Clinical outcome following post-chemotherapy retroperitoneal lymph node dissection in men with intermediate- and poor-risk nonseminomatous germ cell tumour

Authors


Brett S Carver, Department of Urology, Sidney Kimmel Center for Prostate and Urologic Cancers, 353 East 68th Street, New York, NY 10021, USA.
e-mail: carverb@mskcc.org

Abstract

As usual, the urological oncology section is quite full, and covers a range of topics. Authors from New York describe their considerable experience with retroperitoneal lymph node dissection after chemotherapy, and make some important recommendations about the procedure in men initially diagnosed with intermediate- and poor-risk nonseminomatous germ cell tumours. There are several papers on prostate cancer, and on renal and bladder cancer. As can be seen, the role of chemotherapy in prostate cancer is receiving increasing attention.

OBJECTIVE

To evaluate the outcome in patients treated with chemotherapy and retroperitoneal lymph node dissection (RPLND) after an initial diagnosis of International Germ Cell Cancer Collaborative Group (IGCCCG) intermediate- and poor-risk metastatic nonseminomatous testicular germ cell tumour (NSGCT), as the integration of chemotherapy and surgery in managing advanced NSGCT continues to develop.

PATIENTS AND METHODS

Between 1989 and 2003, 157 patients initially diagnosed with IGCCCG intermediate- and poor-risk NSGCT had RPLND after chemotherapy at the authors’ institution, with a median follow-up of 36 months. Progression-free probability (PFP) and disease-specific survival (DSS) were estimated using the Kaplan–Meier method. Cox proportional hazards regression analysis was used to assess the prognostic significance of risk factors for disease progression after RPLND.

RESULTS

In all, 68 (43%) and 89 (57%) patients were assigned as intermediate- and poor-risk, respectively. At the time of RPLND the median residual retroperitoneal mass was 3.0 cm and 29 (19%) men had elevated serum tumour markers (α-fetoprotein, human chorionic gonadotrophin, or both). Retroperitoneal residual masses were completely resected in 147 (94%) patients; retroperitoneal histology revealed fibrosis in 73 (47%), teratoma in 63 (40%) and viable GCT in 21 (13%). The 5-year overall DSS and PFP were 81% and 70%, respectively. Patients with poor-risk NSGCT were at no greater risk of disease progression than those with intermediate-risk NSGCT. In a multivariate analysis, residual mass size, incomplete surgical resection and the presence of teratoma and viable germ cell cancer independently predicted disease progression after RPLND.

CONCLUSIONS

Patients with advanced NSGCT have long-term freedom from disease progression when chemotherapy is combined with resection of residual masses. Our data suggest that the tumour response to chemotherapy, coupled with complete resection of all residual masses, predicts long-term freedom from disease progression.

Abbreviations
(NS)GCT

(nonseminomatous) germ cell tumour

STM

serum tumour marker

DSS

disease-specific survival

IGCCCG

International Germ Cell Cancer Collaborative Group

RPLND

retroperitoneal lymph node dissection

PFP

progression-free probability

TMT

teratoma with malignant transformation.

INTRODUCTION

In patients with advanced germ cell tumours (GCTs), the introduction of platinum-based combination chemotherapy regimens has resulted in cure rates approaching 80%[1]. A better understanding and use of serum tumour markers (STMs) and integration of adjunctive surgical resection of residual masses has contributed significantly to the comprehensive care of these patients. Nonetheless, ≈ 20% of patients presenting with metastatic disease eventually die from GCT. Currently therapeutic goals are directed at obtaining maximum treatment efficacy in those destined to fail, while at the same time reducing treatment-related toxicity in those considered most likely to have a complete response.

Several published reports have assessed the use of clinical prognostic variables that might predict the outcome of patients with metastatic GCT [2–4]. In general, the level of STMs and the extent and sites of metastatic disease have been found to predict the response to chemotherapy, and ultimately the probability of survival. In an effort to achieve a uniform and standardized predictive model, the International Germ Cell Cancer Collaborative Group (IGCCCG) published a prognostic factor-based staging system [5]. Among patients with advanced testicular nonseminomatous GCT (NSGCT), three prognostic groups are identified on the basis of STM levels and sites of metastatic involvement. Of these patients, ≈ 40% fall within the intermediate- or poor-risk categories, achieving overall 5-year survival rates of 79% and 48%, respectively. We think that the integration of risk-based chemotherapy and aggressive adjunctive surgery is essential in the overall care of patients with advanced NSGCT. Currently, standard therapy in patients with intermediate- and poor-risk metastatic NSGCT at the authors’ institution consists of four cycles of conventional-dose bleomycin, etoposide and cisplatin followed by retroperitoneal lymph node dissection (RPLND) [1]. In selected cases surgery is used in those with elevated STMs. In addition, all extra-retroperitoneal residual masses are resected regardless of histological findings at RPLND. In the current study, we analysed our contemporary results in patients with intermediate- and poor-risk NSGCT managed with chemotherapy and RPLND.

PATIENTS AND METHODS

Between 1989 and 2003, 157 patients with NSGCT, classified as either intermediate- or poor-risk by IGCCCG criteria, had RPLND after first- or second-line chemotherapy. Patients with primary extragonadal NSGCT were excluded from the study. Information was obtained from our institutional database that is updated prospectively. The study was approved by the local Institutional Review Board.

All patients were staged before induction chemotherapy, with STMs, CT of the abdomen and pelvis, and either chest CT or chest X-ray. Nadir STM levels after orchidectomy and before chemotherapy were used in the assignment of IGCCCG risk groups. After first-line chemotherapy, patients were re-staged with STMs and imaging as above. Patients were considered to have elevated STMs if the levels of α-fetoprotein or hCG were above the upper limit of normal (>15 ng/mL and >2.2 U/mL, respectively). Patients with persistently elevated STMs or progression of disease went on to receive second-line chemotherapy before RPLND. All patients had RPLND after chemotherapy, and with rare exceptions all men treated for advanced NSGCT at our institution are recommended to undergo RPLND after chemotherapy and normalization of the STMs, irrespective of residual mass size, and in selected cases where STMs remained persistently elevated. The size of the retroperitoneal mass before and after chemotherapy was assessed by the maximum transverse diameter on CT imaging. Complete resection was defined as complete excision of all residual retroperitoneal masses.

Progression-free probability (PFP) and disease-specific survival (DSS) were estimated using the Kaplan–Meier method. For survival analysis, the follow-up was started at the time of RPLND. Factors associated with progression after RPLND were evaluated in univariate and multivariate analyses using the Cox proportional hazards regression, with P < 0.05 considered to indicate statistical significance, and hazard ratios with 95% CIs reported for the Cox regression model.

RESULTS

The patients’ clinical features are listed in Table 1; the median (range) α-fetoprotein and hCG levels before chemotherapy were 624 (0–92 000) ng/mL and 708  (0–860 000) IU/mL, respectively, and the retroperitoneal mass was 7.5 (0–18) cm. After first-line chemotherapy 107 patients had RPLND and 50 required second-line chemotherapy before RPLND. After chemotherapy the median α-fetoprotein level declined to 3.6 (0–1820) ng/mL and the median hCG declined to 0 (0–101) IU/mL. In all, 29 (19%) patients had persistently elevated STMs before RPLND. The median residual retroperitoneal mass was 3.0  (0–15) cm; residual extra-retroperitoneal masses were present in 74 patients.

Table 1.  The clinical and patient characteristics before RPLND
VariableN (%) patients
  • *

    Patients with clinical stage I NSGCT who relapsed on surveillance and subsequently received chemotherapy.

  • †BEP, bleomycin, etoposide, cisplatin; VIP, etoposide, ifosfamide, cisplatin; VAB-6, vinblastine, bleomycin, cisplatin, cyclophosphamide, dactinomycin; EP, etoposide, cisplatin; PVB, cisplatin, vinblastine, bleomycin ; VeIP, vinblastine, ifosfamide, cisplatin.

Clinical stage*
 I  3 (2)
 IIa  1 (1)
 IIb 10 (6)
 IIc 25 (16)
 III 118 (75)
IGCCCG risk group
 Intermediate 68 (43)
 Poor 89 (57)
Chemotherapy regimen
 First-line157 (100)
  BEP107 (68)
  VIP 16 (10)
  VAB-6 10 (6)
  EP  9 (6)
  PVB  4 (3)
  VeIP  3 (2)
  Other  8 (5)
 Second-line 50 (32)
Elevated STM at RPLND 29 (19)
Residual retroperitoneal mass, cm
 None  5 (3)
 <2 41 (26)
 2–5 54 (34)
 5–10 14 (9)
 >10 33 (21)
Extra-retroperitoneal  residual mass 74 (47)

All 157 patients had RPLND and the histological findings are also outlined in Table 2. Among these, all retroperitoneal masses were completely resected in 147 (94%) patients. Those requiring second-line chemotherapy had a significantly higher incidence of viable GCT in the retroperitoneum (P < 0.001). Teratoma with malignant transformation (TMT) was found in seven (4%) patients. In 46 patients with residual masses of ≤ 20 mm, 15 (33%) had teratoma and two (4%) had viable GCT. In all, 74 patients had evidence of residual disease outside the retroperitoneum. Those with pulmonary, mediastinal, neck and supraclavicular masses all had a complete surgical resection, either concurrently with RPLND, or as staged procedures. Among patients with liver lesions all but three had a complete resection. In the three cases of unresected liver masses, biopsy at the time of RPLND showed only fibrosis or necrosis. Overall histological findings in extra-retroperitoneal residual masses were fibrosis in 49 (66%), teratoma in 17 (23%) and viable GCT in eight (11%).

Table 2.  The pathological characteristics in resected retroperitoneal masses
Variable (n)N (%) patients
All patients (157)
Fibrosis73 (47)
Teratoma alone63 (40)
 mature52 (33)
 immature 4 (70
 TMT 3 (40
Viable GCT21 (13)
 viable cancer alone 11 (7)
 viable cancer ± teratoma10 (6)
First-line chemotherapy only (107)
Fibrosis46 (43)
Teratoma53 (50)
Viable GCT 8 (8)
Second-line chemotherapy (50)
Fibrosis27 (54)
Teratoma10 (20)
Viable GCT13 (26)

The median (interquartile range) follow-up for survivors was 36 (16–76) months; 12 patients (8%) had progressive disease immediately after RPLND. Thirty patients (19%; 15 each in the intermediate- and poor-risk groups) relapsed at a median of 5 months. Figure 1 summarizes the clinical outcome of the patients according to treatment algorithm. The 5-year PFP (95% CI) for all patients was 70 (63–78)%, and the 5-year probability of DSS was 81 (74–88)%. There was no significant difference in disease progression among intermediate- and poor-risk patients (5-year PFP 68% vs 73%, respectively, P = 0.5). Twenty-five patients (16%; 11 in the intermediate- and 14 in the poor-risk group) died from progressive testicular cancer at a median of 18 months. Similar to the findings with PFP, poor-risk patients did not have a higher likelihood of death from disease than intermediate-risk patients (5-year DSS 81% vs 80%, respectively, P = 0.9, Fig. 2a). In a subset analysis, patients receiving only first-line chemotherapy did not differ in their probability of disease progression after RPLND from those requiring second-line chemotherapy (5-year PFP 71% and 67%, respectively, P = 0.3). However, the probability of DSS was significantly better in patients receiving only first-line chemotherapy (5-year DSS 85% vs 72%, P = 0.02, Fig. 2b).

Figure 1.

The clinical outcome for men with intermediate- and poor-risk metastatic NSGCT managed with chemotherapy and RPLND. NED, no evident disease.

Figure 2.

The probability of DSS after RPLND stratified: a, by IGCCCG risk criteria (solid line, intermediate-risk; dashed line, poor-risk); and b, by the need for second-line chemotherapy (solid line, first-line chemotherapy; dashed line, second-line chemotherapy.

In a univariate analysis of clinical and pathological variables, residual retroperitoneal mass size, incomplete surgical resection and the histological finding of teratoma or viable GCT in the retroperitoneum were predictors of disease progression and relapse (Table 3). IGCCCG risk stratification, the need for second-line chemotherapy, and elevated STMs at the time of RPLND were not significant predictors of disease progression. When analysed in a multivariate logistic regression model, incomplete resection of residual retroperitoneal masses, size of residual retroperitoneal masses and the finding of teratoma and viable GCT at RPLND independently predicted disease progression and relapse (Table 3).

Table 3. 
Univariate and multivariate analyses of clinical and pathological variables associated with relapse and disease progression after RPLND
VariablesHazard ratio (95% CI)P
  • *

    Continuous variable (per cm).

Univariate
Clinical
Size of mass*:
 before chemotherapy 1.05 (0.97–1.13)0.2
 after chemotherapy 1.15 (1.05–1.25)0.002
Elevated STMs at RPLND 1.30 (0.63–2.73)0.5
IGCCCG risk group:
 poor vs intermediate 0.80 (0.44–1.47)0.5
Second-line chemotherapy 1.40 (0.75–2.62)0.3
Pathological
Resection
 incomplete vs complete 7.85 (3.51–17.55)<0.001
Retroperitoneal histology
 teratoma vs fibrosis 2.49 (1.16–5.320)0.02
 viable GCT vs fibrosis 6.61 (2.84–15.43)<0.001
Multivariate
Retroperitoneal histology
 teratoma vs fibrosis 3.34 (1.43–7.79)0.005
 viable GCT vs fibrosis 7.26 (2.81–18.80)<0.001
Resection
 incomplete vs complete10.42 (3.89–27.91)<0.001
Size of mass after chemotherapy,  cm 1.14 (1.03–1.25)0.01

DISCUSSION

Overall cure rates for patients with advanced NSGCT have approached 80% over the last two decades. While effective cisplatin-based chemotherapy has been held largely responsible for these trends, the role of RPLND after chemotherapy remains essential. It is well recognized that the prognosis after surgery is primarily predicated on histological features of resected residual masses. Initial reports of post-chemotherapy residual mass histology indicated the presence of fibrosis, teratoma and viable GCT in about a third of cases each [6]. Current series report necrotic debris or fibrosis in 40–50%, teratoma in 35–40% and viable malignant cells in 10–15% of patients [7]. Such a decrease in the proportion of viable malignant cells is generally attributed to stage migration and use of more effective chemotherapy regimens. In the present study of patients with intermediate- and poor-risk NSGCT, the distribution of retroperitoneal histology at RPLND is consistent with other contemporary series.

In an attempt to obviate adjunctive surgery, many criteria and statistical models have been proposed to predict the presence of fibrosis; nonetheless, there is a persistent false-negative prediction of ≈20%[8]. Recently, Oldenburg et al.[9] showed that after chemotherapy a third of retroperitoneal masses of <20 mm contained either teratoma or viable GCT. In the present series, a third of patients with residual masses of ≤20 mm had evidence of teratoma and 4% had viable GCT.

While many advocate a policy of surveillance for such small residual masses, the consequence of residual viable GCT is disease progression. Furthermore, the biological potential of unresected teratoma remains unpredictable. Despite a benign histological appearance, teratoma can grow, obstruct and invade local structures or become entirely unresectable. It might also undergo malignant transformation into chemo-resistant sarcoma or carcinoma. Lastly, late relapse (>2 years) is associated with unresected teratoma and is also highly chemo-resistant [10].

Given the uncertainty in predicting the histology of residual masses after chemotherapy in patients with metastatic NSGCT, we routinely use RPLND on normalization of STMs. However, in a few selected patients with persistently elevated STMs, residual masses are also resected. Moreover, as the reported rate of discordant histology between the retroperitoneum and extra-retroperitoneal residual masses is 29–46%, all sites of disease are resected.

The present results underscore the essential role of surgery after chemotherapy in the comprehensive management of patients with metastatic NSGCT. We report an overall 5-year PFP and DSS of 70% and 81%, respectively, when all sites of residual disease are resected. In the multivariate analysis the only preoperative indicator of disease relapse was the size of residual retroperitoneal masses after chemotherapy.

Contrary to reports by Fizazi et al.[11], the IGCCCG risk category did not affect the overall prognosis after RPLND, but it is possible that a higher proportion of men initially with IGCCCG poor-risk disease had rapid disease progression and never had RPLND, and therefore were not included in the current study. The use of second-line chemotherapy before RPLND did not predict a higher probability of relapse after surgery. However, among this group, the DSS was significantly lower. Such a finding suggests that while relapse occurs with a similar frequency in both groups, patients who receive second-line chemotherapy before RPLND are less likely to be salvageable at subsequent relapse. Nevertheless, the overall outcome strongly supports RPLND in patients who have received second-line chemotherapy.

Several investigators emphasized the importance of complete resection of residual masses after chemotherapy [11–13]. We showed similarly, in a multivariate analysis, that complete resection of residual retroperitoneal masses independently predicted progression-free survival. As outlined previously, while the consequence of incompletely resected viable GCT is certain disease progression, residual teratoma behaves less predictably. In the present series, among the 63 patients with retroperitoneal teratoma, seven (11%) had evidence of malignant transformation. Unlike GCTs, TMT is unresponsive to conventional cisplatin-based chemotherapy regimens and complete surgical resection remains the mainstay of therapy. Our experience with TMT was previously reviewed by Motzer et al.[14]. Finally, in the multivariate analysis the finding of residual retroperitoneal teratoma, viable GCT and the size of the retroperitoneal mass after chemotherapy independently predicted disease relapse.

In conclusion, in patients with advanced NSGCT, long-term cancer control is achieved by instituting cisplatin-based systemic chemotherapy coupled with aggressive resection of all sites of residual masses. Our data suggest that there is no significant difference in outcome between patients with IGCCCG intermediate- and poor-risk disease undergoing RPLND. Also, relapse rates were similar after RPLND for men receiving first- or second-line chemotherapy, but DSS was significantly less for men who relapsed after second-line chemotherapy. In the multivariate analysis, the size of residual retroperitoneal mass after chemotherapy, incomplete surgical resection, and the histological finding of retroperitoneal teratoma and viable GCT, independently predicted disease relapse after RPLND for men with IGCCCG intermediate- and poor-risk NSGCT.

CONFLICT OF INTEREST

None declared.

Ancillary