Management of residual mass in nonseminomatous germ cell tumors following chemotherapy

Abstract

Advanced stage nonseminomatous testis cancer is commonly treated with chemotherapy and surgical resection. Patients with retroperitoneal residual masses >1cm following induction chemotherapy with normalized tumor markers should undergo a post-chemotherapy retroperitoneal lymph node dissection. Post chemotherapy retroperitoneal residual mass less than 1 cm with normal markers may be considered as complete response, although the possibility of residual teratoma and viable germ cell tumor are not definitively ruled out. Excellent long term disease free survival following surveillance may justify this option as the treatment of choice in this cohort of patients.

Introduction

Modern risk-adapted systemic chemotherapy with or without postchemotherapy surgery for patients with disseminated testicular nonseminoma germ cell tumor (NSGCT) results in superb outcomes [Rabbani et al. 1998]. Depending on the stage at diagnosis, 20–50% of patients who undergo induction chemotherapy for metastatic germ cell tumor (GCT) have significant residual retroperitoneal disease [Hendry et al. 2002]. All patients who harbor residual masses ≥1 cm after chemotherapy should undergo postchemotherapy retroperitoneal lymph node dissection (PC-RPLND) due to the significant risk of finding mature teratoma in 40–45% of cases and viable GCT in 10–15% of patients [Heidenreich et al. 2009]. Patients with teratoma in the PC-RPLND specimen have a disease-free survival of 80% while those with viable GCT have a decreased chance of survival [Carver et al. 2007b; Suzuki et al. 1999]. Low-risk patients who obtain a complete serologic remission and no or minimal radiographic residual after chemotherapy can be safely observed without adjunctive regional surgery [Ehrlich et al. 2010; Kollmannsberger et al. 2010] A low risk of relapse is reported in these patients and those who relapse remain curable [Saxman et al. 1998; Debono et al. 1997].

Indications for PC-RPLND

PC-RPLND for residual GCT was first reported over 30 years ago [Comisarow and Grabstald, 1976]. The traditional indication for PC-RPLND has been normalized serum tumor markers with residual retroperitoneal mass. In rare cases, patients who fail induction and salvage chemotherapy and have elevated tumor markers with a resectable mass are candidates for a PC-RPLND. Patients who have normalized their markers with complete resolution of retroperitoneal adenopathy are considered at low risk of relapse and generally do not require surgery, although this has been the topic of some debate for over 20 years [Saxman et al. 1998; Debono et al. 1997].

It is often difficult to measure the exact size of the residual nodal tissue following chemotherapy, since nodes are often matted together in the retroperitoneum. Moreover, the imaging criteria of the normal retroperitoneum are rarely discussed and there is no definite consensus on nodal size criteria [Fossa et al. 1989]. It is therefore incumbent on the expert surgeon to review the prechemotherapy and postchemotherapy imaging to determine the extent and size of residual disease. The European Germ Cell Cancer Consensus Group [Krege et al. 2008], Indiana University [Beck and Foster, 2009] and Oregon/UBC study [Kollmannsberger et al. 2010] suggest that patients who achieve remission (residual retroperitoneal lesion <1 cm) do not require PC-RPLND. The group from Indiana University reported their long-term experience with 141 patients who achieved a complete remission (CR; defined as normalization of serum tumor markers and residual mass <1 cm) who underwent surveillance following induction chemotherapy. Median follow up was 15.5 years. Overall 12 patients (9%) recurred at a median of 11 months (6 in the retroperitoneum), eight of whom had no evidence of disease at a median follow up of 11 years. Four patients died, all within 12 months of induction chemotherapy. Importantly, no patients recurred with pure teratoma. Five patients had late relapse (two of which were in the retroperitoneum) all of whom had no evidence of disease at long-term follow up. The sole predictor of relapse and cancer-specific survival (CSS) was International Germ Cell Consensus Classification (IGCCC) risk classification. The presence of teratoma in the orchiectomy specimen was neither predictive of relapse nor CSS. In the Oregon/UBC series, there were 276 patients of whom 161 had CR or partial response (PR; defined as residual disease <1 cm), 151 with good-risk disease. Median follow up was 4.3 years and there were 10 relapses (6%), 8 of which were in the retroperitoneum and were cured with PC-RPLND. One was cured with chemotherapy and PC-RPLND and the other with relapse in the lung was cured with chemotherapy alone. All patients relapsed with IGCCC good-risk disease and the CSS was 100%. In the Princess Margaret Hospital experience, 129 patients underwent surveillance postchemotherapy after CR for a median of 7 years with only one patient dying of disease (99% CSS) and 92% recurrence-free survival [Kakiashvili et al. 2009]. The rationale to observe residual mass <1 cm in postchemotherapy NSGCTs are (1) postchemotherapy microscopic teratoma may be generally biologically benign, (2) the small relapse rates of observed cases do not justify surgery in all patients achieving CR and (3) patients that do relapse still remain curable. However, some authors advocate retroperitoneal lymph node resection in all cases of NSGCT irrespective of residual mass size after chemotherapy [Sheinfeld, 2003]. Their basis for this recommendation is twofold: (1) all patients with metastatic disease require chemotherapy to treat active cancer and surgery to eradicate microscopic teratoma, as these two components may comprise up to 50% of postchemotherapy residual masses [Carver and Sheinfeld, 2009]; (2) unresected microscopic teratoma has an unpredictable biologic behavior that may ultimately have a deleterious effect on outcome [Beck and Foster, 2009]. Microscopic teratomas, however, may remain clinically insignificant; in which case, resection would probably only add morbidity without improving outcome. Late relapse or malignant transformation of teratoma is very rare and not invariably fatal [Geldart et al. 2006]. Radiation exposure is another important issue to be considered as more strict retroperitoneal imaging is necessary in patients who undergo surveillance in comparison with those who undergo PC-RPLND. There is limited high-level evidence for the long-term impact of surveillance radiation exposure in this setting. Although some centers still recommend PC-RPLND for all testis cancer patients with retroperitoneal disease who responded to chemotherapy, other experts agree that PC-RPLND should be performed only if lymph nodes ≥1 cm persist following chemotherapy, even though the possibility of teratoma or viable cancer in smaller lesions cannot be ruled out. PC-RPLND should be performed 4–6 weeks following the last dose of chemotherapy in order to allow patients to recover their blood counts [Kakiashvili et al. 2009].

Patients who have an elevated yet stable serum tumor marker, teratoma in the primary tumor, and a postchemotherapy cystic mass in the retroperitoneum may also be candidates for PC-RPLND rather than salvage chemotherapy. Beck and colleagues showed that cystic teratomas may contain variably elevated levels of alpha feto-protein (AFP) and human chorionic gonadotropin and postulated that a leak into the bloodstream could explain the elevated, yet stable serum markers in this situation [Beck et al. 2004]. Teratomas can rarely invade the vena cava and present with a tumor thrombus that should not be mistaken for a deep venous thrombosis [Moore et al. 2006]. Resection of these masses along with tumor thrombectomy or even resection of the inferior vena cava is often curative [Duty and Daneshmand, 2009].

Prediction model

Prediction models have been used to refine the selection criteria of candidates for surgery. Ideally patients with postchemotherapy residual mass but high probability of benign tissue can continue surveillance, whereas patients with normal appearing postchemotherapy retroperitoneum but relatively low probability of benign pathology may benefit from resection [Vergouwe et al. 2007]. Over the past 15 years, several statistical models have been generated in an attempt to accurately predict the presence of necrosis in resected retroperitoneal specimens. Toner and colleagues identified four independent predictors of necrosis in the RPLND specimen: residual mass <1.5 cm, >90% shrinkage after chemotherapy, and pretreatment AFP and lactate dehydrogenase (LDH) levels. Multivariate regression analysis with these variables was able to accurately predict necrosis in 83% of patients, but the false-negative rate was still high at 20% [Toner et al. 1990]. Using logistic regression, Steyerberg and colleagues found the most significant predictors of necrosis in PC-RPLND pathology to be teratoma-negative primary tumor, normal serum tumor markers before chemotherapy, a small (<20 mm) residual mass and a major radiographic regression (>90%) after chemotherapy [Steyerberg et al. 1995]. The authors also used an international dataset comprising >550 patients from six different study groups to identify six variables as predictors of necrosis: the absence of teratoma in the orchiectomy specimen, normal pretreatment AFP and human chorionic gonadotropin levels, elevated LDH levels before chemotherapy, a small mass before or after chemotherapy, and a large reduction in volume after treatment. The predictive model reliably distinguished necrosis from viable cancer or teratoma, but was much less sensitive in distinguishing teratoma from viable cancer [Steyerberg et al. 1998]. Even an update of this model in >1000 patients was unable to predict necrosis for the whole cohort of patients with reliable accuracy [Vergouwe et al. 2007]. Several radiologic characteristics of residual masses have been related to different histologic findings in residual masses. CT parameters, such as attenuation and density, and magnetic resonance (MR) imaging parameters have been studied but have been proved not to be very informative [Steyerberg et al. 2000].

Imaging studies

There are currently no reliable imaging techniques to accurately identify patients who have residual necrosis/fibrosis, teratoma or viable cancer postinduction chemotherapy [Vergouwe et al. 2001]. Modern CT scan remains the most useful imaging tool in these patients. MRI cannot reliably predict viability of the residual mass after chemotherapy [Hogeboom et al. 1993]. 18-Fluoro-deoxyglucose positron emission tomography (FDG-PET) has no benefit over CT in primary staging. A postchemotherapy residual mass is the most common indication for FDG-PET [Karapetis et al. 2003]. In NSGCTs, PET scans cannot differentiate teratoma from necrosis/fibrosis, but has been used to predict fibrotic residual mass in patients with no teratoma in their primary tumor [Spermon et al. 2002]. Although a positive PET scan is highly suggestive of residual viable cancer, false-negative rates of up to 40% have been reported in a prospective trial [Pfannenberg et al. 2004]. FDG-PET has significant limitations in the detection of very small retroperitoneal lesions and mature teratoma components of any size. False-positive or equivocal results may be seen in association with inflammatory disease, postoperative changes or idiopathic reasons. Repeat PET scans are recommended 2–3 months later in equivocal cases [Karapetis et al. 2003]. PET scans have more utility in patients with disseminated seminoma treated with chemotherapy where there is a higher sensitivity and specificity for determination of residual viable disease.

Extent of surgery

PC-RPLND is a technically demanding operation requiring expertise in retroperitoneal surgery, vascular techniques, and detailed understanding of abdominal and retroperitoneal anatomy. The boundaries of surgical resection have always been the topic of discussion and controversy regarding templates still exists. Although modified nerve-sparing templates may be appropriate for lower stage disease, some authors have demonstrated the presence of tumor outside these templates in the postchemotherapy setting [Carver et al. 2007a].

In a Memorial Sloan-Kettering study 7–32% of men had teratoma or viable cancer outside the boundaries of a modified template [Carver et al. 2007a]. This study had a median follow up of 45 months and additional extratemplate relapses might be potentially identified with longer follow up. The incidence of extratemplate retroperitoneal metastasis for men with intratemplate residual masses less than 1 cm, 1–2 cm, and 2–5 cm were 8%, 18%, and 29% respectively. A total of 80% of patients with extratemplate retroperitoneal disease had teratoma components and 20% had viable GCT [Carver et al. 2007a].

Wood and colleagues demonstrated an 8% incidence of contralateral spread of disease among 113 patients with bulky retroperitoneal disease undergoing postchemotherapy surgery [Wood et al. 1992]. Similarly, Fossa and colleagues [Fossa et al. 1989] and Rabbani and colleagues [Rabbani et al. 1998] reported 5.7% and 2.6% incidences of teratomatous residues outside the boundaries of a modified template dissection, respectively. They concluded that a modified template surgery could be indicated in patients with no palpable residual disease, left-sided primaries, and right-sided testis cancer with absence of mature teratoma or vital cancer in the residual masses [Heidenreich et al. 2009].

Although the standard of care in patients with normal levels of tumor markers and a retroperitoneal residual mass is to remove the mass with bilateral RPLND, several alternatives have been suggested in order to maximize preservation of ejaculatory function [Ozen et al. 2001].

Aprikian and colleagues removed the mass initially and performed either bilateral or limited RPLND according to the results of intraoperative frozen section analysis. No relapses occurred in the operative field. They showed that all the teratomas and viable cancers were located within the residual masses. No viable tumor cells or teratoma were found in the accompanying lymph node specimens [Aprikian et al. 1994]. The surgical approach should also be adapted to the size and location of the mass. Most masses can be accessed via a midline approach whereas larger masses and those requiring suprahilar dissections are best approached either through a thoracoabdominal or midline incision extended to the costochondral margin. Thoracoabdominal approaches afford the possibility of surgical resection of ipsilateral lung lesions. Although simultaneous PC-RPLND and thoracic resections are feasible, more complex mediastinal masses are probably best approached in a staged manner to reduce complication rates. Findings of fibrosis in the retroperitoneal specimen should not preclude thoracic resections since up to 20% of patients can have teratoma or viable cancer in the chest [McGuire et al. 2003].

Desperation RPLND

Patients with elevated or rising tumor markers who have undergone induction chemotherapy, failed salvage chemotherapy, and have resectable retroperitoneal disease may be candidates for the so-called ‘desperation PC-RPLND’. These patients often have chemotherapy resistant disease and surgery may afford the only chance for cure. PC-RPLND in this setting is technically arduous, often involving removal of adjacent organs and is usually associated with significantly lower survival rates. Despite elevated markers, up to 50% of patients will harbor mature teratoma or necrosis/fibrosis in the surgical specimen [Beck et al. 2007]. In patients with viable cancer in the resected specimen, up to one third will have long-term disease-free survival [Beck et al. 2005]. Incomplete resection portends a poor prognosis and patients should be carefully selected for extensive surgical procedures.

Laparoscopic PC-RPLND

Laparoscopic RPLND (L-RPLND) is a technically demanding procedure that should be undertaken by experienced laparoscopic surgeons at dedicated centers, familiar with retroperitoneal anatomy and vascular repair techniques in the event of any complication or open conversion [Basiri et al. 2010]. Primary L-RPLND was introduced as a staging tool for clinical stage I patients with long-term oncological data continuing to mature [Neyer et al. 2007]. After chemotherapy, L-RPLND has been limited to small-volume masses, with initial reports describing significant morbidities during and after surgery. Rassweiler and colleagues converted seven of nine stage II patients from L-PC-RPLND to open PC-RPLND [Rassweiler et al. 1996], while Palese and colleagues reported that two of seven patients in the Johns Hopkins experience required open conversion [Palese et al. 2002]. Furthermore, several complications including significant vascular injuries, transection of external iliac artery, renal artery thrombosis and hematoma are reported during L-PC-RPLND [Albers, 2002]. Although L-PC-RPLND may be technically feasible especially for unilateral template dissection, there is lack of objective evidence to establish long-term oncological equivalence to open surgery, as well as morbidity and cost-effectiveness [Calestroupat et al. 2009].

Adjunctive surgery

Adjunctive surgery is required in at least 20% of patients undergoing PC-RPLND due to the presence of large residual masses around vital structures and resultant severe desmoplastic reaction. The most common adjuvant procedure includes a left nephrectomy and occasionally en bloc vena caval and/or aortic resection with graft placement [Stephenson et al. 2006]. Following chemotherapy, approximately 50–70% of patients will have residual disease in the retroperitoneum and up to 35% will have radiographic evidence of extraretroperitoneal (ERP) masses including liver, lung or neck [Carver and Sheinfeld, 2009]. Complete surgical resection of all residual ERP masses is indicated as approximately 50% of patients will harbor teratoma or viable cancer at these sites and the histologic discordance between the retroperitoneal and ERP sites is approximately 30% [Carver and Sheinfeld, 2009]. It might be acceptable to perform simultaneous surgical resection of multiple disease sites in selected patients. However, the combined procedures must be technically feasible with acceptable morbidity, otherwise a staged approach should be used [Carver and Sheinfeld, 2009]. According to Indiana experience, 35% of patients undergoing combined procedures had perioperative complications, which was significantly higher compared to patients undergoing a PC-RPLND alone [Tognoni et al. 1998].

Complications of PC-RPLND

Complication rates following PC-RPLND are higher than for primary RPLND ranging from 7% to 30% with a mortality rate of about 1% [Mosharafa et al. 2004; Baniel and Sella, 1999]. The most significant source of morbidity in the postchemotherapy group is pulmonary toxicity (8–10%) commonly related to prior bleomycin, wound infection (5%), small bowel obstruction (2%), chylous ascites (2%) and neurological injuries (1%) [Baniel and Sella, 1999].

Tumor size, location and postchemotherapy desmoplastic reaction, along with inferior preoperative status due to chemotherapy exposure, are probably the main reasons for the higher morbidity rate [Baniel and Sella, 1999]. Fortunately, the complication rate appears to have decreased with time due to improved surgical technique and perioperative management [Winter et al. 2009]. We have recently adopted an extraperitoneal approach using a midline incision in select patients undergoing PC-RPLND with a significant reduction in complications rate and hospital stay (unpublished data).

Despite a reasonable number of studies investigating the influence of cancer treatment on sexual function, the question remains open as to whether and to what degree testicular cancer patients are at risk for sexual and fertility morbidity. Retrograde ejaculation is a well-known complication after bilateral RPLND, with a consequent impact on fertility. To address the problem of anejaculation, modified template dissection and/or nerve-sparing techniques have been developed with a 95% success rate [Pettus et al. 2009]. Coogan and colleagues emphasized that nerve-sparing surgery could be performed in only 20% of patients during PC-RPLND [Coogan et al. 1996]. A modified nerve-sparing PC-RPLND, especially in patients with small residual mass could enhance the rate of postoperative antegrade ejaculation up to 80% without oncologic compromise [Miki et al. 2009; Coogan et al. 1996]. Excellent antegrade ejaculation and a low relapse rate attributable to the surgical technique (3%) was reported by Ozen and colleagues after removal of the residual mass only in postchemotherapy surgery, although ‘lumpectomies’ are highly discouraged by experts due to high in-field relapse rates [Ozen et al. 2001].

Impaired spermatogenesis (hypospermia) in patients with testicular cancer may also lead to a decrease in paternity rate at diagnosis and after orchiectomy. Only half of patients who attempt paternity are successful after treatment with no increased risk for congenital abnormalities or late effects in the offspring [Hartmann et al. 1999].

Pathology

The histopathologic findings in postchemotherapy surgical specimens determine the need for further treatment and surveillance protocol. In recent series, the incidence of viable cancer is decreasing most likely due to optimized chemotherapy regimens and a better selection of patients for surgery. Pathologic findings in patients with advanced NSGCT after induction therapy are necrosis in approximately 40–50%, teratoma in 35–40%, and viable carcinoma in 10–15% of specimens [Sheinfeld et al. 2007]. Following salvage chemotherapy, the finding of viable carcinoma increases to about 50%, teratoma in 40%, and necrosis/fibrosis in only 10% [Fox et al. 1993]. Patients with viable disease resected at postchemotherapy surgery are generally recommended to receive two postoperative cycles of cisplatin-based therapy, with two-thirds remaining disease free in the long term [Nichols et al. 1991]. Patients with completely resected very low-volume disease may also be observed. However, patients with unresectable disease, partial resection, or elevated tumor markers should be considered for full salvage chemotherapy.

Teratoma was initially thought to follow a benign course when present in the retroperitoneal area but this is only the case for children. Although the early recognition and resection of teratoma have been accompanied by an excellent prognosis, the untreated disease may have a lethal potential by continued local growth, obstruction or invasion to adjacent structures or undergoing malignant transformation into chemoresistant sarcoma or carcinoma. The incidence of malignant transformation is approximately 3–6% in men undergoing PC-RPLND after induction chemotherapy but increases to 12–18% in men undergoing reoperative PC-RPLND for late relapse [Carver et al. 2007b]. Late relapse (>2 years) is generally associated with unresected teratoma in the retroperitoneum that is highly chemoresistant [Shayegan et al. 2007]. The incidence of residual microscopic teratoma in the retroperitoneum after achieving a CR to chemotherapy is approximately 20–25% but the clinical relapse rate appears to be roughly 5%. Those who relapse on observation remain curable and there is no evidence that immediate surgery on all CRs would improve survival rate [Beck and Foster, 2009]. Teratoma in the orchiectomy specimen is associated with the presence of teratoma or viable GCT in the retroperitoneum after chemotherapy in up to 85%. Conversely, half of patients with no teratoma in the primary specimen had teratoma in the retroperitoneum [Beck et al. 2002].

The chromosomal study of the stromal cells from residual retroperitoneal fibrous masses after chemotherapy suggests that these stromal cells are derived from the same tumor progenitor cells as the pre-existing metastatic testis tumor. This might challenge the concept that postchemotherapy fibrosis lesions in the retroperitoneum are benign. It also supports the practice of surgical removal of any residual masses after chemotherapy for GCTs [Cheng et al. 2007].

Prognosis

Overall cure rates for patients with advanced NSGCT have increased to 80% over the last two decades. While effective platinum-based chemotherapy has been held largely responsible for these trends, the role of RPLND after chemotherapy remains essential.

In a multivariate analysis, Shayegan and colleagues showed that residual tumor size, incomplete surgical resection, the histological finding of retroperitoneal teratomas, and viable cancer independently predicted progression-free survival [Shayegan et al. 2007]. Also Fizazi and colleagues identified three independent prognostic variables for survival in patients who underwent PC-RPLND: complete resection, good risk classification, and <10% viable malignant cells [Fizazi et al. 2001]. The finding of residual teratoma in the retroperitoneum after chemotherapy has been assumed to be a favorable prognostic factor. However, several series have reported a high frequency of recurrences [Carver et al. 2007b]. The histologic finding of teratoma with malignant transformation in the PC-RPLND specimen has also been associated with a poor prognosis [Carver et al. 2007b].

Conclusion

Patients with advanced NSGCT have long-term disease-free survival when chemotherapy is combined with resection of residual retroperitoneal lymphadenopathy. There is no clinical, radiologic, or serologic parameters available to predict with sufficiently high accuracy the histopathology of retroperitoneal residual mass after chemotherapy. PC-RPLND is recommended in patients with residual masses >10 mm with considerable possibility of adjuvant surgeries. Nerve-sparing techniques are feasible in selected patients to preserve antegrade ejaculation with excellent oncologic outcomes.