Germ cell tumors of the testis have an overall survival rate greater than 90% as a result of a successful multidisciplinary approach to management. Late relapse affects a subset of patients however, and tends to be chemorefractory and the overall prognosis is poor. Surgery is the mainstay in management of late relapse but salvage chemotherapy can be successful. In this review, the clinical presentation and detection of late relapse, clinical outcomes, and predictors of survival in late relapse and the importance of a multidisciplinary treatment approach for successful management of late relapse are discussed.
Key points
- •
Late relapse affects approximately 1% to 6% of patients with germ cell tumors and may develop after more than 30 years, necessitating long-term follow-up of patients with germ cell tumor.
- •
The most common site of late relapse is the retroperitoneum, possibly related to suboptimal treatment of the retroperitoneum at the time of initial diagnosis and management.
- •
The histology of late recurrence is distinct, with more aggressive biology, including somatic malignant transformation, than that found at retroperitoneal lymph node dissection for primary germ cell tumor.
- •
The overall prognosis of late relapse is poor, and treatment requires an experienced multidisciplinary team, and most often requires surgery.
Introduction
Germ cell tumors (GCTs) of the testis have favorable prognosis, with overall survival rates higher than 90% for all stages when properly managed by a multidisciplinary approach. Because most GCT relapses occur within 24 months, it was historically considered that 2-year disease-free survival was equivalent to cure. As the experience with GCT accrued, it was noted that late relapse (LR), defined as relapse after a 2-year disease-free interval in the absence of a second primary tumor, was a distinct clinical entity. The incidence of LR in patients who have achieved complete response (CR) is reported to be 1% to 6%, although because most reports are from referral centers, it is difficult to precisely define the incidence. In a pooled analysis of 7 studies comprising 5880 patients, the estimated incidence of LR is 3.2% in nonseminomatous GCT (NSGCT) and 1.4% in seminoma, and based on referral patterns, the incidence of LR may be increasing. Factors associated with LR of GCT include the presence of bulky adenopathy at the time of initial presentation and teratoma at postchemotherapy retroperitoneal lymph node dissection (RPLND).
LR is a distinct clinical entity based on clinical presentation, pattern of histology, and poor overall survival, and the successful management of LR is one of the important challenges of contemporary management of testicular GCT. In this review, the data regarding the clinical presentation and detection of LR, clinical outcomes, and predictors of survival in LR are presented, and the importance of a multidisciplinary treatment approach for successful management of LR is discussed.
Introduction
Germ cell tumors (GCTs) of the testis have favorable prognosis, with overall survival rates higher than 90% for all stages when properly managed by a multidisciplinary approach. Because most GCT relapses occur within 24 months, it was historically considered that 2-year disease-free survival was equivalent to cure. As the experience with GCT accrued, it was noted that late relapse (LR), defined as relapse after a 2-year disease-free interval in the absence of a second primary tumor, was a distinct clinical entity. The incidence of LR in patients who have achieved complete response (CR) is reported to be 1% to 6%, although because most reports are from referral centers, it is difficult to precisely define the incidence. In a pooled analysis of 7 studies comprising 5880 patients, the estimated incidence of LR is 3.2% in nonseminomatous GCT (NSGCT) and 1.4% in seminoma, and based on referral patterns, the incidence of LR may be increasing. Factors associated with LR of GCT include the presence of bulky adenopathy at the time of initial presentation and teratoma at postchemotherapy retroperitoneal lymph node dissection (RPLND).
LR is a distinct clinical entity based on clinical presentation, pattern of histology, and poor overall survival, and the successful management of LR is one of the important challenges of contemporary management of testicular GCT. In this review, the data regarding the clinical presentation and detection of LR, clinical outcomes, and predictors of survival in LR are presented, and the importance of a multidisciplinary treatment approach for successful management of LR is discussed.
Clinical presentation and detection of late relapse
Detection of Late Relapse
LR is most commonly detected during routine follow-up of patients with GCT by detection of increased serum α-fetoprotein (AFP) or human chorionic gonadotropin (HCG) levels, radiographic findings, or clinical symptoms. In the 3 largest series of patients with LR, serum AFP is more commonly increased than HCG ( Table 1 ). In the pooled analysis by Oldenburg and colleagues, this trend holds true: of 426 patients with LR, 207 (49%) had increased serum AFP and 100 (24%) had increased serum HCG levels.
| Series | Total Number of Patients | Number of Patients with | ||
|---|---|---|---|---|
| Clinical Symptoms (%) | Increased AFP Levels (%) | Increased HCG Levels (%) | ||
| George et al, 2003 | 83 | 55 (66) | 43 (52) | 8 (10) |
| Sharp et al, 2008 | 75 | 45 (60) | 36 (48) | 18 (24) |
| Dieckmann et al, 2005 | 122 | 36 (30) | 48/92 (52) | 23/78 (29) |
Clinical symptoms including back pain or palpable abdominal mass may indicate LR in patients with GCT, although many patients present without symptoms. Dieckmann and colleagues reported on 122 patients with LR, of whom 30% were symptomatic and 61% were detected incidentally at routine follow-up. In the Indiana and Memorial Sloan Kettering Cancer Center (MSKCC) series, more than 60% of patients were symptomatic at time of LR. Early detection of LR before the onset of clinical symptoms may be critical to outcomes, because the presence of clinical symptoms at LR was associated with poorer cancer-specific survival (CSS) (hazard ratio [HR] = 4.9; 95% confidence interval [CI], 1.6–15.2) and overall survival (HR = 3.74; 95% CI, 1.3–11.3) in the MSKCC and Indiana series, although this was not confirmed in 2 other series.
Time to Late Relapse
The median time to LR from CR to primary treatment is approximately 4 to 7 years. Approximately 40% to 50% of LRs occur within 2 to 5 years of initial therapy and 50% to 60% present after more than 5 years. More than 30% of LRs may present more than 10 years after CR, with the longest reported time to relapse 37.7 years. The recognition that many LRs can present more than a decade after initial treatment underscores the critical importance of lifelong follow-up in patients with GCTs. Guidelines for follow-up are provided by the National Comprehensive Cancer Network, which recommends follow-up, including physical examination, serum tumor markers (STMs), and imaging, with decreased frequency over time corresponding with a decreased risk of relapse and annual follow-up starting at year 5. Computed tomography (CT) is the standard for surveillance imaging, although there is some concern that radiation exposure from surveillance CT scans may increase the risk of secondary malignancy. Use of magnetic resonance imaging and low-dose CT has been considered to lower the risks of radiation exposure; however, these techniques are unproved in surveillance of GCT. Risk-stratified surveillance has been suggested as a means to minimize radiation exposure, because the risk of recurrence decreases over time.
Patterns of Late Relapse
The retroperitoneum is the most common site of LR, accounting for approximately 45% to 70% of the LRs, regardless of initial histology, treatment of primary, or clinical presentation ( Table 2 ). Other sites of LR can include the lungs, mediastinum, retrocrural space, neck/supraclavicular nodes, liver, brain, and pelvis. Multiple sites of LR have been reported in 36% to 52% patients. Factors associated with development of LR include increased tumor burden at the time of primary therapy, absence of primary RPLND, and the presence of teratoma after primary therapy.
| Series | Total Number of Patients | Number of Patients with Late Relapse in | |||
|---|---|---|---|---|---|
| Retroperitoneum (%) | Lungs (%) | Mediastinum (%) | Neck (%) | ||
| George et al, 2003 | 83 | 39 (47) | 21 (25) | 8 (10) | 7 (8) |
| Sharp et al, 2008 | 75 | 54 (72) | 15 (20) | 9 (12) | 8 (11) |
| Dieckmann et al, 2005 | 122 | 71 (58) | 8 (7) | 14 (11) | 10 (8) |
| Gerl et al, 1997 | 25 | 12 (48) | 2 (8) | 8 (32) | 1 (4) |
Clinical outcomes and predictors of survival in late relapse
Disease recurrences more than 24 months after CR to primary treatment in chemotherapy-naive patients tend to show similar outcomes to patients presenting de novo with metastatic GCT. LR after CR to treatment including chemotherapy portends a worse prognosis. Because of the disparate clinical outcomes based on primary treatment, we have advocated for defining LR only in the postchemotherapy setting. For the purposes of this review, we consider LR based on initial management of GCT and disease stage at presentation.
Late Relapse in Patients Initially Managed with Surveillance
Surveillance is a recognized management strategy for patients with clinical stage (CS) I GCT, and despite a low risk of relapse, most patients are salvaged. In a recent analysis by Kollmannsberger and colleagues, 173 of 1344 (13%) patients with CSI seminoma relapsed overall, with 43 (3.2%) relapsing after 2 years, with all LRs either radiosensitive or chemosensitive. In a separate analysis of 638 patients followed with surveillance for CSI seminoma from 4 centers, there were 38 relapses after 2 years (6% LR rate), with the latest relapse at 12 years. Risk factors for relapse included size of primary tumor, rete testis invasion, and lymphovascular invasion (LVI), although these risk factors have not been validated in other studies. In a single-institution series at Princess Margaret Hospital, 203 men with CSI seminoma were managed by surveillance, and the risk of relapse at 5 years was 4%.
The rate of LR after surveillance for CSI NSGCT is low, with some series having no cases of LR. In the recent report by Kollmannsberger and colleagues, of 1139 patients with CSI nonseminoma, 221 (19%) relapsed on surveillance, including 18 (1.6%) after 2 years. Seventeen of the 18 LRs were cured with standard therapy. In another study of 233 patients who underwent surveillance for CSI NSGCT, 3% (7 patients) had LR and all patients were successfully salvaged with chemotherapy with or without RPLND for disease-specific survival of 100% at median follow-up 52 months (range 3–136 months). With appropriate surveillance and early detection, it is possible to salvage most cases of LR.
Rice and colleagues recently reported on a series of patients with LR after surveillance for CSI NSGCT referred to Indiana University. The median time to diagnosis of LR in 28 patients was 48.5 months but occurred as late as 27 years. Thirteen of 28 patients presented with increased STMs, with increased HCG levels in 9 of them. In contrast, AFP level is more commonly increased in all patients with LR. The retroperitoneum was the dominant site of relapse in these patients, with 26 of 28 patients (94%) experiencing retroperitoneal relapse. The rate of LR in patients with CSI GCT is unknown, because these patients were largely referred from other institutions. Eight of 9 patients who had primary RPLND for LR in this series had viable malignancy, including 2 with secondary somatic malignancy and 1 with teratoma. Of the 19 patients who had postchemotherapy RPLND for LR of CSI GCT, 5 had viable malignancy, 4 had teratoma, and 10 had fibrosis in the RPLND specimen. Of the 13 patients with viable malignancy at RPLND (primary and postchemotherapy), there were 3 with secondary somatic malignancy and 4 died of their disease at median follow-up 21 months (range 1–237 months). Although 86% had no evidence of disease (NED) at last follow-up, these results highlight the potential of even CSI GCT to be lethal and underscores the need for diligent surveillance protocols.
Late Relapse After Adjuvant Radiotherapy or Chemotherapy for Clinical Stage I Seminoma
The LR rate after adjuvant therapy for CSI seminoma is very low. Only 4 of 2466 (0.2%) of patients had relapses after 3 years in results from 3 pooled randomized trials of 20 versus 30 Gy radiotherapy, para-aortic versus dogleg radiotherapy, or radiotherapy versus single-dose carboplatin. However, median follow-up of these studies is short, potentially limiting conclusions about LR.
In studies with longer follow-up, the overall relapse rate after adjuvant radiotherapy for CSI seminoma is reported to be 1.4% to 6.9%. In a series of 194 patients treated with adjuvant radiotherapy for CSI seminoma at Princess Margaret Hospital, 11 experienced relapse with 2 relapses after more than 4 years. In a recent study from Mayo Clinic, in patients who received adjuvant radiotherapy for CSI seminoma, only 3 of 199 patients relapsed (1.5%) with 1 LR (median follow-up 13.7 years; range 0.1–37 years).
Powles and colleagues reported an analysis of patients who received adjuvant single-agent chemotherapy for CSI seminoma. In this study of 199 patients with median follow-up of 9 years (range 0.1–20.1 years), 4 patients (2%) had LR. All patients with LR were successfully salvaged with cisplatin chemotherapy.
Late Relapse in Patients Initially Presenting with Low-Stage Nonseminomatous Germ Cell Tumors Managed with Retroperitoneal Lymph Node Dissection or Chemotherapy
A standard option for treatment of CSI/IIa NSGCT is bilateral infrahilar RPLND, and infield recurrences are rare after properly performed RPLND. However, the retroperitoneum is the most common site of LR in patients with low-stage NSGCT primarily managed with RPLND, suggesting that an inadequately controlled retroperitoneum is a contributing factor and may be caused by inadequate initial management of the retroperitoneum with incomplete removal of lymphatic tissue as a result of surgical technique or modification of surgical templates. Of 19 patients initially treated with primary RPLND and referred to MSKCC for LR from 1990 to 2004, 11 (58%) had received chemotherapy (either adjuvant or for early relapse) before LR. Twelve of the 19 (63%) patients relapsed in the retroperitoneum, including 8 in the primary landing zone, suggesting that LR could have been prevented by proper management of the retroperitoneum at initial treatment. Five-year CSS after LR in patients treated with primary RPLND was 58% (95% CI, 31%–68%) in these patients. In a series of 35 patients with LR reported by Indiana University, primary RPLND was performed in 31 (89%), including 19 (61%) who relapsed in the retroperitoneum despite having received adjuvant chemotherapy for pathologic stage II disease. These results impress the importance of proper surgical management of the retroperitoneum and imply that 2 cycles of chemotherapy cannot compensate for inadequate RPLND.
Eggener and colleagues reported that the use of modified RPLND templates would result in residual extratemplate retroperitoneal disease in 3% to 23% of patients depending on the template used. Carver and colleagues showed that for men undergoing postchemotherapy RPLND for NSGCT, extratemplate disease is found in 4% to 32% of patients, depending on which modified template was used. In a series of 369 patients who had resection of postchemotherapy masses only (not full template), there were 37 (10%) with retroperitoneal relapse.
A single cycle of bleomycin plus etoposide plus cisplatin (BEP) chemotherapy for patients with CSI NSGCT has been advocated by some, with a reduction in relapse risk by approximately 90% compared with surveillance. In a recently published update of this series with median follow-up 7.9 years, there were no LRs in 255 patients with stage CSIa NSGCT without LVI and relapses in 8 of 258 (3.2%) patients with LVI including 3 with LR. Because even 2 cycles of BEP do not eliminate the risk of LR in the adjuvant setting after early relapse, it is likely that as the data from this trial mature, there will also be a predictable increase of LR.
Late Relapse in Patients Initially Presenting with Advanced Stage Germ Cell Tumors
Patients presenting with metastatic GCT (ie, CSIIb or higher) are typically managed with cisplatin-based chemotherapy alone or combined with surgery, although some have advocated treating low stage II patients with surgery alone. In the Indiana series there was a relapse rate of 35% in patients with low stage II disease treated with surgery alone but no relapses in patients with stage II disease who received 2 cycles of adjuvant cisplatin-based chemotherapy. In patients with pathologic stage II NSGCT who received primary RPLND and 2 cycles of adjuvant chemotherapy, there were no LRs in 87 patients with median follow-up of 8 years in the MSKCC series.
Ehrlich and colleagues reported a series of 141 patients with metastatic disease who had CR to induction chemotherapy alone, with an overall 9% rate of relapse, including 5 patients (3.5%) with LR (from 3 to 13 years) at median follow-up time 15.5 years, including 2 in the retroperitoneum. As mentioned earlier, the most common site of LR in patients treated with induction chemotherapy is the retroperitoneum. Sharp and colleagues analyzed 75 patients referred to MSKCC from 1990 to 2004 for LR and of the 45 treated with induction chemotherapy, 34 (76%) developed LR in the retroperitoneum, including 17 of 18 who did not have postchemotherapy RPLND. The 5-year CSS rate after LR in the 45 patients with LR after induction chemotherapy was 53% (95% CI, 36%–68%).
Using a combined regimen of induction chemotherapy and postchemotherapy RPLND seems effective at reducing the risk of relapse, with 5-year progression-free survival of 98%. Kondagunta and colleagues reported LR in 3 of 282 (1.1%) International Germ Cell Cancer Collaborative Group (IGCCCG) good-risk patients treated with 4 cycles of etoposide plus cisplatin induction chemotherapy. In a series of 418 patients reported by Gerl and colleagues who received cisplatin-based induction chemotherapy with or without surgery for advanced GCT, the cumulative risk of LR was 1.1% at 5 years and 4.0% at 10 years. Previous early relapse and poor-risk disease (Medical Research Council criteria) were associated with higher risk of LR.
It has proved difficult to predict the presence of fibrosis and necrosis after induction chemotherapy. Even in patients with complete radiographic response to induction chemotherapy, viable malignancy or teratoma can be found in up to 30% of patients. Complete resection of teratoma can convey several advantages to the patient. Teratoma is chemoresistant and may grow and invade adjacent structures or become unresectable. Furthermore, there is a risk of malignant transformation to non–germ cell malignancies, including sarcoma or carcinoma. The incidence of malignant transformation in the postchemotherapy setting is 3% to 6%. These tumors do not respond to cisplatin-based chemotherapy regimens like GCTs, and complete surgical resection is the best treatment option for these patients. Because it remains difficult to accurately predict fibrosis/necrosis and eliminate the risk of viable GCT or residual teratoma, a subset of patients who undergo induction chemotherapy alone or undergo incomplete RPLND may remain at risk for LR.
Histopathology of Late Relapse
The histology of the late recurrence is distinct, with more aggressive biology than that found at RPLND for primary GCT ( Table 3 ). Whereas teratoma is found in approximately 20% to 30% of node-positive primary RPLND specimens and approximately 40% of postchemotherapy RPLND specimens, teratoma is found in up to 60% of patients with LR. The incidence of somatic malignant transformation of teratoma is around 20% for patients with LR, which is significantly increased compared with the rate noted in the MSKCC primary RPLND series (0.4%; 2 of 550 patients) and postchemotherapy series (18 of 532 patients; 3%). The incidence of viable GCT in LR is approximately 60% to 70% and significantly higher than the 10% found at postchemotherapy RPLND. Yolk sac tumor is the most common nonteratomatous germ cell element in LR, but all other types can be found. Because of the high rate of teratoma, somatic malignant transformation and chemoresistant viable GCT, late recurrence is not likely to respond to chemotherapy and is best treated with surgical excision. Attempts to define the molecular basis for chemoresistance have been unsuccessful.
Related posts:
Epidemiology and Diagnosis of Testis Cancer
Intratubular Germ Cell Neoplasia of the Testis, Bilateral Testicular Cancer, and Aberrant Histologies
Management of Low-Stage Testicular Seminoma
The Role of Postchemotherapy Surgery in Germ Cell Tumors
Chemotherapy for Good-Risk Nonseminomatous Germ Cell Tumors
Desperation Postchemotherapy Retroperitoneal Lymph Node Dissection for Metastatic Germ Cell Tumors
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
