| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
a Department of Surgery, Cardiothoracic Division, Indiana University School of Medicine, Indianapolis, Indiana
b Department of Medicine, Biostatistics Division, Indiana University School of Medicine, Indianapolis, Indiana
c Department of Medicine, Oncology Division, Indiana University School of Medicine, Indianapolis, Indiana
Accepted for publication September 13, 2007.
* Address correspondence to Dr Kesler, Indiana University Department of Surgery, Cardiothoracic Division, Barnhill Drive EM #212, Indianapolis, IN 46202 (Email: kkesler{at}iupui.edu).
| General thoracic surgery:
The Annals of Thoracic Surgery CME Program is located online at http://cme.ctsnetjournals.org. To take the CME activity related to this article, you must have either an STS member or an individual non-member subscription to the journal.
|
 |
Abstract |
|---|
 Top Abstract IntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
Methods: A total of 158 patients (mean age, 29 ± 8 years) who underwent postchemotherapy operations for PMNSGCT were reviewed and multiple variables analyzed.
Results: Ten (6%) operative deaths occurred, nine of which were attributed to respiratory failure, and 26 (18%) patients experienced postoperative complications, including 9 with respiratory failure. None of 17 recent patients who received chemotherapy regimens that did not contain bleomycin experienced pulmonary complications (p = 0.12 vs patients who received bleomycin). Operative survivors were followed up a median of 34 months (range, 1 to 194 months). Multivariable analysis demonstrated that the postchemotherapy pathologic category of complete necrosis vs teratoma), persistent germ cell or nongerm cell cancer, and elevated serum tumor markers after operation were independently predictive of survival.
Conclusions: Operative risks for PMNSGCT appear to be improved with the use of chemotherapy regimens that do not contain bleomycin. Patients pathologically demonstrating complete tumor necrosis in the residual mass after chemotherapy have excellent long-term survival, with decreasing survival after resection of teratoma and persistent germ cell or nongerm cell cancer. Patients pathologically demonstrating persistent germ cell or nongerm cell cancer have poor but possible long-term survival, which justifies an aggressive surgical approach in patients who are deemed operable.
The treatment of nonseminomatous germ cell tumors of testicular origin with cisplatin-based chemotherapy, followed by surgical resection of residual disease, represents one of the most successful paradigms of cancer therapy [1]. It has been well established, however, that primary mediastinal nonseminomatous germ cell tumors (PMNSGCT), although histologically identical to their more common testicular counterparts, have a distinctly inferior overall prognosis and therefore have been designated as poor-risk nonseminomatous germ cell tumors along with other subsets of testicular nonseminomatous germ cell tumors [2, 3].
After cisplatin-based chemotherapy for PMNSGCT, a residual mass (RM) in the mediastinum is usually present and only infrequently pathologically demonstrates complete tumor necrosis [4, 5]. Surgical removal of any RM after chemotherapy has therefore been advocated. We reviewed our 25-year institutional experience in the cisplatin era to determine surgical risks and predictors of survival after surgery for PMNSGCT with implications for future treatment.
|
Material and Methods |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
Eleven (7%) patients had undergone a prior attempt at postchemotherapy resection at outside hospitals. At the time of diagnosis, 94% (143 of 152) presented with elevated serum tumor markers (STM), either alphafetoprotein (AFP) or β-human chorionic gonadotropin (βHCG; Table 1). A pretreatment biopsy was done in 136 patients (86%). The biopsy specimens either confirmed at least one of three nonseminomatous germ cell tumors subtypes or patients were treated on the basis of a serologic diagnosis, or both. Radiographic evidence of extramediastinal metastatic disease was present in 46 (29%) patients at the time of diagnosis. The distribution of prechemotherapy pathology and metastases are given in Table 1.
|
With the poor sensitivity and specificity of STMs for viable malignancy in the RM after chemotherapy as well as the unsatisfactory results of second-line chemotherapy for PMNSGCT, our institutional practice during the past decade has been that surgery should be undertaken if the residual disease is deemed operable after first-line chemotherapy, regardless of STM status [4, 8]. Seven patients had operations before completing four chemotherapy cycles owing to a growing mass with rapidly declining STMs. Fifty-six (35%) patients presented to surgery with elevated STMs. In 24 of these patients, AFP levels were greater than 100 ng/mL or βHCG levels were greater than 100 µIU/mL. At the time of operation, 19 patients had rising STMs.
Surgical Procedures
Although our specific surgical approach has evolved over time, the basic premise of a complete en bloc removal of the RM, thymus, and surrounding involved structures has been consistent. The details of our current surgical approach to the RM after chemotherapy has been described elsewhere [9, 10]. In summary, the operation is typically delayed for 4 to 5 weeks after cisplatin completion, which allows functional status and bone marrow recovery.
The approach is planned according to the size and location of the RM and can be sternotomy, posterolateral thoracotomy, bilateral anterior thoracotomies with transverse sternotomy, or the "clam shell" incision. Cardiopulmonary bypass circuits are routinely available in case cardiac or great vessel involvement is encountered requiring bypass support.
The operation for PMNGCT is technically demanding because preoperative chemotherapy renders surrounding mediastinal tissues fibrotic, obscuring normal anatomic planes. The effectiveness of cisplatin-based chemotherapy for germ cell cancer, however, also usually results in extensive tumor necrosis that is more marked around the periphery. This finding usually allows a complete resection, which minimizes operative morbidity by preserving critical structures that abut but are not densely adherent to or directly involved with the RM, such as lung, great veins, phrenic nerves, and occasionally, cardiac chambers where the pericardial barrier has been violated.
An extrapleural dissection is considered sufficient if the RM abuts but does not invade the chest wall. If the RM is simply adhering to the visceral pleura of either lung without invasion, removing a small rim of lung parenchyma with the RM is usually adequate to obtain a tumor free margin. Frank invasion of the RM into pulmonary parenchyma or hilum usually requires formal anatomic resection. Similarly, phrenic nerves can usually be separated from an adjacent RM with scalpel dissection, although dense adherence or direct involvement requires en bloc removal. Diaphragmatic plication is performed only if an ipsilateral lobectomy or pneumonectomy is not required with phrenic nerve resection.
If only one (usually the left) brachiocephalic vein is removed en bloc with the RM, venous reconstruction is usually not performed because upper extremity venous insufficiency in these cases is typically temporary and minor. If both brachiocephalic veins are removed with the RM, then unilateral brachiocephalic reconstruction—preferably the right—is performed using an externally-stented polytetrafluoroethylene (PTFE) vascular prosthesis. The superior vena cava is similarly reconstructed with an externally stented PTFE vascular prosthesis, and autologous pericardium is used to patch partial superior vena cava defects. Right atrial and partial pulmonary artery defects are repaired with thin-walled PTFE prosthetic patches.
Intraoperative frozen section analyses of surgical margins are obtained in cases where critical structures abutting the RM are preserved or visibly close surgical margins exist. When required, the timing of pulmonary metastasectomy is individualized based on several factors, including the surgical approach to the residual mass, the magnitude of pulmonary resection required to remove the RM, and the magnitude of pulmonary resection for metastasectomy.
Efforts are made postoperatively to minimize intravenous fluid administration, particularly for patients who have received bleomycin. Most postchemotherapy NGCT patients present for operation with a baseline sinus tachycardia, which is not treated with fluid or pharmacologic blockade if blood pressure and urine output remain adequate. In patients who had elevated STMs at operation, the STMs were typically measured before hospital discharge and at the 1-month follow-up visit.
Routine long-term follow-up was left to the discretion of the referring facility; however in general, chest roentgenograms and STMs were recommended every 6 months for the first 5 years then yearly thereafter for most patients. For patients pathologically demonstrating teratoma in the RM, body scans were also recommended along with STMs measurement because surgery for local or distant teratoma recurrence has a relatively high cure rate.
Statistical Analysis
The 
2 and Fisher exact tests, as appropriate, were used to assess associations between postoperative respiratory failure (acute respiratory distress syndrome/pneumonia) and operative mortality, with both magnitude of pulmonary resection and bleomycin use for all patients. Kaplan-Meier analysis was used to calculate overall survival in operative survivors by postchemotherapy pathologic category of the RM (necrosis, teratoma, persistent germ cell, or nongerm cell cancer, or both) and if persistent germ cell or nongerm cell cancer, or both, was pathologically present, by the percentage of viable malignancy in the RM (<50% or 
50%). Cox proportional hazard models were used to examine the association between overall survival and variables given in the Appendix. Variables that were significant at the alpha 0.20 level in univariate analyses were included in a stepwise multivariable model. For this model, Cox proportional hazard regression was used with an 
level of 
0.10 for entry and 0.05 for staying in the model.
|
Results |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
|
All gross disease was removed with tumor-free margins in 156 patients, whereas 2 patients earlier in our series had positive microscopic margins (pleura and superior vena cava) for viable malignancy on the final pathologic analysis. Pulmonary metastasectomy was required in 19 patients, and 16 have undergone staged extrathoracic metastasectomy for synchronous or metachronous disease, including bone in 5, cervical lymph node in 4, and central nervous system in 3. All patients were deemed free of disease after metastasectomy.
Postchemotherapy Pathology
A benign postchemotherapy pathologic category of necrosis (n = 40) or teratoma (n = 53) was found in 93 (59%) of patients in this series. The other 65 patients (41%) pathologically demonstrated malignancy: 49 had persistent germ cell cancer, and 26 had nongerm cell cancer (sarcoma, 21; carcinoma, 5), which included 10 patients who had both persistent germ cell and nongerm cell cancer. In 27 patients with persistent germ cell or non-germ cell cancer, or both, the RMs contained 50% or more viable cancer, whereas 24 patients had less than 50% viable cancer. Patients with pathologic evidence of persistent germ cell cancer in the RM were typically given two additional cycles of cisplatin and etoposide after recovery. We do not believe that postoperative radiation is of value, even in cases with degenerative nongerm cell cancer present in the RM, provided that complete resection has been achieved.
Mean estimated volume of the RM calculated by pathologic measurements (length x width x height [cm3]) revealed significantly smaller mean (±SE) RM in patients pathologically demonstrating complete necrosis only (430 ± 71, p = 0.02) compared with patients with teratoma (1414 ± 409) or persistent germ cell or nongerm cell cancer, or both, (1016 ± 205), although the variance was large.
Of 93 patients with benign pathology, 26 (28%) had preoperative elevated STMs, and 18 of 41 patients (44%) with pathologic evidence of persistent germ cell cancer presented to surgery with normal STMs, as did 16 of 24 patients (66%) with nongerm cell cancer. Of the 19 patients with rising STMs at the time of operation, 14 had pathologic evidence of persistent NSGC cancer with or without nongerm cell cancer, and 3 patients demonstrated benign disease (teratoma, 2; necrosis, 1). In one of these patients with rising STMs, only nongerm cell cancer was present. Finally, the pathologic findings of the 7 patients who received less than four chemotherapy cycles before surgical referral due to an increasing mediastinal mass and normalization of STMs included 3 patients with teratoma, 3 with degenerative nongerm cell cancer, and 1 demonstrating persistent germ cell cancer.
Postoperative Complications
Postoperative complications occurred in 36 patients (23%), including 19 with respiratory failure requiring prolonged (>48 hours) mechanical ventilatory support, 7 with alveolar air leak exceeding 7 days, and 4 with delayed pericardial effusion with tamponade. The rate of respiratory failure was significantly higher (p = 0.02) in patients who underwent formal anatomic pulmonary resection (21% lobectomy/pneumonectomy) than in those who did not undergo an anatomic pulmonary resection (8%). The 19 (14%) patients who experienced postoperative respiratory failure all received bleomycin-containing chemotherapy regimens compared with none of the 17 patients who did not receive bleomycin (p = 0.12), including 9 patients who underwent bilobectomy (n = 2) and lobectomy (n = 7). Of the patients who developed postoperative respiratory failure, 9 (47%) died. An additional patient died secondary to a pulmonary embolism for a total of 10 (6%) operative deaths. With respect to the operative mortality and the magnitude of pulmonary resection, 7 of 59 patients (12%) who underwent a formal anatomic resection died (lobectomy, 3; bilobectomy, 2; pneumonectomy, 2) compared with only 3 of 99 (3%) who required either no or less than a lobar pulmonary resection (p = 0.04).
Overall Survival
Of the 148 operative survivors, 89 patients were alive and well and 3 are alive with disease after a median 34 months (range, 1 to 194) of follow-up. Fifty-six patients have died, 49 of germ cell cancer and 4 of nongerm cell cancer causes (3 due to leukemia and 1 secondary to pulmonary fibrosis). Three patients died of unknown causes. Survival depended on pathology, with patients pathologically demonstrating tumor necrosis only in the RM having excellent long-term survival (Fig 1). Patients with an RM that contained teratoma had intermediate long-term survival. The decline in survival over time in the teratoma category is partly reflective of three deaths secondary to leukemia in this series. Survival was poor in patients with pathologic evidence of persistent germ cell or nongerm cell cancer, or both, in the RM, with most deaths secondary to recurrent disease occurring within 2 years of surgery. For patients with persistent germ cell or nongerm cell cancer pathology, or both, the survival rate was 33% when most of the RM contained viable malignancy compared with 59% for those with less than 50% of viable malignancy in the RM (log-rank p = 0.22; Fig 2).
|
|
|
|
|
Comment |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
The lower rate of benign histology after cisplatin-based chemotherapy for PMSNGCTs compared with testicular NSGCTs has prompted exploration of different chemotherapeutic strategies.
Bleomycin-containing chemotherapy regimens have traditionally been the standard of care for patients with poor-risk NSGCTs, including PMNSGCTs. Although diffusion capacity is routinely measured during bleomycin therapy and bleomycin is discontinued if a decrease in diffusion is identified, subclinical pulmonary toxicity, which is exacerbated by large thoracic surgical procedures not infrequently requiring significant pulmonary resection, seems likely [17]. A 14% rate of postoperative pulmonary failure in the otherwise young and healthy patients in our series who received bleomycin preoperatively tends to support this speculation.
A recent phase III intergroup study including 181 NSGCTs patients with poor-risk disease found statistically equivalent survival in patients who received VIP compared with standard BEP [7]. Although not statistically significant in this study, the observation that recent patients who received VIP preoperatively and did not experience postoperative pulmonary complications is encouraging and supports our current use of nonbleomycin regimens for PMNSGCTs.
These present data confirm a previous report from our institution that the pathologic category of the RM is independently predictive of survival [4]. Other reports have also found heterogeneous survival for PMNSGCT patients. Memorial Sloan Kettering Cancer Center reported a series of 49 PMNSGCT patients, 32 of whom underwent surgical resection of residual disease after platin-based chemotherapy [5]. Complete tumor necrosis was identified in 12% of surgical specimens, whereas teratoma and viable cancer were found in 66%. The overall 2-year survival for their series was 38%; however, an 81% survival rate was found in patients demonstrating pathology of necrosis or teratoma. The significantly better in survival of patients demonstrating complete necrosis compared with patients with teratoma in our large series is of interest because both are considered benign histology and cannot be completely explained by the 3 patients in the teratoma group who died of leukemia. We can only speculate that the pathologic finding of complete tumor necrosis may be a better surrogate for all microscopic metastatic cancer being successfully treated by chemotherapy.
A large multicenter review of extragonadal NSGCT patients including 287 with PMNSGCT reported an overall 5-year survival of 45% [18]. Two-year survival varied widely in that study, from 34% in the subset of patients who presented visceral metastases to 84% in younger patients without metastases and normal βHCG at the time of diagnosis. That review indicated only 49% of PMNSGCT patients underwent postchemotherapy operations. Mature teratoma and complete necrosis was present in 26% and 37% of patients, respectively; however, the pathology of any excised residual disease was not analyzed with respect to survival outcome. Although we did find similar trends toward inferior survival in patients presenting nonpulmonary metastases and elevated βHCG, these variables did not reach statistical significance in this study of operable and operated patients.
A controversial area has been the role of surgical treatment in the presence of elevated STMs after chemotherapy. A clear role for second-line salvage chemotherapy before considering surgical treatment has been established for testicular NSGCTs refractory to first-line chemotherapy [19]. Normalization of STMs for PMNSGCT patients before considering an operation was historically deemed to be important. In contrast to testicular primaries however, second-line salvage chemotherapy has not been effective for PMNSGCT. A multicenter study by Hartmann and coworkers [20] reviewing the results of second-line chemotherapy for 142 patients with extragonadal NSGCTs, which included 79 PMNSGCT patients, found only 11% long-term disease-free survivors. Moreover, a mediastinal NSGCTs origin was independently predictive of a negative response to second-line chemotherapy in their study.
A combination of the inferior results for second-line chemotherapy, the poor sensitivity/specificity of STMs after chemotherapy for residual malignant or benign disease, and the potential for surgery to salvage cases with residual persistent or degenerative NSGCTs, or both, has led us to recommend surgery if operable after first-line chemotherapy, even if STMs are rising [4, 8]. This approach appears to be further justified, particularly in light of the long-term survival after salvage surgery to remove persistent germ cell or nongerm cell cancer, or both, when less than 50% of the RM contains viable cancer in our series. Patients who demonstrate progressive disease and cannot undergo operation after or during first-line therapy represent extremely unfortunate situations. We believe that high-dose platin-based chemotherapy with rescue by autologous stem cells, which can now be harvested through peripheral blood, needs to be further investigated in this subset.
This retrospective study has several limitations. In particular, there are limitations extrapolating these data to general populations of PMNSGCT patients. Some PMNSGCT patients will, unfortunately, demonstrate radiographic and serologic progression of malignant disease during chemotherapy, which renders them inoperable and diminishes overall survival rates. In contrast, very select patients may have a complete radiographic and serologic response after chemotherapy and are therefore not referred for surgical evaluation. Although we estimate less than 10% of PMNSGCT cases in either category, because many of the patients in our series were diagnosed and received chemotherapy at outside facilities, we cannot accurately determine patient numbers from our institutional data.
Biases may also exist with respect to our preoperative, operative, and postoperative statistics because our institution is a referral center for NSGCTs. Although we believe that the incidence of local mediastinal recurrence with the surgical approach as described is low, accurate statistics regarding the sites of recurrence are unknown. Finally, these long-term survival data could also be skewed, positively or negatively, owing to an institutional selection bias for patients to undergo surgical therapy.
In conclusion, primary mediastinal NSGCTs are a challenging subset of NSGCTs and survival outcome depends on both successful chemotherapy and surgical treatment. We currently believe non-bleomycin-containing regimens will reduce operative risks in this regard. New chemotherapy strategies that reduce the incidence of persistent germ cell or nongerm cell cancer, or both, need continued investigation. Although overall survival is inferior to NSGCTs of testicular origin, survival after successful treatment for PMNSGCTs is heterogeneous, and favorable subsets can be identified. An aggressive but balanced surgical approach after cisplatin-based chemotherapy can result in long-term survival even in patients with persistent germ cell or nongerm cell cancer, or both, and is warranted in these otherwise young and healthy patients.
|
Appendix |
|---|
AFP = alphafetoprotein; βHCG = β-human chorionic gonadotropin.
|
Acknowledgments |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
