Preoperative catheter spinal angiography and embolization of cervical spinal tumors: Outcomes from a single center

Objective The existing literature regarding preoperative cervical spinal tumor embolization is sparse, with few discussions on the indications, risks, and best techniques. We present our experience with the preoperative endovascular management of hypervascular cervical spinal tumors. Methods We performed a retrospective review of all patients who underwent preoperative spinal angiography (regardless of whether tumor embolization was performed) at our institution (from 2002 to 2012) for primary and metastatic cervical spinal tumors. Tumor vascularity was graded from 0 (tumor blush equal to the normal adjacent vertebral body) to 3 (intense tumor blush with arteriovenous shunting). Tumors were considered “hypervascular” if they had a tumor vascular grade from 1 to 3. Embolic materials included particles, liquid embolics, and detachable coils. The main embolization technique was superselective catheterization of an arterial tumor feeder followed by injection of embolic material. This technique could be used alone or supplemented with occlusion of dangerous anastomoses of the vertebral artery as needed to prevent inadvertent embolization of the vertebrobasilar system. In cases when superselective catheterization of the tumoral feeder was not feasible, embolization was performed from a proximal catheter position after occlusion of branches supplying areas other than the tumor (“flow diversion”). Results A total of 47 patients with 49 cervical spinal tumors were included in this study. Of the 49 total tumors, 41 demonstrated increased vascularity (vascularity score > 0). The most common tumor pathology in our series was renal cell carcinoma (RCC) (N = 16; 32.7% of all tumors) followed by thyroid carcinoma (N = 7; 14.3% of all tumors). Tumor embolization was undertaken in 25 hypervascular tumors resulting in complete, near-complete, and partial embolization in 36.0% (N = 9), 44.0% (N = 11), and 20.0% (N = 5) of embolized tumors, respectively. We embolized 42 tumor feeders in 25 tumors. The most commonly embolized tumor feeders were branches of the vertebral artery (19.0%; N = 8), the deep cervical artery (19.0%; N = 8), and the ascending cervical artery (19.0%; N = 8). Sixteen hypervascular tumors were not embolized because of minimal hypervascularity (8/16), unacceptably high risk of spinal cord or vertebrobasilar ischemia (4/16), failed superselective catheterization of tumor feeder (3/16), and cancellation of surgery (1/16). Vertebral artery occlusion was perfo