81: Cardiopulmonary Bypass for Extended Thoracic Resections

Bronchogenic carcinoma remains the most common cause of cancer death in both men and women in the United States. These tumors can exhibit local progression and invasion before metastatic spread has occurred, which does not preclude resection with curative intent. Any contiguous structure within the chest may be involved, with chest wall involvement being the most common. Other potential sites of local invasion include the left atrium, aorta, superior vena cava, vertebral bodies, diaphragm, and esophagus. The increased potential for morbidity and mortality are well documented for these complex extended resections, making appropriate patient selection crucial. Long-term prognosis depends on accurate pretreatment staging to assist in selection of therapy and complete resection. Cardiopulmonary bypass (CPB) may be necessary to allow surgical resection of central, locally advanced malignancies because they involve, or are close to, the heart and/or the great vessels. CPB serves as an alternative to conventional ventilation during extended resections providing oxygenation and hemodynamic support. This chapter will review the role of CPB for the extended resection of lung cancer, as well as the clinical and technical considerations and expected surgical outcomes.

For those patients presenting with locally advanced lung cancers, long-term outcome is primarily dependent on the ability to obtain a complete (R0) resection. For non–small-cell lung cancers (NSCLCs) with direct invasion into the mediastinum, Martini et al.¹ found a survival rate of 30% if an R0 resection was obtained in contrast to 14% if the resection was incomplete. Likewise, for tumors invading the heart or the great vessels, 5-year survival rate ranges between 23% and 40% with complete resection versus 0% with incomplete resection.^2,3

The operative approach is ultimately based on the tumor anatomy, the need for vascular reconstruction, and the urgency with which circulatory support is initiated. To optimize outcomes, one must maintain a flexible strategy with regard to arterial and venous cannulation sites, need for aortic clamping, cardioplegia requirements, and deairing options.

When an injury to a major vascular structure (i.e., pulmonary artery, superior vena cava) occurs during a thoracotomy, CPB support may be required. In the setting of a right thoracotomy approach, cannulation can be achieved via the ascending aorta and the right atrium. In the setting of a left thoracotomy approach, cannulation can be obtained via the descending thoracic aorta and main pulmonary artery.⁴

If the groin is accessible, systemic venous drainage can be achieved by placing a long venous cannula into the right atrium through the femoral vein. In the emergent setting, decompression of the heart with the ability to control blood loss and return shed blood is usually all that is required to enable primary or patch repair of the injury to the central vascular structure.

In the elective setting, standard median sternotomy is the surgical approach used to address lesions involving the central pulmonary arterial system. Standard ascending aorta and right atrial cannulation are used for CPB institution. Reconstruction with pulmonary homograft or autologous pericardium can be readily achieved. In lesions involving the left atrium, median sternotomy is usually satisfactory; however, one must be alert to the issues of deairing and ensure appropriate means of preventing systemic air emboli. Additionally, tumor emboli can occur, and care must be taken to limit tumor manipulation before cardiac decompression and adequate circulation control.⁵ If the left thoracotomy is the surgical approach used, it is useful to access the left femoral vein with a small caliber catheter prior to turning the patient to the side.

Cancer progression is a multistep process including proliferation, migration, vascular invasion, and angiogenesis. The effects of operative stress on cancer cell proliferation have been shown in several in vitro models.^6–8 One such study demonstrated that lung cancer cell proliferation was better in human serum obtained immediately following lung resection when compared with serum drawn preoperatively.⁸

The immunosuppressive effects of extracorporeal circulation have been clearly elucidated. The physiologic stress of CPB has been shown to alter the circulating levels of certain cytokines and growth factors.^9–11 There are also inhibitory effects to both cell-mediated and humoral immunity.^12,13 In a study performed in patients undergoing cardiac surgery with and without CPB, those undergoing CPB had a significantly decreased inhibitory capacity of serum for cancer cell proliferation.¹⁴ Previous series of extended resections have demonstrated an earlier recurrence rate after resection with CPB.^3,15

Centrally Advanced Tumors

Locally advanced tumors that involve the central pulmonary vasculature or the heart (T4 lesions) are classically considered to be unresectable. Achieving a tumor-free proximal margin or satisfactory proximal vascular control may not be possible with standard (non-CBP) techniques. A small but definable subset of such patients will benefit from surgery if CPB is used to facilitate these complex resections. Accurate preoperative evaluation, including aggressive staging, must be performed to exclude the presence of occult metastatic disease, determine the patient's physiologic fitness, and establish the limits of resection to achieve the optimal long-term survival for each individual patient. Since these tumors are often larger and more centrally located, preoperative imaging should include PET and CT scanning. There should be consideration for magnetic resonance imaging (MRI) of the brain to rule out occult disease. Mediastinoscopy must be performed not only for identifying nodal disease, but also as an assessment of the extent of airway and pulmonary artery involvement.

The role of induction chemotherapy prior to resection of T4 malignancies remains undefined. There are some small series in which preoperative chemotherapy improved resectability.^16–18 In one series, induction chemotherapy prevented 20% of patients from proceeding on to resection due to toxic side effects.¹⁹ Furthermore, there is evidence that induction therapy increases surgical morbidity and mortality.^20–22 In patients undergoing extended surgical resection, induction therapy followed by surgery is associated with an increase in morbidity and mortality when compared to surgery alone.²⁰ Therefore, decisions for preoperative chemotherapy should be tailored on a case-by-case basis.

Most thoracic surgeons are reluctant to perform pulmonary resections with patients on CPB. Several authors^23–26 have reviewed the results and safety of combined cardiac and pulmonary procedures requiring CPB. Their opinions are varied, and several authors have expressed concerns for the adverse effects of CPB on hemostasis and pulmonary function. Others^4,27,28 with significant institutional experience have written more extensively on the subject, describing the advantages, disadvantages, and parameters for patient selection when CPB is used as an adjunct to conventional thoracic surgical techniques.

Byrne et al.⁴ reviewed a decade of experience at Brigham and Women's Hospital and Massachusetts General Hospital in Boston. Between January 1992 and September 2002, CPB was used in 14 patients during planned curative resection of locally advanced thoracic malignancies. In 8 of the 14 patients, CPB use was planned to facilitate resection. In the remaining six patients, CPB was required as an emergent therapy to manage central vascular injury. Indications for planned CPB included tumor involvement of the left atrium, pulmonary artery, and superior vena cava. Complete resection was achieved in 12 patients (86%). There was one operative death from pulmonary embolism. Complications included low cardiac output state (5), stroke (1), pulmonary edema (1), and reoperation for bleeding (3). The overall 1-, 3-, and 5-year survival rates were 57%, 36%, and 21%, respectively. The authors concluded that although CPB is rarely required for thoracic malignancy resection, in appropriate circumstances, it can be used with low morbidity and mortality and may be lifesaving if the surgery is complicated by a central vascular injury. They also concluded that the ability to perform a complete resection influences ultimate survival. In addition, optimal outcome depends on careful patient selection with use of radiographic imaging and thorough intraoperative inspection.

Vaporciyan et al.²⁸ reported the University of Texas MD Anderson Cancer Center's experience from January 1995 to July 2000 using CPB for resection of metastatic or noncardiac primary malignancies that extended directly into the heart. This series included 19 patients, 11 of whom underwent surgery for curative intent. Complete resection was achieved in 10 of these patients. There were two deaths in the group operated on for palliation. Major complications occurred in the majority of patients (58%) and included acute respiratory distress syndrome, mediastinal hematoma, and pneumonia. The overall 1- and 2-year survival rates were 65% and 45%, respectively. The authors concluded that the use of CPB has a role in selected patients with these central thoracic malignancies if there is confidence that complete resection can be achieved.

Bronchogenic carcinoma of the right upper lobe can invade the mediastinal pleura and on rare occasions invade the superior vena cava (Fig. 81-1). Involvement of the superior vena cava may also occur as a consequence of metastatic nodal disease, which, when present, is a uniformly poor prognostic indicator. There is an increasing experience in extended resections of pulmonary malignancy with en bloc resection of the superior vena cava. Several authors^2,15,29–31 have suggested a benefit in selected patients, but there is still uncertainty regarding a consistent benefit.

Technical Considerations for Vena Cava Resection and Reconstruction

Optimal exposure to the superior vena cava is achieved through median sternotomy. Bronchogenic tumors are approached through the right chest, which provides excellent visualization of the superior vena cava and the atrium, but limited access to and control of the left brachiocephalic vein. Two techniques can be used for resection and reconstruction of the superior vena cava. If less than a third of the circumference of the vena cava is involved, partial resection with primary or patch closure using autologous pericardium can be performed if there are concerns regarding stenosis. If there is greater involvement of the superior vena cava, complete resection and reconstruction will be required. Reconstruction can be performed with autogenous venous or prosthetic grafts. Options for autogenous graft replacement include the jugular vein, the superficial femoral vein, or spiral saphenous vein. This option is more limited because the graft diameter must be as great as the brachiocephalic vein for a satisfactory outcome.

Dartevelle³⁰ has described the use of synthetic polytetrafluoroethylene vascular graft as the material of choice for complete venous replacement. Before the work of Dartevelle, prosthetic graft replacement of the superior vena cava was thought to be a surgical contraindication because of the high rate of thrombosis and infection, as well as the deleterious physiologic effects of clamping the superior vena cava. Specific interventions such as maintaining adequate cerebral perfusion pressure, monitoring cerebral oxygenation, limiting clamp time with efficient reconstruction, paying attention to the prevention of bacterial contamination, using anticoagulation judiciously, and appropriately using shunts or bypass are all required to ensure a satisfactory outcome. Contraindications to superior vena caval resections included complete venous obstruction and involvement by a malignant that is otherwise unresectable.

Results of Combined Pulmonary and Superior Vena Caval Resection

Dartevelle described his experience with combined pulmonary and complete caval resection in 14 patients with NSCLC.³⁰ These extensive tumors required carinal pneumonectomy in six patients, extended pneumonectomy in seven patients, and one patient received a lobectomy. Six of the 14 patients had ipsilateral mediastinal nodal disease. Major complications were seen in three patients which included bronchopleural fistula in two and extrapericardial cardiac herniation in one. The mortality rate was 7.1%, and 5-year survival rate was 31%. A more recent review of 29 patients undergoing superior vena cava resections over a 25-year period revealed better 5-year survival in patients without carinal involvement (43.2%) and in those with non–squamous-cell carcinoma (51%).³²

Thomas et al.²⁹ reviewed their institutional experience in 15 patients, four of whom required complete caval resection. There was one death in the postoperative period and a complication rate of 20%. The authors observed a median survival of 8.5 months with two local recurrences, and the 1-, 2-, and 5-year survival rates were 46.7%, 32%, and 24%, respectively. The authors concluded that extended resection, when feasible, is justified.

Spaggiari et al.³¹ reported similar observations in their report on 25 patients in which seven patients had complete resection of the superior vena cava with graft interposition. Ipsilateral mediastinal nodal disease was observed in 56%. Complete resection was achieved in 80% of patients with a perioperative complication rate of 36% and mortality rate of 12%. The observed median survival was 11.5 months, and the 5-year actuarial survival rate was 29% with four long-term survivors. The data were inadequate to differentiate survival according to nodal status, but the authors commented that the survival trends favored node-negative patients. The authors recommended mediastinoscopy in all patients with superior vena caval involvement to exclude N₂ patients.

Given the anatomic continuity between the pulmonary hilum and the pulmonary veins, the left atrium is the most common cardiac chamber involved by bronchogenic carcinoma. Left atrial invasion can occur by one of the two mechanisms. The first involves contiguous invasion from the base of the pulmonary vein, whereas the second is via direct invasion by the primary tumor or metastatic lymph node.

The first description of left atrial resection was by Allison in 1946.³³ Initial attempts at this extended resection involved combined pulmonary and left atrial resection with the use of a side-biting vascular clamp.³⁴ Although the series are small, some of these patients did show long-term benefit. However, with more extensive left atrial involvement and the risk of tumor embolization, this approach may be inappropriate.

Technical Considerations for Left Atrial Resection and Reconstruction

CPB offers the advantage of complete radical resection and reconstruction on an arrested heart. Alternatively, resection and repair can be performed on a beating heart with a clamp applied to a more extensive area of the left atrium. The approach is through a median sternotomy. Standard ascending aortic and bicaval cannulation can be utilized. The left atrium can be opened after aortic cross-clamping and cardioplegic arrest or after the induction of hypothermic ventricular fibrillation in order to avoid air embolism.^35,36

In cases when a vascular clamp is utilized, there is a risk of massive hemorrhage with clamp dislodgement. One distinct advantage of CPB is that vascular clamp dislodgement and injuries to the left atrial wall during resection and reconstruction can be addressed in a more controlled manner. Reconstruction can typically be performed with autologous or bovine pericardium.

Results of Combined Pulmonary and Left Atrial Resection

In a series of 19 patients undergoing simultaneous pulmonary and left atrial resection, Ratto et al.³⁷ reported a 5-year survival rate of 14% with a median survival of 25 months. With complete resection, others have found 5-year survivals ranging from 16% to 23%.^34,38 The majority of patients require pneumonectomy along with radical resection to achieve a tumor-free margin. The presence of N2 disease portends a dismal prognosis and should preclude surgical resection.

Left-sided lung cancers may involve the descending thoracic aorta. It is often difficult to determine the presence and extent of aortic invasion on imaging studies (Fig. 81-2). If the fat plane between the aorta and the tumor is absent or there is tumor abutment involving greater than 90 degrees of the aortic circumference, invasion should be suspected. However, ultimate determination of involvement of the aorta generally is made at the time of exploration. Aortic invasion by bronchogenic carcinoma is usually limited to the adventitia. As with lesions involving the central pulmonary vasculature, these T₄ tumors with full-thickness involvement of the thoracic aorta are classically deemed unresectable. Experience with combined pulmonary and aortic resection is limited and often anecdotal. However, some authors^8–10 have described their experiences with these combined resections. These reviews suggest a favorable impact on survival in highly selected patients; however, the few available reports in the literature make it difficult to draw more generalized conclusions.

Technical Considerations for Aortic Resection and Reconstruction

Different technical methods are needed to manage the aorta that is locally invaded with lung cancer. The options include resection of the tumor in a subadventitial plane, partial resection with patch reconstruction, and complete segmental aortic resection with tube graft reconstruction. The use of a shunt prosthesis has been described between the ascending and descending aorta.^2,15 Venous cannulation can be obtained through the main pulmonary artery or femoral veins with arterial cannulation of the descending aorta. A right radial arterial line should be utilized to gage the adequacy of upper body perfusion. The aorta can then be cross-clamped between the innominate and left common carotid artery.³⁹ If there is involvement of tumor of the aortic arch proximal to the left common carotid artery, hypothermic circulatory arrest with selective cerebral perfusion may be required.^40,41 Hybrid approaches such as endovascular stent grafting of the thoracic aorta followed by resection of the aortic adventitia and media have been described.⁴²

Results of Aortic Resection and Reconstruction

Nakahara et al.¹⁵ reported their results in three patients in whom the aorta was resected at the time of pulmonary resection. One patient was a midterm survivor; however, the other two succumbed to metastatic disease within a year. In 1994, Tsuchiya et al.² reported their experience in 28 patients at the National Cancer Hospital in Tokyo. Resection in a subadventitial plane was used in 75% of patients. In those 21 patients, 10 had complete resections, and all patients with incomplete resections did poorly. The authors commented that peeling the adventitia off the aorta is inadequate for lung cancer invading the aorta. Of the seven patients who had a complete aortic resection with tube graft reconstruction, four developed recurrent disease. Only one of the seven patients who were managed with complete aortic resection was a long-term survivor. Klepetko et al.³⁹ reported similar observations in their experience in seven patients undergoing full-thickness aortic resection with patch or tube graft reconstruction. Perfusion was supported with CPB in six of the operations without apparent untoward events. Long-term survival was achieved in two of the seven patients. Horita et al.⁴⁰ reported that only one patient remained alive for more than 5 years following resection of the aortic arch under hypothermic circulatory arrest. These observations regarding pulmonary malignancies with direct aortic invasion support the observation that complete resection and long-term cures are rare, but possible, in appropriately selected patients.

The management of bronchogenic carcinomas that invade vital structures within the mediastinum and chest requires sound surgical judgment and a pragmatic approach. Many of these lesions are metastatic at the time of presentation, and a thorough evaluation, including surgical staging of the mediastinum, must be performed to ensure that patients who are offered surgery have a realistic expectation of benefit. Since the majority of data regarding these subgroups of patients are case series and anecdotal reports, conclusions are difficult to draw. However, resections and the appropriate use of CPB to provide a margin of safety for these extended operations do offer a potential for curative therapy in an appropriately selected patient. Prognosis for these patients is otherwise dismal.

This chapter by Byrne et al. is a very important comment on our technical advances in resecting T4 lung cancer. It is clear from the preceding sections that partial or en bloc resections are possible, and may even confer improved survival upon this subgroup of patients. Although all the reported series are small, the literature does support the value of resection. Using an aggressive approach up front offers the patients the best option for cure. One caveat in my own experience is to exclude patients with either N2 or N3 disease as their long-term survival is limited. By performing a mediastinal staging procedure on all these patients, one can determine in whom the risks of these aggressive operations are most worthwhile.