Patients and methods
- Top of page
- Patients and methods
This study began as a collaboration between oncologists in Christchurch and Auckland, and the first patient was registered in February 1980. Subsequently patients were entered in the study from the other four regional oncology centres in New Zealand (Dunedin, Wellington, Palmerston North and Hamilton). This report deals with 248 patients (median age at presentation 29 years, range 16–77) with clinical stage I NSGCTT who were registered in the study between 21 February 1980 and 16 April 1997. The data are based on the follow-up to 20 May 1997; the median (range) follow-up was 53 (1–185) months. The number of patients entered from each centre was: Auckland 104, Christchurch 66, Wellington 19, Dunedin 30, Palmerston North 21 and Hamilton eight. Patients were usually registered within one month of orchidectomy but the orchidectomy date was taken as the baseline point for follow-up.
To be eligible for the study the patient had to have their testicular tumour identified histologically as a NSGCTT, classed as malignant teratoma intermediate (MTI), malignant teratoma undifferentiated (MTU), malignant teratoma trophoblastic (MTT) or teratoma differentiated (TD), according to the British Testicular Tumour Panel Classification . Patients with seminoma were also included if their serum βhCG before orchidectomy was ≥300 IU/L. Patients with seminoma and elevated levels of AFP were considered to have NSGCTT and were eligible for the study.
The presence or absence or vascular and/or lymphatic invasion (VLI) in the primary tumour was noted, this phenomenon having been identified by ourselves and others as the prognostic factor of greatest significance [3,8,11,12]. VLI were grouped together because of the recognized difficulty experienced by histologists in differentiating vascular channels and lymphatic vessels in such patients. The histological subtype, the extent of the primary tumour (T stage), the presence of abnormal levels of serum AFP before orchidectomy and the presence or absence of VLI in the primary tumour were related to relapse and analysed statistically. The chi-square test was used for univariate comparisons. A multivariate logistic regression analysis was not possible because cell frequencies were sparse.
In the first few years of the study, patients with primary tumour stages T1–T3 were entered, but since 1989 only T1 and T2 tumours have been included, as study group members decided that tumour invading the spermatic cord (T3) warranted chemotherapy from the outset. The other inclusion requirements were a normal physical examination, normal chest X-ray, normal CT of chest, abdomen and pelvis and normalization of serum AFP, βhCG and LDH if abnormal before orchidectomy. Patients were required to give informed consent.
Surveillance included clinical examination, estimation of blood tumour markers and chest X-ray monthly in the first year, every second month in the second year and at 3–4-monthly intervals in the third year. In addition, CT of abdomen and pelvis was repeated every 2–3 months in the first year, every 4 months in the second year and thereafter as clinically indicated. After 3 years the follow-up was at the discretion of the clinician, but participants were encouraged to continue with follow-up for at least 10 years.
- Top of page
- Patients and methods
The histological subtype was MTI in 113, MTU in 92, MTT in 25 and TD in 14. In addition there were four patients with pure seminoma and a serum βhCG of >300 IU/L before orchidectomy. The primary T stage was T1 in 225 cases, T2 in 19 and T3 in four. Serum AFP and/or βhCG were raised to abnormal levels before orchidectomy in 167 of 248 patients (67%); AFP alone was elevated in 101 of 246 patients (41%) and βhCG alone abnormally elevated in 118 of 245 patients (48%). VLI was present in 92 of 243 (38%) patients in whom it was defined.
Subsequent relapse of disease was recorded in 70 of the 248 patients (28%); 45 of these (64%) were detected in the first 6 months of follow-up, 60 (86%) in the first year and 69 (99%) in the first 28 months. One relapse was detected at 74 months in a man who presented at 32 years of age with a mixed germ cell tumour (MTU and seminoma). VLI was noted and both the AFP and βhCG were elevated before orchidectomy. His surveillance was unremarkable until 74 months when, after developing right flank discomfort, he was found to have an obstructed right kidney. The tumour markers were not elevated and CT of the abdomen showed ureteric obstruction but no lymphadenopathy. However, at laparotomy, a 1.5 cm node was resected and this contained pure seminoma. The ureter was obstructed by intraluminal seminoma. He was treated with six cycles of carboplatin and has remained in remission; this case has been reported in detail elsewhere .
Six contralateral clinical stage I testicular tumours (five seminoma and one MTU) were detected during the follow-up. All patients were treated with orchidectomy and those with seminoma also had adjuvant abdominal radiation. All these patients are currently free of disease at a median follow-up of 55 months since diagnosis of the second tumour.
Sixty-seven of the 70 relapsed patients received treatment with a cisplatin or platinum analogue-based drug combination, supplemented in 15 cases with RPLND. Of the three patients who did not have chemotherapy, one was found at laparotomy to have differentiated teratoma which was removed and no chemotherapy was given. Another patient at relapse refused to have treatment and died from his disease. The third patient had a suspicious abdominal CT which was under review when he committed suicide. At necropsy he was found to have a single small involved retroperitoneal node.
Six of the 248 patients died; two from non-neoplastic causes, one from a myocardial infarction with no evidence of relapse and the other by suicide, as described above. Another refused treatment at relapse and died. Three of the relapsed and treated patients, all with VLI, also died from their disease. One who had MTI and seminoma in his primary tumour relapsed at 5 months with a solitary pulmonary metastasis. This disease responded to four cycles of cisplatin-containing chemotherapy but the pulmonary metastasis redeveloped. It was treated by surgical excision and two further courses of chemotherapy. Despite this, he relapsed again in the chest and was treated with radiation, with resolution. However, he relapsed for the fourth time 6 months later; a multidrug combination achieved only a partial response and his disease progressed until he died. The second patient had MTI and relapsed in the spleen at 12 months. After splenectomy he was treated with four cycles of cisplatin, vinblastine and bleomycin and achieved a complete remission, but one year later he developed hepatic metastases. Further courses of cisplatin and etoposide did not induce remission and he died from his disease. The third patient had MTT and relapsed at 4 months with low-volume lung disease. He entered a complete remission after three cycles of cisplatin-containing treatment, but relapsed 2 months later with pulmonary and brain metastases. He had radiotherapy to the whole brain, intrathecal methotrexate and multiple cycles of combination chemotherapy involving many agents. However, his disease progressed and he died.
The method by which the relapse was detected is given in Table 1; relapse in three patients was detected clinically before any other evidence became apparent (two had a supraclavicular node and one splenomegaly). Tumour markers and CT of the abdomen were prominent in detecting relapse but on four occasions the plain chest X-ray was the first site of suspicion. In each of these four cases the plain X-ray abnormality was confirmed by CT of the chest, but CT of the abdomen and pelvis was within normal limits and the tumour markers were not abnormal. Of 48 patients with positive tumour markers at the original diagnosis, 31 were positive at relapse; of 22 originally negative, 11 were positive and 11 negative at relapse.
Table 1. The first evidence for detecting relapse
The histological subtypes, T stage of the primary tumour, presence or absence of AFP in the blood before orchidectomy and the presence or absence of VLI in the primary tumour were documented and related to relapse; these features are shown in Table 2.
Table 2. Tumour status and relapse
A univariate sequential comparison of MTI, MTU and MTT with all other histology revealed no statistically significant difference between these subtypes in terms of relapse; 10 of 25 patients (40%) with MTT relapsed, but this was not statistically significant (P=0.17). The presence of VLI in the primary tumour predicted relapse; 42 of 92 (46%) patients with this finding relapsed, compared with only 26 of 151 (17%) of those in which the finding was absent (P<0.001).
Compliance, in terms of attendance at clinics, was less than planned in 30 of the 248 patients (12%). The median (range) age of the noncompliant patients was 26 (17–46) years. Six of the 30 (20%) patients not complying relapsed; one of these patients at relapse refused treatment and subsequently died. The others have remained well and disease-free. At a median (range) follow-up of 53 (1–185) months, 242 patients (98%) remained free of disease; four have died from disease, including one who refused treatment when his disease relapsed, and another two died from other causes.
- Top of page
- Patients and methods
Clinical stage I NSGCTT can be managed after orchidectomy by adjuvant abdominal radiation or adjuvant RPLND. With both of these techniques there is a 10–15% failure rate, mostly because of relapse outside the abdomen. In addition, both radiotherapy and RPLND have adverse effects, notably bone marrow damage in the case of radiotherapy and ejaculatory problems after node dissection . Thus, when CT of the abdomen, reliable tumour markers and effective chemotherapy became available, many clinicians felt that orchidectomy followed by surveillance and chemotherapy at relapse was a valid management option.
In 1987, the UK Medical Research Group reported upon a retrospective study of risk factors for predicting relapse in clinical stage I nonseminomatous tumours. They found that invasion of testicular vascular channels, invasion of testicular lymphatics, the presence of undifferentiated (embryonal) cells and the absence of yolk sac elements were significantly important in this respect . Many other groups have reached similar conclusions with some variations, but there is overwhelming agreement that VLI in the primary tumour is the outstanding factor, notwithstanding the difficulty in differentiating vascular and lymphatic involvement [3,8,11,12,15].
In the patients in the present study in whom there was VLI, the relapse rate was 46%, which was significantly greater (P<0.001) than in those who did not have such involvement (17%). There was no increased rate of relapse with any particular histological subtype, although there was a trend towards a greater relapse rate with MTT. Neither raised pre-orchidectomy AFP nor more advanced stage (T2) predicted a higher relapse rate (Table 2). Although only four patients with involvement of the cut end of the cord were entered in this study, three relapsed and subsequently we have elected to treat such patients with chemotherapy.
There have been three major criticisms of the policy of orchidectomy and surveillance only for patients with clinical stage I NSGCTT, i.e. noncompliance which may result in inadequate surveillance, the occurrence of late relapses and cost. Noncompliance is a common finding in a proportion of patients with any illness; it would be surprising if some patients with NSGCTT did not fail to comply, as they are generally young men who are often socially mobile. This problem was assessed in the present 248 patients; 30 of them did not comply (12%), i.e. failed, without adequate reason, to attend two consecutive clinic appointments. No special identifying characteristics were noted in these patients; their median age was 26 years (range 17–46). Six of these 30 patients relapsed (20%), whereas 64 of the 218 (29%) compliant patients developed a relapse; these relapse rates are not significantly different (P=0.286). One of the noncompliant patients refused to have treatment at relapse and subsequently died. All other noncompliant patients with relapse were treated successfully. An additional patient in this group developed a contralateral seminoma which was a clinical stage I tumour and was treated successfully by orchidectomy and adjuvant radiation. The overall outcome of the noncompliant patients was no different from the whole cohort. It may be reasonable to assume that the patient who relapsed and refused treatment would have refused the offer of adjuvant chemotherapy or RPLND at the outset. In our experience, noncompliance is also seen in patients who are offered, are receiving or have had chemotherapy, and is a reflection of the complex psychosocial situation in which this group of young men find themselves. We would expect those who recommend other modes of treatments to find a similar noncompliance rate.
Of the 70 relapses, 66 (94%) were identified within 24 months; one patient’s disease relapsed at 74 months, as described earlier. Others have also noted late relapses [16,17] and this has led critics to state that RPLND is the primary treatment of choice, as it removes the major potential for relapse in a site which is the most difficult and costly to monitor. We do not agree that RPLND is a superior treatment, as late relapses also occur in those who have undergone RPLND as their original treatment. Indeed, the group from the Indiana University Medical Centre reported 81 late relapses (>2 years), of which 35 occurred in patients who originally had clinical stage I disease; 31 of these had undergone RPLND supplemented in 14 by immediate adjuvant chemotherapy. The other 46 late relapses occurred in patients who had presented with Stage II or III disease. That report would suggest that, irrespective of the primary mode of treatment, there will be late relapses . In these circumstances it would seem unfair to single out surveillance for criticism when late relapse is also seen after RPLND and adjuvant chemotherapy.
The first evidence of relapse (Table 1) was most commonly the elevation of the serum tumour markers, either alone or in combination with an abnormality on abdominal CT. However, the tumour markers were often inconsistent between presentation and relapse. Thus, the measurement of tumour markers is important, irrespective of whether or not they were originally abnormal. In three patients the clinical examination was the first evidence of relapse (in two the detection of a supraclavicular node and in a third the detection of splenomegaly). This is contrary to the observation of Lander et al. , who found that clinical examination was unhelpful in detecting relapse in their 54 patients and they recommended that follow-up consultations could be reduced. In another four patients the plain chest X-ray detected disease not apparent from tumour markers or abdominal CT. This is discordant with the findings of Gels et al. , who noted that the chest X-ray detected relapse in only two of their 154 patients; they concluded that the chest X-ray could be omitted from the surveillance protocol. However, their observations must be placed in perspective; their patients had CT of the chest every 3 months in the first year, every 6 months in the second and at the end of the third year of observation. We believe that a clinical examination and chest X-ray are useful features of a surveillance programme, and continue to recommend monthly clinical review in the first year with a chest X-ray and tumour markers. In the second year we recommend tumour marker evaluation every second month, supplemented by a clinical examination, chest X-ray and abdominal CT every 4 months; in the third year we recommend a clinical examination every 3–4 months supplemented by tumour markers, chest X-ray and abdominal CT (if appropriate). After 3 years, the follow-up should be at the discretion of the clinician, but we recommend 6-monthly follow-up to 5 years and annual follow-up to at least 10 years. Tumour markers and a clinical examination would be undertaken at these times but radiological examinations only as indicated by clinical circumstances. This programme is intensive and generates significant cost, the third criticism of such a management policy. It is possible that less intensive programmes could result in similar outcomes, but this will require formal testing.
Of the present cohort of 248 patients, 242 (98%) were alive and disease-free at the time of the last follow-up, 17 years after the initiation of this study. Two patients died from unrelated causes and, if they are excluded, the overall disease-specific survival at 17 years is 242 of 246 (98.4%); the median follow-up for the whole group is 53 months. These results are similar to those of an earlier report, when the study included only 115 patients with a disease-specific survival of 98.3% at a median follow-up of 36 months. The later results confirm our earlier impression that surveillance of Stage I NSGCTT is an appropriate management technique . These results are similar to those reported by others who practise surveillance alone after orchidectomy for clinical Stage I NSGCTT [4,5,16]. They are comparable with the 2-year disease-specific survival rates after primary RPLND .
Both Pont et al.  and Cullen et al.  reported their experiences with adjuvant chemotherapy for high-risk clinical stage I cases. These groups treated those patients judged to be at high risk of relapse with two cycles of adjuvant bleomycin, etoposide and cisplatin (BEP). Pont et al. reported a 93% survivorship with two deaths, one from progressive testicular cancer and another from lung cancer . Cullen et al. reported two deaths in their 114 patients at a median follow-up of 4 years. One of these patients was reclassified as having an adenocarcinoma of the rete testis and the other had cisplatin-resistant NSGCTT . Forty-two of the present relapsing patients had VLI and thus would have been regarded as having a high risk of relapse. Three of these patients died despite appropriate treatment, but 39 of 42 (93%) achieved a durable complete remission with conventional combination chemotherapy. The results of orchidectomy and surveillance with treatment of subsequent relapsed cases in our hands achieved similar results to those of Pont et al.  and Cullen et al. . It appears that the outcome of these two management protocols in terms of disease-free survival are comparable.
Orchidectomy with surveillance and treatment of relapse produces results that are comparable to those achieved by orchidectomy plus RPLND. As the outcome in terms of survival appears to be equal across the three management techniques, the adverse effects of these protocols should be compared. RPLND has the disadvantage that the patients are hospitalized for 10–14 days and full recovery usually takes several weeks. Virtually all patients lose ejaculatory function after traditional bilateral suprahilar node dissection . Nerve-sparing node dissection may evade this problem, but may be less useful in preventing relapse . As noted previously, some of these patients will relapse and need to undergo chemotherapy . This group of patients will need to be followed closely over the next few years after the operation.
Those patients selected for adjuvant chemotherapy because of their high risk of relapse require two cycles of combination chemotherapy. Only half these patients can benefit, as only these would in fact have relapsed. Thus half the patients would be unnecessarily exposed to the potential toxicities of cytotoxic drugs, which is undesirable. The treatment usually recommended is two cycles of BEP . Whilst careful management should preclude most short-term problems, impaired hearing due to auditory toxicity may still reduce the quality of life and even single doses of bleomycin can precipitate pulmonary fibrosis. Late adverse effects may include the development of second neoplasms from etoposide. Vascular toxicity from bleomycin, perhaps exacerbated by cisplatin, is usually limited to small-vessel disease, but cardiac ischaemia has been reported and may assume importance as the patient cohort ages. Most reported second malignancies after etoposide have occurred when the drug has been used at a dosage of ≥2 g/m2 and in combination with drugs that inhibit DNA repair . This dosage would not be achieved by patients having two cycles of conventional BEP (720 mg/m2 ) but safe levels of etoposide exposure are unknown . Klepp et al.  favour adjuvant chemotherapy for patients with clinical stage I and VLI in the primary tumour, and proposed two adjuvant courses of cisplatin, vinblastine and bleomycin. They point out that they are proposing an adjuvant treatment not containing etoposide, to avoid or reduce the risk of leukaemia in these patients. We do not routinely recommend adjuvant chemotherapy for patients regarded to be at high risk of relapse, although the option is discussed with these patients. For patients electing to undergo adjuvant therapy it may be preferable to offer two courses of adjuvant cisplatin, vinblastine and bleomycin, rather than BEP.
There are reports of the psychological effects that cancer and its treatments have on young patients [9,23]. Proponents of RPLND and selective adjuvant chemotherapy for high-risk cases propose that their management protocols enable follow-up visits to be minimized, thus avoiding such psychological trauma. We question this; relapse is still possible and patients treated by either of these techniques require a relatively intense follow-up. Indeed, the follow-up schedule recommended by the MRC group  for those patients who have had adjuvant chemotherapy is almost identical with the recommendations that we offer to our surveillance patients, i.e. 26 outpatient visits with tumour marker tests and chest X-rays, and a similar number of CT scans over the first 5 years of follow-up. It is unlikely that follow-up after adjuvant chemotherapy would be any less stressful psychologically than that imposed upon the patients undergoing surveillance.
Thus, there are three modes of management available for this highly curable malignancy. Each technique produces an excellent survival outcome, but each has its merits and each its problems. Therefore, selection of the appropriate management must depend upon the views of the clinician and wishes of the patient. There are significant disadvantages to surgery or the use of etoposide in the adjuvant setting. Selective RPLND has reduced but not eliminated ejaculatory problems. The risk of etoposide can be eliminated by replacing this drug, but the use of vinblastine instead may enhance neurotoxicity. In conclusion, we believe that orchidectomy and surveillance with combination chemotherapy at relapse has the least overall adverse effects upon this cohort of patients and is the optimal treatment.