| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Eur J Cardiothorac Surg 2007;31:173-180. doi:10.1016/j.ejcts.2006.11.007
Copyright © 2007, European Association for Cardio-Thoracic Surgery. Published by Elsevier B.V. All rights reserved
a Department of Cardio-Vascular and Thoracic Surgery, Cliniques universitaires Saint-Luc, Université catholique de Louvain, Brussels, Belgium
b Department of Pathology, Cliniques universitaires Saint-Luc, Université catholique de Louvain, Brussels, Belgium
c Department of Oncology, Cliniques universitaires Saint-Luc, Université catholique de Louvain, Brussels, Belgium
Received 16 August 2006; received in revised form 31 October 2006; accepted 7 November 2006.
* Corresponding author. Address: Cardio-Vascular and Thoracic Surgery Unit, Cliniques universitaires St-Luc, Université catholique de Louvain, Avenue Hippocrate 10, B-1200 Brussels, Belgium. Tel.: +32 2 764 61 07; fax: +32 2 764 89 60. (Email: Poncelet{at}chir.ucl.ac.be).
 |
Abstract |
|---|
 Top Abstract 1. Introduction2. Material and methods3. Results4. CommentsAppendix AReferences |
|---|
Key Words: Chest wall • Lung • Cancer • Mediastinum
|
1. Introduction |
|---|
|
TopAbstract1. Introduction2. Material and methods3. Results4. CommentsAppendix AReferences |
|---|
For decades, the cornerstone for the surgical treatment of pulmonary metastatic disease has been both the feasibility of complete excision of pulmonary lesions and the control of primary disease, and those pre-requisites remain unquestioned. Initially reserved for tumor for which neither chemotherapy nor radiotherapy could offer satisfactory results, pulmonary metastasectomy has now been increasingly proposed for a broad spectrum of tumor etiologies as a supplementary tool to improve patient survival.
Among the numerous variables that have been looked upon as the prognostic factors for survival, the most cited are disease-free interval (DFI)[6], number of metastasis [4,6], tumor histology [7], nodal (N) stage [4], type of pulmonary resection [7,8], and completeness of resection [6].
Similarly, chest wall resection has for long been performed in thoracic surgical oncology for both primary and secondary tumors. However, most of the published series have focused on both the technical aspects of such procedures and the survival and predicting variables [9–13]. Those studies are generally a mixture of primary and secondary tumors, largely due to a very few patients suffering from such neoplasms even in ‘large-volume’ thoracic surgery centers. This case mix obviously weakens any conclusion that can be drawn on survival and prognostic indicators. In this study, we describe our experience with a homogeneous group of patients with exclusively secondary thoracic tumors, and whether these were located in the lung parenchyma or on the chest wall or even in the mediastinum. To our knowledge, this is the first study that directly compares parenchymal and non-parenchymal thoracic metastatic disease, which we believe is relevant in an era of multimodality treatment for the vast majority of tumors, and irrelevant to the preferential thoracic site of implantation [13,14].
Over the years, our center has developed an aggressive approach toward patients presenting with thoracic metastatic disease, combining chemotherapy and/or radiotherapy and surgery, including iterative procedures, in an attempt to improve patient survival. This study summarizes a single-institution experience and is a retrospective analysis of all consecutive patients who had been operated on for thoracic metastases, including those located on chest wall and in the mediastinum. Our purpose was to evaluate patient's disease-free and overall survival, as well as to identify variables that might influence prognosis in order to better select patients to whom such aggressive bi- or tri-modalities approach would benefit.
|
2. Material and methods |
|---|
|
TopAbstract1. Introduction2. Material and methods3. Results4. CommentsAppendix AReferences |
|---|
Preoperative examinations for clinical staging of patients generally included CT scan of the chest and the abdomen. Most of the CT examinations were performed at our institution on a helical scanner (SR 7000, Philips, Eindhoven, The Netherlands). CT examinations were performed after an intravenous injection of 60–100 cc of contrast medium (Meglumine ioxitalamaat; Telebrix 35, Guerbet, Aulnay-sous-bois, France) injected at a rate of 1–1.5 cc/s. For patients who had their chest and abdomen CT scan performed at another institution as part of their initial evaluation and were then referred to our department for surgical treatment, the CT scans were reviewed by an experienced radiologist. In the recent years of the study and if appropriate, depending on the pathology, a PET-FDG scan (ECAT EXACT HR, CTI, Knoxville, TN) was obtained to assess the number of metastatic sites prior to resection (47/93 patients). Overall, PET-FDG scan was performed on 34 patients with epithelial tumors, 9 patients with sarcoma, and 2 patients each with choriocarcinoma and melanoma. All but five patients (42/47) had negative mediastinal lymph node staging by PET-FDG. Importantly, the five patients with positive mediastinal LN staging by PET-FDG had only single station nodal disease and were, therefore, considered suitable for curative resection.
In addition, serum tumor markers were collected according to the primary tumor in all cases (Carcino-embryonic Antigen (CEA) for colorectal, NSE for neuroendocrine, CA 15.3 and CA 125 for breast, ß-HCG and 
-feotoprotein for teratoma, chromogranin for carcinoid, and calcitonin for thyroid tumors). Due to the small number of patients included in this study and for meaningful statistical analysis, only the subgroup of patients with colorectal metastatic disease will later be analyzed with regard to CEA level as a prognostic marker for long-term survival.
Resection was classified as R0 (macro- and microscopically complete), R1 (microscopically incomplete), and R2 (macroscopically incomplete) [16].
The Committee on Human Rights in Research (Institutional Review Board) of Cliniques universitaires Saint-Luc approved this study.
2.2 Selection of predictor variables
The predictor variables selected for their potential impact on prognosis were age, sex, pathology of tumor, nodal status at the primary surgery, disease-free interval, number of metastasis, presence of metastases other than thoracic metastasis, type of resection, uni- or bi-lateral disease, size of metastasis, completeness of resection, nodal status at the time of metastasectomy, and finally, the number of thoracic repeat metastasectomies. For the subgroup of patients with colorectal metastatic disease, preoperative CEA level was also analyzed as a prognostic marker for long-term survival.
For patients with synchronous metastatic disease (n = 8), DFI from primary tumor was assigned as 0 month.
2.3 Data collection and patient follow-up
All the information necessary for the study was collected from the patient's operative reports, the hospitalization charts, and our thoracic surgery database. All follow-ups were completed from January 2006 until March 2006 and were done through contacts with the referring oncologist, the primary physician, or the patient's family when appropriate. Follow-up was 100% complete (93/93 patients) until completion of the study as on March 10, 2006.
2.4 Statistical analysis
For the statistical analysis, primary histologies were grouped as follows: epithelial regrouping lung (n
= 4), colorectal (n
= 26), pancreas (n
= 1), renal (n
= 11), and breast (n
= 9) and thymic carcinoma (n
= 2) as well as any endocrine tumor (n
= 9), sarcoma (including all histological subtypes) (n
= 21), teratomas (n
= 6), and melanomas (n
= 4).
Survival was calculated from the date of the first metastasectomy to the last date of follow-up (or death). For disease-free survival (DFS) estimates in patients presenting recurrences, the number of days was calculated from the date of the thoracic resection to the first documentation of either locoregional or distant recurrence.
For analysis of descriptive statistics, Chi-square or Fisher's exact test was used as appropriate.
Kaplan–Meier log-rank test and Cox (univariate and multivariate) regression models were used to compare survival and identify predictors of survival. All preoperative predictors were included in the univariate analysis, whereas variables with a p-value equal or <0.10 in the univariate analysis were selected for the multivariate analysis. A backward conditional method was used for variable selection by the Cox multivariate regression model. All statistical analyses were performed with SPSS version 11.5 software (SPSS Inc.).
|
3. Results |
|---|
|
TopAbstract1. Introduction2. Material and methods3. Results4. CommentsAppendix AReferences |
|---|
Out of the 93 patients, 8 presented with synchronous and 85 presented with metachronous metastatic disease. The percentage of patients with a DFI of less than 12 months, 12–24 months, 24–36 months, and above 36 months were 18%, 21.5%, 14%, and 46.5%, respectively.
On the basis of pathologic assessment, a single metastasis accounted for 58% of the cohort (54/93), two metastases were present in 17% (16/93), and three or more metastases in 25% (23/93). The maximum number of resected metastases was 16.
The clinico-pathological data of the entire cohort are summarized in Table 1 .
|
and/or Interleukin-2. For adrenal neuroendocrine carcinoma, we used mithotane as adjuvant treatment. For sarcoma patients, chemoradiation therapy was always used for treatment of the primary site as well as recurrences using cisplatin–adryamicin combination. Finally, teratoma patients were given bleomycin, etoposide, and cisplatin as a part of neoadjuvant treatment.
3.2 Operative strategies and residual (R) status
The surgical approach consisted in 109 muscle-sparing postero-lateral thoracotomies, 9 sternotomies, 8 VATS, and 8 miscellaneous. There were 89 lung parenchymal resections, 19 chest wall resections, 12 chest wall resections extended to the lung parenchyma, and 14 mediastinal tumor resections. Lung resections consisted of 58 wedge resections (65%), 4 segmentectomies (4.5%), 21 lobectomies (23.5%), and 6 pneumonectomies (6.7%).
Extensive parenchymal and chest wall resections were performed on 12 patients (13%) (wedge in 7, lobectomies in 3, and pneumonectomies in 2, respectively).
Indication for surgery for mediastinal tumor resection were bone metastasis invading the anterior mediastinum in nine patients (breast carcinoma, n = 4; sarcoma, n = 3; and hepato-carcinoma and atypical carcinoid, n = 1 each), isolated mediastinal metastasis in two patients (recurrent thymoma and germ cell tumor, n = 1 each), and finally lymph node metastasis in three patients (adrenal carcinoma, thyroid medullary carcinoma, and Non Small Cell Lung Cancer (NSCLC) in one patient each).
Mediastinal lymph node dissection was performed in 49 out of 93 patients (52.7%) and was performed in all cases of major parenchymal resections (n = 27), as well as in minor parenchymal resection with colorectal or renal cell metastatic disease. Nodal stations were labeled according to the American Thoracic Society guidelines.
Resection was classified as R0 (macro- and microscopically complete) [16] in 70 patients (75.2%), R1 (microscopically incomplete) in 14 patients (15%), and R2 (macroscopically incomplete) in 9 patients (9.7%). Two patients, both with germ cell tumors after induction chemotherapy, were classified as ypTxNxMx as their thoracic metastasis were found to be completely sterilized.
3.3 Morbidity and mortality
Overall 30-day mortality was 1.5% (2 patients out of 134 procedures). The first patient died of right ventricular failure on post-operative day (POD) 12, whereas the second developed acute respiratory distress syndrome (ARDS) and died on POD 22. The two deaths occurred in patients with bilateral metastatic disease after the second sequential thoracotomy and bilateral (bi-)lobectomies. The first thoracotomy had been performed 12 and 7 days prior, respectively. A third patient died during her hospitalization, for an overall in-hospital mortality of 2.2%. That patient had undergone sequentially a sternotomy for resection of a metastasis from a choriocarcinoma invading the left atria, which required cardio-pulmonary bypass, and subsequently a right thoracotomy for a right lower lobectomy. She died from multi-organ failure on POD 66. Seven additional major morbidities (5.2%) were encountered in the early post-operative period. Four were treated surgically (complete pneumonectomy for lobar infarction 1, diaphragmatic hernia repair 1, empyema drainage 1, and superficial wound revision 1). Three were treated conservatively (prolonged air leak, ventricular tachycardia, and late chronic broncho-pleural fistula in one patient each).
3.4 Survival
A total of 52 deaths (55.9%) were encountered during follow-up. Among those deaths, 45/52 (86.5%) were caused by progression of disease and 7/52 only (13.5%) were cancer unrelated. There were three early deaths as described above and four late deaths (cardiac, liver failure, stroke, and sepsis in one patient each) (Table 2
).
|
|
|
The survival of patients with epithelial tumors (59% and 42% at 3-year and 5-year, respectively, median: 38 months) and sarcoma (47% at 3-year and 5-year, median: 30 months) did not differ when those two largest groups were compared. Survival for patients with teratoma and patients with melanoma were 50% at 3-year and 5-year for both groups, median survival time of 8 and 30 months, respectively.
For DFI of less than 12 months, 12–24 months, 24–36 months, and more than 36 months, the 3-year (and 5-year) survival were 53%, 54%, 59%, and 55%, (47%, 42%, 51%, and 41%), respectively. There was no statistical difference in survival between these four groups.
Fig. 3 illustrates the overall survival of patients with one-time metastasectomies (n = 65) and those with repeat metastasectomies (n = 28). Their 3-year (and 5-year) survival were 47% and 81% (31% and 73%), respectively. This difference in survival was highly significant (p = 0.001) in favor of those who had undergone repeat metastasectomies.
|
Analysis of the relapse rate among the two major pathological subgroups, epithelial and sarcoma showed a relapse rate of 82% and 90%, respectively (p = 0.30).
Mean time to recurrence did not differ significantly between the two most representative subgroups: 31 months for epithelial tumors and 24 months for the sarcoma group (p = 0.22). For patients who had undergone a complete resection, the relapse rate was 75% (53/70 patients) whereas patients with microscopic tumor burden (R1) remaining after resection had a relapse rate of 100% (23/23 patients) (p = 0.005). Recurrence was both thoracic and extra-thoracic in the majority of those patients (51/76 or 67%), whereas it was thoracic only in 25/76 patients.
Among the 45 patients with recurrent thoracic metastasis, 28 had only thoracic metastasis and they were subjected to 35 redo-metastasectomy (one reoperation in 16 patients, two in 7 patients, and three or more in 5 patients).
3.6 Relative risk of death in univariate analysis
All patients who underwent metastasectomy were included for univariate and multivariate analysis of relative risk (RR) of death (n
= 93) (Table 3
).
|
The distribution of metastasis (bilateral vs unilateral), pathologic subtypes, pathological Nodal status (pN) at metastasectomy, sex, DFI, other metastatic site, and age were not significant.
For the variable ‘type of resection’, we sought to determine the effect on survival for patients treated for metastasis not involving the lung parenchyma (non-pulmonary vs pulmonary resection) (p = 0.23). Thereafter, we analyzed the effect of the extent of parenchymal resection (wedge/segmental vs anatomic resection). Again, no significant difference in survival was observed (p = 0.47).
In the subgroup of patients with colorectal metastatic disease (n = 26), bilateral lesions was also found to be a significant predictor for death (p = 0.02). We also analyzed the RR of death according to the CEA level in this subgroup of patients since it is a well-established prognostic factor [6]. The median preoperative value was 4.3 ng/ml. Among long-term survivors (n = 13), the median value was 3.3 ng/ml, whereas among the non-survivors (n = 13), the value was increased to 6.5 ng/ml. With our limited number of patients, this 2-fold increase did not result in a statistical difference (Student's t-test, p = 0.45).
3.7 Relative risk of death in multivariate analysis
For the multivariate analysis, using the Cox regression model, all variables with a p-value equal or less than 0.10 were included (Table 4
).
|
We performed similar analysis separately on each of the two largest cohorts of patients, the sarcoma group (n = 21) and the colorectal group (n = 26). In patients with colorectal metastatic disease, bilateral lesions were also found to be a significant predictor for death, with a hazard ratio (HR) of 3.7 (95% CI: 1.1–12.8).
|
4. Comments |
|---|
|
TopAbstract1. Introduction2. Material and methods3. Results4. CommentsAppendix AReferences |
|---|
In many aspects, such as the operative approach, the type of pulmonary resection performed, the percentage of repeat metastasectomies, and the completeness of resection, our series reflects the landmark data of the International Registry for Lung Metastasis (IRLM) [6]. In our series, the majority of procedures were open thoracotomies. This is in agreement with the most published series [4,6,7] and might be explained by our mean number of resected metastasis (2.5 metastasis/pa) as well as their deep intraparenchymal location. Video-assisted thoracic surgery (VATS) has been advocated for metastatic lung wedge resection with satisfactory results [17,18]. However, as shown by McCormack et al. [19] in a prospective study, the inability of VATS to accurately evaluate the number of lung metastasis has prevented us from applying this type of approach to the majority of patients, and we currently advocate VATS for metastasis diagnostic purposes and in high-risk patients with solitary metastasis on high-resolution CT scan and/or PET-FDG scanning. Whenever synchronous bilateral metastatic disease was present, we elected a two-staged unilateral approach instead of a median sternotomy because of the combined reasons that we were aware of: the poor sensitivity of first- and second-generation CT scans in the detection of pulmonary metastasis [19], and we believed that a thorough exploration of the left lower lobe was unfeasible with the sternotomy approach, even though this point is still a matter of debate [20–22]. Obviously, with technological improvements and the introduction of the last-generation spiral CT scans, the size threshold for nodule detection has currently decreased to the 5 mm limit, while sensitivity and specificity for metastatic disease using those new tools still mandate complete surgical exploration [23,24]. It is likely that co-registration of CT scan with PET-FDG scanning will improve the specificity, but only for those tumors that are known to be highlighted with Fluoro-18-deoxyglucose.
Our 30-day mortality of 1.4% is in agreement with the operative mortality reported by the IRLM (1.3%) and again in our series, the highest mortality was observed in the subgroup of patients in whom formal parenchymal resections (lobectomies or pneumonectomies) were performed.
In terms of survival, our 3-year and 5-year survival rates of 51% and 36%, respectively, for the entire cohort compares favorably to the Registry and other published series [6–8,25]. When we considered patients completely resected (R0), the respective 3-year and 5-year survival rates were 63% and 51%. Unfortunately, our analysis of 13 variables for predictors of survival did not bring up all the previously published variables [4,6–8]. Obviously, the limited number of patients in this series has greatly hampered the statistical power of our analysis.
This being formulated, our strongest predictors for survival remained well-established predictors, such as the number of metastases, the completeness of resection, and to a lesser degree, the size of metastasis. In our multivariate analysis, having more than two metastases, being incompletely resected, or having a metastasis greater than 3 cm increased the likelihood of death by 2.7-fold, 2.1-fold, and 1.9-fold, respectively. When primary tumor pathology was considered, there was a trend for improved survival in teratoma and miscellaneous tumors whereas melanoma patients had the worst prognosis.
One third of our patient population (30/93) had undergone a curative resection of metastasis at another site (liver mostly) prior to thoracic resection, which is much higher than what was described in the Registry (446/5277 patients or 8%). However, the ‘other than thoracic metastasectomy’ variable did not prove to be a significant prognostic predictor (3-year survival of 51.1% and 50% with and without other site of metastasectomy, respectively, p = 0.88). Not surprisingly and secondary to ‘positive’ selection, patients with repeat thoracic metastasectomies had a 3-year survival rate of 81%, whereas the others had a survival rate of 47% (p = 0.001), again supporting an aggressive approach toward eligible patients with multiple sites or recurrent metastatic disease. The latter figures are in agreement with the 3-year survival rate observed by Robert et al. [7] in the 63 patients of their published series who had undergone repeat metastasectomies. In our study, we were unable to show that a shorter disease-free interval was negatively correlated to survival. Since most of our patients with DFI of less than 12 months (n = 17) had either sarcoma or teratoma, it is possible that the established efficacy of adjuvant chemotherapy has been responsible for an overall increased survival and thus have overshadowed the true impact of DFI on survival in this series. Also, the median DFI in the Registry report was much shorter (19 months) than in our series (median DFI = 31 months), which may also have influenced the results of our analysis.
Our survival analysis between parenchymal and non-parenchymal metastatic disease did not reveal statistical differences between the two groups, bearing in mind that the latter group included only 20 patients. Our survival rate in this group that included both metastatic disease to the mediastinum and to the chest wall were 72% and 64% at 3-year and 5-year, respectively. These results compares favorably with the study by Pastorino et al. [9] and more recently by Pfannschmidt et al. [11] in which both authors reported a 5-year survival rate of 38%.
In conclusion, this study demonstrates that with an aggressive surgical approach as part of a multimodal approach, satisfactory mid- and long-term survival can be achieved in patients with both parenchymal and non-parenchymal thoracic metastatic disease. In every case, only patients with anticipated complete surgery should be offered this invasive treatment modality.
|
Appendix A |
|---|
|
TopAbstract1. Introduction2. Material and methods3. Results4. CommentsAppendix AReferences |
|---|
|
Acknowledgments |
|---|
|
References |
|---|
|
TopAbstract1. Introduction2. Material and methods3. Results4. CommentsAppendix AReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  B. A. Whitson, S. S. Groth, R. S. Andrade, L. Garrett, A. Z. Dudek, J. Jessurun, and M. A. Maddaus Extension of survival by resection of asynchronous renal cell carcinoma metastases to mediastinal lymph nodes. J. Thorac. Cardiovasc. Surg., May 1, 2008; 135(5): 1022 - 1028. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 |
