337 Sarcoidosis and Testicular Carcinoma Daniel Rayson, M.D. Patrick A. Burch, M.D. Ronald L. Richardson, M.D. Division of Medical Oncology, Mayo Clinic and Mayo Foundation, Rochester, Minnesota. BACKGROUND. Case reports and small case series have described patients with sarcoidosis and testicular carcinoma. The goals of this study were to determine the incidence of sarcoidosis in patients with testicular carcinoma seen at the Mayo Clinic between 1950 and 1996 and to examine the relation between these two diseases, particularly in regard to the therapy and follow-up of patients with testicular carcinoma. METHODS. A computerized search of the patient data base at the Mayo Clinic was conducted to identify all patients seen between 1950 and 1996 who had a diagnosis of a malignant testicular tumor and sarcoidosis. RESULTS. A total of 14 patients were identified. The median age at diagnosis of testicular carcinoma was 31.5 years. Eleven patients presented with Stage I disease and 3 with Stage II disease. Twelve patients had carcinoma diagnosed before sarcoidosis. The median age at the time of diagnosis of sarcoidosis was 36.5 years. Nine patients presented with radiographic Stage I disease, four with radiographic Stage II disease, and one with extrapulmonary disease. The estimated cumulative incidence of sarcoidosis in patients with testicular carcinoma seen at the Mayo Clinic between 1976 and 1995 was 617.3 per 100,000. CONCLUSIONS. These data suggest that, compared with other solid tumors, testicular carcinoma has the strongest association with sarcoidosis. The observed incidence represents an approximate 100-fold increase compared with a general population of young white men. Although an etiologic relation is possible, the problems of access and surveillance bias should be addressed in prospective studies. Cancer 1998;83:337– 43. © 1998 American Cancer Society. KEYWORDS: epidemiology, sarcoidosis, testicular carcinoma, incidence, association. Presented in part at the 66th Annual Meeting of the Royal College of Physicians and Surgeons of Canada (RCPSC), Vancouver, British Columbia, Canada, September 24-28, 1997. The authors thank J. P. Utz, M.D., Division of Pulmonary and Critical Care Medicine and Internal Medicine, and S. J. Jacobsen, M.D., Section of Clinical Epidemiology, for their review of the article. Dr. Rayson’s current address: Daniel Rayson, M.D., London Regional Cancer Centre, 790 Commissioners Road East, London, Ontario, Canada N6A 4L6. Address for reprints: Dr. Patrick A. Burch, Division of Medical Oncology, Mayo Clinic, 200 First Street SW, Rochester, MN 55905. Received September 26, 1997; revision received January 14, 1998; accepted January 14, 1998. © 1998 American Cancer Society T he relation between sarcoidosis and malignant disease remains controversial. In an initial report by Brincker and Wilbek1 in 1974, the incidence of malignant disease in patients with respiratory sarcoidosis reported to the Danish Institute of Clinical Epidemiology was higher than expected because of an increased incidence of lymphoma and lung carcinoma. No other solid tumors occurred at a higher than expected rate in that study. Subsequent review of the reported cases challenged these results, and the debate is ongoing.2– 4 The strongest evidence for a relation that goes beyond chance association involves lymphoma and sarcoidosis. Brincker5 proposed the existence of a sarcoidosis-lymphoma syndrome after finding that malignant lymphoproliferative diseases developed 5.5 times more often than expected in middle-aged patients with chronic active sarcoidosis. Solid tumors that have been purported to occur with a higher than expected frequency in patients with sarcoidosis include lung and breast carcinoma, and it has been suggested that immunologic abnormalities resulting from sarcoidosis might permit the development 338 CANCER July 15, 1998 / Volume 83 / Number 2 were seen between 1950 and 1996. A total of 14 patients are described, and to our knowledge represent the largest case series reported in the literature. METHODS FIGURE 1. Biopsy specimen obtained at mediastinoscopy, demonstrating noncaseating epithelioid granulomas. of certain solid malignancies.3,6 However, solid data from large surveillance studies are lacking to support an etiologic relation.3,7 This review was prompted by a case of a 45-yearold man with a history of classic seminoma treated with orchiectomy followed by whole abdominal irradiation in 1994. In 1996, a chest radiograph showed extensive right hilar and paratracheal lymphadenopathy. He underwent mediastinoscopy, and biopsies revealed noncaseating epithelioid granulomas consistent with sarcoidosis (Fig. 1). Several case reports and small case series have described patients with sarcoidosis and testicular malignancy. Compared with other solid tumors, these cases have been unique in that, in most instances, the diagnosis of malignancy preceded that of sarcoidosis.4 The goals of this study were to determine the incidence of sarcoidosis in patients with malignant testicular tumors who were seen at our institution between 1950 and 1996 and to examine the relation between these two diseases, particularly in regard to the therapy and follow-up of patients with testicular carcinoma. This study describes all Mayo Clinic patients with a malignant testicular tumor and sarcoidosis who The medical record archive at the Mayo Clinic is a comprehensive source of information regarding the medical care of all patients seen at the clinic. A computerized search of the medical index at the Mayo Clinic was used to identify all male patients with a diagnosis of sarcoidosis and testicular tumor who were seen between 1950 and 1996. With the use of the same computer data base, the total number of cases in which either sarcoidosis or testicular carcinoma was reported as a diagnosis among all men seen at the Mayo Clinic between 1976 and 1995 was ascertained. Estimated cumulative incidence rates were calculated based on the number of cases of each disease evaluated during this period. All charts were reviewed for clinical history, tumor characteristics, radiologic and pathologic evaluation, treatments, and follow-up and outcome information. MEDLINE was searched for the key words testicular tumor, malignant and sarcoidosis to identify all previous cases described in the world literature from 1966 to the present. RESULTS Of the 21 patients identified, 2 were females with ovarian germ cell tumors, 1 patient had lymphoma, 1 had metastatic prostate carcinoma, and 1 had a benign testicular condition. Two patients had a primary mediastinal germ cell tumor diagnosed concurrently with pulmonary sarcoidosis, which likely represented a local sarcoid-like tissue reaction. These seven patients were excluded from the analysis. The 14 remaining patients (Table 1) to our knowledge represent the largest case series reported of testicular malignancy and sarcoidosis. The detailed characteristics of the patients are presented in Table 2. The median age at diagnosis of testicular carcinoma was 31.5 years (range, 18 –51 years). Eleven patients presented with Stage I disease (confined to the testicle) and 3 presented with Stage II disease (metastatic to retroperitoneal lymph nodes). No patients had supradiaphragmatic disease (Stage III) at presentation. Six patients had pure seminomatous tumors, three had nonseminomas, and five had mixed histologic findings. Of the 14 patients with primary testicular malignancy, 12 had their tumor diagnosed before sarcoidosis, and only 2 patients had sarcoidosis before carcinoma. For the 12 patients who had carcinoma before sarcoidosis, the median interval between diagnoses Sarcoidosis and Testicular Carcinoma/Rayson et al. TABLE 1 Characteristics of 14 Patients with Testicular Malignancy and Sarcoidosis Seen at the Mayo Clinic Between 1950 and 1996 Testicular carcinoma Age at diagnosis (yrs) Median: 31.5 Range: 18–51 Stage at diagnosis (no.) I: 11 II: 3 III: 0 Histology (no.) Pure seminoma: 6 Nonseminoma: 3 Mixed: 5 Metastatic sites (no.) None: 9 Retroperitoneal lymph nodes: 4 Lung (?): 1 Sarcoidosis Age at diagnosis (yrs) Median: 36.5 Range: 20–53 Radiographic stage of disease (no.) Stage I: 9 Testicular carcinoma (no.) Stage II: 4 Stage III: 0 Extrapulmonary: 1 Extrapulmonary manifestations (no.) Skin: 4 Ocular: 1 Hypercalcemia: 1 was 59.5 months (range, 24 –288 months). The median duration of follow-up for this group was 124 months (range, 24 – 420 months). All patients with primary testicular malignancy were without evidence of malignancy at last follow-up. The median age at diagnosis of sarcoidosis was 36.5 years (range, 20 –53 years). Nine patients presented with radiographic Stage I disease (limited to lymph node involvement), four patients presented with radiographic Stage II disease (involving lung parenchyma and mediastinal lymph nodes), and one patient had extrapulmonary disease. No patients presented with radiographic Stage III disease (parenchymal involvement alone). In one patient (Case 3), the diagnosis was made at autopsy. All patients but one had tissue confirmation of sarcoidosis. Five patients had manifestations of extrapulmonary disease. Four patients had skin involvement (Cases 7, 8, 12, and 14), one had ocular involvement (Case 12), and one had hypercalcemia (Case 9). Only one patient required treatment of pulmonary sarcoidosis (Case 2). All other treatments were administered for extrapulmonary manifestations; these therapies included topical corti- 339 costeroids (Case 14), hydroxychloroquine (Case 14), ocular corticosteroids (Case 12), and systemic corticosteroids with pamidronate for hypercalcemia (Case 9). In all cases, disease either stabilized or underwent spontaneous regression with no significant pulmonary morbidity or death due to progressive sarcoidosis. Forty-five previous cases have been reported in the world literature. Twenty-one of the patients presented with a diagnosis of testicular tumor before sarcoidosis,8 –19 8 had an antecedent diagnosis of sarcoidosis,16,17,19,20 and 16 had concurrent diagnoses.16 –19,21–28 DISCUSSION The diagnosis of sarcoidosis is made on the basis of a compatible clinical history, exclusion of other causes of granulomatous disease (such as tuberculosis, fungal infections, or berylliosis), and tissue confirmation of noncaseating granulomas.29 Diagnostic tests such as determination of the serum angiotensin-converting enzyme level, gallium scanning, and bronchoalveolar lavage with analysis of the lymphocyte count and T4/T8 subsets have low specificity and are not recommended for routine evaluation in suspected cases of sarcoidosis, although they may be used as supportive evidence in making the diagnosis. The Kveim test, although highly specific and sensitive, requires precise methodologic standardization and largely has been replaced by direct tissue analysis as the standard for confirmation of the diagnosis of sarcoidosis.30 Published incidence and prevalence rates of sarcoidosis vary widely among nations and among different racial and geographic groups within the same country.31.32 In the U. S., data from large epidemiologic studies estimate the incidence of sarcoidosis to vary between 4 and 7.6 per 100,000 white males and to be approximately 10-fold higher among black males.33–35 However, the true incidence likely is considerably higher because of the high frequency of asymptomatic disease and because in the majority of cases the disease resolves spontaneously and thus never comes to medical attention.35 Higher incidence rates have been reported in persons with increased access to medical care and among persons who undergo routine chest radiographic screening for employment or immigration purposes.36 With use of the same computer data base, the total number of cases in which either sarcoidosis or testicular carcinoma was reported as a diagnosis among all men seen at the Mayo Clinic between 1976 and 1995 was ascertained. There were 1839 diagnoses of sarcoidosis and 1782 diagnoses of malignant testicular carcinoma during that period. Of the 14 patients with primary testicular carcinoma and sarcoidosis re- 36 51 29 23 22 28 37 42 27 34 34 18 43 29 Case 1 2 3 4 5 6 7 8 9 10 11 12 13 14 T S S S E S, E T, E ? S, E S, E, T, C S, A S S S Histology None RP lymph nodes None None ?Pulmonary nodule RP lymph nodes None None RP lymph nodes RP lymph nodes None None None None Metastatic sites Orchiectomy, RPLND Orchiectomy, RPXRT Orchiectomy, RPLND, chemotherapy Orchiectomy, RPXRT Orchiectomy, RPXRT Orchiectomy, chemotherapy Orchiectomy, RPLND, Chemotherapy Orchiectomy Orchiectomy, RPLND Orchiectomy, chemotherapy Orchiectomy, RPLND Orchiectomy, RPXRT Orchiectomy mediastinal, RPXRT Orchiectomy, RPXRT Treatment Recurrence in cervical lymph node, chemotherapy; NED NED NED NED NED NED NED NED NED Died of CAD, NED testicular carcinoma Recurrence, contralateral testicle; orchiectomy and XRT NED Died of lymphoma; NED testicular carcinoma NED Outcome Age at diagnosis, (yrs) Sarcoidosis Symptoms 38 45 30 108 36 37 20 35 32 49 45 29 24 36 24 91 60 144 36 80 28 Rash None None None Rash, eye pain Hypercalcemia None Rash Skin nodules None None Yes Yes Yes No Yes Yes Yes Skin Skin Yes Yes Yes 59 Dyspnea Autopsy 49 Yes Biopsy Testicular malignancy before sarcoidosis 288 (autopsy) 53 None 24 Sarcoidosis before testicular malignancy 85 29 None Interval between diagnoses, (mos) I I II II I I I Extrapulmonary I I I I II II Radiographic stagea Topical steroids, hydroxychloroquine None None None Ocular steroids Steroids pamidronate None None None None None None Steroids None Treatment Progressive, asymptomatic disease, then regression Not available Not available Stable Improved Stable Regression Not available Regression Regression Regression Not available Stable Regression Outcome 420 30 24 48 204 104 66 144 60 161 273 293 106 427 Total followup (mos) a S: seminoma; RPXRT: retroperitoneal external beam irradiation, NED: no evidence of disease; CAD: coronary artery disease; A: anaplastic, RPLND: retroperitoneal lymph node dissection; XRT: external beam irradiation; E: embryonal; T: teratoma; C: choriocarcinoma; RP: retroperitoneal I indicates lymph node involvement only, whereas II indicates lymph node and lung parenchymal involvement. Age at diagnosis, (yrs) Testicular carcinoma TABLE 2 Therapy and Outcome of 14 Patients with Testicular Malignancy and Sarcoidosis Seen at the Mayo Clinic Between 1950 and 1996 340 CANCER July 15, 1998 / Volume 83 / Number 2 Sarcoidosis and Testicular Carcinoma/Rayson et al. ported in our series, in 12 the diagnosis of testicular carcinoma was made during this interval and in 11 the diagnosis of sarcoidosis was made during the same period. The estimated cumulative incidence of sarcoidosis in the patients with testicular carcinoma seen at the Mayo Clinic during this period was 617.3 per 100,000. Conversely, the estimated cumulative incidence of testicular carcinoma in the patients with sarcoidosis was 652.5 per 100,000. Both these figures are considerably higher than expected from population-based data, with the overall incidence of testicular carcinoma being 4.5 per 100,000 men in 199037 and the incidence of sarcoidosis being 4 to 7.6 per 100,000 men.33–35 Our data are consistent with those reported by Toner and Bosl,16 who estimated that the incidence of sarcoidosis in patients with testicular germ cell tumors seen at the Memorial Sloan-Kettering Cancer Center between 1980 and 1989 was 510 per 100,000. Testicular malignancy is the most curable of solid tumors, with 3-year survival rates for all stages of disease combined increasing from 89% to 96% between 1978 and 1987.37 Retroperitoneal lymph node dissection after orchiectomy had been the mainstay of treatment for disease localized to the testicle or with evidence of spread to the retroperitoneum.38 However, with the advent of effective platinum-based chemotherapy, outcomes for patients with Stage I disease who undergo orchiectomy without retroperitoneal lymph node dissection (reserving systemic chemotherapy for recurrence) have been equal to those for patients who undergo initial retroperitoneal lymph node dissection.39 – 42 The efficacy of platinum-based chemotherapy for even advanced, widely metastatic disease makes the detection of recurrent disease of the utmost importance and has led to intensive surveillance schedules for patients treated with curative intent. Accordingly, many institutions have adopted a policy of close observation for reliable patients treated with orchiectomy alone.41 This intensive surveillance makes diagnosis of an asymptomatic case of sarcoidosis much more likely, and thus surveillance bias is an important factor in the interpretation of this case series on the relation between sarcoidosis and testicular malignancy. However, it is of interest that the median interval between the diagnosis of sarcoidosis and testicular carcinoma in this series was nearly 5 years (59.5 months), which is beyond the period of the most intensive surveillance. Therefore, this finding somewhat militates against surveillance bias being the sole factor behind the high incidence observed in the current study. The patients presented in this case series did not differ significantly from those reported in the literature with regard to either testicular carcinoma or sar- 341 coidosis. The median age at diagnosis of testicular carcinoma in our case series was 31.5 years. This conforms to the mean age at diagnosis of testicular carcinoma as ascertained by the Surveillance, Epidemiology, and End Results data base of 32 years, with the diagnosis occurring between the ages of 20 and 44 years in 83.5% of all cases.37 The median age at diagnosis of sarcoidosis in our series was 36.5 years, which is consistent with the data regarding age at presentation reported in the literature. The incidence of sarcoidosis is highest in the young adult population; up to 70% of all patients are age , 40 years at presentation.32 In the Rochester, Minnesota, population-based study, the peak incidence in males occurred between ages 30 –39 years.35 All 14 patients in this series were without evidence of recurrent testicular carcinoma at last follow-up, a finding consistent with the 5-year relative survival rates of 98 –99% for patients with local and regional disease between the years 1986 and 1991.43 The four patients who were treated with adjuvant systemic chemotherapy in our series did not experience any unusual toxicities. All patients had either stabilization or spontaneous regression of sarcoidosis at last followup, and only one patient required treatment of pulmonary disease. This is in agreement with findings in numerous reported series, in which only approximately 10% of patients with sarcoidosis required treatment for progressive respiratory disease, and has led to close observation as the treatment of choice for most stable cases because of the high rate of spontaneous regression.30 The high estimated cumulative incidence rates reported in this study and that by Toner and Bosl16 must be interpreted cautiously in regard to a possible etiologic association between the two conditions. Possibly, these figures represent the fortuitous association of two uncommon conditions, both of which occur in the same age group and one of which (sarcoidosis) is discovered incidentally because of intensive surveillance after curative treatment of testicular carcinoma. A spurious correlation may exist between these two diseases because of the compounding of the independent probability of two diagnoses coexisting in a patient population drawn from a tertiary care center, a phenomenon known as Berkson’s bias.44 The emphasis on detecting recurrent testicular carcinoma is attributable to the efficacy of salvage chemotherapy. Recognition of the general patterns of tumor recurrence is important so as to avoid undertreatment of curable disease while avoiding overtreatment of a benign condition. Of interest, of the five reported cases in which patients received chemotherapy for sarcoidosis, one with a fatal outcome,20 four 342 CANCER July 15, 1998 / Volume 83 / Number 2 have been reported since 1990. Undoubtedly, other cases have gone unreported. In general, isolated mediastinal lymph node recurrence of testicular carcinoma is uncommon. In a series of 42 patients with recurrent nonseminoma, Gels et al.41 found that only 2 patients (4.8%) had recurrence solely in the mediastinum. Isolated supradiaphragmatic lymph node recurrence in pure seminoma is somewhat more common because of the use of adjuvant abdominal irradiation in patients with Stage I and II disease. Mediastinal involvement by sarcoidosis is very common, with up to 50% of all cases diagnosed as radiographic Stage I disease (involving mediastinal and hilar lymph nodes only).9 In cases of apparent mediastinal recurrence with negative tumor markers, tissue confirmation is important to avoid potentially serious therapeutic error, but it still may lead to diagnostic dilemmas. Sarcoid tissue reactions can occur within a primary tumor, in the lymph node basin draining a malignancy, or even in distant lymph node sites. These reactions, which morphologically are identical to the noncaseating granulomas of systemic sarcoidosis, occur in up to 4.4% of all carcinomas.45 We found two cases with concurrent diagnoses of sarcoidosis and a mediastinal germ cell tumor that are likely examples of this reaction. Further study of the relation between these two entities should eliminate this confounding situation by excluding cases that are likely to represent local sarcoid tissue reactions. Subsequent study of the possible etiologic association between sarcoidosis and testicular malignancy should be based on large, prospective surveillance programs in which screening for sarcoidosis is performed in two cohorts of young men, one of which is undergoing surveillance for recurrent testicular carcinoma and one of which undergoes the same intensity of diagnostic testing (without tumor markers) at equal intervals. Retrospective reviews of case series run the risk of overestimating the strength of a possible relation because of inherent access and surveillance biases. These two diseases possibly have an etiologic association, but the inherent biases of retrospective reviews limit the strength of this conclusion. The relation between testicular carcinoma and sarcoidosis remains an interesting and clinically important field of study, but any definite statement regarding a causal relation awaits future investigation. REFERENCES 1. Brincker H, Wilbek E. The incidence of malignant tumours in patients with respiratory sarcoidosis. Br J Cancer 1974;29: 247–51. 2. Romer FK. Sarcoidosis and cancer–Da critical view. In: Jones Williams W, Davies BH, editors. Eighth International Conference on Sarcoidosis and Other Granulomatous Diseases. Cardiff: Alpha Omega Publishing, 1980:201–5. 3. Brincker H. Coexistence of sarcoidosis and malignant disease: causality or coincidence? Sarcoidosis 1989;6:31– 43. 4. Rømer FK. Sarcoidosis and cancer. Lung Biol Health Dis 1994;73:401–5. 5. Brincker H. The sarcoidosis-lymphoma syndrome. Br J Cancer 1986;54:467–73. 6. Suen JS, Forse MS, Hyland RH, Chan CK. The malignancysarcoidosis syndrome. Chest 1990;98:1300 –2. 7. Brincker H. Solid tumors preceding or following sarcoidosis. Med Pediatr Oncol 1987;15:82– 8. 8. Trump DL, Ettinger DS, Feldman MJ, Dragon LH. ‘Sarcoidosis’ and sarcoid-like lesions. Their occurrence after cytotoxic and radiation therapy of testis cancer. Arch Intern Med 1981;141:37– 8. 9. Gefter WB, Glick JH, Epstein DM, Miller WT. Sarcoidosis: a cause of intrathoracic lymphadenopathy after treatment of testicular carcinoma. AJR Am J Roentgenol 1982;139:820 –1. 10. Fossa SD, Abeler V, Marton PF, Lien HH, Hoie J. Sarcoid reaction of hilar and paratracheal lymph nodes in patients treated for testicular cancer. Cancer 1985;56:2212– 6. 11. Urbanski SJ, Alison RE, Jewett MA, Gospodarowicz MK, Sturgeon JF. Association of germ cell tumours of the testis and intrathoracic sarcoid-like lesions. Can Med Assoc J 1987; 137:416 –7. 12. Biglino A, Cariti G, Musset M, Gioannini P. Pulmonary sarcoidosis associated with Leydig cell testicular neoplasm. Chest 1988;94:428 –9. 13. Parr MJ, Williams MV. Sarcoidosis mimicking metastatic testicular tumour. Br J Radiol 1988;61:516 – 8. 14. Sieber PR, Duggan FE. Sarcoidosis and testicular tumors. Urology 1988;31:140 –1. 15. Droz JP, Ruffie P, Piot G, Ghosn M, Caillaud JM, Elias D. Sarcoidosis and testicular germ cell tumor. Case report. Scand J Urol Nephrol 1990;24:171–3. 16. Toner GC, Bosl GJ. “Sarcoidosis, sarcoid-like lymphadenopathy,” and testicular germ cell tumors. Am J Med 1990;89: 651– 6. 17. Looijen R, Hoekstra HJ, Sleijfer DT, Postmus P, Oosterhuis JW, de Boer WJ. Sarcoid reaction mimicking intrathoracic dissemination of testicular cancer. Cancer 1990;66:2221–3. 18. van Klaveren RJ, Festen J, Lacquet LK, Cox AL. Gegeneraliseerde sarcoidose in combinatie met een maligniteit bij 3 patienten; de waarde van mediastinoscopie. Ned Tijdschr Geneeskd 1992;136:533–5. 19. Leatham EW, Eeles R, Sheppard M, Moskovic E, Williams MP, Horwich A, et al. The association of germ cell tumours of the testis with sarcoid-like processes. Clin Oncol (R Coll Radiol) 1992;4:89 –95. 20. Colebunders R, Bultinck J, Servais J, Denis L. A patient with testis seminoma, sarcoidosis, and neutropenic enterocolitis. Hum Pathol 1984;15:394 – 6. 21. Geller RA, Kuremsky DA, Copeland JS, Stept R. Sarcoidosis and testicular neoplasm: an unusual association. J Urol 1977;118:487– 8. 22. O’Connell M, Powell S, Horwich A. Sarcoid-like lymphadenopathy in malignant teratoma. Postgrad Med J 1983;59: 108 –10. 23. Blacher EJ, Maynard JF. Seminoma and sarcoidosis: an unusual association. Urology 1985;26:288 –9. Sarcoidosis and Testicular Carcinoma/Rayson et al. 24. Heffner JE, Milam MG. Sarcoid-like hilar and mediastinal lymphadenopathy in a patient with metastatic testicular cancer. Cancer 1987;60:1545–7. 25. Bassoulet J, du Chatelard P, L’Her P, Natali F, Morel C, Cosnard G, et al. Cancer du testicule, sarcoidose et neurinome. Ann Urol (Paris) 1990;24:219 –23. 26. Van Hoef ME, Schornagel JH. Testicular cancer with enlarged mediastinal lymph nodes: a diagnostic pitfall. Neth J Med 1990;37:202– 6. 27. Kok TC, Haasjes JG, Splinter TA, ten Kate FJ. Sarcoid-like lymphadenopathy mimicking metastatic testicular cancer. Cancer 1991;68:1845–7. 28. Eng C. Thoracic adenopathy: metastatic seminoma or sarcoid? Hosp Pract (Off Ed) 1992;27:208 –10. 29. Mitchell DN, Scadding JG, Heard BE, Hinson KF. Sarcoidosis: histopathological definition and clinical diagnosis. J Clin Pathol 1977;30:395– 8. 30. Teirstein AS. Sarcoidosis, a ship in search of a sail. Lung Biol Health Dis 1994;73:745– 61. 31. James DG. Epidemiology of sarcoidosis. Sarcoidosis 1992; 9(2):79 – 87. 32. Siltzbach LE, James DG, Neville E, Turiaf J, Battesti JP, Sharma OP, et al. Course and prognosis of sarcoidosis around the world. Am J Med 1974;57:847–52. 33. Sartwell PE, Edwards LB. Epidemiology of sarcoidosis in the U.S. Navy. Am J Epidemiol 1974;99:250 –7. 34. Cooch JW. Sarcoidosis in the United States Army, 1952 through 1956. Am Rev Resp Dis 1961;84(5 Pt 2):103– 8. 35. Henke CE, Henke G, Elveback LR, Beard CM, Ballard DJ, Kurland LT. The epidemiology of sarcoidosis in Rochester, Minnesota: a population-based study of incidence and survival. Am J Epidemiol 1986;123:840 –5. 36. Hennessy TW, Ballard DJ, DeRemee RA, Chu CP, Melton LJ 37. 38. 39. 40. 41. 42. 43. 44. 45. 343 III. The influence of diagnostic access bias on the epidemiology of sarcoidosis: a population-based study in Rochester, Minnesota, 1935–1984. J Clin Epidemiol 1988;41:565–70. SEER Cancer Statistics Review; 1973–1990. Bethesda (MD): National Institutes of Health, National Cancer Institute, 1993. Swanson DA, Johnson DE. M. D. Anderson experience with surveillance for clinical stage I disease. In: Johnson DE, Logothetis CJ, Von Eschenbach AC, editors. Systemic therapy for genitourinary cancers. Chicago: Year Book Medical Publishing, 1989:304 –11. Duchesne GM, Horwich A, Dearnaley DP, Nicholls J, Jay G, Peckham MJ, et al. Orchidectomy alone for stage I seminoma of the testis. Cancer 1990;65:1115– 8. Patel SR, Richardson RL, Kvols L, Zincke H. Observation after orchiectomy in clinical stage I nonseminomatous germ cell tumor of testis. Mayo Clinic experience. Am J Clin Oncol 1990;13:379 – 81. Gels ME, Hoekstra HJ, Sleijfer DT, Marrink J, de Bruijn HW, Molenaar WM, et al. Detection of recurrence in patients with clinical stage I nonseminomatous testicular germ cell tumors and consequences for further follow-up: a singlecenter 10-year experience. J Clin Oncol 1995;13:1188 –94. Warde P, Gospodarowicz MK, Panzarella T, Catton CN, Sturgeon JF, Moore M, et al. Stage I testicular seminoma: results of adjuvant irradiation and surveillance. J Clin Oncol 1995; 13:2255– 62. Wingo PA, Tong T, Bolden S. Cancer statistics, 1995. CA Cancer J Clin 1995;45:8 –30. Berkson J. Limitations of the application of fourfold table analysis to hospital data. Biometrics 1946;2:47–53. Brincker H. Sarcoid reactions in malignant tumours. Cancer Treat Rev 1986;13:147–56.