Critical Appraisal of Extent of Resection of Clival Lesions Using the Expanded Endoscopic Endonasal Approach

Background: The expanded endoscopic endonasal approach has become a well-accepted technique for accessing clival lesions. Cadaveric anatomical studies have demonstrated the unobstructed view of the clivus and ventral brainstem that may be achieved with this approach. Critical neurovascular structures, primarily the paraclival internal carotid artery (ICA) segments, are generally considered to represent the lateral limits of safe exposure. To determine what true anatomic limitations exist using an endoscopic endonasal approach in patients with actual clival lesions, a detailed analysis of the postoperative imaging following resection of these lesions was performed.

Objective: The purposes of this study are to (1) present a critical evaluation of our experience using an expanded endoscopic endonasal approach to clival lesions and (2) evaluate, based on the location of residual tumor, the anatomic limitations to the approach.

Methods: A retrospective review of all patients undergoing an endoscopic endonasal operation at UCLA Medical Center from 2008 through 2011 was performed. Nineteen patients with lesions involving the clivus were identified. Extent of resection was determined by pre- and postoperative tumor volumes. Results were divided into gross total resection (GTR), >95% resection, and <95% resection.

Results: Lesions included nine chordomas, five invasive pituitary adenomas, one adenocarcinoma, one meningioma, one adenoid cystic carcinoma, one fibrous dysplasia, and one leiomyosarcoma metastases. Mean patient age was 54.6 years. Mean initial tumor volume was 26.2 cm³. Three patients underwent planned subtotal resections. Of the remaining patients GTR was achieved in 8/16 (50%), >95% in 5/16 (31%), and <95% in 3/16 (19%). Average initial tumor volume for the group with <95% resection was 39 cm³. Complications included three postoperative CSF leaks, two new VIth nerve palsies, one temporary worsening of a VIth nerve palsy, and one new V2 partial injury. Residual tumor occurred most commonly with tumor extension posterior and lateral to the clival and intracavernous ICA segments. Other limitations included caudal tumor invasion at the level of the occipital condyle and inferior extension to the bottom one third of the dens.

Conclusion: The endoscopic endonasal approach represents a safe and effective technique for the resection of clival lesions. Despite excellent overall visualization of this region, we did find that adequate exposure of the most lateral and inferior portion of large tumors is often difficult to obtain, even with the angled endoscope, and may represent the anatomic limitations to this approach. Knowledge of these limitations allows us to determine preoperatively which tumors are best suited for an endoscopic endonasal route and which may be more appropriate for an open skull base or combined technique.