Postchemotherapy resections of residual masses from metastatic non-seminomatous testicular germ cell tumors*

Annals of Oncology 8: 531-538, 1997. © 1997 Kluwer Academic Publishers. Printed in the Netherlands.

Original article Postchemotherapy resections of residual masses from metastatic non-seminomatous testicular germ cell tumors* I T . Hartmann,1 H.-J. Schmoll,2 M. A. Kuczyk,3 M. Candelaria1 & C Bokemeyer1 1 Department ofHemalology/Oncology/Immunology. UKL- Medical Center 11, Eberhard-Karls-University. Tubingen, 2Department of Hemalology/Oncology. Martin-Luther-Unirerstty, Halle/Wittenberg, 3Department of Urology, University Medical School, Hannover, Germany

Summary Purpose: To analyse the frequencies of histological findings, predictive factors for the presence of undifferentiated tumor and variables influencing the survival of patients with nonseminomatous germ cell tumors who underwent secondary resection of residual masses after cisplatin-based combination chemotherapy. Patients and methods: 134 patients with a median age of 26 years (15^47) undergoing at least one surgical intervention at Hannover University Medical School were included. One hundred nine patients had received first-line chemotherapy and 25 underwent surgery after second-line chemotherapy. Results: After first-line chemotherapy the distribution of histological findings was 52% necrosis, 27% differentiated teratoma and 21% undifferentiated tumor for 82 patients with marker negative PR (PRm-). Incompletely resected mass and failure to achieved complete tumor marker normalisation were significantly associated with the finding of undifferentiated tumor. Five-year progression-free survival rates according to histological findings were 78%, 67% and 66% for necrosis, differentiated teratoma and undifferentiated tumor. Patients with undifferentiated tumor in the resected specimen routinely received postoperative additional chemotherapy. Factors asso-

ciated with a worse overall survival were progressive disease within three months, persistent AFP elevation prior to surgery, prechemotherapy elevated LDH levels or mediastinal lymph node involvement at primary diagnosis. In 8 of 27 patients (30%) undergoing multiple resections at different sites a dissimilar histology was found. In the 25 patients operated after salvage chemotherapy undifferentiated tumor was found in 80%. A five-year survival of 44% compared to 80% after firstline chemotherapy was achieved. Conclusions: Resection of residual tumors after first-line chemotherapy remains essential in the treatment of metastatic testicular cancer. Undifferentiated tumor may still be found in 20% of patients despite achieving PRm- after first-line chemotherapy. Necrosis is found in only 50% of marker normalized patients after first-line and approximately 30% after secondline chemotherapy. Future studies have to prove whether the combination of clinical prognostic factors and the use of PETscanning will allow to spare subsets of patients from secondary resection. Key words: metastastic non-seminomatous testicular cancer, postchemotherapy residual mass, secondary surgical intervention, outcome analysis, predictive factors for histological findings

Following chemotherapy, the resected residual masses may histologically contain either necrosis, differentiated Metastatic testicular cancer has become a model for a teratoma or even undifferentiated tumor. For patients curable neoplasm. With the introduction of cisplatin- with undifferentiated tumor, the secondary resection based combination chemotherapy, approximately 50% may prolong survival depending on the biological agto 80% of patients presenting with disseminated non- gressiveness of the disease, the completeness of surgery seminomatous germ cell tumors will achieve long-term and the effectiveness of further additive treatment [5]. survival [1]. Surgical resection of residual tumor masses The resection of differentiated teratoma may prevent is an essential part of the treatment [2]. In most patients, local tumor growth and late disease recurrence with residual masses are located in the retroperitoneal space possible malignant transformation [6-8]. In case of and the resection is performed as a retroperitoneal necrosis, patients will have no improvement of their lymph node dissection or lumpectomy. Additionally, prognosis by the operation beyond the diagnosis of a residual pulmonary metastases will be considered for pathologic CR. Identification of those patients prior to postchemotherapy resection. These surgical procedures the surgical intervention would be of benefit in order to may be associated with long-term morbidity in some prevent surgery-related complications [9-13]. Only the patients, even if improved techniques are used [3, 4]. excision of all residual tumors provides a definitive histological diagnosis [15, 16]. The current analysis investigates the outcome of patients with metastatic non-seminoma* Presented in part at the 21st ESMO-meeting, Vienna, Austria, Notous germ cell tumors who underwent resections of vember 1-5, 1996. Introduction

532 Table 1 Characteristics of patients undergoing postchemotherapy resection of residual masses from metastatic testicular cancer after cisplatin-based combination chemotherapy between 1978-1995 at diag-

Table 2 Treatment of patients undergoing secondary resection for residual masses from metastatic testicular cancer. Treatment group

Treatment group

Median age (yrs) [range] Histology of primary tumor Malignant teratoma undifferentiated Malignant teratoma trophoblastic Malignant teratoma intermediate Teratoma differentiated Missing (AFP positive) Location of the primary tumor Gonadal Extragonadal (retroperitoneal) Unknown Indiana University Classification Minimal Moderate Advanced Tumor marker elevation prior to chemotherapy HCG AFP both markers Localisation of metastases prior to chemotherapy Retroperitoneal Lungs Liver Mediastinal/supraclavicular lymph nodes Other sites

first-line (n = 109)

102 6 1

secondline n pts (%)

50 (46%)

6 (20%)

26 (24%)

12(48%)

22 (20%)

6 (24%)

9 (8%)

1 (4%)

second-line (" = 25)

25(15-44] 28(20-47] 57(53%) 23(21%) 21(19%) 7 (6%) 1(1%)

first-line n pts (%)

14(56%) 4(16%) 6(24%) 1(4%) 24 (96%) 1 (4%)

15(14%) 35 (32%) 59(54%)

1(4%) 8 (32%) 16(64%)

53(49%) 49 (45%) 37(34%)

12(48%) 11 (44%) 10(40%)

95 (87%) 24 (96%) 55 (50%) 17(68%) 11 (10%) 2 (8%) 20(18%) 3(12%) 10(9%) 2 (8%)

residual masses at a single or at multiple localisations after cisplatin-based first-line or second-line chemotherapy for metastatic disease. Patients and methods Between 1978 and 1995, 134 patients underwent secondary resection after cisplatin-based combination chemotherapy from metastatic nonseminomatous germ cell tumors at Hannover University Medical School. One hundred nine (81%) of them were operated after first-line chemotherapy and 25 (19%) after salvage chemotherapy for relapsed disease after previous CR. The median age at diagnosis was 25 years for the patients receiving first-line chemotherapy (range 15-44) and 28 years (range 20-47) for patients receiving second-line chemotherapy. One hundred twentysix patients had tumors of testicular origin, six had retroperitoneal extragonadal germ cell tumors, whereas no patients with primary mediastinal tumors were included. In two cases no primary tumor were found, but a considerable elevation of serum J3-HCG and AFP indicated the presence of a NSGCT. Histology was classified according to the British Tumor Panel [17]. The staging investigations prior to and after completion of chemotherapy consisted of physical examination, laboratory testing including serum tumor markers (5-HCG, AFP and LDH, chest X-ray and abdominal/thoracic CT scans and of all other initially involved sites. Patients were classified according to the Indiana University Classification [18]. Characteristics including histology of the primary tumor, number of metastatic sites, extent of disease and tumor markers for both patients undergoing surgery after firstand second-line chemotherapy are summarized in Table 1. All patients had received cisplatin containing standard-dose combination chemotherapy. Details of the different treatment protocols are

Type of chemotherapy PEB Cisplatin Etoposide Bleomycin Cisplatin PVB Vinblastine Bleomycin PEI Cisplatin Etoposide Ifosfamide PEBOI Cisplatin Etoposide Bleomycin

(20mg/m 2 ,d 1-5) (100mg/m 2 ;d 1-5) (30mg;d2, 9, 16) (20mg/m 2 ;d 1-5) (0.2mg/kg;d 1,2) (30mg;d2, 9, 16) (20mg/m 2 ;d 1-5) (75mg/m 2 ,d 1-5) (1.2g/m 2 ,d 1-5) (50mg/m 2 ,d 1-3) (170mg/m 2 ;dl-3) (15mg/nr;d 1,8,

15,22) Vincristine (2mg;d 1,8, 15,22) (5g/m 2 ;d 15) Ifosfamide Type of surgical intervention Retroperitoneal lumpectomy 92 (84%) Thoracotomy 36 (33%) Laminectomy 2 (2%) Resection of hepatic lesions 4 (4%) Cervical lymphadenectomy 3 (3%) Craniotomy 1 (1%)

18(72%) 12(48%) 1 (4%) 2 (8%)

summarized in Table 2 [19-22]. None of the included patients was treated with radiation therapy. Complete remission (CR) was defined as a complete disappearance of all clinical, radiological and biochemical evidence of disease. A partial response with marker normalisation (PRm-) was defined as any decrease in tumor size and normalisation of the tumor markers. PR without marker normalisation (PRm+) was defined as a radiological response > 50% in size but no complete normalisation of p^HCG and/or AFP at the end of chemotherapy. Progressive disease was defined as either residual lesions increasing in size or occurrence of new lesions and/or elevation of tumor markers at repeated controls. Size criteria for performing a retroperitoneal intervention was a residual mass > 1 cm and at thoracic sites > 0.5-1 cm in diameter at CTscan. To restrict the extent of the surgical intervention, retroperitoneal operations were routinely performed as a lumpectomy. In case of undifferentiated tumor, all lymph nodes even without enlargement were resected in this localisation (formal RPLND). Immunohistochemical tumor marker stains were used in some patients if undifferentiated tumor was suspected. One hundred thirty-eight surgical interventions were performed in 109 patients after first-line chemotherapy, 27 patients were operated at two localisations and two patients at three sites. The type of surgical intervention is outlined in Table 2. Surgery was carried out within three months after the end of chemotherapy. Multiple surgical interventions were performed within a 3-6-week interval, except for five patients with one-stage operations. The histological findings of postchemotherapy surgery were classified as necrosis, differentiated teratoma or undifferentiated tumor. In case of undifferentiated tumor, patients were scheduled to receive two additional cycles of cisplatin/ifosfamide-based chemotherapy, mainly the VIP-regimen [21]. Complete resection (R0) was defined as excision of all residual tumors with clear surgical margins. Patients were classifed as incompletely resected by evidence of a makroscopically visuable residual mass (R2) or if the surgical margins were not clear in histological examination (Rl). During the first two years, follow-up examinations were performed three-monthly. The interval was thereafter prolonged to six months in the third year and to annually from the fourth year on. For all living patients, the current status as of August 1995 was gained. The KaplanMeier method [23] was used to determine the overall and relapse-free

533 survival distributions assessed from the date of the last surgical intervention. The overall survival calculation used death due to any reason as the endpoint. In the first-line group the relapse-free period ended with surgery-related death in one case and treatment/disease unrelated death in another patient (travel accident). Apart from the results of the histological examination at postchemotherapy resection, the extent of disease at primary diagnosis, localisation of the tumor residuals prior to surgery, pre- and postchemotherapy tumor marker status, time interval to relapse and the completeness of resection were examined for their impact on survival in univariate analysis. Comparisons were made using the log-rank test. Cox regression was applied to model the simultaneous effects of several variables. Significance for entry was calculated from a likelihood ratio (LR) statistic [24]. A forward stepwise selection method was used with P < 0 05 as an entry criterion. The calculated odds ratio (OR) may be interpreted as relative risk. Multivariate logistic regression analyis was performed to estimate the probability of undifferentiated tumor to be found at postchemotherapy resection based on the combination of preoperative variables The multivariate analysis included all factors with P < 0 05 identified in preceeding univariate analysis [25]. All statistical analyses were performed using SPSS for Windows 6.0 software (Microsoft Corp., Redmond, WA, USA).

Results /. Resections after first-line chemotherapy (n pts —109) 1. Histological findings At the time of surgical intervention 82 patients (75%) had partial remissions with marker normalisation, 11 patients were PRm+ (10%) and for 12 patients (11%) progressive disease without marker normalisation was documentated. Four patients had a partial remission and unknown marker status. Patients with progressive disease and increasing tumor markers after chemotherapy included in this analysis underwent 'desperation surgery' for lack of a effective salvage chemotherapy in the years before 1988. Necrosis, differentiated teratoma and undifferentiated tumor were found in 46%, 22% and 32% of all patients. In the subgroup of 82 patients with P R m - the distribution was 52% necrosis, 27% differentiated teratoma and 21% undifferentiated tumor. Among 11 patients with PRm+, three of seven patients showing a decrease of tumor markers but failed to completely normalize prior to operation (marker slope) had undifferentiated carcinoma, whereas in all four remaining patients with rising tumor markers shortly after chemotherapy undifferentiated tumor was found. Of 12 patients classified as progressive disease, eight (67%) had undifferentiated tumor; differentiated teratoma was found in two (17%) and necrosis in the other two patients (17%). A clear correlation between response status after chemotherapy and the histological results was found (P = 0.001). The estimated five-year progression-free survival rates for all 109 patients according to histological findings at postchemotherapy resection were similar with 78% for necrosis, 67% for differentiated teratoma and 66% for undifferentiated tumor. In 21% of retroperitoneal resections mature teratoma was found (n - 92 operations) compared to 33% at thoractomy (n - 36 operations). There was a tendency for undifferentiated carcinoma to

be found more often in the retroperitoneal area (33%) compared to the lungs (17%) (P = 0.06). 2. Resection grade and histological findings In 82 patients (75%) complete resection of residual tumors was achieved, while in 24 cases resections were incomplete (n.e. = 3 pts). In patients with multiple resections the worst result per patient was taken into account. No difference was found in obtaining a complete surgical excision with respect to different tumor localisations, particularly comparing the retroperitoneal space and the lungs (76% versus 71%). Patients with 'advanced disease' at initial diagnosis more frequently received only incomplete resections (P = 0.04). Incomplete resection was statistically associated with the persistence of undifferentiated tumor since 22 of 24 patients with incomplete resections had undifferentiated tumors (P — 0.002). Remarkably, these patients had only a slightly decreased two-year survival rate of 83% following surgery plus additive chemotherapy. The two-year survival rate for patients with resection of necrosis or mature teratoma were 93% and 90%, respectively. In 13 of 25 patients (52%) who relapsed, the recurrence occurred at the initial resection area, two of them (8%) with evidence of new metastases at other sites. In 12 patients (48%), the relapse was located only outside of the initially resected area. Eight of the 13 patients (62%) who relapsed after postchemotherapy surgery at the previously operated localisation have had a complete resection. 3. Predictive factors for histological findings Neither elevated tumor markers at initial diagnosis, extent of disease, number and localisation of the metastases nor the histology of the primary tumor at diagnosis - especially presence of teratomatous elements - did influence the histological findings at secondary resection. In multivariate logistic regression analysis incompleteness of surgery was associated with a high likelihood for the persistence of undifferentiated tumor (odds ratio = 4.6) (Table 3). Decreased but not normalized values for (3-HCG and AFP possessed a higher relative risk for undifferentiated tumor (OR = 1.39). Patients with progressive disease or rising tumor markers (n - 16) were excluded from this analysis because, as mentioned above, remission status after chemotherapy was highly predictive for the finding of undifferentiated tumor. 4. Survival At the end of the study period, 92 of 109 patients (84%) were alive, 78 (72%) of them without evidence of disease and 14 patients (12%) with active disease. The median observation time was 35 months (range 1-169). The estimated five-year overall and relapse-free survival rates were 80% and 66%, respectively. Fifteen patients had died of metastatic testicular cancer during the observation period. Surgery-related mortality was 1%. Early relapse within three months after secondary resection, persisting AFP, elevated LDH at diagnosis or initial

534 Table 3. Predictive factors the histological findings at secondary resection after first-line chemotherapy

Table 4. Prognostic factors for overall survival in patients with resection of residual tumor masses after first-line chemotherapy.

Univanate analysis to discriminate undiff. vs. diff. teratoma or necrosis

Univariate analysis

Incomplete resection grade Pre-surgery elevated both AFP and p-HCG values Pre-surgery elevated AFP values Pre-surgery elevated LDH values Pre-surgery elevated p-HCG values Not -

P< 0.0001 P = 0.001 P = 0.005 P = 0.03 P = 0.04

significant Histology of primary tumor Elevated tumor markers at diagnosis Elevated LDH values at diagnosis Extent of disease Localisation of residual mass (thoracic vs. abdominal)

-

Log rank

Initial mediastinal lymph node involvement Pre-chemotherapy elevated LDH levels Pre-surgery elevated AFP levels Early recurrence (< 3 months)

P P P P

= = = =

0.02 0.01 0.005 0.0001

Not significant - Histology of primary tumor - Grade of resection - Age at diagnosis - Histology at resection - Pre-surgery B-HCG levels - Pre-chemotherapy tumor marker - Number of surgical interventions - Initial extent of metastatic disease

Logistic regression analysis Predictive factors for undifferentiated tumor

Signif. of LogLR

df

Pre-surgery elevated AFP and P-HCG values

0.049

1

Factors associated with undifferentiated tumor

Signif of LogLR

Incomplete resection grade (RI/RII)

0.0000

Odds ratio (OR)

Multivariate Cox regression analysis Variable

Sig.

df

OR

(95% CI)

1.39

Early recurrence ( < 3 months)

0.00001

1

19.2

(5.8-63.5)

df

Odds ratio (OR)

N = 96 pts included in multivariate analysis.

1

4.59

N = 89 pts included (pts with progressive disease or rising tumor markers were excluded from this analysis).

massive mediastinal lymph nodes involvement at primary diagnosis were identified as negative prognostic factors for survival (Table 4). Patients with residual masses at several localisations undergoing multiple resections had an estimated five-year survival of 68% in contrast to 82% for patients operated only at one site (« = 82) (/> = 0.14). 5. Patients with resections at multiple sites Necrotic tissue was found in 18 (67%), differentiated teratoma in four (15%) and undifferentiated tumor in five patients (19%) at the first site, and in 18 (67%), one (4%) and eight patients (30%) at the second surgically resected site, respectively. Two patients underwent a third surgical intervention yielding undifferentiated tumor and necrosis. In total, 10 patients (37%) had undifferentiated tumor in at least one resected specimen. A correlation between the completeness of resection and the histological finding was detectable, since only incomplete resections were possible in 9 of 12 patients revealing undifferentiated tumor (P - 0.006). Eight of 27 patients (30%) showed dissimilar histological findings at the different sites. Among 18 patients who underwent both excision of retroperitoneal and thoracic residuals, a less favorable finding was found in two patients at the retroperitoneal site and in one at thoracotomy. Considering all surgical interventions, necrosis was more frequently found at thoracotomy (n = 15/21) compared to retroperitoneal surgery (n = 17/27) (P = 0.05).

Of two patients who underwent bilateral thoracotomy, one showed necrosis at both sites and the other had dissimilar findings with differentiated teratoma and undifferentiated tumor. All resections of hepatic lesions revealed necrotic tissue (n = 4). The histology was identical at three retroperitoneal resections and but once undifferentiated carcinoma was found. //. Resections after second-line chemotherapy (n pts = 25) Twenty-five patients underwent postchemotherapy resections after second-line cisplatin-based combination chemotherapy. Only 40% had reached a normalisation of elevated P-HCG and/or AFP after second-line chemotherapy prior to the operation. In 14 patients, complete resection of residual masses was possible, among them all patients with marker normalisation. Twenty patients (80%) had undifferentiated tumors in their resected specimens. The distribution of histological findings in the remaining five patients was necrosis in four (16%) and differentiated teratoma in one (4%). Comparing the survival data of these patients to those operated after first-line chemotherapy, the prognosis was significantly worse with a calculated five-year survival of 44% (P < 0.0001). In case of undifferentiated tumor in the specimen, further chemotherapy seemed to have no influence on survival since 13 patients with post-surgery chemotherapy and seven without had a comparable five-year survival (38% versus 43%). Complete resection of undifferentiated tumor resulted in a five-year survival of 50% versus 30% in cases with incomplete resections (n.s.). Of nine patients with completely resected undifferentiated carcinoma, additionally receiving postopera-

535 Table 5 Literature reports of overall and disease-free survival from patients with secondary resection after 1st line chemotherapy for metastatic testicular cancer. Investigator

n (pts)

Disease/progression -free survival (%)t

Overall survival (%)

Follow-up (months)

nee

td

ut

all

nee

td

ut

all

95

58 -

89 72

92 93

86 92

58 27

94 78

55(1-102) 36 (5-96)

56

85 80 -

93 -

88 -

41 56

80 87 -

60' 60" 36(24-60) 36

Fossa, 1989 [35] Mulders, 1990 [36]

101 55

92 -

Hendry, 1993 [12] Steyerberg, 1993 [38] Aprikian, 1994 [3] Fox, 1993 [6]

231 86 40 43

-

Gerl, 1995 [37]

111

77

80

62

77

85

83

77

83

60"

Current series

109

78

67

66

72

90

83

77

84

60°

Comments

Incl. 15 pts with seminoma, 10 pts without complete marker normalisation

Follow-up of surviving pts, postoperative ctx (DDP based) = 74%-DFS (27 pts) Postoperative ctx (DDP/Ifo/Vbl or Eto) in case of ut Postoperative ctx (DDP/Ifo based) in case of ut

Abbreviations: - = not available; nee = necrosis, td = differentiated teratoma; ut = undifferentiated tumor; ctx = chemotherapy; DDP = cisplatin, Ifo = ifosfamide, Eto = etoposide, Vbl = vinblastin. " Estimated five-years survival [23].

tive chemotherapy, four remained disease-free at two years.

tiated tumor from teratoma, a higher prechemotherapy LDH level, a larger postchemotherapy mass and a smaller relative shrinkage of the mass were of value, based on 299 patients in the second analysis. These Discussion models were useful to predict necrosis, but were less valuable for the distinction of undifferentiated tumor Postchemotherapy resection is an established part of the from mature teratoma. treatment in metastatic non-seminomatous germ cell The proposed model may be helpful for the decision tumors if residual tumor masses are radiologically de- to operate patients immediately or to follow their residtectable. About 15% to 20% of those abnormalities will ual tumors closely, dependent on the calculated risk histologically contain undifferentiated tumor, 45% to score. However, there is a high incidence of dissimilar 50% necrotic tissue and 30% to 40% mature teratoma, histological findings in resected masses from different as recently confirmed in a report including nearly a localisations in the same individuum, as demonstrated thousand patients from different institutions [25]. In our in patients requiring more than one resection. Different series either undifferentiated tumor or differentiated histological results were reported ranging between teratoma were found in 54% of patients. In case of either 25%-47% in five studies and in our series 8 of 27 necrosis or differentiated teratoma, no further treatment patients (30%) had dissimilar histological findings is necessary. It would be useful to identify patients har- [30-34]. A possible explanation might be that the sysbouring necrosis in residual masses prior to the opera- temic chemotherapy may have different effects on each tion in order to prevent surgery-related morbidity and metastatic subpopulation, depending on pharmacologimortality. On the other hand, surgical resection of differ- cal characteristics of the applied drugs, the size of entiated teratoma is useful in order to prevent recurrent the metastases, blood delivery of the affected organ teratoma or a subsequent malignant transformation and local growth conditions. Since the pretreatment [26-28]. Undifferentiated tumor in the resected speci- characteristics included in the calculation of Steyermens may indicate a poor prognosis and additional berg et al. were certainly identical within one single chemotherapy is often used in these patients. patient, the observed variation of histological findings Recently, Steyerberg et al. performed a meta-analysis among different sites demonstrates that the prediction based on 556 patients with non-seminomatous germ cell remains difficult. However, relative shrinkage of the tumors operated after first-line chemotherapy in order masses was not evaluated in our patients and in all other to develop a statistical model for the prediction of the series. histology of residual masses [29]. Independent predictive Survival dala of a total of 776 patients with metastatic factors for necrosis were the absence of teratomatous non-seminomatous germ cell tumors from eight larger elements in the primary tumor, prechemotherapy nor- series undergoing surgical resections after first-line cismal AFP/p-HCG values, elevated LDH, small pre- or platin-based chemotherapy is summarised in Table 5. postchemotherapy masses or a shrinkage of the masses The data reported, demonstrate that between 72%-89% < 70% during chemotherapy. To distinguish undifferen- of patients will remain disease-free at three years after

536 surgery. The finding of necrosis and/or differentiated teratoma in the resected specimens resulted in a 78%92% and 67%—95% long-term survival rate, respectively [3, 6, 12, 35-38]. In 1981, Einhorn et al. reported that patients with undifferentiated tumor at postchemotherapy resection have a poor prognosis with a long-term survival in only 2 of 22 patients. Four of the 22 patients had received cisplatin-based additive chemotherapy after surgery and both long-term survivors were in that group [9]. In the meantime, other investigations have confirmed that the use of cisplatin-based additive chemotherapy may improve the outcome of patients after the resection of vital carcinoma with 56% to 70% of patients reported to be disease-free at three years [6, 11, 37, 39]. Three investigations demonstrated clear differences for disease-free survival between patients receiving postoperative chemotherapy or no further treatment. In Tait's investigation among 12 patients treated postoperatively, six remained disease-free, in contrast to one of four patients without additive chemotherapy [11]. Fox reported an overall disease-free survival of 56% in 43 patients (follow-up: 36 months), while the 27 patients with postoperative cisplatin-based chemotherapy had a disease-free survival of 74% [6]. Donohue et al. from the same institution found 19 of 27 patients (70%) being continuously disease-free with postoperative chemotherapy in contrast to none of seven patients without [39]. In our series, a five-year disease-free survival of 66% was calculated with additive chemotherapy being routinely used in case of undifferentiated carcinoma in resected specimen. Surprisingly, no statistically significant difference in survival between patients with incompletely and completely resected undifferentiated tumors was found in our patients in contrast to several reports [12, 13, 37, 38, 40]. This might be partly explained by the use of extended postoperative chemotherapy for patients (3—4 cycles) with incomplete resections. In the series of Gerl et al., only one recurrence was observed at a site of an incompletely resected mass, although incomplete resection predicted a worse outcome [37]. We found 13 patients who relapsed locally at their previous resection site and only five of them had had non-complete surgical resections. The extent of initial tumor burden correlates inversely with the frequency of complete resections [13, 37] and, as expected, the duration of the relapse-free interval was considered to be important for outcome. A short time to relapse (<3 months) was associated with a worse subsequent survival as shown before by Donohue and other investigators [25, 37, 38,41]. The overall survival of the patients being resected after first-line chemotherapy with respect to the number of surgical interventions and of patients after secondline chemotherapy is shown in Figure 1. There was no statistical significant survival difference detectable between patients with only single or with multiple resections after first-line chemotherapy. Patients after second-line chemotherapy had the worst prognosis reflecting the lower effectiveness of chemo-

12

24

36

60

Months —

2nd line n=25

—*— 1st line-single n=82 —•— 1st line-multiple n=27 Figure 1 Kaplan-Meier plots of overall survival after post-chemotherapy surgery with respect of number of surgical intervention (single versus multiple) and type of chemotherapy (first-line versus salvage treatment).

therapy. Undifferentiated tumor was found in 80% these patients [6, 39, 42]. Although the use of additive chemotherapy seems to increase the chance for long-term cure in case of surgical resection of vital carcinoma after firstline chemotherapy, the effect of additive chemotherapy was not demonstrable in the salvage chemotherapy setting. This was in accordance to Donohue et al. who have noted the same findings in a large series of 93 patients undergoing resections after second-line chemotherapy [39]. Among 40 patients in his series who had 'desperation surgery' defined as localised resectable markerpositive disease, 25% remained continuously diseasefree. In our series, 3 of 15 patients (20%) achieved longterm survival with surgery used in a comparable situation. With a long-term survival of approximately 40% in the whole group of our patients with salvage chemotherapy, the surgical removal of residual tumors appears to be a useful approach. However, due to the small number of patients available for the analysis after secondline chemotherapy, no solid conclusions can be drawn. Thus, we were not able to detect a significant survival difference between incomplete and complete resections as reported by Donohue et al. However, in contrast to our series, these authors had defined incomplete resection as either unclear tumor margins at operation or the occurrence of a rapid local relapse despite clear surgical margins. The latter event occurred in 24 of 36 patients, retrospectively classified as incompletely resected, thus selecting a subset of patients with a worse prognosis on the basis of the clinical course. The management of patients with multiple residual masses at different localisations after chemotherapy is controversly discussed. Most investigators advise the excision of all tumor residuals [5, 11, 32, 33, 40]. A recently published investigation has shown that one-

537

stage resection of neck, chest and abdominal residuals may be a safe approach in selected patients at an experienced center [31]. No differences in the complication rate was found in our patients if surgery was carried out either as a sequential (« = 22) or as a one-stage procedure (n = 5). Dissimilar histological results at different anatomical localisations were detected in 30% of patients in accordance to other investigations and a tendency towards more unfavorable histologies in the retroperitoneal compared to the pulmonary residuals was found (P = 0.05) [30-34, 43]. Recently, positron emission tomography with marked 2-fluoro-2-deoxyglucose molecules (FDG) is under investigation in patients with metastatic testicular cancer. PET-FDG imaging may be useful for the detection of residual carcinoma. However, the distinction between necrosis and differentiated teratoma may still remain difficult since both tissues have a relatively low metabolic activity [44]. It may be speculated for patients with multiple residual masses that the combination of a negative result at PET-examination at all sites and necrotic tissue found at the retroperitoneal resection may identify those patients who can be spared from resections at other sites. Another important advance for PET-FDG examination could be the identification of patients with undifferentiated tumor prior to the operation. Despite marker normalisation, we found undifferentiated tumor in 17 of 82 patients (21%) after first-line chemotherapy. In summary, secondary resection of tumor residuals after first- and second-line chemotherapy is still an essential part of the treatment of metastatic testicular cancer. More accurate prognostic models and the use of PET imaging might further help to identify residual undifferentiated tumors and may guide decisions about the operative procedure particularly in patients with technically feasible operations.

References 1 2. 3.

4.

5.

6.

7.

8

Einhorn LH. Treatment of testicular cancer: A new and improved model J Clin Oncol 1990; 8: 1777-81. Peckham M Testicular cancer. Rev Oncol 1988; 1: 439-53. Apnkian AG, Herr HW, Bajorin DF et al. Resection of postchemotherapy residual masses and limited retroperitoneal lymphadenectomy in patients with metastatic testicular nonseminomatous germ cell tumors. Cancer 1994; 74. 1329-34. Wahle GR, Foster RS, Bihrle R et al. Nerve sparing retroperitoneal lymphadenectomy after primary chemotherapy for metastatic testicular carcinoma. J Urol 1994; 152: 428-30. Toner GC, Panicek DM, Heelan RTet al. Adjunctive surgery after chemotherapy for nonseminomatous germ cell tumors: Recommendations for patient selection. J Clin Oncol 1990, 8: 1683-94. Fox EP, Weathers TD, Williams SD et al. Outcome analysis for patients with persistent nonteratomatous germ cell tumor in postchemotherapy retroperitoneal lymph node dissections. J Clin Oncol 1993; 11- 1294-9 Gelderman WA, Schraffordt-Koops H, Sleijfer DTet al. Results of adjuvant surgery in patients with stage III and IV nonseminomatous testicular tumors after cisplatin-vinblastine-bleomycin chemotherapy. J Surg Oncol 1988; 38: 227-32. Roth BJ, Greist A, Kubilis PS et al. Cisplatin-based combination

9.

10.

11.

12.

13.

14.

15.

16

17.

18.

19.

20.

21.

22. 23. 24. 25.

26. 27.

28.

29.

chemotherapy for disseminated germ cell tumors. Long-term follow-up. J Clin Oncol 1988; 6: 1239-47. Einhorn LH, Williams SD, Mandelbaum 1 et al. Surgical resection in disseminated testicular cancer following chemotherapeutic cytoreduction. Cancer 1981; 48: 904-8. Geller NL, Bosl GJ, Chan EY. Prognostic factors for relapse after complete response in patients with metastatic germ cell tumors. Cancer 1989; 63. 440-5. Tait D, Peckham MJ, Hendry WF et al. Post-chemotherapy surgery in advanced non-seminomatous germ-cell testicular tumours: The significance of histology with particular reference to differentiated (mature) teratoma. Br J Cancer 1984; 50 (5): 601-9. Hendry WF, A'Hern RP, Hetherington JWet al. Para-aortic lymphadenectomy after chemotherapy for metastatic non-seminomatous germ cell tumours: Prognostic value and therapeutic benefit. Br J Urol 1993, 71: 208-13 Jansen RL, Sylvester R, Sleyfer DTet al. Long-term follow-up of non-seminomatous testicular cancer patients with mature teratoma or carcinoma at postchemotherapy surgery. EORTC Genitourinary Tract Cancer Cooperative Group (EORTC GU Group). EurJ Cancer 1991; 27. 695-8 Fossa SD, Qvist H, Stenwig AE et al Is postchemotherapy retroperitoneal surgery necessary in patients with nonseminomatous testicular cancer and minimal residual tumor masses 7 J Clin Oncol 1992, 10: 569-73. Donohue JP, Rowland RG, Kopecky K et al. Correlation of computerized tomographic changes and histological findings in 80 patients having radical retroperitoneal lymph node dissection after chemotherapy for testis cancer. J Urol 1987; 137: 1176-9. Pugh RCB. Testicular tumours-introduction In Pugh RCB (ed): Pathology of the Testis Oxford: Blackwell Scientific Publications 1976; 139-62. Birch R, Williams S, Cone A et al. Prognostic factors for favorable outcome in disseminated germ cell tumors. J Clin Oncol 1986; 4: 400-7. Einhorn LH, Donohue JP. Combination chemotherapy in disseminated testicular cancer. The Indiana University experience. Semin Oncol 1979; 6: 87-93. Williams SD, Birch R, Einhorn LH et al. Treatment of disseminated germ-cell tumors with cisplatin, bleomycin, and either vinblastine or etoposide. N Engl J Med 1987; 316: 1435-40. Harstnck A, Schmoll HJ, Kohne-Wompner CH et al. Cisplatin, etoposide, ifosfamide, vincnstine and bleomycin combination chemotherapy for far advanced testicular carcinoma. Ann Oncol 1991; 2.197-202. Loehrer PJ Sr, Lauer R, Roth BJ et al. Salvage therapy in recurrent germ cell cancer. Ifosfamide and cisplatin plus either vinblastine or etoposide. Ann Intern Med 1988; 109: 540-6. Kaplan EL, Meier P. Non-parametric estimation from incomplete observations. J Am Stat Assoc 1958; 53' 457-81. Cox DR. Regression models and life tables (with discussion) J R Stat SocSerB 1972; 34: 187-220. Spiegelhalter DJ. Probabilities prediction in patient management and clinical trials Stat Med 1986; 5. 421-33. Steyerberg E\V, Keizer HJ, Stoter G et al. Predictors of residual mass histology following chemotherapy for metastatic non-seminomatous testicular cancer. A quantitative overview of 996 resections. EurJ Cancer 1994; 30: 1231-9. Logothetis CJ, Samuels ML, Tnndade A et al. The growing teratoma syndrome. Cancer 1982; 50: 1629-35. Molenaar WM, Oosterhuis JW, Meiring A et al. Histology and DNA contents of a secondary malignancy arising in a mature residual lesion six years after chemotherapy for a disseminated nonseminomatous testicular tumor. Cancer 1986; 58: 264-8. Ahlgren AD, Simrell CR, Tnche TJ et al. Sarcoma arising in a residual testicular teratoma after cytoreductive chemotherapy. Cancer 1984; 54: 2015-8. Steyerberg EW, Keizer HJ, Fossa SD et al. Prediction of residual retroperitoneal mass histology after chemotherapy for metastatic

538

30.

31

32.

33.

34.

35.

36.

37.

38

nonseminomatous germ cell tumor: Multivariate analysis of individual patient data from six study groups. J Clin Oncol 1995; 13: 1177-87. Gerl A, Clemm C, Schmeller N et al Sequential resection of residual abdominal and thoracic masses after chemotherapy for metastatic non-seminomatous germ cell tumours. Br J Cancer 1994; 70- 960-5. Brenner PC, Harry W, Morse MJ et al. Simultaneous retroperitoneal, thoracic, and cervical resection of postchemotherapy residual masses in patients with metastatic nonseminomatous germ cell tumors of the testis J Clin Oncol 1996; 14. 1765-9. Tiffany P, Morse MJ, Bosl G, et al. Sequential excision of residual thoracic and retroperitoneal masses after chemotherapy for stage III germ cell tumours. Cancer 1986; 57: 978-83. Qvist HL, Fossa SD, Ous S et al. Post-chemotherapy tumour residuals in patients with advanced nonseminomatous testicular cancer. Is it necessary to resect all residual masses? J Urol 1991; 145: 300-3. Mandelbaum I, Yaw PB, Einhorn LH et al. The importance of one-stage median sternotomy and retroperitoneal node dissection in disseminated testicular cancer. Ann Thorac Surg 1983; 36. 524-8. Fossa SD, Ous S, Lien HH, Stenwig AE. Post-chemotherapy lymph node histology in radiologically normal patients with metastatic nonseminomatous testicular cancer. J Urol 1989; 141. 557-9. Mulders PFA, Oosterhoff GON, Boetse C et al. The importance of prognostic factors in the individual treatment of patients with disseminated germ cell tumours Br J Urol 1990; 66: 425-9. Gerl A, Clemm C, Schmeller N et al. Outcome analysis after post-chemotherapy surgery in patients with non-seminomatous germ cell tumours. Ann Oncol 1995; 6: 483-8. Steyerberg EW, Keizer HJ. Zwartendijk J. et al. Prognosis after

39.

40

41. 42. 43.

44.

resection of residual masses following chemotherapy for metastatic nonseminomatous tesitcular cancer: A multivariate analysis. Br J Cancer 1993; 68: 195-200. Donohue JP, Fox EP, Williams SD et al. Persistent cancer in postchemotherapy retroperitoneal lymph-node dissection: Outcome analysis. World J Urol 1994; 12 (4): 190-5. Harding MJ, Brown IL, MacPherson SG et al. Excision of residual masses after platinum based chemotherapy for nonseminomatous germ cell tumours. Eur J Cancer Clin Oncol 1989; 25: 1689-94. Josefsen D, Ous S, Hoie J et al. Salvage treatment in male patients with germ cell tumours. Br J Cancer 1993, 67: 568-72. Peckham MJ. Horwich A, Easton DF, Hendry WF. The management of advanced testicular teratoma. Br J Urol 1988; 62. 63-8. Steyerberg E, Keizer HJ, MessemerJEet al Residual pulmonary masses after chemotherapy for metastatic nonseminomatous germ cell tumor. Cancer 1997; 79: 345-55. Stephens WS, Gonin R, Hutchins GD, Einhorn LH. Positron emission tomography evaluation of residual radiographic abnormalities in postchemotherapy germ cell tumour patients. J Clin Oncol 1996; 14: 1637-41.

Received 23 December 1996, accepted 22 April 1997.

Correspondence lo: C. Bokemeyer, MD Dept. of Hematology, Oncology, Immunology UKL - Medical Center II Eberhard-Karls-University Tubingen Otfried-Muller-Str. 10 D-72076 Tubingen Germany