainer, M., Kiihrer, I. and Kratzik, C.: Testicular cancer-state of the art. Wien Clin. Wochenschr., 106(2): 37, 1994. 14. Gaetje, R. and Popp, L. W.: Is differentiation of benign and malignant cystic adnexal masses possible by evaluation of cysts fluids with respect to color, cytology, steroid hormones, and tumor markers? Acta Obst. Gynec. Scand., 73:502, 1994. 15. Alles, A. J . , Warshaw, A. L., Southern, J . F., Compton, C. C. and Lewandrowski, K. B.: Expression of CA 72-4 (TAG 72) in the fluid contents of pancreatic cysts. A new marker to distinguish malignant from benign neoplasms and pseudocysts. Ann. Surg., 219: 131, 1994. 16. Rubin, D., Warshaw, A. L., Southern, J . F., Pins, M., Compton, C. C. and Lewandrowski, K. B.: Expression of CA 15.3 protein in the cyst contents distinguishes benign from malignant pancreatic mucinous cystic neoplasms. Surgery, 115: 52, 1994. 17. Barlett, N. L., Freiha, F. S. and Torti, F. M.: Serum markers in germ cell neoplasms. Hematol. Oncol. Clin. N. Amer., 5: 1245,

1991. 18. Riither, U., Rothe, B., Grunert, K., Bader, H., Sessler. R . , Nunnensiek, C.. Rassweiler, J., Ldthgens, M., Eisenberger, F. and Jipp, P.: Role of human chorionic gonadotropin in patients with pure seminoma. Eur. Urol., 2 6 129, 1994. 19. Paus, E., Fossh, S. D., Risberg, T. and Nustad, K.: The diagnostic value of human chorionic gonadotropin in patients with testicular seminoma. Brit. J. Urol., 59: 572, 1987. 20. Rastel, D., Ramaioli, A,, Cornillie, F. and Thirion, B.: CYFRA 21-1, a sensitive and specific new tumor marker for squamous cell lung cancer. Report of the first European multicentre evaluation. CYFRA 21-1 Multicentre Study Group. Eur. J. Cancer, 3 0 A 601, 1994. EDITORIAL COMMENT This paper is interesting, but it does not help very much in the few instances in which one is unable to tell preoperatively if the scrota1 lesion is intratesticular or not. The authors’ observations on hydrocele fluid were all made a t the time of inguinal exploration. Perhaps analysis of these tumor markers could be useful on aspirates if the diagnosis were still in doubt despite negative peripheral tumor markers, ultrasonography and even magnetic resonance imaging. These diagnostic dilemmas are rare but may occur. However, to aspirate a hydrocele when there is suspicion of a testicular tumor goes against the tenant of not violating the scrotum. Even if a n aspirate were to be performed as a part of a study, the investigator would still face a dilemma because, a s shown in this paper, testicular tumors do not always result in a detectable tumor marker (P-hCG or AFP) in the hydrocele fluid. In the case of cytokeratin-19 fragments, although this marker was absent in the serum of patients with testis tumor, it was not always present in hydrocele fluid either. This paper would be improved by the observation of whether or not detection of 1 or more of the markers in hydrocele fluid would correlate with persistence in elevated serum levels post-orchiectomy, a s we all recognize that the persistence of a tumor marker post-orchiectomy is of extreme prognostic importance. Nevertheless, the above comments are not meant in a negative vein, because it is through studies of this sort that we can continue to make advancement against this disease, which is lethal in some despite the great therapeutic advances achieved with cisplatin based regimens. David G. McLeod Urology Service Walter Reed Arniv Medical Center Washington, D. C