Fax: +44 (208) 643 8809
Metastatic nonseminomatous germ cell tumors of the testis
Results of elective and salvage surgery for patients with residual retroperitoneal masses †
Version of Record online: 15 MAR 2002
Copyright © 2002 American Cancer Society
Volume 94, Issue 6, pages 1668–1676, 15 March 2002
How to Cite
Hendry, W. F., Norman, A. R., Dearnaley, D. P., Fisher, C., Nicholls, J., Huddart, R. A. and Horwich, A. (2002), Metastatic nonseminomatous germ cell tumors of the testis. Cancer, 94: 1668–1676. doi: 10.1002/cncr.10440
This work was undertaken by The Royal Marsden National Health Service Trust, which received a proportion of its funding from the National Health Service Executive; the views expressed in this publication are those of the authors and not necessarily those of the National Health Service Executive.
- Issue online: 15 MAR 2002
- Version of Record online: 15 MAR 2002
- Manuscript Accepted: 29 NOV 2001
- Manuscript Revised: 27 NOV 2001
- Manuscript Received: 26 MAR 2001
- Institute of Cancer Research
- the Bob Champion Cancer Trust
- Cancer Research Campaign
- National Health Service Executive
- testicular neoplasms;
- para-aortic lymphadenectomy;
A mass may persist in the para-aortic region after patients undergo chemotherapy for metastatic, nonseminomatous germ cell tumor of the testis (NSGCT). Retroperitoneal lymphadenectomy removes the mass, which may contain residual active malignancy, and allows histologic assessment of the effectiveness of the chemotherapy. Whereas some have favored early, elective removal of such masses, others have chosen to observe them, reserving salvage surgery for patients who experience disease recurrence. A retrospective analysis was undertaken to define the outcome in these two groups of patients.
After receiving chemotherapy for metastatic NSGCT, 442 men underwent lymphadenectomy for residual masses (measuring ≥ 1 cm in greatest dimension) between 1976 and 1999, inclusive. Three hundred thirty men underwent elective surgery within 3 months of the completion of chemotherapy, and 112 men underwent salvage surgery after receiving reinduction chemotherapy for tumor recurrence.
The residual mass was removed completely in 87% and 72% of patients in the elective and salvage lymphadenectomy groups, respectively; was removed with difficulty and possibly incompletely in 9% and 21% of patients, respectively; and was definitely removed incompletely in 4% and 7% of patients, respectively. The operative mortality rate was 0.9% in the elective surgery group and 1.8% in the salvage surgery group. There was malignant teratoma, undifferentiated in 8.5% of patients in the elective surgery group and in 49% of patients in the salvage surgery group (P < 0.001). Differentiated teratoma and necrosis/fibrosis were present in 66.0% and 25.4% of patients in the elective surgery group, respectively, and in 38.4% and 12.5% of patients in the salvage surgery group, respectively. The authors were unable to produce a clinically useful model to predict the presence of necrosis/fibrosis only in either group. The 5-year recurrence free and overall survival rates were 83% and 89%, respectively, in the elective surgery group and 62% and 56%, respectively, in the salvage surgery group. For the salvage surgery group, the completeness of surgical excision and the presence of undifferentiated teratoma were of overriding importance for overall survival. A variety of other patient-related, tumor-related, and surgery-related factors also were significant in the final model for the elective surgery group.
The current results demonstrate the low level of morbidity that can be obtained, even in the salvage surgery group, and the importance of complete surgical resection in this setting. Because it is not possible to predict with sufficient accuracy which patients will have favorable pathology (necrosis/fibrosis), the authors continue to recommend elective surgery for all suitable men with residual masses after they receive first-line chemotherapy. Cancer 2002;94:1668–76. © 2002 American Cancer Society.
The introduction of cisplatin-containing combination chemotherapy has dramatically improved the outcome of patients with metastatic testicular teratoma. However, although 70–75% of men will achieve complete radiologic remission,1 the remaining 25–30% of men will have a residual mass, usually sited in the para-aortic area. These masses may contain residual undifferentiated teratoma (in a small proportion of men), differentiated teratoma, or necrosis/fibrosis only.2 The addition of radiotherapy in a small, uncontrolled series of observations did not produce any reduction in the proportion of patients with residual, active malignancy.3 Our usual practice has been to remove such residual masses 4–6 weeks after the completion of four cycles of chemotherapy, provided that the serum marker levels had returned to normal, and the results of this policy have been analyzed from time to time.4–8 If the surgery was done when the serum markers were still elevated, then there was a high likelihood that the excised mass would contain active tumor; however, even with normal marker levels, 14% of patients had residual malignancy.8 Thus, normalization of serum markers was no guarantee that the residual mass was tumor free.9
During the same period, patients were referred who had recurrent disease after receiving chemotherapy elsewhere, after which para-aortic masses had not been removed but had simply been kept under surveillance. Further reinduction chemotherapy was required before the para-aortic masses were removed. This wait-and-see policy offered an apparently attractive alternative approach, because some patients were able to avoid surgery.10, 11 However, the overall response to reinduction chemotherapy was not as good compared with the overall response to primary chemotherapy, and surgical clearance seemed to be more difficult technically. Therefore, the outcome of patients who underwent surgery in this salvage setting has been recorded and compared with the outcome of the larger group of patients who underwent early elective retroperitoneal lymphadenectomy (RPLND).
MATERIALS AND METHODS
After receiving chemotherapy for metastatic nonseminomatous germ cell tumor of the testis (NSGCT), 442 men age 15–64 years (median, 28 years) underwent lymphadenectomy by a single surgeon for residual masses (measuring ≥ 1 cm in greatest dimension) between 1976 and 1999, inclusive. The distribution of surgeries for each year is shown in Figure 1. Three hundred thirty men underwent elective surgery within 3 months of completing chemotherapy, most often using 4 courses of cisplatin, vinblastine, and bleomycin (PVB);12 bleomycin, etoposide, and cisplatin (BEP);13 or more intensive schedules for patients with a poor-risk presentation.14 One hundred twelve men underwent salvage surgery after completing reinduction chemotherapy using second-line agents for disease recurrence.15–17
All patients were assessed in the Joint Testicular Tumor Clinic approximately 1 month before surgery, and the surgeon and oncologist carefully reviewed previous treatment. Serum markers were rechecked, and scans were reviewed and repeated if necessary to define the exact position of the residual mass in relation to the great vessels and surrounding normal structures. Patients and their relatives were advised of the anticipated risk of surgery, with particular regard to possible damage to surrounding structures or adjacent organs, such as the ipsilateral kidney, which may have been densely adherent to the mass. Risk of loss of ejaculation was assessed in the light of past experience.18
Para-aortic lymphadenectomy was done through a long midline abdominal incision in 383 patients and through a thoracoabdominal incision in 53 patients; a loin or iliac incision was used in 6 patients. At operation, the mass was exposed in the retroperitoneum and carefully separated from the great vessels. The ipsilateral ureter was freed from its lateral aspect, and the mass was then removed with a margin of normal connective tissue, staying as far as possible in established planes of cleavage. Excision was classified as complete, difficult/?complete, or incomplete according to the surgeon's satisfaction that all abnormal tissue had been removed. No attempt was made to remove macroscopically normal lymph nodes. Sympathetic nerves were preserved carefully whenever possible in all patients who underwent surgery after 1984, when the importance of nerve-sparing surgery in ejaculatory function was appreciated. After surgery, all patients were followed in the Joint Testicular Tumor Clinic with regular clinical examination, measurement of tumor markers, and imaging, as required.
All data were recorded prospectively in the Bob Champion Unit testicular cancer research data base. Categoric data were examined using the chi-square test, with Fisher exact test used where appropriate. Survival and recurrence free survival were calculated using the Kaplan–Meier method.19 Survival was measured from the date of para-aortic lymphadenectomy for all patients. The log-rank test20 was used to test differences between groups. Multivariate Cox regression analysis21 was used to identify prognostic groups that influenced survival and recurrence free survival. Stepwise logistic regression analysis was used to identify factors that could predict necrosis in the resected specimen.
The residual masses measured ≤ 3 cm in the greatest transverse dimension in 125 patients (28%), 4–8 cm in greatest dimension in 215 patients (49%), and ≥ 9 cm in greatest dimension in 102 patients (23%). The size distribution was similar (P = 0.47) in the elective surgery group and the salvage surgery group.
Completeness of Surgical Excision
Excision was complete in 288 patients (87%) in the elective surgery group and in 81 patients (72%) in the salvage surgery group. Difficulty in completing the excision was experienced in 30 patients (9%) in the elective surgery group and in 23 patients (21%) in the salvage surgery group, mainly due to dense adherence of the mass to surrounding vital structures. Excision was obviously incomplete in 12 patients (4%) in the elective surgery group and in 8 patients (7%) in the salvage surgery group. Thus, excision was complete significantly more often in the elective surgery group (P = 0.001).
The complications due to surgery are shown in Table 1. The ureter was involved with tumor and required repair or resection in 5 patients, whereas the bowel was involved and required resection or repair in 12 patients, including 2% of elective procedures and 6% of salvage procedures. Four patients underwent resection of liver metastases at another institution. Lesions in lung or mediastinum were resected either synchronously or at another time in 30 patients.
|Complication||No. of patients (%)|
|Elective surgery||Salvage surgery||Overall|
|Nephrectomy||31 (9)||19 (17)||50 (11)|
|Aorta/iliac artery damage||11 (3)||7 (6)||18 (4)|
|Vena cava damage||13 (4)||6 (5)||19 (4.3)|
|Postoperative complications||14 (4)||6 (5)||20 (4.5)|
|Operative mortality||3 (0.9)||2 (1.8)||5 (1.1)|
Information on ejaculation after surgery was available for 255 patients. Among 65 men who underwent surgery before 1984, 66% had normal ejaculation compared with 81% of 190 men who underwent surgery after 1984, when a nerve-sparing technique was adopted whenever possible. There were similar proportions of men with normal ejaculation in the elective and salvage surgery groups. Forty-seven patients successfully fathered children after undergoing surgery.
The histologic findings related to tumor size are shown in Table 2. Teratoma was differentiated in 261 patients (59%), including 66% of patients in the elective surgery group and 38% of patients in the salvage surgery group. Only necrotic or fibrous tissue was found in 98 patients (22%), including 25% of patients in the elective surgery group and 13% of patients in the salvage surgery group. Sarcomatous tissue was found after 3 elective procedures (1%) and after 7 salvage procedures (6%).
|Histology||Size of tumor mass (cm)|
|Elective surgery||Salvage surgery|
|< 3||4–8||> 9||< 3||4–8||> 9|
Serum markers were elevated at the time of surgery in 54% of patients with undifferentiated tumor in the excised specimens, including 32% of patients in the elective surgery group and 65% of patients in the salvage surgery group. For patients with normal markers at the time of surgery, only 5.3% had undifferentiated tumor after an elective procedure; this rate was 16% after salvage procedures.
Logistic Regression Model Predicting Necrosis/Fibrosis
For the elective surgery group alone, multivariate logistic regression analysis was used to produce a model predicting necrosis/fibrosis. Smaller size of excised mass, fewer days on chemotherapy, and absence of differentiated teratoma at orchidectomy were significant predictors of necrosis/fibrosis in the excised surgical specimen. The model had a sensitivity of 27.8% and a specificity of 92.3% (Table 3). This means that clinical management based on this model would have avoided 35 operations, although 15 of these patients would have harbored differentiated or undifferentiated teratoma.
|Observed||Logistic regression model prediction||Predictive value|
|Residual teratoma||180||15||Specificity, 92.3%|
Survival and Recurrence Free Survival Models
The median follow-up was 6.4 years at the time of analysis. The 5-year recurrence free survival rate (Fig. 2) was 83% (95% confidence interval [95% CI], 78–87) for the elective surgery group, with an overall 5-year survival rate (Fig. 3) of 89% (95% CI, 84–92%). For the group of men who underwent a salvage surgical procedure, the recurrence free survival rate was 62% (95% CI, 50–72), with an overall 5-year survival rate of 56% (95% CI, 44–66%). Univariate and multivariate analyses predicting outcome were performed for the elective surgery and salvage surgery groups. For the elective surgery group, a wide range of patient-related (increasing age), tumor-related (raised markers presurgery, presence of undifferentiated teratoma in resected specimen, size of residual mass, and other organ involvement), and surgical features (completeness of resection, bilateral surgery, and synchronous surgery) were related to recurrence free survival on univariate analysis (Table 4). Nerve-sparing RPLND was not related to outcome. Many of these factors may be expected to be interrelated, and, on multivariate analysis, the final model included patient age, raised markers presurgery, presence of undifferentiated teratoma, and size of mass ≥ 9 cm together with the need for bilateral surgical resection (Table 4). For overall survival, very similar features were seen on univariate analysis (Table 4) but included nephrectomy rather than synchronous surgery. On multivariate analysis, the final model included age, presence of undifferentiated teratoma, the need for bilateral resection, and other organ involvement but also included nephrectomy. The presence of undifferentiated tumor and the need for bilateral resections together with increasing age were the common adverse factors in the multivariate analysis for both recurrence free and overall survival. These factors probably reflect initial chemoinsensitivity, extent of disease, and possibly an increase in difficulty of later effective salvage treatment with increasing age. The fact that tumor size and raised markers prior to surgery were not included in the overall survival model may reflect the possibility of later effective salvage treatments for some patients, even though they may have an increased risk of recurrence. The presence of nephrectomy in the survival model, but not in the recurrence free survival model, conversely, may indicate that patients who undergo nephrectomy are more difficult to salvage with chemotherapy if they develop recurrent disease.
|Factor/group||Elective||Salvage||Elective only||Salvage only|
|Univar P value||Multivariate||Univar P value||Multivariate||Univar P value||Multivariate||Univar P value||Multivariate|
|HR||P value||HR||P value||HR||P value||HR||P value|
|Completeness of excisiona|
|Difficult||30||22||0.001||—||0.05||< 0.001||—||0.279||0.001||2.740||0.008||< 0.001||3.640||< 0.001|
|MTU||28||55||0.001||2.359||0.016||< 0.001||5.770||< 0.001||0.005||—||0.051||0.001||3.143||0.005|
|Tumor size (cm)c|
|Other organ involvement|
|Markers prior to RPLND|
For men in the salvage surgery group, univariate analysis showed a more restricted group of factors: The presence of undifferentiated teratoma, the need for bilateral surgery, and, most importantly, complete excision were related to recurrence free survival (Table 4). Only completeness of excision remained (presence of undifferentiated teratoma was of borderline significance; P = 0.051) in the multivariate model, indicating the overwhelming importance of excising all active tumor in patients who demonstrate histologic failure after receiving salvage chemotherapy. For overall survival, the factors on univariate analysis were the same as the factors for recurrence free survival with the addition of patient age and size of excised mass. All of the factors remained in the final model for overall survival except for the size of the excised mass (Table 4). The fact that patient age appears to influence survival but not recurrence free survival may reflect an age-related effect on the treatment possibilities if patients develop further recurrent disease or, indeed, may reflect long-term effects resulting in noncancer deaths.
Radiologic clearance of metastatic disease may be incomplete after chemotherapy for NSGCT. Thus, there is a dilemma: Should the oncologist wait and see what happens to the mass and the markers or arrange for the mass to be excised? We have analyzed our experience over the past 25 years to determine which clinical and histopathologic features relate to outcome with these different management approaches.
In the elective surgery group, primary chemotherapy had eliminated malignancy completely, leaving only scar tissue in 25% of patients. Only 8% of patients had residual undifferentiated teratoma in this group, with sarcoma in 1% of patients. Serum markers remained elevated at the time of surgery in approximately 33% of these patients with active tumors, and the remaining patients had normal serum markers; thus, the risk of leaving undifferentiated malignancy after primary chemotherapy, when serum markers are normal, was 7.0%. This proportion was similar at 7.3%, even if markers had been elevated before chemotherapy. In approximately 66% of the patients in the elective surgery group, there was residual differentiated teratoma, which contained a wide variety of tissues, ranging from a few apparently innocent cysts to a mélange of variegated epithelia set in a sea of primitive mesenchyme. Sometimes, these can continue to expand as the so-called growing teratoma syndrome.22 It is known that this tissue is unstable, and there are many examples of late recurrences,23–29 sometimes occurring in lymph nodes as small as 1.7 cm in greatest dimension.30 Teratoma may change into frankly malignant tumor, although sometimes not for some time after the primary treatment was completed.22, 31
Patients in the salvage surgery group who underwent lymphadenectomy after recurrence and reinduction second-line chemotherapy comprised a totally different population: almost 50% of patients had residual undifferentiated teratoma in the excised specimens, and 7% were sarcomatous. It is evident that chemotherapy was not as effective in this group of patients compared with patients in the elective surgery group. The patients who required second-line chemotherapy may have been selected either because they had less chemosensitive tumors or because chemoresistance may have been induced by previous treatment. Sarcomatous degeneration, in particular, is notoriously chemoresistant.32, 33 In these patients, lymphadenectomy clearly plays an important therapeutic role in getting rid of chemoresistant tissue, which must be excised as thoroughly as possible. This is not simply an information-gathering exercise, because almost 50% of patients had active malignancy. Only 13% of patients had scar tissue only, whereas the remaining patients had differentiated teratoma.
There is no general agreement on the minimum size of postchemotherapy mass that merits excision. Our practice has been to recommend excision of all masses measuring ≥ 1 cm in greatest dimension. Recent prediction models based on international data sets of more than 700 patients have indicated that masses measuring < 1 cm in greatest dimension have a 25% chance of containing mature teratoma and a 5% chance of harboring malignant disease.34 Using a decision analysis model, the estimated gains in 5-year survival were 4.3% and 2.7% after excision of masses measuring 1–2 cm and < 1 cm in greatest dimension, respectively.35 Our practice has been to excise the residual mass as completely as possible but not extend the dissection to the surrounding macroscopically normal lymph nodes. After undergoing complete excision, only 3% of our patients who had a recurrence had it in the abdomen. This limited approach to lymphadenectomy has gained general acceptance,36 although some surgeons still prefer to use a modified template.37 We attempted to construct a model to predict the presence of necrosis/fibrosis only because it can be argued that these patients had undergone unnecessary surgery. However, although, by using a full range of clinical and pathologic parameters, we could produce a model with high specificity (92.3%), i.e., correctly identifying those without necrosis/fibrosis only, it had low sensitivity (27.8%), i.e., correctly predicting necrosis/fibrosis only, and was not judged clinically useful. Other workers also have tried to construct prediction models. Vergouwe and colleagues38 studied 278 patients, together with a confirmatory group of 276 patients, who were treated with chemotherapy prior to undergoing para-aortic lymphadenectomy. Those authors found that the histologic make-up of the primary tumor, pretreatment tumor marker revels, the size of residual mass, and reduction in mass size with treatment gave some indication of the likely pathology of the residual mass. This model had a specificity of 82% and a sensitivity of 44%; however, once again, it could not form the only basis for the decision whether to remove a mass.
It is noteworthy that both of these models shared three common factors (initial tumor histology, pretreatment tumor marker levels, and size of residual mass). Reduction in mass size was not measured in our study, but this potentially important factor will be assessed in the future.
Excision of postchemotherapy residual masses requires careful preoperative localization by computed tomography scanning, so that the optimum surgical approach can be planned relating the site of the mass to surrounding vital structures. The relation of the mass to the renal vessels is of particular importance. Metastases in the liver,39 neck,40 or lungs41 may require synchronous or staged removal. Metastases in the chest can be excised at the same time as abdominal lymph nodes.42 However, in general, we have favored a separate, staged approach to widespread disease rather than trying to do too much through one incision, which may provide suboptimal access to difficult areas of tumor. The significance of leaving a little differentiated teratoma behind is not clear, although there is evidence that this tissue probably is unstable.22, 31 Currently, we believe that every effort should be made to remove it, a view that is shared by others.43
The incidence of residual malignant teratoma, undifferentiated in this series was similar to the incidence reported from other centers (Table 5). 44–52 The striking difference in the incidence of undifferentiated teratoma (embryonal carcinoma) between the elective surgery group and the salvage surgery group was statistically highly significant (8% vs. 49%, respectively; P < 0.001). Clearly, although there may be a choice whether to remove a residual mass after primary chemotherapy, it is abundantly clear that every effort should be made to remove a residual mass after second-line chemotherapy has been given for patients with recurrent disease. Under these circumstances, complete excision offers a highly significant survival advantage.53 The risk after primary chemotherapy is subtler, largely relating to the high incidence of mature teratoma. There are numerous examples on record of malignant degeneration in such tissue that sometimes occurs many years after primary treatment.
|Study||No. of patients||Histology of residual mass (%)|
|Bracken et al.44||60||25 (42)||14 (23)||17 (28)|
|Staehler et al.45||65||23 (35)||25 (39)||17 (26)|
|Harding et al.46||42||15 (36)||14 (33)||9 (21)|
|Williams et al.47||25||13 (52)||9 (36)||3 (12)|
|Mulders et al.48||55||31 (56)||12 (22)||12 (22)|
|Herr et al.49||122||57 (47)||48 (39)||17 (14)|
|Kulkarni et al.50||67||18 (27)||29 (47)||20 (30)|
|Donohue et al.52||870||218 (25)||452 (52)||200 (23)|
|Elective||330||84 (25)||218 (66)||28 (8)|
|Salvage||112||14 (13)||43 (38)||55 (49)|
|Total||442||98 (22)||261 (59)||83 (19)|
The technical aspects of removal of these masses has been described elsewhere, and the potential hazards have been well documented.54–56 The problems encountered relate first to the size of the mass and, second, to the involvement of adjacent structures. Our overall operative mortality was 1%, rising to almost 2% in patients who underwent salvage lymphadenectomy. This is difficult and dangerous surgery that should be done in specialized centers. Nevertheless, we believe that this is less dangerous than leaving a mass that may contain active tumor. In our view, these results support a policy of early and elective excision of all substantial postchemotherapy masses rather than pursuing a wait-and-see policy, which leads to the need for second-line chemotherapy in a significant proportion of patients with the possibility of salvage surgery. This is also the view reached after analysis of 870 patients who were treated in the United States.52
- 7Para-aortic lymphadectomy after chemotherapy for testicular tumour. Br J Urol. 1988; 62: 470–471., , , .
- 14Intensive induction chemotherapy for poor risk tumours. In: HorwichA, editor. Testicular cancer—investigation and management. 2nd ed. London: Chapman and Hall Medical, 1996: 271–285..
- 15Salvage therapy and autologous bone marrow transplantation. In: HorwichA, editor. Testicular cancer—investigation and management. 2nd ed. London: Chapman and Hall Medical, 1996: 307–318., .
- 20Asymptotically efficient invariant procedures. J R Stat Soc (A). 1972; 135: 185–206., .
- 21Analysis of survival data. New York: Chapman and Hall, 1984:, .
- 31Teratoma following chemotherapy for non-seminomatous germ-cell tumor: a clinicopathologic correlation. Proc ASCO. 1983; 2: 139., , .
- 55Complications of lymph node dissection. In: SkinnerDG, editor. Complications of urologic surgery. Philadelphia: WB Saunders, 1976: 422–435., .