Is thoracoscopic surgery justified to treat pulmonary metastasis from colorectal cancer?☆

Abstract

We retrospectively analyzed 143 consecutive patients undergoing pulmonary resection for metastasis from colorectal cancer, either through thoracotomy or thoracoscopy from 1987 to 2005. Patients with incomplete resection were excluded. Patients were divided into two groups, based on whether or not they underwent thoracoscopy (n=72) or open thoracotomy (n=71) at the first pulmonary metastasectomy. Two patients undergoing thoracoscopy died postoperatively (one from pulmonary thromboembolism and one from gastrointestinal bleeding). Factors influencing postoperative recurrence-free or overall survival were multiple pulmonary metastasis and history of liver metastasis by univariate analysis, and multiple pulmonary metastasis, hilar or mediastinal nodal metastasis, larger diameter of the pulmonary metastasis, and surgery by wedge resection by multivariate analysis. Five-year recurrence-free rates after the first pulmonary metastasectomy were 34.4% in thoracoscopy and 21.1% in thoracotomy, respectively (P=0.047). Overall 5-year survival rates were 49.3% in thoracoscopy and 39.5% in thoracotomy, respectively (not significant). We found no significant difference in the survival rates between the thoracotomy and thoracoscopy groups, even with elimination of the patients with multiple pulmonary metastases in both groups. We suggest that thoracoscopic surgery for pulmonary metastasectomy from colorectal cancer may be justified if the surgical treatment is indicated.

1. Background

Since thoracic surgeons appreciated that thoracoscopic surgery was feasible for safely resecting pulmonary lesions, they have performed pulmonary metastasectomy through thoracoscopy in cases where the pulmonary nodules are identified by chest computed tomography (CT) [1]. Its relative advantages, however, have not been properly assessed in terms of the postsurgical prognosis. We previously reported a retrospective analysis on patients with pulmonary metastasis from various pathological types of malignant neoplasms, where pulmonary metastasectomy through thoracoscopy appeared to be feasible, because the postsurgical prognosis of patients undergoing pulmonary metastasectomy either through open thoracotomy or thoracoscopy was not significantly different [2].

In this study, we focused on the surgical outcomes of patients with pulmonary metastases from colorectal carcinoma to minimize the diverse characteristics of pulmonary metastases from other organs.

2. Patients and methods

We retrospectively enrolled consecutive patients with a history of colorectal carcinoma who underwent pulmonary resection for pulmonary metastasis between 1987 and 2005. Patients who had undergone macroscopically incomplete resection of the pulmonary metastases at the time of first resection were excluded from this study.

Patients for whom the surgical resection of pulmonary metastases was indicated in this study were those who had resectable pulmonary metastases, up to ten in number, from colorectal cancer without uncontrollable tumor foci in other organs and who were in reasonably good general physical condition for pulmonary resection. The pulmonary metastasectomy through thoracoscopy has been started in our institution since 1996. Thoracoscopic surgery had been favorably performed on patients who had a solitary or a small number of pulmonary metastases for several years, however, multiple metastasis and location in deep lung parenchyma have not been contraindications for thoracoscopic metastasectomy. The patients were divided into the following two groups: those in whom thoracoscopic surgery was performed (thoracoscopy group) and those who had conventional open thoracotomy (open group).

The patients had undergone helical chest CT (5 mm slice) with contrast medium before the chest surgery. The number, the size, and the location of pulmonary metastases were determined. A pulmonary solid nodule, occasionally with a cavity, was suspected to be a pulmonary metastasis if the diameter of the nodule was 5 mm or greater, there was no calcification, or the nodule appeared de novo or enlarged in serial CT films even if its diameter was <5 mm.

The pulmonary metastasis was resected by pulmonary wedge resection, when it was located in the periphery of the lung, and by segmentectomy or lobectomy, when it was deep in the pulmonary parenchyma. Thoracoscopic surgery was performed for both solitary and multiple metastases. Recently, lobectomy or segmentectomy for the resection of pulmonary metastases located deep within the lung parenchyma has often been performed through thoracoscopy. For detection of small nodules located deep in the parenchyma, we often performed CT-guided percutaneous marking by a hookwire preoperatively. We did not perform the hilar or the mediastinal lymphadenectomy at pulmonary metastasectomy, instead, we performed sampling (excisional biopsy) of the hilar or the mediastinal lymph nodes which were suspected to be involved by the cancer.

Patients visited the outpatient clinic of our hospital every 3–6 months. A whole-body CT scan and serum titer for carcinoembryonic antigen (CEA) were performed to detect recurrence of the colorectal carcinoma. The observation time was terminated on 1 April 2007.

The present study focused on two outcome variables: overall survival and recurrence-free survival. Recurrence-free survival time was defined as the time between the initial pulmonary metastasectomy and the last follow-up or recurrence of the cancer. If the patient was cured by the following metastasectomy after the pulmonary recurrence, the patient's recurrence-free survival time was censored at the last follow-up period. If a patient died without cancer recurrence, the patient's recurrence-free survival time was censored at the time of death.

The log-rank test and Cox proportional hazards model were used to individually examine the relationship between recurrence/survival and various potential prognostic factors. The prognostic factors included age, gender, site of the primary cancer (colon or rectum), staging of the primary cancer (stage I, II/stage III, IV), presence of liver metastasis, serum level of CEA, disease-free interval (DFI), number of pulmonary metastases, maximal diameter of the largest pulmonary metastasis, location of the pulmonary metastases, extent of surgical resection, and usage of thoracoscopy during metastasectomy. We performed a χ²-test for the bivariate analysis of categorical data and a Fisher's t-test for the bivariate analysis of numeric data. The data were considered significant when the P-value was <0.05.

Prior to the study, the research review board at our institution examined and approved our research protocol in light of the Declaration of Helsinki. All patients provided informed consent prior to surgery.

3. Results

We enrolled 143 patients that underwent 199 pulmonary metastasectomies (87 open thoracotomies and 112 thoracoscopies) from 1987 to 2005. Two patients who died postoperatively (one from pulmonary thromboembolism and one from gastrointestinal bleeding) after thoracoscopic surgery were included in this study. The first pulmonary metastasectomy was performed through open thoracotomy in 71 patients (open group) and thoracoscopy in 72 patients (thoracoscopy group). Twenty-two thoracotomy patients and 21 thoracoscopy patients had multiple pulmonary metastasectomies. The observation times were 46.7±54.3 months in the open group and 34.4±22.0 months in the thoracoscopy group (P=0.076).

Table 1 shows the distribution of the risk factors in the two groups of patients. Older patients with fewer numbers of metastases, smaller diameter lesions, and lesions located at the periphery of the lung that were able to be removed by wedge resection were more frequently observed in the thoracoscopy group. During the observation period, 36 patients in the open group and 23 patients in the thoracoscopy group died.

Fig. 1 gives Kaplan–Meier estimates of the recurrence-free survival rates and the overall survival rates of patients in the open group and thoracoscopy group. The overall 5-year survival rates were 39.5% in the open group and 49.3% in the thoracoscopy group. The overall survival rate was significantly higher in the thoracoscopy group (P=0.047). The recurrence-free 5-year survival rates were 21.1% in the open group and 34.4% in the thoracoscopy group. The recurrence-free survival rate was not significantly different between the two groups (P=0.064).

Other factors influencing the postoperative prognosis were also analyzed by univariate analysis in Table 2 . Patients with multiple pulmonary metastasis showed significantly lower recurrence-free survival than those with solitary metastasis (P<0.001). Liver metastasis was also a factor which negatively affected the recurrence-free survival (P=0.011).

To eliminate the confounding factor of surgical approaches, we performed Kaplan–Meier estimates of the recurrence-free survival rates and the overall survival rates of patients with solitary pulmonary metastasis in the open group and the thoracoscopy group, respectively (Fig. 2 ). There was no significant difference in the overall survival rate or recurrence-free survival in each group.

Multivariate analysis showed that factors negatively influencing overall postoperative survival were the maximal diameter of the largest pulmonary metastasis (P<0.001, odds ratio=1.049 (by 1 mm), 95% of confidence intervals; 1.024–1.076) and surgical resection by wedge resection (P=0.026, OR=4.24, 95% CI; 1.19–15.1). Factors negatively influencing postoperative recurrence-free survival were multiple pulmonary metastases (P=0.017, OR=1.80, 95% CI; 1.11–2.92) and positive hilar or mediastinal lymph nodal involvement (P=0.012, OR=3.47, 95% CI; 1.32–9.17). Thoracoscopic surgery was not a factor influencing postoperative prognosis, as determined by multivariate analysis (P=0.29 in overall survival, P=0.18 in recurrence-free survival).

4. Comments

The feasibility of pulmonary metastasectomy through thoracoscopy has been recognized by thoracic surgeons, and some clinical studies have indicated that the surgical outcome of thoracoscopic resection might be justified because of its lower invasiveness. In this study we demonstrated that the surgical outcomes of pulmonary metastasectomy through thoracoscopy were not significantly different from those of pulmonary metastasectomy through conventional open thoracotomy, even on patients with solitary pulmonary metastasis. Nonetheless, we should consider the problems of thoracoscopic metastasectomy when comparing pulmonary resections through conventional open thoracotomy.

One problem is whether the safety margin from the tumor rim can be adequately set through thoracoscopy or not. Due to the limited visual field, thoracoscopic surgery has been shown to violate the safe surgical margin for the resection of intrapulmonary malignant neoplasms. Port-site implantation of the intrathoracic tumor was also one of the problems associated with thoracoscopic surgery.

We have reported elsewhere that the rates of local recurrence between patients undergoing metastasectomy through thoracoscopy and through open thoracotomy were not significantly different [2]. We have been fortunate to have no cases in this study with any apparent positive surgical margin or port-site tumor implantation, as we have carefully captured the intrapulmonary tumor with fixing forceps before wedge resection with staplers and wrapped the specimen with a plastic bag prior to removing it from the body.

In this study, we showed that patients undergoing lobectomy or segmentectomy had a better postsurgical survival rate than patients undergoing wedge pulmonary resection, as determined by a multivariate analysis. In general, the surgical procedure does not influence the postsurgical prognosis of pulmonary metastasis. However, intraoperative assessment of the lymph nodes might be considered at the time of pulmonary metastasectomy, because 8 to 17% of patients with pulmonary metastases also have positive mediastinal nodes [3].

Another problem is whether chest CT can properly detect all metastatic foci in the lung or not. Because of the limited access from the outside, bimanual palpation of the whole lung is impossible through thoracoscopy. Chest CT is a crucial preoperative diagnostic procedure for identifying the number, location, and size of the pulmonary metastases in thoracoscopic surgeries. Small nodules are often missed by the CT. Retrospective and prospective studies have shown that intraoperative finger palpation through a thoracotomy can detect additional small pulmonary nodules which were not detected by the chest CT. The authors of these studies postulated that pulmonary metastasectomy would be incomplete without bimanual palpation in thoracoscopy [4,5].

More pulmonary metastases were often found at open thoracotomy following thoracoscopic metastasectomy [6]. Small pulmonary metastases could not be identified completely by a preoperative helical CT [7,8].

The evolution of CT has enabled us to detect very small nodules preoperatively, which has, in turn, increased the rate of false-positive intrapulmonary nodules in patients suspected to have pulmonary metastases. We recently reported that more than 40% of pulmonary nodules equal to or smaller than 5 mm in diameter which were detected by CT in patients suspected of having pulmonary metastasis were not pulmonary metastases [9]. Thorough bimanual palpation also potentially overestimates the number of pulmonary metastases, thus resulting in excessive pulmonary resections. The postsurgical prognosis was not significantly different between the thoracoscopy group and patients receiving thoracoscopy followed by confirmatory open thoracotomy [10].

We suggest that patients with pulmonary metastasis might often be amenable to undergo multiple surgeries, because some small metastatic foci at the initial pulmonary resection are too small to be detected by CT and finger examinations. These may appear later and prompt the consideration of a second pulmonary metastasectomy. Some studies have shown the feasibility of repeated pulmonary resections for recurrent pulmonary metastases from various malignant neoplasms. The postoperative survival rate of patients undergoing second pulmonary metastasectomy was higher than that after a single surgical intervention for lung metastases [11]. Patients who are persistently free of disease at the primary location but who have recurrent, resectable metastatic disease of the lung are likely to benefit from operation a second, third, or even fourth time [12]. It was also reported that multiple sequential pulmonary metastasectomies were feasible and that the 5-year survival for two metastasectomies was 60% [13]. Thoracoscopic procedures are advantageous in repeated pulmonary metastasectomies, because respiratory function, pain control [14] and quality of life [15] after thoracoscopic pulmonary resection are better preserved than after open thoracotomy.

5. Conclusion

We found that the rates of postoperative death and 5-year-survival were not significantly different between patients who underwent thoracotomy vs. thoracoscopy. We suggest that thoracoscopic surgery for pulmonary metastasectomy from colorectal cancer may be justified if the surgical treatment is indicated.

References

Conference discussion

Dr. H.-B. Ris (Lausanne, Switzerland): You have had 143 patients, 72 underwent thoracoscopy and 71 thoracotomy. In which way were these two groups really comparable? You told us that patients with thoracoscopy were older and had more solitary nodules than those undergoing open surgery. How did you make the choice between thoracotomy and thoracoscopy?

Dr. Nakajima: It is a retrospective study from historical data in our institution. We performed only thoracotomy for metastasectomy 10 years ago or earlier. At that time, of course, we performed open thoracotomy for solitary metastasis. But now I usually perform thoracoscopic surgery as much as possible; Even if the tumor is located central portion of the lung, we like to perform a VATS lobectomy instead of the open thoracotomy and lobectomy.

Dr. Ris: If you have a patient presenting several years after the initial breast cancer treatment with a solitary lung nodule, how can you be sure that you don't face a lung cancer primary. Have you looked at how many of these solitary nodules were really metastases and not a new primary of the lung?

Dr. Nakajima: Your question treats another problem. In another study, we found that if the diameter is smaller than 5 mm, the sensitivity of a pulmonary metastasis was <60%. So there are many other pathologies of the small nodules.

In this retrospective study, we focused on patients with pulmonary metastasis from colorectal cancer which was pathologically determined.

Dr. R. Santosham (Chennai, India): In multiple pulmonary metastasis through the thoracoscopic approach in the central lesions, how were you able to decide where they are? Because a tactile feel is quite important. How were you able to decide?

Dr. Nakajima: When the pulmonary metastasis is multiple and located in the central portion?

Dr. Santosham: Yes. How were you able to locate them without feeling them?

Dr. Nakajima: If the nodule is very small and located deep in the parenchyma, we like to perform the preoperative marking using a hook wire through computed tomography and perform the thoracoscopic surgery nowadays.

Dr. D. Waller (Leicester, UK): You showed that wedge resection was a negative predictor of survival, yet you also suggested that repeat metastasectomy may be possible. Are you advocating performing thoracoscopic lobectomy for solitary metastases? And if so, doesn't that limit your ability to perform repeat metastasectomy if you have already performed lobectomy?

Dr. Nakajima: From our study of the multivariate analysis, the prognosis after lobectomy is higher, better than wedge resection. But it is a result of our retrospective study, and we usually perform the wedge resection for the peripherally located tumor through thoracoscopy nowadays. So I think that it should be studied further in the future.

Dr. Y. Lee (Seoul, South Korea): Some articles highlight the lymph node dissection when you perform the metastasectomy. Did you perform the radical lymph node dissection when you performed the metastasectomy or open thoracotomy?

Dr. Nakajima: Our policy is to perform the wedge resection through thoracoscopy and only lymph node sampling. Even if the tumor is located deeply in the parenchyma, we would perform thorascopic lobectomy and lymph node sampling, not dissection. Because I believe that positive lymph node metastasis of the hilum or the mediastinum negatively affect postoperative survival rate of the patients with pulmonary metastasis. But if we perform the hilar and the mediastinal lymph node dissection, I'm not sure whether it contributes to postoperative long survival or not.

Dr. R. Stanbridge (London, UK): I'm a little confused whether you're presenting only solitary metastases or multiple metastases resections. But I think you are presenting both groups. That's the first question.

Second question, did you have any difference in positive margins in the thoracoscopic group especially with multiple resections?

Dr. Nakajima: Previously I analyzed the risk of the positive margin for cancer through thoracoscopic surgery for the metastasectomy, and I found that the rate of positive margin of the staple line either by thoracoscopy or open thoracotomy was not different from each other in our study, in our institution. The rate of local recurrence was not different in either of the two groups.