PATIENTS AND METHODS
- Top of page
- PATIENTS AND METHODS
- CONFLICT OF INTEREST
This was a retrospective study of RPLND after first-line chemotherapy in patients with testicular or primary retroperitoneal nonseminomatous GCTs (NSGCTs) and initial retroperitoneal involvement. In all, 533 patients with a diagnosis of testicular tumour at any stage or of primary retroperitoneal NSGCT were seen at the Centre Léon Bérard (CLB) between 30 June 1992 and 31 December 2002. Among these, 219 patients with RPLN involvement were treated with platinum-based chemotherapy; 151 had RPLND for residual masses. The 68 others patients achieved either a clinical complete response (cCR) or a partial response (PR), but patients refused RPLND or had progressive disease during or within 4 weeks of the last cycle of first-line chemotherapy. A CR was defined as the disappearance of all metastatic lesions on thoraco-abdomino-pelvic CT, and the normalization of serum tumour marker (STM) levels for ≥4 weeks after the last cycle of chemotherapy. A cCR was obtained by chemotherapy alone and associated with the disappearance of all radiological evidence of metastatic lesions. A pathological CR was the complete resection of inactive (necrosis, fibrosis and teratoma) residual disease; a surgical CR was the complete resection of active (malignant component) residual disease. A PR was defined as a decrease of >50% in the sum of the largest perpendicular diameters of all measurable lesions either with normalization (PR–) or >90% decrease of STM levels (PR+). Stable disease was <50% shrinkage of metastatic lesions with a >90% decrease of STM levels. All other patients had progressive disease and were not considered for analysis in this study.
The clinical files of all patients were reviewed. The various areas involved were identified on abdominal-pelvic CT. An area was considered involved when a LN was >1 cm at the diagnosis on CT. The topography of all RPLNs was recorded according to Donohue et al., as was their size. Areas of RPLN involvement were as follows: paracaval, precaval, gonadal vein, inter-aortocaval, pre-aortic, para-aortic, ipsilateral iliac, contralateral iliac and interiliac.
The theoretical indication for RPLND was defined in our institution in 1992 according to the recommendations of the MSKCC , as resection of all residual masses whatever the initial size of the node, and systematic resection whatever the response to chemotherapy when the initial LN size is >3 cm. According to the MSKCC [5,6], right modified radical RPLND includes the lymphatic zones: paracaval, precaval, inter-aortocaval, pre-aortic, right gonadal vein and right iliac; left modified radical RPLND includes the lymphatic zones: precaval, inter-aortocaval, pre-aortic, para-aortic, left gonadal vein, and left iliac; and full bilateral RPLND involves all LN areas described above. Unilateral modified RPLND is generally indicated, except in patients with palpable macroscopic contralateral LN involvement.
RPLND was considered as conforming with recommendations when the theoretical extent of resection of all LN areas, as defined for each type of RPLND, was respected. Resection was defined as complete when the surgeon considered the resection of all macroscopic lesions as complete and when no lesion was detectable on abdominal-pelvic CT at 1–2 months after surgery. However patients with residual disease outside the retroperitoneum had surgical resection according to standard indications.
Patient files were reviewed as follows. All patients seen during the study period were examined by two medical oncologists (J.P.D., A.F.) who reviewed all clinical, pathological and radiological data for treatment decision-making. Patients were treated completely at the CLB or referred to tertiary treatment centres. Whatever the hospital, all patients were seen at the CLB at key moments of decision making. In all, 103 patients were referred to their surgeon (urologists or visceral surgeons) and 48 were operated on by the same surgeon (M.R.) at the CLB, either at the request of the patients or their referring urological surgeon. For the purpose of this study, all clinical files and original radiological reports were reviewed by two junior medical oncologists (E.T., H.B.) under the control of a senior medical oncologist (A.F.). Surgical reports were reviewed jointly by the junior medical oncologist (E.T.) and a senior oncology surgeon (P.M.) who had not operated on the patients. However, the surgical reports were analysed on the basis of the original report made by the surgeon who had operated on the patient. All original material was collected into a database constructed.
The event-free survival (EFS) was calculated according the method of Kaplan-Meier, from the first day of chemotherapy to disease progression or the date of last follow-up. The log-rank test was used to compare survival probabilities.
Patients were followed every 3 months during the first year, then every 6 months during the next 3 years, and once a year thereafter for 10 years. At each visit, they had a physical examination, STM levels were determined (hCG, α-fetoprotein, AFP, and lactate dehydrogenase, LDH) and thoraco-abdominal CT. The CNS was imaged (CT or MRI) when patients had neurological symptoms. The last follow-up of patients was updated at the 31 July 2008.
- Top of page
- PATIENTS AND METHODS
- CONFLICT OF INTEREST
In all, 151 patients had RPLND after first-line chemotherapy; their characteristics are described in Table 1. The median (range) size of the largest RPLN was 40 (8–240) mm; 110 (73%) patients had initial RPLNs of >30 mm, and half of the 151 patients had other metastatic sites.
Table 1. The characteristics of the 151 patients at diagnosis
|Characteristic||Median (range) or n (%)|
|Age, years|| 27.3 (15–67)|
|Primary site|| |
| Testis||144 (95)|
| Left|| 74 (54)|
| Right|| 70 (46)|
| Retroperitoneum|| 7 (5)|
|Histological pattern subtype in primary site|| |
| Embryonal carcinoma|| 116 (77)|
| Teratoma|| 82 (54)|
| Yolk sac|| 59 (39)|
| Seminoma|| 49 (32)|
| Choriocarcinoma|| 32 (21)|
|Size of RPLNs, cm|| |
| ≤3|| 41 (26)|
| >3|| 110 (64)|
|Other metastatic site(s)|| |
| Yes|| 77 (51)|
| No|| 74 (49)|
|Metastatic sites|| |
| Lung|| 60 (40)|
| Mediastinal lymph nodes|| 23 (15)|
| Retrocrural lymph nodes|| 20 (13)|
| Supra-clavicular lymph nodes|| 19 (12)|
| Cervical or axillary lymph nodes|| 11 (7)|
| Liver|| 7 (4)|
| Brain|| 1 (0.6)|
| Other|| 3 (2)|
|STM level|| |
| Unknown|| 0|
| Normal|| 44 (29)|
| Elevated|| 107 (71)|
| Unknown|| 1 (0.5)|
| Normal|| 52 (34.5)|
| Elevated|| 98 (65)|
| Unknown|| 17 (11)|
| Normal|| 72 (47)|
| Elevated|| 62 (42)|
|IGCCC prognostic group|| |
| Good|| 62 (41)|
| Intermediate|| 48 (32)|
| Poor|| 30 (20)|
| Unclassified|| 11 (7)|
All patients received platinum-based chemotherapy (a combination of bleomycin, etoposide and cisplatin in 84% of cases), except one who received a combined regimen with carboplatin. Two patients achieved a cCR, 134 a PR– and 15 a PR+.
The median interval between the first day of the last cycle of chemotherapy and surgery was 38 (17–114) days for all patients, except three who were operated at 153, 252 and 334 days. These three patients initially refused the RPLND and finally accept RPLND because of the persistence of residual masses. RPLND was performed by 67 different surgeons; 46 operated on only one patient, 14 on two, three on three, three on four, one on seven, and one surgeon operated on 48 patients. Theoretically, according to the MSKCC recommendations [5,6], from the initial location of their RPLNs, 107 patients should have had a full bilateral RPLND, 34 a left modified unilateral template, and 10 a right modified unilateral template.
Of the 151 patients included in the study, 75 had a conforming RPLND and 76 had a non-conforming RPLND. Conformity was therefore equal to half for all patients, at 100% for the surgeon who performed 48 RPLND and 26% for others; resection was complete in 93% and 75% of conformed and non-conformed cases, respectively (Table 2).
Table 2. Conformity of RPLND and completeness of resection, and the relationship with relapse
|Complete resection, n||70||58||128|
| LN ≤3 cm||14||24|| 38|
| LN >3 cm||56||34|| 90|
|Incomplete resection|| 5||18|| 23|
| LN ≤3 cm|| 0|| 3|| 3|
| LN >3 cm|| 5||15|| 20|
|Relapse in the retroperitoneum, n (%)|| || || |
| N RP relapses|| 2|| 3|| 5|
|Incomplete resection|| 5||18|| 23|
| N RP relapses|| 1|| 9|| 10|
| N RP relapses|| 3||12|| 15|
The pathological examination of RPLND specimens showed that 73 (48%) patients had teratoma, 62 (41%) had necrosis, 13 (9%) had viable GCTs and three (2%) had a malignant transformation of teratoma. Among the 73 patients with teratoma components in the residual mass, 53 had teratoma in the primary tumour.
At the same time as RPLND, 10 patients had vascular surgery (aortic replacement and/or vena cava resection), six had an inguinal orchidectomy, five a right or left hepatectomy or hepatic metastasectomy, five retrocrural surgery, three a segmentary colectomy, two a supra-clavicular LN resection, two a nephrectomy, and one a pulmonary wedge resection. Fourteen patients had surgical resection of residual disease outside the retroperitoneum at a different time than RPLND.
Three patients had repeated RPLND because the resection was incomplete; pathological examination of the LNs showed leiomyosarcoma in one and teratoma the other two. Histological examination of the second RPLND revealed no leiomyosarcoma component in the first patient and confirmed the presence of teratoma in the other two. Notably, only one patient had a laparoscopic RPLND.
Complications within 30 days of RPLND were lymphocele in 22 patients, chylous ascites in seven, chylothorax in one, intestinal occlusion in seven, infection in four, haemorrhagic complications in three, iliac deep venous thrombosis in two, and renal artery thrombosis, leg ischaemia and a foreign body in one each. More importantly, one patient died from intra-abdominal bleeding 10 days after RPLND.
The data on ejaculation morbidity were known for 117 patients; 44 (38%) had definitive retrograde ejaculation after RPLND, and no patient had a nerve-sparing RPLND.
The median (range) follow-up was 77 (1.3–186.5) months. Twenty-eight patients (18.5%) of the 151 relapsed after RPLND at a median of 6.1 (1.3–93.2) months. Nine relapses were reported in the group of patients with conformed RPLND and complete resection (group A, 70, relapse rate 13%) vs 19 in the 81 with a conformed RPLND but incomplete resection or nonconforming RPLND with complete or incomplete resection (group B, relapse rate 23%; Table 2). Fifteen of the 28 patients relapsed in RPLNs (two in group A and 13 in group B; Table 3). All retroperitoneal recurrences were inside the boundaries of template resection. Fourteen of these 15 patients initially had LNs of ≥5 cm and all had active disease except one who had growing teratoma at the relapse. However, half of the patients with active residual retroperitoneal disease after chemotherapy relapsed in the retroperitoneum.
Table 3. Patients with RPLN relapse
|Patient||Conformity||Completeness||Histology of RPLND||Size, mm||Relapse at site of RPLND||Treatment after RPLND||Outcome|
| 7||No||No||Teratoma + embryonal carcinoma||190||Yes||Chemo + HDCT||NED|
|10||No||No||Embryonal carcinoma|| 20||Yes||Chemo + HDCT||NED|
|30||No||No||Teratoma + malignant transformation PNET|| 50||Yes||Chemo PNET||DOD|
|39||No||No||Embryonal carcinoma||120||Yes||Progressive disease||DOD|
|111||Yes||Yes||Necrosis + Teratoma|| 90||Yes||Surveillance||NED|
|128||No||Yes||Embryonal carcinoma|| 73||Yes||4 VeIP||DOD|
|151||No||No||Embryonal carcinoma|| 58||Yes||Surveillance||DOD|
|160||No||No||Teratoma + embryonal carcinoma|| 70||Yes||Surveillance||DOD|
In all, 18 patients died (12%), five in group A (7%) and 13 in group B (16%); 17 (11%) died from disease progression and one died 10 days after RPLND. At the date of the last follow-up, 132 patients (87.4%) were alive with no disease, 1 (0.6%) is alive with progressive disease and 18 patients (12%) had died. The EFS probability at 10 years was 85% for group A and 72% for group B (Fig. 1).
Figure 1. EFS: A, the 70 patients with compliant and complete RPLND; B, the 81 patients with compliant but incomplete resection or with noncompliant complete or incomplete RPLND.
Download figure to PowerPoint
- Top of page
- PATIENTS AND METHODS
- CONFLICT OF INTEREST
The treatment strategy for advanced NSGCTs is cisplatin-based chemotherapy and complete resection of residual disease [10–12]. Standard chemotherapy is a combination of cisplatin, etoposide and bleomycin . The number of cycles administered is based on the International Germ Cell Cancer Consensus Group (IGCCCG) classification .
The resection of residual masses is indicated because histological studies have shown that 45% of patients have teratoma, 45% have necrosis and 10% have active residual disease . Different authors have examined the criteria for indicating surgical resection of residual disease, i.e. retroperitoneal and mediastinal lymph nodes , lung metastases  and liver metastases [16,17]. The retroperitoneum is the most frequent site of residual disease . Different studies have tried to determine what should be the indications and extension for RPLND, based on initial histology, initial size and postchemotherapy shrinkage of RPLNs . Nevertheless, the most widely recognized criteria for indications of RPLND are based on the MSKCC study . Patients who are likely to benefit from RPLND are those with initial RPLNs of ≤3 cm and presence of radiologically assessable residual LNs after chemotherapy, and those with initial LNs of >3 cm whatever the response to chemotherapy.
These criteria published by Toner et al. were used in our retrospective study. The European Germ Cell Cancer Consensus Group  recommends the resection of any residual mass after chemotherapy. The recommendations of the National Comprehensive Cancer Network are identical . RPLND was not indicated in the European Association of Urology guideline in case of cCR after chemotherapy whatever the initial LN size .
However, the data from the MSKCC and from international studies are based on CT techniques used in the 1980s and early 1990s. Recent improvement in the quality of CT might change the evaluation of the response to chemotherapy in the future.
RPLND techniques were initially developed with curative intent at a time when no effective chemotherapy was available. Skinner et al. developed an extensive supra- and infra-renal RPLND using the thoraco-phreno-abdominal approach. Afterwards other investigators developed less invasive techniques [23,24]. Bilateral RPLND became the standard of care during the 1970s [2,7]. This approach allowed the treatment of stage II disease by RPLND followed or not by chemotherapy . It was also the standard treatment of residual disease after chemotherapy. Further technical developments include nerve-sparing procedures [26,27] and limitation of RPLND templates [28,29]. In our study no patients had nerve-sparing RPLND and 38% had retrograde ejaculation .
The extension of postchemotherapy RPLND has been controversial for many years. The MSKCC approach to limit the template of RPLND was the only method developed using an experimental approach. In a first step, Wood et al. reported a retrospective analysis to study the location of residual disease that required complete resection, i.e. active disease and teratoma. They concluded that all patients with right testicular GCTs require bilateral RPLND; conversely those with left testicular cancer, with no contralateral macroscopic involvement, only require a modified left RPLND. This procedure, tailored to the extension of the tumour, was validated in a prospective study by Aprikian et al. and Herr . However pathological examination of frozen sections has not yet been introduced in routine practice; the method was only recently validated in a large series of patients .
More recently, Heidenreich et al.[33,34] recommended full bilateral RPLND in bulky residual disease after chemotherapy, as did Carver et al.. We based our retrospective study on the criteria defined in the studies described above [5,6]. Only half of the present patients had been operated on according to these recommendations; however, only 15 had a retroperitoneal recurrence, all of them in the boundaries of template resection. Carver et al. reported that 7–32% patients had teratoma or viable GCT outside the boundaries of a modified template. Steiner et al. reported, in a series of 102 patients, one recurrence inside the boundaries of left unilateral template dissection.
Retroperitoneal relapse was more frequent if the RPLND did not conform to the standards (13/15 patients). It was also more frequent when resection was incomplete (10/15 patients) than when it was complete (5/15 patients). Clearly, incomplete resection is more frequent when the RPLND does not conform to the recommendations, and is associated with a relapse rate of two-thirds. The analysis in Table 2 shows that up to 10 patients might have benefited from a conformed RPLND.
Because there were relatively few patients in each study group, the results cannot be evaluated statistically. Nevertheless some links are apparent between the different groups, as shown in Table 2; there is a strong relationship between conformity of the procedure and completeness of surgery, and between RPLN size and both conformity and completeness of surgery. Table 2 also shows that there is a strong relationship between relapse and completeness of surgery and, to a smaller extent, conformity of the procedure, two factors which are linked, as mentioned above. The number of recurrence was not correlated with conformity when the resection was complete.
Thus the major question is whether tumour size is a prognostic factor, as it might be a limitation to proper resection of residual tumour according to the skill of the surgeon. The present study cannot answer this question and no previous data have been published in GCTs. However, it was reported that the outcome of poor-risk patients with testicular GCTs is linked to the number of patients treated in the institution . In rectal cancer, different studies of practice and even a randomized trial have shown that experience in rectal cancer surgery is an independent prognostic factor of outcome . The rate of conformity to standards according to the number of patients operated by surgeons reported in our study supports this observation.
It is certainly not the only factor involved. Retroperitoneal relapses generally occur in patients with bulky disease; all but one of the present patients had a maximum RPLN diameter of ≥5 cm. Notably, the bulk of the tumour is not a prognostic factor per se in the IGCCCG classification. Nevertheless, neither conformity of the procedure nor completeness of surgery appear sufficient to prevent retroperitoneal relapse. The management of active residual disease is also controversial. The most important prognostic factors of relapse and survival in this case is completeness of surgery and percentage of active disease .
In conclusion, this is, to our knowledge, the first analysis of the quality of RPLND after chemotherapy according to recommendations. The possible limitations are that the study was retrospective and the relatively few relapses. Nevertheless, only half of the patients had a RPLND that conformed to the recommendations, with complete resection in most cases (84%). Sixty-seven surgeons operated on 151 patients in 10 years. Thirteen of 15 patients who relapsed in the retroperitoneum had incomplete or/and nonconforming RPLND. Half of the patients had active disease at RPLND.
RPLND after chemotherapy is more likely to be incomplete when the residual disease is ≥5 cm. There are arguments in favour of complete surgery based on a standardized procedure. All of these results support the implementation of RPLND in a tertiary centre.