- To examine the management and outcomes of patients with stage I seminoma and to relate these to overall treatment burden.
etoposide and cisplatin
bleomycin, etoposide and cisplatin
There has been an increasing incidence worldwide of testicular germ cell tumours, such that in 2008 there was an estimated 52,000 cases . Approximately 50–60% of patients have seminoma and these now represent most of this increased incidence . As 80% of patients with seminoma present with clinical stage I disease, stage I seminomas now represent the largest proportion (>50%) of newly diagnosed patients with testicular cancer. Historically, radiation therapy (RT) was used as routine management in the adjuvant setting after radical orchiectomy, with ∼5% relapse rate (usually outside of the irradiated volume) . The introduction of cisplatin-based chemotherapy (ChT), the recognition of the late adverse effects of RT , improvements in diagnostic imaging and the experience gained with surveillance in stage I non-seminoma have all led to an increase in the use of surveillance as a management strategy in stage I seminoma. Although patients who are managed with surveillance have a relapse rate of ∼15%, these relapses are successfully salvaged with subsequent treatment [3, 5-7]. Surveillance has become accepted as a standard, and increasingly preferred, management option over the past 10 years in stage I seminoma [8-10]. Despite this, there remain concerns that: 1) while an expectant strategy leads to the avoidance of unnecessary treatment in most patients and thereby the problem of treatment-related toxicity, this strategy might lead to an excess overall burden of treatment; 2) surveillance may result in an increased use of multiple treatment methods, particularly in those who relapse [11-13]; and 3) surveillance is associated with the need for repeated imaging of the retroperitoneal lymph nodes for at least 5 years which is not required for patients who have adjuvant RT (although surveillance of other potential sites of relapse are still required).
To address these issues, we reviewed the experience at a single institution and compared the treatment burden of patients with stage I seminoma managed by surveillance and adjuvant RT.
After approval from our institutional ethics review board, data were gathered from a prospectively maintained database. We retrospectively identified 764 patients with stage I seminoma managed at the Princess Margaret Hospital, between 1981 and 2004 to allow for a minimum of 5 years of follow-up. Patients managed with surveillance were initially part of a phase II study, which began in the 1980s, and this then became the standard approach offered to patients from the early 1990s. The choice of surveillance or adjuvant RT was determined by patient (and physician) preference and the decision was not based on any risk adaptation. The surveillance and adjuvant RT protocols used have evolved since the beginning of the time period under study and have previously been published [6, 14]. The follow-up schedule of both management strategies was identical except for the omission of CT imaging of the abdomen and pelvis for patients who received adjuvant RT. No patients received adjuvant carboplatin ChT.
All patients who relapsed were identified and data were extracted, including the relapse date, site, and detailed specifics of treatment at relapse, as well as information regarding any subsequent relapse. Locoregional relapse was defined as infradiaphragmatic retroperitoneal nodal disease (para-aortic, pelvic, inguinal adenopathy), while distant relapse was defined as supradiaphragmatic nodal disease (supraclavicular, mediastinal or other distant adenopathy) or systemic disease (lung, liver, soft tissue, bone). The treatment for first locoregional relapse in patients managed with surveillance was generally infradiaphragmatic RT alone, where gross nodal disease, e.g. a solitary enlarged node, was treated to a higher dose than the surrounding nodal regions at risk of harbouring microscopic disease. The exception was where combination ChT, usually etoposide and cisplatin, with bleomycin (BEP), or without bleomycin (EP), was used for more extensive locoregional nodal recurrence (e.g. solitary nodal disease >5cm or multiple grossly involved nodes involving both para-aortic and pelvic region) or for contraindication to RT.
Our institution has previously published the long-term outcomes of patients with stage I seminoma between 1981 to 1991 . In that study, a total of 364 patients with stage I seminoma were reviewed, of whom 194 were treated with adjuvant RT and 172 were placed on surveillance. The present study updates these results with an additional 13 years of experience. Our group has also recently reported on the value of serum tumour markers in patients with stage I disease, although with a slightly different time frame from that of the current study, and that study included patients who underwent surveillance only .
The optimum follow-up strategy for patients with stage I seminoma has not been determined and has evolved over the past 30 years. The follow-up for the patients in the present cohort consisted of 20 ambulatory care visits over 10 years with CT of the abdomen and pelvis at each visit, with chest X-rays and blood evaluation performed less frequently (Fig. 1). The follow-up schedule for patients who received adjuvant RT was similar, except that CT was omitted. Of the 764 patients, 122 (16%) patients were lost to follow-up by the 3rd year.
Adjuvant RT consisted of megavoltage radiation using parallel-opposed fields with appropriate shielding. Traditionally, the treatment volume consisted of the para-aortic nodes and the ipsilateral ilac nodes, by encompassing a volume from the top of T11 to the bottom of the obturator foramen, but from 2003 onwards, the inferior border of the RT field was placed at the cranial aspect of the acetabulum. Treatment of gross nodal disease at relapse was as above, with the addition of a boost to the involved nodes.
We calculated total treatment burden after initial radical orchiectomy according to the management strategy by identifying episodes of treatment. In our primary analysis, surveillance was assigned a treatment burden value of 0; a value of 1 unit (or episode of treatment) was assigned to one course of RT, to surgery (node dissection) and to one course of ChT. A secondary analysis of treatment burden was also performed by assigning a treatment burden value of 1 to each individual cycle of ChT, e.g. three course of BEP were assigned a value of 3 and four cycles of EP were assigned a value of 4. We also performed a tertiary analysis by assigning a treatment burden value of 0.5 to surveillance.
The late effects of treatment were not included in the consideration of treatment burden for the present study. We chose to keep the treatment burden calculation as simple as possible, in recognition of the fact that there may be many ways of calculating this but there is no standardized methodology.
Overall survival probabilities were produced using the Kaplan–Meier method. The probability of relapse over time for patients in the surveillance and RT groups was produced using the cumulative incidence function. sas version 9.2 was used for descriptive analyses and to estimate the probability of all-cause mortality.
Of 764 patients with stage I seminoma, 484 followed a surveillance strategy and 280 underwent adjuvant RT; their management is shown in Fig. 2. The actuarial overall survival rates for all patients were 98 and 95% at 5 and 10 years, respectively. The 5- and 10-year overall survival rates for the surveillance group were 98.6 and 97.7%, respectively, and were similar to those of the adjuvant RT group, which were 97.2 and 91.4%. The median follow-up for the surveillance and adjuvant RT groups was 6.6 and 8.5 years, respectively. The use of adjuvant radiotherapy declined over the period studied and, more recently from 2000 to 2004, 4% of patients underwent adjuvant RT (Fig. 3).
In the surveillance group, 72 (15%) patients relapsed (Fig. 4) with a median (range) time to relapse of 14 (3–108) months. Of these, 57 of the relapses occurred within 2 years, 10 relapses after 2–5 years and 5 relapses occurred after 5 years. Seventy-one patients had locoregional nodal relapse and one patient had distant relapse in the mediastinum in addition to inguinal node relapse. Of these, initial salvage treatment consisted of RT in 56 patients, ChT in 15 patients and one patient had retroperitoneal lymph node dissection because of a previous contralateral stage I seminoma treated with adjuvant RT before the period under study. The median (range) RT dose was 35 (25–37) Gy. Of the 15 patients who received ChT, 13 were treated with EP, one patient was treated with BEP and one patient was treated with other platinum-based ChT (cyclophosphamide, vinblastine, actinomycin, bleomycin, cisplatin). Six patients (five of whom had been treated with RT and one with ChT) subsequently developed second relapse after initial salvage treatment at a median (range) time of 16 (11–26) months from orchiectomy; five had disease in the mediastinum and the other had nodal disease in the left supraclavicular fossa. Further ChT in these patients consisted of EP in four patients, BEP in one patient and BIP (bleomycin, ifosphamide, cisplatin) in the one patient who had received ChT for initial relapse. Overall, 3% of patients in the surveillance group received ChT and five patients (1%) had both RT and ChT at some point during their management.
Of the 280 patients who received adjuvant RT, 14 (5%) relapsed at a median (range) time of 15 (5–72) months (Fig. 4). Of these, 11 relapses occurred within 2 years, two relapses at 2–5 years and one relapse after 5 years. Ten patients who relapsed with supradiaphragmatic disease (mediastinum, supraclavicular fossa, lung) were treated with ChT and, of four patients who had isolated inguinal relapse, three were treated with local RT and one with surgical excision during the early 1980s. EP was used in seven patients while three had other platinum-based ChT before (B)EP became standard. There was no second relapse in any of the patients in the adjuvant RT group. Overall, ChT was required in 10 (3.6%) patients in the adjuvant RT group and all of these patients had both RT and ChT. In this group of patients a total of seven developed a contralateral testis tumour: four in the surveillance group and three in the adjuvant RT group.
At last follow-up 24 patients had died (eight in the surveillance group and 16 in the RT group). Of these, only one patient had died from seminoma and all others had died from unrelated causes (no patients died from treatment-related causes). The one patient that died from seminoma was on surveillance and recurred shortly after surgery. He had two recurrences and died 2 years after surgery. He received ChT for each recurrence.
When one course of ChT (e.g. 3 x BEP) was considered to be one treatment episode, this resulted in 78 treatment episodes in the surveillance group, or 0.16 treatment episodes per patient based on a denominator of 484 surveillance patients. In the adjuvant RT group, there were 294 treatment episodes or 1.05 treatment episodes per patient. In the secondary analysis, assuming one cycle of ChT (e.g. 1 x BEP, 1 x EP) to be equivalent to one treatment episode, the result was an increase in the total number of treatment episodes to 151, while the number of treatment episodes per patient was 0.31 for the surveillance group. Similarly, in the adjuvant RT group, this resulted in 320 treatment episodes or 1.14 treatment episodes per patient (Fig. 5).
In a tertiary analysis, by assigning an arbitrary value of 0.5 for surveillance, the numbers of treatment episodes per patient in the surveillance group were calculated to be 0.66 (assuming one course of ChT to be equivalent to one treatment episode) and 0.81 (one cycle of ChT being equivalent to one treatment episode). In both cases, the number of treatment episodes per patient in the surveillance group was lower than that for adjuvant RT (1.05 vs 1.14).
The present study compared the treatment burden in patients with stage I seminoma managed either with surveillance or adjuvant RT over a 25-year period, and updates the results and long-term outcomes of patients with stage I seminoma treated at our institution . The relapse rate in the two cohorts was similar to that reported in the literature [16-18]. The treatment burden, as represented by treatment episodes per patient, was considerably lower in the surveillance group, by a factor of 1.4–6.6, depending on definition, than in the adjuvant RT group, with similar survival rates.
Multiple non-randomized studies have shown that surveillance is a safe strategy and that relapse can be salvaged effectively with RT or combination ChT [5, 19, 20]; however, there has been some reluctance to adopt this strategy as the standard of care despite the presence of surveillance as an option for management in international consensus guidelines [8-10, 21]. A similar situation has arisen in other areas of oncology, such as in breast cancer, where there was an initial reluctance to adopt breast-conserving therapy ; there was minimal increase in its use despite a large body of evidence demonstrating its efficacy. An explanation for this in seminoma may have been the rarity of the disease, resulting in a lack of familiarity with surveillance and concerns regarding the safety of not offering adjuvant treatment. In some countries, adjuvant RT has been replaced by adjuvant ChT (single-agent carboplatin) [23, 24] with the philosophy that the latter is associated with less acute toxicity and possibly less late toxicity, and is therefore more acceptable; however this change in adjuvant treatment does not address the issue of the lack of a robust prognostic model to identify those patients who actually need treatment.
One concern expressed with regard to surveillance in patients with seminoma is the issue of whether such patients, when relapse occurs, will do so with a requirement for more intensive treatment than would have been the case if adjuvant therapy had been used. This issue of so-called ‘treatment burden’ has been examined in a single study of 164 patients undergoing surveillance, of whom 22 (13%) relapsed. Of these, six of 13 patients treated with RT and two of nine treated with ChT for relapse had a second relapse resulting in an overall second relapse rate of 36% . It should be noted that this was a small study and the majority of patients were managed in the 1980s with older imaging techniques during their surveillance. By contrast, our study shows that the proportion of patients who required ChT was similar in the surveillance and the adjuvant RT groups; thus, the initial treatment burden for patients was reduced when surveillance was used as the initial management strategy and subsequent treatment burden did not appear to be any greater than for patients in the adjuvant RT group.
A large proportion of our patients underwent RT as salvage treatment at first relapse and we observed a 9% rate of second relapse, consistent with that seen in two Danish surveillance series, where the incidence rates of second relapse, after RT had been used as the treatment for initial relapse, were 7% and 11% [5, 17]. These relapse rates appear to mirror those seen in patients who present with de novo stage IIA/B seminoma treated with RT [25-28]. Patients who subsequently received ChT salvage treatment for second relapse appear to have a greater risk of second malignancy compared with those salvaged with RT or ChT alone, as suggested by epidemiological studies [29-31]. Avoidance of both RT and ChT use in relatively young patients is worthwhile to avoid long-term toxicity, but this should be balanced against the toxicity of combination ChT if a policy of ChT were to be applied for all patients at first relapse.
The present study did not address the option of single-agent carboplatin as adjuvant therapy. Randomized data have shown a 5.3% relapse rate with adjuvant carboplatin vs 4% with adjuvant RT at 5 years . We did not treat any patients with carboplatin in this cohort but, with our methodology, adjuvant carboplatin treatment would result in all patients receiving at least one episode of treatment and the burden of treatment (as measured in the present study) would presumably be similar to that of adjuvant RT and higher than that of surveillance. We acknowledge that assigning a treatment burden value of 1 for carboplatin is arbitrary; however, this analysis was to illustrate the concept that adjuvant treatment has a higher burden than surveillance and was not intended to compare carboplatin with other forms of ChT.
One potential advantage of carboplatin discussed in the MRC/EORTC study  has been that of a lower risk of primary germ cell tumours in the contralateral testis with adjuvant carboplatin as opposed to RT. Although the results are promising, longer follow-up has shown that the contralateral risk is not completely eliminated and the possibility that later occurrences of contralateral primary tumour may be seen with longer follow-up .
At our institution combination ChT is typically used for more extensive locoregional nodal recurrence. Interestingly, we performed a hypothetical calculation of treatment burden, assuming that all patients who developed a first locoregional relapse were to be treated with three cycles of BEP ChT. In this analysis, the number of treatment episodes per patient in the surveillance group was 0.45, more than twofold lower than that for adjuvant RT (1.15). Assuming three cycles of BEP is equivalent to one treatment episode, the number of treatment episodes per patient is 0.15 for surveillance and 1.05 for adjuvant RT.
Although the present study included a large cohort of patients from a prospectively collected database, it is limited by its retrospective nature. In addition, over the time period studied, there was an evolution of management policy. A proportion of patients was lost to follow-up and did not contribute to the estimate of relapse; however, this should not have substantially affected our conclusions, as rates of relapse in our study were similar to others in the literature.
Our definition of ‘treatment burden’ was intended to quantify the number of treatments that a patient undergoes to remain relapse-free from seminoma, thereby determining which management strategy results in patients receiving more therapy, whether it be RT, ChT or surgery; therefore, in the present study, a greater number of treatments (overall) may signify a higher burden on the patient. Nevertheless, we acknowledge that this definition may not result in equivalent toxicity profiles and/or quality of life and that the perception of burden may vary significantly with each patient, especially in regard to the toxicities for the different treatment methods. Our scoring system also did not account for factors that would contribute to the burden of surveillance, such as the additional abdominal CT scans and any potential psychological impact of not having treatment. Patient input would be helpful, but this was beyond the scope of the present study, given its retrospective nature. Similar issues may, however, be faced by patients, regardless of the post-orchiectomy management as adjuvant therapy does not completely eliminate risk of relapse or the need for follow-up in these patients. Interestingly, a recent abstract on the cost of adjuvant management in stage I seminoma has suggested that the Medicare reimbursement for surveillance in the USA ($10,643) is similar to that of other adjuvant treatments (para-aortic RT $11,678, dog-leg radiation $9662), with carboplatin being the lowest ($7870) . It should be noted, however, that these costs may differ in different healthcare systems.
Our surveillance schedule may include more frequent visits and imaging than that of other centres. More recently we have moved to a less intensive schedule in keeping with others in the published literature; however, with less frequent surveillance, it may be assumed that the more extensive disease at relapse seen with less frequent imaging may lead to more intensive treatment for salvage [5, 34].
We acknowledge that the dose of radiation used in the adjuvant setting as opposed to that used in definitive treatment is generally lower (20–25 Gy vs 30–36 Gy) and was not included in the treatment burden calculation. The overall RT treatment time in our institution is the same regardless of whether the intent is adjuvant or definitive RT, as the extra dose to gross nodal disease is delivered concurrently. In a curable disease such as seminoma, the late effects of RT or ChT are important and factoring these into any treatment burden calculation might be useful; however, although many patients (particularly those treated with RT) were treated >20 years ago, the relatively small number of patients who required treatment in the surveillance group and the median follow-up for both groups (<10 years) was too short for a meaningful assessment of this.
In conclusion, as defined in this study, treatment burden in patients with stage I seminoma managed with a surveillance policy was lower than in patients treated with adjuvant RT. This study supports surveillance as the preferred management after radical orchiectomy. The selective use of RT as salvage therapy at first relapse for patients managed by surveillance results in a similar requirement for combination ChT to that for patients receiving adjuvant RT.