Patients with head and neck cancers and associated postirradiated carotid blowout syndrome: Endovascular therapeutic methods and outcomes
Article Outline
Purpose
This study retrospectively evaluated the technical and hemostatic outcomes of reconstructive and deconstructive endovascular management in patients with head and neck cancers associated with carotid blowout syndrome (CBS).
Methods
Twenty-four patients with head and neck cancers with CBS involving the main trunk of carotid artery underwent endovascular therapy. This included reconstructive management with self-expandable stent grafts to preserve the diseased carotid artery in 11 patients and deconstructive management with balloons, coils, or acrylic adhesives to occlude the diseased carotid artery in 13 patients. Based on clinical severity and therapeutic priority, we classified CBS in our patients into two groups: acute or impending and threatened. The angiographic severity was graded from 0 to 3. Evaluation of technical outcome included technical success, initial and delayed complications, and patency of stent graft in the reconstructive group. The hemostatic outcome was evaluated by immediate hemostatic result, rebleeding, and duration of hemostasis. Sex, age, clinical and angiographic severities, local wound complications, and location of the pathologic lesion were examined as predictors of the technical and hemostatic outcomes of endovascular management by using Cox regression method.
Results
Technical success and immediate hemostasis were achieved in all patients of both groups. Initial complications during the procedures were encountered in four patients (36.4%) who underwent reconstructive management and in one patient (7.7%) who underwent deconstructive management (P = .142). Delayed complications during the follow-up were seen in one patient (9.1%) with reconstructive management and one patient (7.7%) with deconstructive management (P > .99). Rebleeding occurred in five patients (45.5%) in the reconstructive management group and in three patients (23.1%) in the deconstructive management group (P = .659). The mean duration of hemostasis after initial reconstructive and deconstructive management was 4.0 ± 8.1 and 8.5 ± 10.1 months, respectively (P = .249). Rebleeding was noted in 7 of 11 patients (63.6%) with acute CBS and in 1 of 13 patients (7.7%) with impending and threatened CBS (P = .008).
Conclusion
There is no significant difference in technical and hemostatic outcomes between the reconstructive and deconstructive endovascular management methods. Hemostatic results were influenced by clinical severity. The rebleeding rate is higher in patients with advanced and acute clinical severity.
Carotid blowout, or rupture of the carotid artery (CA), is a life-threatening complication associated with head and neck cancer and its therapy.1, 2 The reported incidence of CA rupture after radical neck dissection is 4.3%.3 Patients with carotid blowout syndrome (CBS) can have a variety of clinical presentations due to rupture of the CA, including acute hemorrhage or exposure of part of the CA.4, 5, 6 Carotid blowout syndrome tends to occur in patients with head and neck cancers and those with radiation-induced necrosis, recurrent tumors, wound complications, or pharyngocutaneous fistulas.1, 2, 3, 4
Emergency surgical management of CBS is often technically difficult to perform in previously irradiated areas and is associated with high neurologic morbidity and mortality rates. The reported average neurologic morbidity and mortality rates associated with surgical management of CBS are 40% and 60%, respectively.7 Deconstructive endovascular therapy, such as with permanent balloon occlusion of the diseased carotid artery, has improved outcomes1, 6, 7; however, as many as 15% to 20% of patients with CBS who are treated with permanent carotid occlusion experience immediate or delayed cerebral ischemia.1, 8 A balloon occlusion test may be performed before threatened CBS is treated definitively, but this test is usually not possible in acute cases. In addition, test occlusion or even positron emission tomography studies may not help in identifying the small subset of patients in whom delayed hemodynamic ischemia develops after the carotid artery is permanently occluded.1, 4, 7, 8, 9 These results highlight the limitations of deconstructive endovascular therapy for patients with CBS.
Stent grafting has the potential to preserve the diseased CA and achieve hemostasis. Some authors report it as promising in treating CBS in patients at risk of CA occlusion8, 10; however, some studies have described unfavorable long-term outcomes.5, 11, 12, 13 Therefore, the purpose of this study was to compare the technical and hemostatic outcomes of endovascular reconstruction using self-expandable stent grafts with those of endovascular deconstruction using balloons, coils, or acrylic adhesive in the management of CBS in patients with head and neck cancers. We also provide an assessment of the clinical severity in our patients compared with their angiographic findings to highlight the importance of early diagnosis and management of CBS.
Materials and methods
Patient population
This retrospective study was exempt from Institutional Review Board approval. Written consent was obtained from each patient or the family before intervention. From 2003 to 2006, 1632 patients with head and neck cancers were treated in our institute, and 56 patients sustained CBS. Two patients could not accept either surgical or endovascular management due to profuse hemorrhage with hypovolemic shock. Two patients were treated with surgical ligation.
The study excluded 28 patients because their pathologic lesions involved the branches of external carotid artery (ECA) that were only eligible for deconstructve management. The study included 24 patients with CBS involving the main trunk of the CA (Table I). We included these patients because the locations of their pathologic lesions could be treated by either reconstructive or deconstructive endovascular therapy. Eleven men with a mean age of 49.2 ± 7.8 years (range, 35-65 years) underwent reconstructive endovascular surgery in which self-expandable stent grafts were placed to preserve the diseased CA. Thirteen patients (10 men and 3 women) with a mean age of 51.4 ± 9.9 years (range, 34-67 years) underwent deconstructive endovascular surgery in which the permanent CA occlusion was done. All patients had received radiation therapy or chemoradiotherapy and presented with various degrees of irradiation-induced change in their head and neck regions. Previous surgical therapy for the malignancy had been done in 8 of the 11 patients in reconstructive group and 10 of the 13 patients in the deconstructive group.
Table I. Baseline characteristics of patients with reconstructive and deconstructive endovascular management
| Variable | Reconstructive (n = 11) | Deconstructive (n = 13) | P |
|---|---|---|---|
| Sex, No (%) | |||
 Male | 11(100) | 10(76.9) | .223 |
| Female | 0(0) | 3(23.1) | |
| Age, mean ± SD (range) years | 49.2±7.8(35-65) | 51.4±9.9(34-67) | .557 |
| Clinical cancer diagnosis, No. | |||
| Nasopharyngeal | 1 | 3 | .596 |
| Hypopharyngeal | 5 | 1 | .061 |
| Laryngeal | 2 | 1 | .576 |
| Tongue | 1 | 1 | 1 |
| Buccal | 3 | .223 | |
| Sinus | 2 | .482 | |
| Parotid gland | 1 | 2 | 1 |
| Tonsil | 1 | .458 | |
| Local complications, No. | |||
| Fistular formationa | 5 | 6 | .973 |
| Wound complicationsb | 10 | 11 | 1 |
| CBS group, No (%) | |||
| Acute | 5(45.5) | 6(46.1) | .973 |
| Impending and threatened | 6(54.5) | 7(53.9) | |
| Angiographic grade, No (%) | |||
| 0 and 1 | 3(27.3) | 5(38.5) | .769 |
| 2 and 3 | 8(72.7) | 8(61.5) | |
| Location of lesion (n = 28) | 13 | 15 | |
| ICA, No (%) | 2(15.5) | 7(46.7) | .115 |
| CAB, No (%) | 5(38.4) | 3(20) | .410 |
| CCA, No (%) | 6(46.1) | 2(13.3) | .096 |
| ECA, No (%) | 0(0) | 3(20) | .226 |
aPharyngocutaneous or aerodigestive. |
bNecrosis, ulceration, sinus tract. |
Clinical severity was used to classify CBS into three types: acute, impending, and threatened1, 7:
Based on therapeutic priority, we further classified CBS in our patients into two groups: (1) acute and (2) impending and threatened. The patients of the former group had a clinical emergency that required emergency treatment. The latter group was composed of patients whose status could be treated more electively.
Locations of pathologic vascular lesions, such as pseudoaneurysms, were recorded as the internal carotid artery (ICA), carotid bifurcation (CBF), ECA, or common carotid artery (CCA). The ECA was designated the main trunk of ECA proximal to the orifice of linguofacial trunk.
If endovascular management was anticipated, a balloon test occlusion was attempted if the patient was hemodynamically stable and not bleeding profusely. A standardized protocol has been described in detail in the literature.1, 7 The indications for reconstructive endovascular therapy were the patients at risk of permanent carotid occlusion, such as incomplete circle of Willis on angiograms (patients 3 and 4), contralateral carotid severe stenosis or total occlusion, intolerance to a balloon occlusion test (patients 8, 10, and 11), or emergency status of the patient precluding an occlusion test (patients 1, 2, 5-7, and 9).7, 9 The deconstructive method was done for patients without the above risks of carotid occlusion (patients 1-5, 7, and 10-12) or for those refused reconstructive management (patients 6, 8, 9, and 13).
Angiographic evaluation
We used a transfemoral arterial approach to obtain a complete neuroangiogram of the supra-aortic arteries and their branches. The angiographic findings were used to classify the severity of vascular injury as grades 0 to 3.14 A grade of 0 meant no angiographic vascular disruption. Grade 1 was defined as focal irregularity or slight focal bulging of the diseased CA, such as a focal weakening in the vascular wall. Grade 2 was defined as a pseudoaneurysm of the injured CA, in which there was a focal CA rupture confined by the integrity of the surrounding tissue. Grade 3 was defined as active extravasation from the completely ruptured CA. We further classified grades 0 and 1 as slight carotid injury and grades 2 and 3 as advanced carotid disruption. The former group had a complete or weakening vascular wall. The latter group had an incomplete or ruptured vascular wall. We compared these angiographic findings with the clinical severities and patient outcomes.
Medication for reconstructive management
Patients with threatened CBS were premedicated with a dual antiplatelet regimen consisting of orally administered aspirin (324 mg) and clopidogrel (300 mg) 1 day before treatment. Patients with acute or impending CBS were prophylactically given intravenous glycoprotein IIb/IIIa receptor inhibitor (Aggrastat; Merck & Co, Inc, West Point, Pa) during the interventional procedure.11 We gave an initial intravenous infusion at 0.4 μg/kg/min for 15 to 20 minutes, followed by continuous infusion at 0.1 μg/kg/min for 4 to 6 hours after the procedure. Approximately 50 to 70 U/Kg of heparin was also given to keep the activated clotting time >250 seconds.
After deployment of the self-expandable stent graft, a dual antiplatelet regimen with aspirin (324 mg) and clopidogrel (75 mg) was begun. One month later, this regimen was changed to aspirin (100 mg) for life-long use. Because patient 4 sustained a brain abscesses after the stent graft deployment, we gave 4-week prophylactic antibiotic therapy to patients 7 through 11.
Reconstructive management
After identifying the pathologic lesions, we exchanged the diagnostic catheter and wire to a 10F or 11F introducer sheath and a .0.35-inch, 300-cm Amplatz exchange wire (Cook, Bloomington, Ind) through the right femoral artery into the cervical ICA. A self-expandable Wallgraft stent graft (Boston Scientific Corp, Natick, Mass) was then advanced along this exchange wire to the CA, where it was appropriately deployed11 (Fig 1). To avoid rebleeding from reconstitution through the branches of ECA, we placed fiber coils in the main trunk of the ECA before deployment of stent grafts if the pathologic lesions were close to the CBF. A control angiogram was obtained immediately and 15 minutes after deployment of the stent graft to confirm appropriate positioning of the stent graft and patency of the CA.11 The reconstructive endovascular management was considered complete when adequate coverage of the pathologic lesion by stent graft was achieved or clinical hemostasis was reached, or both.
Fig 1.
A, Left carotid angiogram in patient 10 of the reconstructive group showed only slight stenosis in the common carotid artery (grade 0, arrow). B, Contrast-enhanced axial computed tomography (CT) of the neck revealed a necrotic sinus tract (arrowheads) close to the left common carotid artery (arrow). C, Reconstructive CT angiography (curved multiplanar reformatted images) of the left carotid artery 1 month later showed patency of the stent grafts. D, Four months later, obvious distal marginal stenosis was noted (arrowheads). Large area of soft tissue necrosis and ulceration (arrows) surrounding the stent grafts caused “floating” of left carotid artery (long arrows).
Deconstructive management
A 7F Shuttle sheath (Cook, Minneapolis, Minn) was placed through the femoral artery to the diseased CCA. Two types of deconstructive management were used in this study: cross occlusion and proximal occlusion. Cross occlusion consisted of deployment of embolic materials from the pathologic lesion or from the CA distal to the pathologic lesion to its proximal site (Fig 2). In such situations, we advanced two microcatheters into the CA. One “distal microcatheter” was advanced to the CA of the pathologic lesion or distal to it, and the other “proximal microcatheter,” mounted with a detachable balloon, was placed in the CA proximal to the lesion.
Fig 2.
A, Right carotid angiograms in patient 8 of the deconstructive group showed a ruptured internal carotid artery with active extravasation (grade 3, arrows). B, Cross occlusion was performed with deployment of two balloons distal and proximal to the pathologic lesions (arrows). A mixture of liquid adhesives was injected in the internal carotid artery between the 2 balloons (arrowheads).
We inflated and deployed the detachable balloon from the proximal microcatheter first. We then deployed microcoils (Target Therapeutics, Fremont, Calif), injected acrylic adhesive (Histoacryl, Braun, Germany), or deployed a premounted balloon through the distal microcatheter to occlude the pathologic lesion and its adjacent CA. Proximal occlusion consisted of placing the embolic materials in the CA proximal to the pathological lesions. At least two detachable balloons were serially inflated and deployed to ensure permanent vascular occlusion.1 Proximal occlusion was used in cases when associated focal carotid stenosis or tortuosity impeded the balloon or microcatheter positioning to cross the lesions. Successful deconstructive management was defined as complete obliteration of the pathologic lesion and the related CA and the achievement of clinical hemostasis.
Outcome evaluation
Immediate postprocedural outcomes were evaluated by the interventional neuroradiologists or by a clinical oncologist, or both. Evaluation of technical outcomes included technical success, initial and delayed complications, and patency of the stent grafts of the reconstructive group. The complications presented during the therapeutic procedures were defined as “initial,” and those presented after the procedures were defined as “delayed.” Hemostatic outcomes were evaluated by the immediate hemostatic results, presence of rebleeding, and duration of hemostasis. Patients in the reconstructive group underwent follow-up contrast-enhanced computed tomography (CT), CT angiography, ultrasonography, or conventional angiography within the first month and then every 2 to 4 months so that patency of the stent grafts could be assessed. Patients in the deconstructive group underwent regular clinical outpatient department follow-up every 6 months. If rebleeding occurred, emergency angiography and interventional management were performed. Follow-up lasted a median of 4 months (mean, 9.34 ± 11.42 [range, 0.1-37] months).
Statistical analysis
We analyzed the technical outcomes and hemostatic outcomes of our patients, including technical success, initial and delayed complications, patency of the stent grafts in the reconstructive group, immediate hemostasis, presence of rebleeding, and duration of hemostasis after initial intervention by using the t test, χ2 test, or Fisher exact test, when appropriate. Age, sex, type of CBS, local wound complication, angiographic severity, location of the pathologic lesion, and type of endovascular management were examined as predictors for outcome analysis by using Cox regression analysis. Clinical severity of CBS was correlated with angiographic severity by using the Fisher exact test. For all analyses, P < .05 was considered statistically significant.
Results
Baseline characteristics
Table I summarizes the baseline characteristics of all patients. Adjustments for sex, age, clinical diagnosis, local wound complication, clinical and angiographic severity, and location of pathologic lesion (including lesions in ICA and other than ICA) did not affect the results.
Clinical and angiographic classification
The correlation between clinical and angiographic severity is shown in Fig 3. Pseudoaneurysms (grade 2) were the most common angiographic findings and were present in all clinical groups. Active extravasation (grade 3) was only noted in patients with acute CBS. All 11 patients with acute CBS had advanced (grade 2 and 3) carotid disruption on angiogram (100%). Clinical severity correlated well with the angiographic severity (P = .002).
Fig 3.
Graph shows correlation of clinical severity (groups: acute, impending, and threatened) with angiographic severity (grades 0, 1, 2, and 3).
Technical outcome
Endovascular management was successfully accomplished in all 24 patients during the initial procedures (Table II, Table III, Table IV). Four initial complications (36.4%) occurred in the 11 patients in the reconstructive group, including acute thromboembolism in three patients, and dissection and a type 3 endoleak in one patient. Only one of these four initial complications was symptomatic (patient 3). A delayed complication in the reconstructive group included septic thrombosis of the stent graft with multiple brain abscesses in one patients (9.1%).
Table II. Summary of reconstructive management of 11 patients with carotid blowout syndrome
| Patient | Technical outcome | Hemostatic outcome | |||
|---|---|---|---|---|---|
| Initial management: stent graft (mm)a | Complications: initial/delayed (time) | Follow-up for stent patency | Time of rebleeding/reintervention | Outcome/timeb (cause of death) | |
| 1 | 8 × 50 | Acute asymptomatic ICA thrombosis, occlusion/none | Not applicable | 0.6 mo, disease progression/none | Died/0.6 mo (rebleeding) |
| 2 | 8 × 30 | None/none | Not applicable | 0.1 mo, inadequate coverage of the lesion/none | Died/0.1 mo (rebleeding) |
| 3 | 8 × 50c | Embolic infarct, treated with thrombolytic therapy/none | 1 mo: patency | None | Died/2 mo (disease progression) |
| 4 | 8 × 50c | None/brain abscess (4 mo) | 2 mo: patency; 4 mo: septic thrombosis | None | Alive/28 mo |
| 5 | 9 × 70 | None/none | 0.5 mo: patency | None | Died/1.5 mo (disease progression, mediastinitis) |
| 6 | 7 × 30 | Transient asymptomatic in-stent thrombosis/none | 3 mo: stenosisd; 6 mo: asymptomatic thrombosis | 0.5 mo, disease progression/9 × 50 Wallgraft | Died/36 mo (lung metastasis) |
| 7 | 8 × 50 | None/none | 3 mo: patency; 3.7 mo: stenosisd; 6 mo: asymptomatic thrombosis | 2 mo, disease progression/8 × 50 Wallgraft; 3 mo/direct percutaneous puncture of ECA for embolization | Alive/27 mo |
| 8 | 9 × 70 | None/none | 0.5 mo: patency | None | Died/1 mo (lung metastasis) |
| 9 | 8 × 50c | Asymptomatic CA dissection treated with 7 × 40 Wallstente/none | 4 mo: stenosisd treated with 7 × 50 Wallstente; 9 mo: patency | 0.02 mo, type 3 endoleak by the Wallstente/9 × 70 Wallgraft, direct percutaneous puncture for embolization; 9 mo, recurrent type 3 endoleak/10 × 70-mm Wallgraft | Died/9.1 mo (transfusion complication, sepsis) |
| 10 | 8 × 50, 9 × 70c | None/none | 4 mo: stenosisd | None | Died/4.5 mo (sepsis, disease progression) |
| 11 | 8 × 50, 9 × 70c | None/none | 4 mo: stenosisd | None | Died/4 mo (UGI bleeding, transfusion complication) |
aWallgraft: Boston Scientific Corporation, Natick, Mass. |
bTime of follow-up. |
cPlus fiber coils in the external carotid artery. |
dDistal marginal stenosis. |
eCarotid Wallstent: Boston Scientific Corporation. |
Table III. Summary of deconstructive management of 13 patients with carotid blowout syndrome
| Patient | Technical outcome | Hemostatic outcome | ||
|---|---|---|---|---|
| Initial management: permanent carotid occlusion (method) | Complications: initial/delayed (time) | Time of rebleeding | Outcome/timea (cause of death) | |
| 1 | NBCA w/prox balloon (B) | None/none | None | Died/10 mo (multiple metastasis, sepsis) |
| 2 | NBCA w/prox balloon (B) | None/none | None | Alive/37 mo |
| 3 | NBCA w/prox balloon (B) | None/none | None | Alive/4 mo |
| 4 | 2 balloons (B) | None/none | None | Died/0.5 mo (radiation myelopathy, sepsis) |
| 5 | 2 balloons (A) | None/none | None | Died/2 mo (recurrent tumor, sepsis) |
| 6 | Coils, NBCA w/prox balloon (B) | None/none | None | Died/5 mo (brain metastasis, sepsis) |
| 7 | 2 balloons (A) | None/none | 3 mo (2 mo: drop off the balloon by débridement) | Died/3 mo (disease progression, rebleeding) |
| 8 | Distal balloon, NBCA w/prox balloon (B) | None/none | None | Died/18 mo (sepsis, disease progression) |
| 9 | 2 balloons (A) | Embolic infarct/none | 3 mo | Died/3 mo (rebleeding) |
| 10 | 2 balloons (A) | None/brain abscess (0.5 mo) | 3 mo | Died/3 mo (recurrent tumor w/rebleeding) |
| 11 | NBCA w/prox balloon (B) | None/none | None | Died/1 mo (disease progression, sepsis) |
| 12 | 2 balloons (A) | None/none | None | Alive/14 mo |
| 13 | 10 fiber coils w/prox balloon (B) | None/none | None | Alive/10 mo |
aTime of follow-up. |
Table IV. Analysis of technical and hemostatic outcomes of reconstructive and deconstructive endovascular management
| Outcome | Reconstructive (n = 11) | Deconstructive (n = 13) | P |
|---|---|---|---|
| Technical outcome, No. (%) | |||
| Initial complication | 4(36.4) | 1(7.7) | .142 |
| Delayed complication | 1(9.1) | 1(7.7) | >.99 |
| Hemostatic outcome | |||
| Rebleeding, No. (%) | 5(45.5) | 3(23.1) | .659 |
| Hemostasis duration, mean ± SD mo | 4.0±8.1 | 8.5±10.1 | .249 |
| Survival time, mean ± SD mo | 11.9±4.8 | 12.2±4.1 | .547 |
Of the 13 patients in the deconstructive group, an initial complication was an acute infarction in the territory of MCA in one patient (7.7%). One patient (7.7%) also presented with brain abscess as a delayed complication after deconstructive management. An initial complication was noted in one of six patients (16.7%) in the deconstructive group with pathologic lesions located in the CA other than ICA and no patients with a lesion located in the ICA (P = .462). Delayed complications were found in two of seven patients (28.6%) with pathologic lesions located in ICA and one of six patients (16.7%) with lesions located in the CA other than ICA (P = .612).
Follow-up imaging performed ≤3 months demonstrated that nine patients (except for patients 1 and 2) in the reconstructive group had patent stent grafts. Varying degrees of distal marginal stenosis were noted in five of the six patients (83.3%) in the reconstructive management group after a 3-month follow-up. One (patient 9) of these five patients was successfully treated with angioplasty and stenting 3 months after the initial intervention. Of the six patients who were followed up longer than 3 months, three (50%) had occluded stent grafts, and two of these (patients 6 and 7) were asymptomatic. Patient 4 presented with septic thrombosis and multiple brain abscesses. The mean duration of stent graft patency was 3.0 ± 2.6 months.
Hemostatic outcome and survival analysis
Immediate hemostasis was achieved in all patients in both groups after the interventional procedures (Table II, Table III, Table IV). Seven episodes of rebleeding were noted in five of the 11 patients (45%) in the reconstructive group. All of them had acute CBS, and rebleeding resulted in death in two of these patients. The other three patients, in whom five episodes of rebleeding from the same diseased CAs occurred, were successfully managed by reintervention. The duration of hemostasis after initial reconstructive management was a mean of 4.0 ± 8.1 months (range, 0.02-9 months). Three of the 13 patients (23.1%) had rebleeding 3 months after initial deconstructive management. All who had rebleeding had undergone proximal occlusion (3 of 5 patients, 60%). The mean duration of hemostasis after initial deconstructive management was 8.5 ± 10.1 months (range, 0.5-37 months).
The mean survival of patients was 11.9 ± 4.8 months in the reconstructive group and 12.2 ± 4.1 months in the deconstructive group. No predictors for survival were found using Cox regression analysis.
Statistical analysis
There were no significant differences in initial and delayed complications, rebleeding rate, duration of hemostasis, and survival time between the reconstructive and deconstructive groups (Table IV). Patients in the deconstructive group who had undergone cross occlusion had a lower rebleeding rate than those who underwent proximal occlusion (0% vs 60%, P = .035).
The rebleeding rates of hemostatic outcome were correlated with clinical severity (Table V). Patients with acute CBS had a higher bleeding rate (P = .008). There were no statistically significant differences in technical outcome and duration of hemostasis with clinical severity.
Table V. Analysis of clinical severity of carotid blowout syndrome and endovascular outcomes
| Variable | Acute (n =11) | Impending Threatened (n = 13) | P |
|---|---|---|---|
| Angiographic grade, No. (%) | |||
| 0 and 1 | 0(0) | 8(61.5) | .002 |
| 2 and 3 | 11(100) | 5(38.5) | |
| Technical outcome, No. (%) | |||
| Initial complications | 4(36.3) | 1(7.7) | .142 |
| Delayed complications | 1(9.1) | 1(7.7) | >.99 |
| Hemostatic outcome | |||
| Rebleeding, No. (%) | 7(63.6) | 1(7.1) | .008 |
| Hemostasis duration, mean ± SD mo | 3.9±5.5 | 8.6±11.4 | .231 |
Discussion
Carotid blowout syndrome in patients with head and neck cancers often results in catastrophic hemorrhage. Although deconstructive endovascular therapy has improved patient outcomes, it carries a risk of cerebral ischemia.1, 6 Reconstructive endovascular therapy for CBS with stent grafts has been proposed; however, recent reports have shown unfavorable durable hemostasis and long-term outcomes.5, 7 In our previous study, we treated eight CBS patients with high risk of carotid occlusion with stent grafts.11 The results were not satisfactory on the basis of their poor technical outcomes (no stent graft patency after 3-month follow-up) and poor hemostatic outcomes (50% rebleeding rate). We therefore ended the application of stent grafts to treat these cancer patients with CBS in emergency or temporary practice.
Recently, however, we found distal marginal stenosis was a common cause of stent graft occlusion and inadequate coverage of the ongoing pathologic lesion by the stent graft was a cause of the rebleeding. With technical improvement to manage these complications, we have seen better recent outcomes than the previous work. Therefore, we designed this study to test if endovascular reconstruction by using self-expandable stent grafts has worse technical and hemostatic outcomes than those of endovascular deconstruction by using balloons, coils, or acrylic adhesive to manage CBS in patients with head and neck cancers. We found the difference of the outcomes between these 2 methods was insignificant. The outcomes were significantly influenced by the clinical severity of CBS of our patients.
At present, no preoperative test is sufficiently accurate to justify the occlusion of a patent CA or predict whether it will cause a neurologic deficit. Autogenous venous or arterial reconstruction has been proposed in the treatment of patients with advanced head and neck cancers with invasion to the carotid system.15 This autogenous tissue reconstruction is valuable in avoiding cerebral ischemic insult when complete resection of the tumor and CA is needed. Compared with surgical autogenous tissue reconstruction, the merits of endovascular stent graft reconstruction include:
The clinical severity of CBS was classified into three groups: acute, impending and threatened.1, 7 This simple clinical classification correlated well with angiographic severity in our study. We favor making this clinical classification as a guide for management of CBS. Typically, patients with acute CBS were diagnosed by clinical and angiographic findings; however, patients with threatened or impending CBS can present with subtle or no angiographic abnormalities (grades 0 and 1).
One of the causes of CBS in a CA without obvious angiographic abnormalities was that the angiogram was obtained in the earliest stage of CBS. The vulnerable CA was intact but had been surrounded by diseased soft tissue such as recurrent tumor or irradiation necrosis. Correlating other image modalities such as CT or magnetic resonance imaging of the head and neck region can help in achieving early diagnosis (Fig 1). Another cause is that the subtle pathologic lesions were not shown on routine biplanar angiograms. We suggest obtaining multiple view angiograms or CT angiograms to aid in the evaluation of patients with CBS without obvious angiographic abnormalities (Fig 1).16, 17
Cerebral ischemic insult is a well-known initial complication of deconstructive management of CBS.1, 18 Reconstructive management, however, also carries the risk of thromboembolism and procedurally related complications such as vascular dissection.11, 13 These complications may be caused by inadequate antithrombotic medication treatment of the patients or product characteristics of the devices. The development of a new design of self-expandable stent graft with a less thrombogenic surface, higher flexibility, and lower profile is indicated to improve technical safety of reconstructive endovascular management.
The mean duration of stent graft patency in this study was 3.0 ± 2.6 months. This unfavorable long-term patency is a limitation of the reconstructive method as a permanent way to manage patients with head and neck cancers and CBS. The causes of poor long-term stent graft patency were:
These technical improvements may help stent grafting to be a good alternative method for patients at high risk for carotid occlusion.
Both endovascular methods were good at achieving immediate hemostasis. Although not statistically significant, deconstructive management provided better hemostatic results than reconstructive management (Table IV). Reconstructive management with stent grafts covered only a segment of the affected CA. The pathologic field in the patients with head and neck cancers can show dynamic changes as a result of complex factors such as tumor recurrence, wound infection, or radiation necrosis. This ongoing pathologic process may progress over the initially treated area (Fig 1). We suggest assessing the extent of the pathologic lesion on CT and the field of previous irradiation before stent graft placement is planned. A long, self-expandable stent graft can fully cover the pathologic field, especially given the continuous shortening of the self-expanding stent graft as it dilates progressively.
This ongoing pathologic process with reconstitution of collateral vessels or recanalization of the thrombosed CAs may also explain the rebleeding in patients in the deconstructive group who underwent proximal occlusion.23 We suggest performing deconstructive endovascular therapy with cross occlusion for these cancer patients with CBS to enhance durable hemostasis if they have no risk of permanent carotid occlusion (Fig 2). Cross occlusion is favored as a permanent deconstructive method to treat CBS because of its low rebleeding rate. Reconstructive endovascular therapy with a stent graft is reserved for patients who are not suitable for the deconstructive method.
For all patients, clinical severity is the significant factor affecting the hemostatic outcome of endovascular management (Table V). Rebleeding occurred more commonly in patients with acute CBS than in those in the impending and threatened group. Poor hemostatic results in patients with acute status included patients with acute CBS who also had a more extensive and complete CA injury or soft tissue injury, or both, than those with impending and threatened CBS and the critical clinical status, and thus poor cooperation of these patients could cause technical difficulty such as the precise positioning of the stent graft deployment, as in patient 2 in the reconstructive group; in addition, patients with acute CBS have complications that may be due to inadequate antithrombotic medication or massive transfusion. We suggest making an early diagnostic and treatment plan, such as CT or angiography with an occlusion test, for patients with head and neck cancers if signs suggestive of CBS are present. Early intervention for patients in stable clinical condition may also improve the outcomes.
A limitation of this study was the small number of patients studied, many of whom had a short survival time, which hindered long-term follow-up. Diverse disease stages and previous treatment of our patients also made statistical analysis difficult. Comparing the two techniques is also difficult because patients who received stent grafts are often those who cannot tolerate carotid occlusion or in whom test occlusion is precluded by their clinical status.
A lack of experience with antithrombotic medications and prophylactic antibiotics might also have affected outcomes of patients in the reconstructive group. Further research with improved prophylactic medications and associated other image modalities such as CT or CT angiography for early diagnosis is invaluable to improve the outcomes.
Conclusion
We found no significant differences in the technical and hemostatic outcomes of reconstructive and deconstructive endovascular management. Clinical severity is a significant factor affecting hemostatic outcome of endovascular management: patients with acute CBS were associated with a higher rebleeding rate than those with impending and threatened CBS. We suggest correlation of clinical findings and other image modalities such as CT to help in early diagnosis of CBS if angiographic abnormalities are subtle. We also suggest early endovascular management for patients with CBS. Deconstructive endovascular management with cross occlusion can be a treatment for durable hemostasis.
Author contributions
References
- . Endovascular therapy for the carotid blowout syndrome in head and neck surgical patients: diagnostic and managerial considerations. AJNR Am J Neuroradiol. 1996;17:843–852
- . Locally recurrent, previously irradiated head and neck cancer: concurrent re-irradiation and chemotherapy, or chemotherapy alone?. J Clin Oncol. 2006;24:2653–2658
- . Radical neck dissection: a 19-year experience. J Laryngol Otol. 1989;103:395–400
- . Endovascular treatment of acute carotid blowout syndrome. J Vasc Interv Radiol. 2000;11:1184–1188
- . Management of carotid “blowout” with endovascular stent grafts. Laryngoscope. 2002;112:428–433
- Recurrent carotid blowout syndrome: Diagnosis and therapeutic challenges in a newly recongnized subgroup of patients. AJNR Am J Neuroradiol. 1999;20:1069–1077
- . Management of carotid artery rupture by monitored endovascular therapeutic occlusion (1988-1994). Laryngoscope. 1995;105:1086–1092
- . Preliminary experience with endovascular reconstruction for the management of carotid blowout syndrome. AJNR Am J Neuroradiol. 2003;24:975–981
- Carotid artery resection: preoperative temporary occlusion is not always an accurate predictor of collateral blood flow. Acta Otolaryngol. 2005;125:196–200
- Carotid stenting for impending carotid blowout: suitable supportive care for head and neck cancer patients?. Palliative Medicine. 2005;19:427–429
- Carotid blowout syndrome in patients with head-and-neck cancers: reconstructive management by self-expandable stent-grafts. AJNR Am J Nueroradiol. 2007;28:181–188
- . Delayed complications of endovascular stenting for carotid blowout. Am J Otolaryngol. 2003;24:417–419
- . Delayed stenosis following placement of polyethylene terephthalate endograft in the cervical carotid artery. J Neurosurg. 2003;98:421–425
- CT in nontraumatic acute thoracic aortic disease: Typical and atypical features and complcations. Radiographics. 2003;23:S93–S110Spec No
- . Resection of the internal carotid artery and replacement with greater saphenous vein: a safe procedure for en bloc cancer resections with carotid involvement. J Vasc Surg. 1996;23:775–782
- . CT angiographic analysis of carotid artery stenosis: comparison of manual assessment, semiautomatic vessel analysis, and digital subtraction angiography. AJNR AM J Neuroradiol. 2007;28:97–103
- . Management of penetrating neck injury: a new paradigm for civilian trauma. J Oral Maxillofac Surg. 2007;65:591–705
- . Endovascular treatment of the carotid artery rupture with massive hemorrhage. J Chin Med Assoc. 2003;66:140–147
- . Brain abscess formation: a delayed complication of carotid blowout syndrome treated by self-expandable stent-graft. AJNR Am J Neuroradiol. 2006;27:1543–1545
- Results of the Jostent coronary stent graft implantation in various clinical settings: procedural and follow-up results. Cathet Cardiovasc Intervent. 2002;56:353–360
- Conformity of carotid stents with vascular anatomy: evaluation in carotid models. AJNR Am J Neuroradiol. 2004;25:604–607
- Recanalization results after carotid stent placement. AJNR Am J Neuroradiol. 2002;23:113–120
- . Antegrade recanalization of completely embolized internal carotid artery after treatment of a giant intracavernous aneurysm: a case report. Surg Neurol. 1999;52:611–616
Competition of interest: none.
PII: S0741-5214(07)02038-1
doi:10.1016/j.jvs.2007.12.030
© 2008 The Society for Vascular Surgery. Published by Elsevier Inc. All rights reserved.
