GERM CELL TUMOR ASSOCIATED PRIMITIVE NEUROECTODERMAL TUMORS

GERM CELL TUMOR ASSOCIATED PRIMITIVE NEUROECTODERMALTUMORS

KRISTEN N. GANJOO, RICHARD S. FOSTER, HELEN MICHAEL, JOHN P. DONOHUE AND

LAWRENCE H. EINHORN*From the Division of Hematology-Oncology, Departments of Medicine, Urology and Pathology, Indiana University Medical Center,

Indianapolis, Indiana

ABSTRACT

Purpose: This retrospective review was done to assess the prognosis and response in patientspresenting with primitive neuroectodermal tumor admixed with germ cell tumor.

Materials and Methods: Of the 40 patients treated at our institution from 1984 to 1999, 15 hadinitial stage I and 25 had initial metastatic disease. Median followup after the diagnosis was 25months (range 4 to 142).

Results: Of the 40 patients 15 presented with clinical stage I disease, including 9 treated withretroperitoneal lymph node dissection and 6 who elected surveillance. Seven of the 9 patients hadnormal lymph nodes and all continuously had no evidence of disease. Two of the 9 patients had lymphnodes involved with teratoma with or without primitive neuroectodermal tumor. Retroperitonealrelapse in 5 of the 6 patients on surveillance was treated with cisplatin based chemotherapy followedby post-chemotherapy retroperitoneal lymph node dissection. Residual primitive neuroectodermaltumor was noted in 4 of the 5 patients and only 3 of 6 are currently without disease at a medianfollowup of 17 months (range 15 to 69). A total of 25 patients presented with metastatic disease, ofwhom 23 underwent cisplatin based chemotherapy. Only 3 patients achieved complete remissionwith chemotherapy alone and 2 of the 3 subsequently relapsed. Of the remaining 20 patients 16underwent post-chemotherapy retroperitoneal lymph node dissection, including 11 with primitiveneuroectodermal tumor in the resected specimen. Two of these 11 patients have continuously had noevidence of disease, while an additional 3 currently have no evidence of disease after further therapy.Teratoma was present in the resected specimen in 5 of 16 patients, of whom 2 have continuously hadno evidence of disease, while an additional 2 currently have no evidence of disease after furthersurgical resection. Therefore, 11 of 25 patients who presented with metastatic disease currently haveno evidence of disease at a median followup of 19 months (range 2 to 111).

Conclusions: Primitive neuroectodermal tumor in the orchiectomy specimen has adverse prognos-tic significance. This condition in the retroperitoneum is potentially curable by retroperitoneal lymphnode dissection but rarely eradicated by chemotherapy. Therefore, we recommend retroperitoneallymph node dissection for all clinical stage I cases with primitive neuroectodermal tumor in theorchiectomy specimen. Patients who present with metastatic primitive neuroectodermal tumorshould be treated aggressively with surgical resection as an integral part of the therapeutic strategy.

KEY WORDS: testis; lymph nodes; metastases; germinoma; neuroectodermal tumors, primitive

Germ cell tumors frequently consist of several cell types,including teratoma. Teratoma may have significant hetero-geneity.1 It may be biologically inert, grow by local extensionor undergo malignant transformation to nongerm cell malig-nancy.2 Teratoma is pluripotent and capable of transformingalong endodermal, ectodermal and mesodermal lines. Prim-itive neuroectodermal tumor develops from the malignanttransformation of teratoma along ectodermal lines, in con-trast to sarcoma, which is mesodermal in origin.3, 4 Primitiveneuroectodermal tumor belongs to a large group of lesionscalled small round blue cell tumors. However, primitive neu-roectodermal tumor arising from germ cell tumor is a distinctentity. It usually does not have the distinct chromosomaltranslocation t(11;22) that is normally present in small roundcell tumors. Some of these neoplasms have an abnormality ofthe isochromosome 12p, as in germ cell tumor, confirming the

germ cell origin of primitive neuroectodermal tumor.5 Prim-itive neuroectodermal tumor most likely arises from terato-matous elements of germ cell tumors.6–8 These tumors areaggressive and usually do not respond to chemotherapy.However, long-term disease-free status may be achieved bythe complete resection of all tumor.

A total of 1,949 retroperitoneal lymph node dissections wereperformed at our institution between 1984 and 1999. Also, wehave previously reported that 2.4% of all testicular germ celltumors treated at our institution contained primitive neuroec-todermal tumor.8 Therefore, this retrospective study representsa single institution series of a rare disease. The clinical outcomeis reported based on stage at presentation. The results of ther-apy and treatment options are also discussed.

PATIENTS AND METHODS

A total of 40 patients with primitive neuroectodermal tumorarising from testicular germ cell tumor were treated at ourinstitution between 1984 and 1999. All histological slides werereviewed at our facility. All patients had adequate staging withplain chest radiography or chest computerized tomography (CT)

Accepted for publication December 21, 2000.Presented at annual meeting of American Urological Association,

Atlanta, Georgia, May 1–6, 1999.Supported in part by the Walther Cancer Institute, Indianapolis,

Indiana.*Requests for reprints: Indiana Cancer Pavilion, Indiana Univer-

sity, 535 Barnhill Dr., RT 473, Indianapolis, Indiana 46202.

0022-5347/01/1655-1514/0THE JOURNAL OF UROLOGY® Vol. 165, 1514–1516, May 2001Copyright © 2001 by AMERICAN UROLOGICAL ASSOCIATION, INC.® Printed in U.S.A.

1514

as well as abdominal and pelvic CT. The majority of patientshad baseline tumor marker studies of b-human chorionic go-nadotropin (b-HCG) and a-fetoprotein (AFP).

Of the 40 patients 15 had clinical stage I disease at diag-nosis. All patients underwent radical orchiectomy followedby retroperitoneal lymph node dissection or surveillance. Iflymph node involvement with any type of germ cell tumorwas documented after pathological examination of the retro-peritoneal lymph node dissection specimen, patients weregiven the option of 2 cycles of adjuvant chemotherapy withetoposide and cisplatin with or without bleomycin. Patientswith negative lymph nodes received no further treatment. Inall cases close followup included a history, physical examina-tion, chest radiography and tumor markers every monthduring year 1, every 2 months during year 2 and once yearlyindefinitely thereafter. Patients who elected surveillance anddid not undergo retroperitoneal lymph node dissection re-ceived more frequent followups with abdominal and pelvicCT every 2 months during year 1 and every 4 months duringyear 2 in addition to a history, physical examination, chestradiography and tumor marker studies. If a patient hadrelapse on surveillance, cisplatin based chemotherapy wasgiven, followed by post-chemotherapy retroperitoneal lymphnode dissection with resection of all residual disease. If apatient had relapse after initial retroperitoneal lymph nodedissection, chemotherapy was administered, followed by sur-gical resection or resection alone.

Of the 40 patients 25 had metastatic disease at presentationwith retroperitoneal disease only, or pulmonary, hepatic or me-diastinal disease. The majority of patients underwent orchiec-tomy and subsequent cisplatin based chemotherapy. Generallyall those with residual tumor underwent post-chemotherapyretroperitoneal lymph node dissection and excision of all resid-ual disease unless disease was surgically unresectable.

RESULTS

The table shows patient characteristics. A total of 15 pa-tients with a median age of 31 years presented with clinicalstage I disease. Pre-orchiectomy tumor marker studies wereavailable in 12 of the 15 cases. Five of 12 patients hadelevated AFP, b-HCG or AFP and b-HCG. All patients un-derwent radical orchiectomy. Of the 15 patients 12 had prim-itive neuroectodermal tumor admixed with other germ celltumor elements in the orchiectomy specimen, while 3 withother germ cell tumor elements but no primitive neuroecto-dermal tumor in the orchiectomy specimen later had relapsewith primitive neuroectodermal tumor.

Nine of the 15 patients who presented with stage I diseaseunderwent post-orchiectomy retroperitoneal lymph node dis-section, including 7 with normal lymph nodes and 6 who werecontinuously without evidence of disease at a median fol-lowup of 26 months (range 4 to 144). One of 7 patients hadrelapse in the pre-sacral area 3 years after retroperitoneal

lymph node dissection and underwent the resection of tera-toma and yolk sac tumor. This patient had relapse again inthe iliac area and underwent chemotherapy and resection ofresidual primitive neuroectodermal tumor. After repeat re-lapse in the abdomen he received radiation therapy withouta response and died of disease.

Two of the 9 patients who underwent retroperitoneallymph node dissection had metastatic tumor, including 1with teratoma only who elected adjuvant chemotherapy andremained without evidence of disease at a followup of 60months. The other patient with teratoma and primitive neu-roectodermal tumor in the lymph nodes required repeat ret-roperitoneal lymph node dissection since he had relapse inthe abdomen 5 months after initial surgery. There was re-lapse again 17 months later in the mesentery and sarcomaresection was done. He currently has no evidence of disease 1month after the last surgical procedure.

Six of the 15 patients who presented with stage I diseaseelected surveillance, including 3 with primitive neuroectoder-mal tumor and teratoma in the orchiectomy specimen, and 3with other germ cell tumor elements. Five of the 6 patients hadrelapse in the retroperitoneum at a median time to relapse of 10months (range 4 to 11). All patients received cisplatin basedchemotherapy and underwent post-chemotherapy retroperito-neal lymph node dissection. In the resected specimen there wasprimitive neuroectodermal tumor in 3 cases, astrocytoma in 1and fibrosis in 1. Two of the 5 patients with relapse aftersurveillance remain without evidence of disease. Three of the 6patients with clinical stage I disease who were on surveillancedied of progressive primitive neuroectodermal tumor, of whom2 had previously had primitive neuroectodermal tumor in theretroperitoneal lymph node dissection specimen and 1 had hadastrocytoma.

There was metastatic disease in 25 cases at presentation,including 18 of metastatic primitive neuroectodermal tumorand 7 of other germ cell tumor elements at the metastatic sitesbut primitive neuroectodermal tumor in the orchiectomy spec-imen (see table). The most common metastatic site was theretroperitoneum. All except 2 patients underwent radical orchi-ectomy at the time of diagnosis. In 12 patients without primi-tive neuroectodermal tumor in the orchiectomy specimen thatlesion was identified in the subsequent surgical pathologicalspecimens. Two patients had an elevated tumor marker afterorchiectomy as the only sign of disease and each underwentprimary retroperitoneal lymph node dissection. One patientwith primitive neuroectodermal tumor in the lymph node spec-imen and multiple relapses did not respond to further therapyand died of disease. In the other patient with normal lymphnodes tumor markers normalized after retroperitoneal lymphnode dissection and he remains without evidence of disease at afollowup of 26 months.

The remaining 23 patients underwent cisplatin based che-motherapy, of whom 3 achieved a complete remission after

Pt. Characteristics Clinical Stage I Disease Metastasis

No. pts. 15 25Median age (range) 31 (20–45) 26 (16–37)Median pre-orchiectomy tumor markers at diagnosis (range):

AFP (ng./ml.) 180 (less than 20–2,282) 1,380 (less than 20–21,838)b-HCG (mIU/ml.) 9 (less than 1.56–15) less than 20 (less than 1.56–31,120)

No. orchiectomy pathology:Primitive neuroectodermal tumor alone 0 1Primitive neuroectodermal tumor with other germ cell elements 12 12Other germ cell elements without primitive neuroectodermal tumor 3 10Fibrosis 0 2

No. metastatic disease sites at initial presentation:Elevated markers 2Retroperitoneum alone 11Mediastinum 2Liver 3Lung 6Bone 1

GERM CELL TUMOR ASSOCIATED PRIMITIVE NEUROECTODERMAL TUMORS 1515

chemotherapy alone and 1 remains with no evidence of dis-ease. The latter patient had an epidural mass, and bonemarrow biopsy positive for teratoma and yolk sac tumor. Heremains in complete remission without any other interven-tion 14 months after diagnosis. The remaining 2 patients hadprimitive neuroectodermal tumor relapse and died of disease.

Of the remaining 20 patients who underwent post-chemotherapy retroperitoneal lymph node dissection 1 refusedfurther treatment and 3 had unresectable disease (multipleliver metastases). Four of the 16 patients also underwent pul-monary wedge resection, lobectomy, posterior mediastinal re-section or liver resection at post-chemotherapy retroperitoneallymph node dissection. There was primitive neuroectodermaltumor in the resected specimen in 11 cases, including 8 withrelapse at a median of 3 months (range 1 to 9). The remaining3 patients currently have no evidence of disease at a medianfollowup of 23 months (range 14 to 38). Relapse involved anabdominal mass in 4 cases, a liver mass in 2, pulmonary me-tastasis in 1 and an orbital mass in 1. Two of 8 patients cur-rently have no evidence of disease, of whom 1 each underwenthigh dose chemotherapy and hepatic lobectomy, and has a fol-lowup of 5 and 19 months, respectively. Five of 16 patients whounderwent post-chemotherapy retroperitoneal lymph node dis-section had teratoma but no primitive neuroectodermal tumorin the resected specimen, including 2 who had relapse withteratoma and primitive neuroectodermal tumor, and 1 whocurrently has no evidence of disease after repeat retroperitoneallymph node dissection. Therefore, 10 of the 16 patients whounderwent post-chemotherapy retroperitoneal lymph node dis-section currently have no evidence of disease, while only 11 of25 (44%) who presented with metastatic disease have no evi-dence of disease at a median followup of 19 months (range 2 to111). There was no primitive neuroectodermal tumor in theoriginal orchiectomy specimen in 15 patients who later hadmetastatic primitive neuroectodermal tumor. The most com-mon germ cell tumor element in the testis of these patients wasyolk sac tumor and mature or immature teratoma.

DISCUSSION

Primitive neuroectodermal tumor arising from germ cell tumorsis a distinct entity. It most likely arises from the teratomatouselements of germ cell tumors.7,8 These primitive neuroectodermaltumors respond poorly to chemotherapy and may only be cured bycomplete surgical resection. They are also more aggressive thantesticular germ cell tumors that do not contain primitive neuroec-todermal tumor. In patients presenting with clinical stage I testic-ular germ cell tumor the cure rate is 95% to 100%. Approximately30% of patients have relapse after radical orchiectomy alone butmost are cured by subsequent cisplatin based chemotherapy.Therefore, patients who choose surveillance over nerve sparingretroperitoneal lymph node dissection have an excellent outcome,provided that they are compliant with followup.9,10 However, the15 patients with testicular germ cell tumor containing primitiveneuroectodermal tumor who presented with clinical stage I diseasein our series had a less favorable outcome since only 11 (73%) hadno evidence of disease. Those who underwent retroperitoneallymph node dissection initially had a better outcome than thosefollowed with surveillance with no evidence of disease in 89% and50%, respectively. Therefore, we recommend that all patients withclinical stage I disease and primitive neuroectodermal tumor inthe orchiectomy specimen should initially undergo retroperitoneallymph node dissection. Patients with relapse on surveillance havea less favorable outcome and are not routinely cured by subse-quent therapy.

Generally patients with metastatic germ cell tumor have a75% cure rate after cisplatin based chemotherapy.11 How-

ever, in our series we noted only 44% long-term survival.This decreased survival was mainly due to unresectable dis-ease, a poor response to chemotherapy and multiple relapses.Therefore, patients who present with metastatic germ celltumor containing primitive neuroectodermal tumor shouldbe treated with initial cisplatin based chemotherapy for thegerm cell tumor elements, followed by the resection of allresidual disease. Evidence of progressive disease while onchemotherapy should lead to early surgical resection, fol-lowed by adjuvant chemotherapy if residual germ cell tumorelements are present in the resected specimen.

Special surgical considerations exist in these primitive neu-roectodermal tumor cases. Teratoma and germ cell elementsrarely recur locally in the retroperitoneum after thorough ret-roperitoneal lymph node dissection. However, even after thecomplete resection of all visible and palpable primitive neuro-ectodermal tumor in the retroperitoneum there is a higherprobability of local recurrence. Most likely microscopic primi-tive neuroectodermal tumor remains in some patients afterretroperitoneal lymph node dissection. Therefore, continuedsurveillance after retroperitoneal lymph node dissection ismandatory since chemotherapy is ineffective and repeat resec-tion may prove therapeutic. Similarly surgeons performingsuch procedures should be willing to alter surgical techniquesintraoperatively to resect aggressively and widely, incorporat-ing vena cava or bowel resection liberally as needed and per-forming aortic resection and replacement as necessary.

CONCLUSIONS

Primitive neuroectodermal tumor represents a highly ma-lignant derivative of germ cell tumors. Standard chemother-apy is largely ineffective for these elements and surgery isthe mainstay of therapy.

REFERENCES

1. Loehrer, P. J., Sledge, G. W., Jr. and Einhorn, L. H.: Heteroge-neity among germ cell tumors of the testis. Semin Oncol, 12:304, 1985

2. Ulbright, T. M., Loehrer, P. J., Roth, L. M. et al: The developmentof non-germ cell malignancies within germ cell tumors. A clini-copathologic study of 11 cases. Cancer, 54: 1824, 1833, 1984

3. Comiter, C. V., Kibel, A. S., Richie, J. P. et al: Prognostic featuresof teratomas with malignant transformation: a clinicopatho-logical study of 21 cases. J Urol, 159: 859, 1998

4. Ahmed, T., Bosl, G. J. and Hajdu, S. I.: Teratoma with malignanttransformation in germ cell tumors in men. Cancer, 56: 860, 1985

5. Motzer, R. J., Amsterdam, A., Prieto, V. et al: Teratoma withmalignant transformation: diverse malignant histologies aris-ing in men with germ cell tumors. J Urol, 159: 133, 1998

6. Little, J. S., Foster, R., Ulbright, T. M. et al: Unusual neoplasmsdetected in testis cancer patients undergoing post-chemotherapy retroperitoneal lymphadectomy. J Urol, 152:1144, 1994

7. Michael, H.: Nongerm cell tumors arising in patients with tes-ticular germ cell tumors. J Urol Pathol, 9: 39, 1998

8. Michael, H., Hull, M. T., Ulbright, T. M. et al: Primitive neuro-ectodermal tumors arising in testicular germ cell neoplasms.Am J Surg Pathol, 21: 896, 1997

9. Hoskins, P., Dilly, S., Easton, D. et al: Prognostic factors in stageI non-seminomatous germ-cell testicular tumors managed byorchiectomy and surveillance: implications for adjuvant che-motherapy. J Clin Oncol 4: 1031, 1986

10. Dunphy, C. H., Ayala, A. G., Swanson, D. A. et al: Clinical stageI nonseminomatous and mixed germ-cell tumors of the testis:a clinicopathologic study of 93 patients on a surveillance pro-tocol after orchiectomy alone. Cancer 62: 1202, 1988

11. Bosl, G. J. and Motzer, R. J.: Testicular germ-cell cancer. N EnglJ Med 337: 242, 1997

GERM CELL TUMOR ASSOCIATED PRIMITIVE NEUROECTODERMAL TUMORS1516