|Home | About | Journals | Submit | Contact Us | Français|
Colon cancer is a systemic disease in 19% of patients and metastasizes most frequently to the liver and the lung. Survival is enhanced with complete surgical resection of pulmonary metastases. Comprehensive restaging and verification of preoperative fitness must precede resection. The operative approach is dictated by the anatomic location of the metastases, whereas the extent of resection remains a balance between complete removal of metastatic deposits while preserving as much lung parenchyma as possible. The presence of metastatic involvement of hilar and mediastinal lymph nodes is ominous. Multidisciplinary care is highly recommended. An evidence-based algorithm for the identification assessment and treatment of patients with pulmonary metastases is proposed.
Colorectal cancer (CRC) is an extremely prevalent cancer, with greater than 1 million cases identified in the world yearly. Of these cases, 19% present with stage IV disease. Colorectal cancer metastases are most commonly found in the liver; lung is the second most common site. Overall survival with untreated stage IV disease is 11.3% at 5 years,1 although in a selected patient population, resection of isolated pulmonary metastases can increase survival rates up to 40% at 5 years.2,3
The first reported case of pulmonary metastasectomy dates back to 1882, when Weinlechner performed pulmonary resection for two incidental lesions found during resection of a sarcomatous chest wall lesion.4 This was followed by a planned resection of pulmonary metastases by Divis5 in 1927. Long-term survival was observed in 1934, as Edwards6 reported 18-year survival after resection of metastatic osteogenic sarcoma. Barney and Churchill7 then reported 23-year survival after resection of pulmonary metastases from renal cell cancer. The first resection of pulmonary colorectal metastases was reported by Blalock8 in 1944. In the intervening years, multiple series have demonstrated that durable survival benefits are obtained with surgical resection of metastatic disease – the current aggregate survival at 5 years stands at approximately 35% (with variations depending on the tissue of origin of the primary tumor).9 In the remainder of this article, the critical issues pertaining to pulmonary metastasectomy for colorectal metastases will be presented.
Patients treated for colorectal cancer undergo regular physical examinations, carcinoembryonic antigen (CEA) serum levels, colonoscopy, and cross-sectional imaging of the abdomen and pelvis as a part of routine surveillance set out in the National Comprehensive Cancer Network (NCCN) guidelines.10 If recurrent or metastatic colorectal carcinoma is suspected, restaging is indicated. This includes colonoscopy, computed tomography (CT) of the thorax/abdomen/pelvis, and the consideration of a positron emission tomography (PET) scan if CT imaging and colonoscopy do not localize the site of recurrence.10 In the case of pulmonary metastases, a multidetector CT scan is considered a first-line test. Present CT technology allows for image acquisition across the entire thorax at 1-mm slice thickness during a single breath hold. Older generation (single detector) CT scans were 100% sensitive for nodules as small as 6 mm, whereas smaller parenchymal nodules were only accurately characterized in 66% of cases. Pleural nodules were found to be more difficult to assess, as only 17% accuracy was achieved for nodules 6 mm or less in size. Increased sensitivity up to 98% can be obtained using intravenous contrast and objectively measuring nodular Hounsfield units.11 PET scans are now the imaging modality of choice for detection of occult metastases (patients with elevated CEA and no evidence of disease on imaging), unsuspected pulmonary or hilar lymph node disease, and for assessing extrathoracic disease burden (Fig. 1).12,13,14
Published estimates suggest that 20% of metastatic nodules are not detectable preoperatively.15 This has relevance for the growing enthusiasm for application of minimally invasive surgical approaches to metastasectomy, in that these approaches do not allow the operating surgeon to bimanually palpate the lung parenchyma. Several strategies have therefore been proposed to enhance intraoperative identification of smaller nodules. These include preoperative placement of hookwires,16,17 methylene blue,18 and most recently, metallic coils19 using image-guidance. The published experience has been small, and the net benefits inconclusive, especially when one considers periprocedural complications (pleuritic pain, pneumothorax, and dislodging of markers). Intraoperative ultrasound,20 image-guided stereotactic navigation,21 and the use of technetium radiotracer scintigraphy22 have also been examined. These techniques may have utility in delineating metastatic burden, but seem to be ideally suited for localizing solitary pulmonary nodules for video-assisted thoracoscopy (VATS) biopsies, rather than for multiple metastasectomies.
It is our practice to routinely secure preoperative diagnosis of pulmonary nodules by fine-needle aspiration biopsy (FNAB), as this facilitates appropriate workup and operative planning. In patients presenting with pulmonary nodules after a long disease-free interval from their primary colorectal cancer resection, and who have specific risk factors for another primary cancer (for example, a significant smoking history or asbestos exposure), it is prudent to determine the tissue of origin. FNAB samples showing positive immunohistochemical staining for TTF1 (thyroid transcription factor 1) represent lung parenchymal origin (91% sensitive, 98% specific); CDX2 staining (caudal type homeobox transcription factor 2) represents colonic tissue origin (83% sensitive, 96% specific).23 This distinction is particularly relevant, as the staging workup and resection of primary lung cancer is significantly different from that necessary for colorectal metastases.
A detailed assessment of the patient's medical status must be undertaken prior to conducting metastasectomy. In addition to general medical fitness, a close examination of cardiopulmonary fitness must be assessed by testing pulmonary function and exercise tolerance. Fitness criteria24 recommended by the American College of Chest Physicians are outlined in Table Table1.1. Smoking status must be determined and smoking cessation advocated if indicated. Details of prior neo- and adjuvant treatments must be sought out25 in the history, as certain chemotherapy regimens are toxic to lung parenchyma (glycopeptides such as bleomycin), whereas others are cardiotoxic (anthracyclines such as doxorubicin). Any history of external-beam radiotherapy should prompt thorough pulmonary function testing. The severity of functional impairment will dictate the extent of pulmonary resection that may be considered. Inclusion of preoperative pulmonary rehabilitation and neuraxial analgesic strategies26 such as epidural or paravertebral blocks serve to optimize postoperative recovery of respiratory function, as do newer surgical approaches such as VATS and muscle-sparing thoracotomy.27
The goal of resection in metastatic CRC is to achieve a complete extirpation of intrathoracic disease without sacrificing excessive amounts of normal lung parenchyma. Guiding principles (see Table Table2),2), outlined by Thomford, Woolner, and Clagett28 in 1965, an extension of earlier work by Alexander and Haight29 in 1947, are still considered30 to provide appropriate patient selection criteria. There should be no doubt that the ultimate goal is a complete (R0 – no gross or microscopic residual disease) resection: following complete metastasectomy, the median survival is 35 months, whereas following incomplete resection, the median survival is only 15 months.3
The surgical approach to R0 resection of pulmonary metastases is dictated by the size, number and distribution of individual tumor deposits. The bilateral thoracosternotomy (“clamshell” incision), popularized initially for lung transplantation and later applied to bilateral metastasectomies, entails bilateral submammary thoracotomies with transverse division of the sternum. This incision provides excellent exposure to the hilum and all lobes of the lung, including both lower lobes.31 However, postoperative complications are significant and include nonunion and displacement of the transverse sternotomy, as well as migration or infection of fixation hardware. Median [midline] sternotomy is commonly used for cardiac surgical procedures and has applicability in pulmonary surgery. The main advantage is the access to both lungs (albeit reduced access to the lower lobes, especially the left), with lower morbidity (decreased intercostal nerve and muscle injury).32 Anterior and lateral thoracotomies currently represent the standard incision for open pulmonary surgery; however, VATS procedures are rapidly gaining popularity. VATS resections utilize laparoscopic-style instruments introduced via trocars, resulting in smaller incisions and reduced postoperative morbidity.33
Data supporting bilateral resections, via clamshell or median sternotomy may have been influenced by the suboptimal imaging modalities of an earlier era (linear tomography and early CT scan technology). The routine necessity for bilateral exploration with bimanual palpation of lung parenchyma has been brought into question by improvements in current-era imaging modalities such as helical multidetector CT scans and PET scans. It may be for these reasons that prior to 1979, lateral thoracotomies were performed for metastasectomy; afterwards, the median sternotomy dominated.34
VATS procedures were reported in the early 1990s and have gained momentum ever since.35 Most published reports suggest equivalent outcomes with open and thoracoscopic approaches; VATS appears to achieve equivalent postoperative and 5-year survival rates,36,37 with reduced perioperative morbidity. Cancer recurrence at VATS port sites has not been widely reported: a single study reports38 four recurrences in 26 cases, with all recurrences occurring within the first 5 postoperative months. A second study39 reports three port site recurrences in 36 cases. These reports encompass results from early VATS procedures – increasing use of wound protectors and specimen extraction bags37 has reduced this risk significantly. The consideration of minimally invasive operative approaches should be based on adherence to identified predictors of a successful VATS approach to pulmonary metastases.15 These parameters, evaluated by Cerfolio and colleagues, include the number, size, and location (central versus peripheral; upper versus lower lobe) of the metastatic deposits, and are summarized in Table Table33.
The choice of operative approach and specifically the use of minimally invasive approaches will likely be the subject of controversy for some time to come or until more data are available. However, Roth,34 and more recently Nakajima36 have published their experiences showing that the operative approach does not appear to influence overall outcome.
Metastases are most commonly located peripherally in the lung parenchyma and more frequently in the lower lobes. This allows for relatively straightforward nonanatomic resection (Fig. 1), as reflected in the frequencies of this type of resection in larger series, such as the International Registry of Lung Metastases (IRLM). The most common pulmonary metastasectomy performed3 was a nonanatomic wedge resection (67% of patients registered in the IRLM), followed by lobar resections (21%), segmentectomies (9%), and pneumonectomy (3%).
Pneumonectomy is rarely performed as the initial procedure for metastasectomy. This is related to several factors including a postoperative mortality rate approaching or exceeding 5%; the lack of conservation of lung parenchyma; and a 5-year survival rate of only 16%.40 Completion pneumonectomy as a second procedure remains controversial.41 In patients who are not operative candidates from a pulmonary reserve standpoint, there is recent evidence that radiofrequency ablation may serve to control metastatic disease,42 although repeated rounds of ablation are likely required.43
Patients whose disease pattern includes both hepatic and pulmonary metastases have been shown to benefit from either sequential or simultaneous liver and lung resections, with a 51% 5-year survival rate.44 Factors predictive of good outcome included younger age, solitary metachronous liver metastases preceding lung metastases, and a long disease-free interval.45 In these cases, assessment by a hepatobiliary surgeon in the context of multidisciplinary cancer care is recommended; close consultation is required to decide between staged or concurrent resection strategies (Fig. 2).
The results of several small series of patients undergoing metastasectomy have been reported – however, we will focus on data from the largest series from 1991–1995, and a more contemporary meta-analysis of published reports from 1995–2006.
The IRLM is a database of pulmonary resections performed for metastases from all cancer types at 18 thoracic centers worldwide over the years 1991–1995, representing an accrual of 5206 separate cases.3 This has generated robust data regarding postoperative outcomes, survival, and the determination of adverse risk factors. The disease-free interval strongly affects outcomes, as patients presenting with early pulmonary metastases (within the first year) had a median survival of 29 months, whereas those presenting with metastases 3 years or more after their primary treatment had a median survival of 49 months. Multivariate analysis of this data allowed for identification of adverse risk factors and survival grouping. Resectability of metastases, the disease-free interval, and number of metastases (single versus multiple) all had significant influence on long-term outcome, with resectability having the most profound impact (Table 4).
The Heidelberg group2 has published a meta-analysis of 14 studies of pulmonary metastasectomies for colorectal cancer, incorporating 1684 patients in series published from 1995–2006. Perioperative mortality ranged from 0–2.5%, confirming that metastasectomy is a safe procedure. The majority of the patients reported underwent unilateral procedures (n=886) involving wedge resections. The overall survival rate in this series was 48% at 5 years.
In this study, peribronchial and mediastinal lymph node metastases were observed in 9.8% of cancer patients with CRC metastases, and had a profound impact on long-term survival. Although overall 5-year survival rate was found to be 48%, the survival rate rose to 60% in the absence of nodal disease. In the presence of peribronchial/hilar nodes, survival was reduced to 17%. No patients with mediastinal lymphadenopathy survived 5 years.46 These findings have been corroborated in another recent meta-analysis.2,30
Preoperative elevation of CEA level (greater than 10 ng/mL47) is also an adverse risk factor. Patients with normal serum CEA levels preoperatively had a 5-year survival rate of 60%, whereas those with high CEA levels had only an 18% survival rate. Other negative prognostic factors include a tumor doubling time of less than 100 days,48 and the presence of several histopathologic markers: presence of p53 staining, vascular invasion, presence of alveolar floating cancer cell clusters, and positive E-cadherin staining have all been related to reduced survival.49,50
There are conflicting data supporting the efficacy of repeat pulmonary metastasectomy for metastatic CRC. Welter's group51 reports that the number of metastases present at the first resection is the most important determinant of overall survival. A Japanese group52 has reported decreased survival with two or more resections, suggesting the first pulmonary metastasectomy is the most effective, while Kim and colleagues53 show that 5-year survival is not different between patients undergoing single or multiple pulmonary resections.
Pulmonary metastasectomy is currently an accepted mode of treatment for metastatic colorectal carcinoma in selected patients. The authors propose an evidence-based approach applicable in this circumstance (Fig. 3).
Suspicion of pulmonary metastases should prompt thorough restaging studies, which include CT of the thorax, abdomen, and pelvis. Consideration should be given for a PET scan for assessment of intrathoracic (hilar or mediastinal) lymphadenopathy as well as occult recurrence at the primary site or elsewhere in the abdomen or pelvis. Patients with isolated thoracic disease should then be further assessed for resection, with the planned extirpative procedure anticipating an R0 resection. The extent of pulmonary resection will further guide the magnitude of preoperative workup: patients requiring a pneumonectomy will necessarily require a greater level of cardiopulmonary reserve as compared with patients requiring wedge resection(s) alone.
For patients whose metastatic disease is unresectable, or whose cardiopulmonary reserve is insufficient, nonresectional treatments are appropriate. There is limited experience in ablative techniques (such as radiofrequency ablation) for lung metastases, but early results are encouraging.42 The role of chemoradiotherapy is not yet defined for pulmonary colorectal metastases because historically, these lesions did not respond well to adjuvant treatments. The current era of chemotherapy10 has shown impressive responses in primary colorectal cancers to such regimens as FOLFOX (fluorouracil, leucovorin and oxaliplatin) and FOLFIRI (fluorouracil, leucovorin and irinotecan), with or without the addition of biologic agents such as bevacizumab (anti-vascular endothelial growth factor; VEGF), cetuximab and panitumumab (anti-epidermal growth factor receptor; EGFR). These have not yet been studied systematically and may actually represent more efficacious treatment for pulmonary metastases. Patients whose functional status is poor may be candidates for palliative approaches to care.
When the metastatic burden and distribution are amenable to curative resection, and the patient is fit for pulmonary resection, the next issue is choice of operative approach. This will be largely dictated by the pattern and location of metastatic disease.
The operative approach best suited for unilateral pulmonary disease is based on principles outlined by Cerfolio and colleagues15 (Table 3), and also on the level of comfort of the operating surgeon with minimally invasive thoracic surgical techniques.
Bilateral pulmonary disease can be approached in a similar manner – the criteria outlined in Table Table11 will determine the suitability of staged VATS resections. If disease is not amenable to VATS, then staged thoracotomies are preferred, followed by median sternotomy or a clamshell approach. Arguments in favor of bilateral access incisions place emphasis on wide exposure to permit a full manual exploration of the thoracic cavity and lung parenchyma, with criticism of the VATS approach being a perceived inability to fully explore the lung. The authors note in their experience of VATS metastasectomies, early mobilization of the inferior pulmonary ligament, coupled with the use of a “three-finger” utility incision (as performed for a thoracoscopic lobectomy) allows for a very satisfactory assessment of the posterior and inferior areas of the lung.
After pulmonary resection for colorectal metastases, follow up should include continued surveillance for local and distant recurrence as outlined in the NCCN guidelines.10 Controlling for pulmonary recurrence and postoperative complications is most effective with contrast enhanced CT scans of the thorax.
Pulmonary metastases from colorectal cancer represent a challenging situation, where metastasectomy carries genuine benefit in carefully selected patients. Key factors to consider include adherence to oncologic principles, technical feasibility of the proposed resection and the overall fitness of the patient to undergo pulmonary resection. Although patient selection and determination of operative fitness must be highly individualized, an evidence-based algorithm to guide patient selection, assessment and operative approach is proposed.