Management of the Neck With Carotid Artery Involvement

Carotid ligation/resection/endovascular stenting

Carotid wall invasion most often arises either from direct extension of a primary head and neck squamous cell carcinoma of the pharynx or from bulky jugular chain lymph node metastasis with extracapsular extension. Although labeled "unresectable" by current AJCC staging criteria, such individuals may be selectively considered for carotid resection as part of primary surgical therapy or a salvage attempt after prior radiotherapy and chemotherapy.

Some recent case series do advocate selective use of carotid resection as part of primary surgical therapy for head and neck carcinomas.[7, 8, 9] These studies suggest better disease-free survival from carotid resection than with nonsurgical therapy for previously untreated patients. This result is not surprising, as primary surgical resection remains the preferred mode of therapy, when deemed feasible, for the types of massive volume tumors that typically lead to carotid invasion. A further rationale for surgery in these untreated patients is that a high proportion of tumors suggested to invade the carotid wall radiographically can ultimately be removed without carotid resection.

Occasionally, benign tumors of the lateral skull base, such as glomus jugulare tumors and schwannomas, as well as various skull base malignancies, may necessitate planned carotid resection.[10] Resection of the internal carotid artery for malignant disease at its entry into the carotid canal can be reconstructed using intracranial bypass but has not been advocated because of unfavorable outcomes.[11]

Clinically, carotid invasion is suggested when a tumor abutting the carotid sheath feels fixed or hypomobile, particularly in the vertical dimension. In the previously untreated neck, carotid arteries that are abutted less than 180o of their circumference by imaging are highly unlikely to require carotid resection. For arteries with greater than 180o of tumor abutment, obliteration of tissue planes between the artery and the tumor on MRI suggests invasion but can be difficult to interpret, particularly in the postradiation salvage setting.

Importantly, involvement of greater than 180° of the carotid circumference is not clearly predictive of histologic invasion,[12] with clinical assessment being of at least equal value. Involvement of the artery 270o or more does accurately predict the surgeon's inability to remove tumor from the artery wall.[13]

In addition, compressive deformity of the artery wall by a closely abutting tumor appears also highly predictive of carotid invasion.[14] In cases in which carotid involvement is feasible based on clinical and radiographic criteria, pursuing surgery requires additional preoperative planning, including involvement of a vascular surgeon in cases in which artery reconstruction is feasible, as well as consideration of angiography with balloon-occlusion testing.

Occasionally, in cases of carotid blowout, emergent carotid ligation and/or resection may be needed without any preoperative testing. In this scenario, reconstruction is still favorable whenever possible, although it may be impractical in a surgical field containing an uncontrolled fistula, which may have caused the rupture itself. Impending rupture is often signaled by minor sentinel bleeding, which can be controlled initially with conservative measures, allowing time for assessment with angiography and consideration of neurointerventional versus open surgical approaches.

In the setting of blowout or sentinel bleed, endovascular stenting may be a useful temporizing or palliative care measure if emergent neurointerventional radiology services are available.[15] However, its use as definitive therapy in patients with long-term survival potential is discouraged.[16, 17] Problems include need for subsequent anticoagulation in a high bleeding-risk setting and intractable bacterial colonization of the stent associated with placement in an infected, radiated wound.

The 3 studies in the table below demonstrate the high morbidity and mortality associated with carotid ligation without reconstruction or preoperative testing. In the largest series, no difference was noted in complications associated with the reason for ligation, which included cancer infiltration, impending rupture, and acute rupture.[18] The incidence of cerebral complications significantly decreased in patients whose common carotids were occluded gradually over 8 days or longer (5.3%), compared with patients with ligation for less than 7 days (30.6%) or those patients with abrupt ligation (42%).

Table. Morbidity and Mortality Associated With Carotid Ligation Without Reconstruction or Preoperative Testing (Open Table in a new window)

The physiology of carotid flow

Preoperative testing and perioperative management of hemodynamics after carotid resection are based on an understanding of cerebral blood flow (CBF) regulation. Under normal physiologic conditions, the average CBF is 50-55 mm/100 g/min, a range that is maintained by the autoregulation capacity of cerebral vasculature. However, in significant hypotension, autoregulation is lost and the CBF fluctuates with arterial blood pressure. Generally, CBF must decrease to 20-25 mL/100 g/min for brain dysfunction to occur. Management of systemic blood pressure can thus be critical for maintaining cerebral perfusion in individuals having undergone carotid resection, even in the absence of immediate posttreatment neurologic sequelae. Delayed onset symptoms and even a cerebral vascular accident (CVA) may develop in patients after carotid occlusion if systemic blood pressure drops.

The timing of permanent brain injury from ischemia has been well characterized in a primate model.[21] Here, the neurologic symptoms that result from obstruction of the middle cerebral artery were partially reversible for up to 3 hours after occlusion. Microscopic infarcts were observed after 15-30 minutes and moderate-to-large infarction 2-3 hours later. After 3 hours, large permanent infarcts developed. With a regional CBF of less than 23 mL/100 g/min, reversible paralysis occurred. With a regional CBF of less than 10-12 mL/100 g/min for 2-3 hours or of less than 17-18 mL/100 g/min during permanent occlusion, the animals developed irreversible neurologic sequelae.

Stump pressure is an important concept for intraoperative decision making in managing cases of sudden rupture or unexpected carotid involvement.[22] Brisk backflow from the distal carotid stump is a reflection of stump pressure, which is regarded as an indicator of adequate collateral blood flow when the carotid is occluded proximally. Though rarely used clinically, this value may be measured with a strain gauge attached to a 19-gauge needle in the experimental setting. Although stump pressures of more than 50-70 mm Hg are considered low risk, caution is still warranted because intraoperative electroencephalogram changes have been demonstrated at higher pressures.[23] In general, the common carotid artery produces higher stump pressures than the internal carotid artery, if the external carotid artery system is intact and thus available to provide backflow. This difference accounts for the significantly higher risks associated with ligation of the internal versus common carotid artery.

Contraindications to surgical management of the neck with carotid artery involvement are based on the patient's comorbidities and ability to tolerate surgery, as well as the technical feasibility of extirpating the tumor. Although few absolute contraindications exist, decision making is heavily influenced by the patient's overall functional status, the anticipated natural course of the tumor, consideration of nonsurgical options, and the patient's level of enthusiasm for surgery given the risk of severe neurologic sequelae or even death.

See the list below:

• Angiography: The initial assessment for risk of stroke during internal carotid occlusion consists of 4-vessel angiography, which establishes the patency of the vessels and the potential availability of collateral flow through the circle of Willis if one carotid is occluded.

• Trial balloon occlusion

  • When collateral flow is present based on angiography findings, temporary preoperative occlusion provides physiologic information on the patient's ability to tolerate transient or permanent occlusion of the carotid to be resected. The original diagnostic trials of carotid occlusion were performed intraoperatively on the common carotid artery using umbilical tape under local anesthesia.[24] Trial balloon occlusion (TBO) performed during angiography has replaced operative trial occlusion. The patient is heparinized while the balloon catheter is placed under angiographic guidance. It is inflated in the internal carotid artery for up to 15-30 minutes while the patient is monitored for development of neurologic signs and symptoms.

  • For multiple reasons, even when the patient tolerates TBO, a cerebral vascular accident (CVA) may develop if simple operative ligation is ultimately performed. First, thrombus is thought to develop in the distal carotid stump, leading to subsequent delayed embolization. Second, intraoperative blood loss and decreased systemic blood pressure under general anesthesia may decrease regional cerebral blood flow (CBF) more than balloon occlusion or ligation alone. For this reason, a hypotensive challenge during TBO may improve the predictive value of the test but is not routinely performed at most centers.[25]

• Flow Scanning

  • Single-photon emission computed tomography scanning

    • Because of the inadequacy of predicting cerebral ischemia with TBO alone, CBF studies during carotid occlusion have been developed. Single-photon emission computed tomography (SPECT) imaging using technetium-99m hexamethylpropyleneamine oxime (Tc-99m HMPAO) provides a semiquantitative comparison of blood flow to each hemisphere. After the patient has tolerated TBO, Tc-99m is injected intravenously, with the balloon kept inflated for an additional 15-30 minutes. Tc-99m is converted to a hydrophilic form inside the brain that is retained for hours. Because the half-life of Tc-99m is 6 hours, SPECT scanning can be delayed until after completion of angiography and TBO.

    • Tc-99m SPECT scanning does not provide an exact measurement of regional CBF. Rather, it is evaluated based on differences in tracer retention between the 2 sides, and adequate CBF is defined as less than a 10% difference between hemispheres.

  • Xenon flow scanning

    • During TBO, CBF may rise or fall in either hemisphere, resulting in a significant difference in CBF between hemispheres. Thus, determination of the absolute CBF may be helpful in overcoming this ambiguity inherent in interpretation of SPECT scans. Xenon-133 scans offer a more quantitative measurement of regional CBF but are technically more difficult than SPECT when used in conjunction with TBO.[5] Although the carotid is occluded, the patient must inhale xenon gas and, unlike with SPECT, undergo simultaneous nuclear medicine scanning because xenon is rapidly absorbed and discharged.

    • The primary difficulty with this technique lies in the need to take the patient from the angiography suite to the nuclear medicine scanner with certainty that the balloon is still in place. In order to perform this procedure optimally, the nuclear medicine scanner needs to be present in the angiography suite. With this technique, adequate CBF is defined as greater than 30 mL/100g/min, although some require a threshold of as high as 40 mL/100g/min to be considered low risk.[26]

    • Regional CBF may also be measured using stable xenon as a CT contrast agent. This method requires transferring the patient to the CT scanning unit with a catheter in place and with the same considerations as in xenon flow scans. The quantitative results from this method have been shown to prevent frequent misinterpretation of flow asymmetries between hemispheres that may lead to a false-positive reading of a SPECT scan.[27]

Preoperative carotid occlusion

Carotid reconstruction cannot be performed in some patients, particularly individuals with the internal carotid artery resected close to the skull base, where sewing a vascular graft to the distal stump may not be feasible. After normal trial balloon occlusion (TBO) and flow testing results, permanent balloon occlusion is a preoperative intervention that may reduce cerebral vascular accident (CVA) incidence over simple ligation in this clinical setting.[6] The underlying principle is that high embolization of the carotid eliminates the standing column of blood present after ligation that is thought to serve as a later source of stump emboli. The method involves angiographic placement of permanent balloons or coils in the carotid siphon region proximal to the ophthalmic artery. Typically, the patient is heparinized, and hemodynamics are closely monitored for 72 hours.

Carotid resection is delayed by 2 weeks to allow for fixation of the coils and to avoid adverse hemodynamic effects from surgery during the vulnerable period immediately following occlusion.

In an early study of this technique, all 8 patients who underwent preoperative permanent occlusion tolerated it without sequelae.[28] However, the application of this technique has still been associated with neurologic complications with protracted intraoperative hypotension; in one case, migration of a balloon was also reported.[2] A further disadvantage of this technique is that it must be applied preoperatively and may thus subject a patient to unnecessary risk if the carotid artery turns out to be uninvolved at the time of surgery.

Permanent balloon occlusion may also be performed without surgery to manage impending carotid rupture. In one series, 22 patients were treated by placing 2 permanent balloons just proximal to the ophthalmic artery and embolizing the internal carotid artery (ICA) down to the level of the carotid bifurcation with liquid biological adhesive (Histoacryl). None of the 22 patients had an immediate complication from the permanent occlusion, although 2 patients developed progressive hemiplegia that began 24 hours later.[29]

Surgical decision making

Although the type of preoperative scanning and precise technique used may vary, patients may be placed into 3 categories based on trial balloon occlusion (TBO) and flow scanning results, as follows:

• High risk - Failed TBO, no cerebral blood flow (CBF) scans obtained

• Moderate risk - Passed TBO, inadequate CBF scan

• Low risk - Passed TBO, adequate CBF scan

Moderate- and high-risk patients usually undergo reconstruction if carotid resection is performed. Although the best management of low-risk patients is less clear, these patients likely also benefit from reconstruction whenever possible. A few patients in the low-risk category undergoing carotid ligation still experience neurologic sequelae, presumptively from inadequately sensitive flow scan workups, perioperative hypotension, or carotid stump emboli. This fact has led some authors to advocate vein graft reconstruction of the artery whenever technically feasible.[1, 2, 30]

Although unusual, neurologic complications may still occur in the face of vein grafting, even in low-risk patients.[31, 32] Such events occur despite heparinization and placement of a temporary shunt to maintain cerebral perfusion during reconstruction, an essential step in high-risk patients. A clot in the graft may be a potential source for an embolic cerebrovascular accident in some cases.

Other operative considerations include attention to the possibility of preserving the external carotid artery, which often requires resection in patients with head and neck cancer. Backflow from an intact external carotid can, in principle, both augment cerebral perfusion and prevent the development of stump emboli. Preservation of the external carotid is, not surprisingly, associated with an approximately 50% decrease in the CVA rate, as apparent from multiple reports.[33, 29, 34, 18, 35, 36]

Lastly, whether or not to reconstruct the carotid must be decided in the larger context of the total resection and reconstruction to be performed. In previously irradiated patients, postoperative exposure of a carotid reconstruction from wound breakdown or contact with fistula drainage may risk lethal hemorrhage and instead bias one's management toward preoperative occlusion of the artery. When reconstruction is performed, attention must be given to adequate protection of the carotid from pharyngeal secretions and coverage with well-vascularized tissue, using pedicled or free tissue transfer to accomplish these ends as necessary.

Neurologic complications

The early cerebral vascular accident (CVA) risk is well described and guides much of the perioperative evaluation and management efforts when carotid resection is considered. However, reports of long-term follow-up in patients with occluded carotid arteries demonstrate a delayed CVA rate as high as 25 times that of the general population. In a report of 814 cases of carotid occlusion performed for intracranial aneurysm, 233 patients developed ischemic symptoms after occlusion.[33] Of these, 79% occurred within the first 48 hours and 10% occurred in the second 48 hours. However, 5 patients had ischemic symptoms at 6 months, 11 months, 12 months, 18 months, and 4 years, respectively. Late ischemic complications have also been confirmed in other reports.[37]

Nerve deficits

Resection of malignant disease that involves the carotid wall typically requires an en bloc resection of other adjacent involved structures, which often include the vagus nerve, the hypoglossal nerve, the spinal accessory nerve, and the cervical sympathetic chain. Particularly when compounded with other deficits, combined vagus and hypoglossal palsies may produce lasting postoperative dysphagia and aspiration, and patients must be apprised of the risks of such disabilities before surgery.

Long-term survival is generally poor in cases of malignant carotid involvement with squamous cell carcinoma;[38] early local recurrence and rapid failure with regional or distant metastatic disease are common. A study by Shinomiya et al reported that in patients with T4 squamous cell carcinoma of the external auditory canal and middle ear, invasion of the brain, carotid artery, and/or jugular vein indicates a poor prognosis. The investigators found that patients with at least one of these factors had a 5-year overall survival rate of 25.5%, versus 65.5% in those without these factors.[39]

However, no universal case against resection of the carotid can be made for involvement of the artery in squamous cell carcinoma, and each patient merits careful consideration within his or her broad clinical context. In fact, data show significant numbers of long-term survivors among previously untreated patients undergoing carotid resection.[7, 8, 9, 40] These favorable surgical outcomes for very biologically aggressive tumors suggest that carotid resection does still have a limited role in the management of the head and neck cancer. Patients require extensive preoperative counseling and must contemplate surgery bearing in mind substantial risks, including those of devastating neurologic sequelae. This risk can be managed significantly by carotid reconstruction whenever possible, but adverse neurologic outcomes from carotid resection cannot be preventedaltogether.