10

Resection of posterior mediastinal lesions

Joseph B. Shrager

INTRODUCTION

The posterior mediastinum is bounded anteriorly by the posterior pericardium and extends posteriorly to the chest wall and laterally to include the costovertebral sulci. Important structures that it contains include the descending thoracic aorta, inferior vena cava and azygous veins, the sympathetic chains and origins of the intercostal nerves at their nerve roots, and the esophagus and associated vagi. Most anatomic systems consider only lesions that are caudal to the fourth thoracic vertebral body to be within the posterior mediastinum, with more cephalad lesions resting within the superior mediastinum.

The majority of posterior mediastinal masses in adults are benign. They can be classified in a clinically useful way according to whether they are cystic or solid on radiographic evaluation. Cystic masses in this region typically represent bronchogenic cysts or esophageal duplication cysts, whereas solid masses are most commonly benign neurogenic tumors (see Figure 10.1) (e.g., schwannomas, neurofibromas, or ganglioneuromas). These neural tumors usually arise from the sympathetic chain or the proximal intercostal nerves, but often the precise anatomical origin does not become clear until operation. Occasionally, one comes across a patient with a pheochromocytoma or paraganglioma, which arise from the randomly located mediastinal paraganglionic cells and may secrete hormones. At initial office evaluation, one should be alert to the presence of hypertension or palpitations, which should dictate measurement of urine metanephrines. Esophageal leiomyomas (benign intramuscular tumors within the esophageal wall) are also generally grouped amongthe posterior mediastinal lesions. The approach to these lesions and to esophageal duplication cysts differs somewhat from the approach to lesions that are unassociated with the esophagus.

In the modern era, posterior mediastinal masses most often come to light as asymptomatic, incidentally identified, radiographic abnormalities. They do less commonly, however, present with signs of infection (in the case of infected cysts), dysphagia, chest pain, cough or dyspnea, or neurological changes (e.g., Horner’s syndrome from tumors involving the upper sympathetic trunk; lower extremity symptoms from dumbbell paravertebral tumors) resulting from mass effect on adjacent structures.

At present, because of the applicability of low morbidity, minimally invasive approaches to the vast majority of posterior mediastinal masses, most authors recommend resection, even when lesions are asymptomatic. I believe that the recommendation about whether to proceed with resection of a posterior mediastinal mass needs to be individualized. Clearly, a symptomatic lesion is best managed by resection, except in unusual circumstances. When an asymptomaticlesion has all of the radiographic characteristics of a benign cyst or tumor (i.e., smooth margins, simple-appearing cyst material, minimally positron emission tomography [PET] positive if PET has been done), the age and general medical condition of the patient are the key considerations. Since these lesions do generally grow—albeit at a fairly slow rate—it is likely that a young patient will eventually develop symptoms due either to impingement on surrounding structures or infection of a cyst. In patients under the age of approximately 60, then, who are otherwise good surgical candidates, even asymptomatic lesions are best managed by resection. For older patients, or those who have substantial comorbidities, it is perfectly reasonable to follow these benign-appearing, asymptomatic lesions with serial radiographic studies and operate only if they begin to grow dramatically, the patient develops symptoms, or a change in radiographic appearance suggests alignancy. In any case, there is no urgency to remove these cysts and tumors promptly. The significance of the very few reports of evelopment of malignancy within a benign posterior mediastinal mass is probably overstated.

SURGICAL PRINCIPLES

Video-assisted thoracoscopic surgery (or robotics) versus thoracotomy

Resection of posterior mediastinal masses may be accomplished by means of either minimally invasive approaches or thoracotomy. The original, simplest, and probably the least costly minimally invasive approach remains video-assisted thoracoscopic surgery VATS). ¹ Robotic approaches are simply VATS approaches that make use of the robot as an “instrument,” which may provide both certain advantages and certain disadvantages. The procedure, apart from the incisions, is essentially the same whether one uses a minimally invasive approach or thoracotomy, and the goal is, of ourse, complete resection. With some exceptions, minimally invasive approaches are considered preferable to thoracotomy in this setting; it has been well established by now that both VATS and robotics result in less postoperative pain and quicker functional recovery than thoracotomy. ² Some surgeons may still argue that minimally invasive approaches may be more likely to leave a patient with microscopic residual disease. In my own experience and that of many others, however, recurrence of these lesions is very rare after VATS/ robotic excision. Given the low recurrence rate and the fact that these masses are almost always benign, the risk-benefit ratio favors VATS/robotics in nearly all cases, in my opinion.

There are, however, several circumstances in which thoracotomy is indicated from the outset. A suggestion of malignancy (in particular, frank invasion of surrounding structures) on preoperative radiography I believe mandates exploration and resection by thoracotomy; in this situation, the potential consequences of positive margins justify the more aggressive approach. This, however, is quite rare for posterior mediastinal masses. The presence of active infection within a cyst is a relative indication for thoracotomy, in that this usually causes substantial obliteration of normal tissue planes and thereby renders VATS/robotic dissection more difficult and possibly hazardous given the more limited instrumentation and angles that can usually be achieved by these approaches. Solid masses larger than approximately 6 cm also call for an open approach, in my opinion: such lesions are typically more difficult to mobilize safely from underlying structures than smaller lesions, they are more likely to be malignant (though this is still rare), and their removal between the ribs is likely to necessitate substantial rib spreading, which may negate the benefit of pure VATS/robotics. Another approach is to attempt VATS/robotics for these larger lesions, but to have a low threshold to convert to thoracotomy if difficulty is encountered. If the dissection can be completed by VATS, one can then typically remove larger tumors by resecting a small portion of rib without sacrificing an intercostal nerve. It is likely that this will result in less pain and earlier recovery than a standard thoracotomy with rib spreading, but this has never been studied to my knowledge.

Other preoperative issues

A patient with a centrally located cyst should undergo bronchoscopy to rule out the rare occurrence of a communication with the bronchial tree. This may be suggested on computed tomography (CT) scans by the presence of an air-fluid level. If a communication is identified, strong consideration should be made to proceed with thoracotomy rather than VATS. When a cyst arises from or abuts the esophagus, the possibility of a communication between the cyst and the esophageal lumen should be similarly investigated. To rule out this also-rare phenomenon, I obtain a barium contrast study during the preoperative workup, followed by intraoperative esophagoscopy at the commencement of the operation. If acommunication is identified, I prefer thoracotomy. Although there are individuals who would still be comfortable proceeding with a minimally invasive approach, with endoscopic suturing after excision of an esophageal duplication cyst with a communication, safe reapproximation of the esophageal mucosa is the paramount consideration in these cases, and, in my view, this is still most reliably carried out through an open approach.

Preoperative investigation with esophagoscopy should also be done to confirm the presence of intact overlying mucosa in cases of suspected leiomyoma of the esophagus. If the mucosa is intact, the possibility of malignancy is essentially ruled out. Simultaneously, endoscopic ultrasonography maybe performed to establish the depth to which the esophageal wall is involved. With a preoperative diagnosis of probable leiomyoma, VATS is the approach of choice in our practice.

So-called dumbbell neurogenic tumors (tumors that invade the neural foramen and have a spinal canal component) are special cases. Any solid mass in the costovertebral sulcus that cannot be clearly separated on CT imaging from the neural foramen should be evaluated by means of magnetic resonance imaging. Although invasion of the neural foramen by tumor is not in itself an indication for thoracotomy, it does necessitate a combined approach with neurosurgical involvement for the intraspinal portion of the procedure. Several versions of such an approach have been described, including a posterior approach via costotransversectomy or extension of a posterior midline incision into a posterolateral thoracotomy, through which both the intraspinal and intrathoracic components of the tumor can be resected. ^{3 − 6} prefer to perform the the operation using the following approach: under a single anesthetic, the eurosurgeons first resect the intraspinal component (laminectomy and intervertebral foraminotomy), then the patient is repositioned to lateral decubitus and we carry out the remainder of the procedure (via VATS or robotics). Failure to diagnose a dumbbell tumor preoperatively and plan an appropriate combined operation with neurosurgeons may lead to inadvertently cutting through tumor from within the chest. This has the potential to result acutely in tumor hemorrhage within the spinal canal and spinal cord compression, with disastrous consequences. In the long-term, the positive margin of resec-tion will result in local recurrence unless recognized and managed appropriately.

Patients with functioning paragangliomas or pheochromocytomas, if hypertensive, should receive approximately weeks of alpha-adrenergic blockade and volume loading, followed by beta blockade. In these cases, before incision, clear discussion should be undertaken between the surgeon and the anesthesiologist about the intraoperative anesthetic and plan for control of the patient’s blood pressure.

There is, in the vast majority of cases, no advantage to preoperative needle biopsy of posterior mediastinal lesions, although this can usually be readily performed either transthoracically, or, in the case of periesophageal masses, transesophageally at the time of endoscopic ultrasound. Only in cases of invasion suggestive of malignancy might a biopsy alter the therapeutic approach. For example, extremely large or invasive-appearing tumors that turn out to be sarcomas may be best treated by preoperative chemotherapy and/or radiation. It is therefore appropriate to obtain a needle biopsy in the unusual cases in which these features are present.

Although VATS/robotics are often excellent approaches to posterior mediastinal lesions, it must be emphasized that one should never hesitate to convert a minimally invasive procedure to a thoracotomy if required. Accordingly, informed consent to undergo thoracotomy should be sought before operation from all patients being treated for posterior mediastinal lesions, even when VATS or robotics is the intended approach. Further, any patient with a tumor encroaching on the neural foramen should understand preoperatively that there is a very rare possibility of spinal cord compro-mise from the operation, as well as of cerebrospinal fluid (CSF) leak.

SURGICAL TECHNIQUES

I will describe the VATS approaches to these tumors, which are very similar to robotic approaches—simply using different tools to access and manipulate the pathology. It is possible that the “wrists” at the end of robotic instruments afford slightly greater facility over currently available VATS instruments in dissecting posterior mediastinal tumors. However, whether this advantage outweighs the disadvantages of robotics—for example, the increased cost, the loss of tactile sense, and the need for surgeons to undergo a second learning curve—is unclear.

VATS resection of neurogenic tumors of the posterior mediastinum

Resection of a solid neurogenic tumor of the posterior mediastinum that does not invade the neural foramen proceeds as follows. ⁷ The figures are drawn from intraoperative photographs of several separate VATS operations that the author has performed.

STEP 1: INTUBATION AND ENDOSCOPY

The patient is intubated with a double-lumen endotracheal tube to allow single-lung ventilation. Preoperative bronchoscopy (for cystic lesions adjacent to airways) or esophagoscopy (for lesions abutting the esophagus) is performed as indicated (see “Other preoperative issues” section).

STEP 2: PATIENT POSITIONING AND PLACEMENT OF PORTS

The patient is placed in lateral decubitus and stabilized with a bean bag so that the operating table can safely be tilted as much as 45 degrees anteriorly. With this degree of tilt, the lung will almost always fall away from the field of vision; thus, there is rarely a need to place an additional port for a lung retractor or to use carbon dioxide (CO₂) gas insufflation. The lack of need for CO₂ allows one to employ a reusable, metal introducer port for the camera and to avoid altogether the more costly, commercially provided, disposable ports.

The port for the camera is placed through an incision that is generally slightly posterior to the anterior axillary line, in the same rib space or one rib caudal to the craniocaudal level of the mass (see Figure 10.2); if it is placed much more anteriorly than the anterior axillary line, the view of these posterior lesions may be obscured by the lung. I prefer a 5 mm, 30-degree camera, which keeps the incision very small and provides less risk of traumatizing the intercostal nerve, yet provides excellent optics. The 30-degree lens provides much greater versatility and visualization around to the “far side” of lesions than a 0-degree lens.

Stay updated, free articles. Join our Telegram channel

Join