A novel classification and management scheme for craniocervical junction disorders with ventral neural element compression

Abstract

OBJECTIVE Craniocervical junction (CCJ) pathologies with ventral neural element compression are poorly understood, and appropriate management requires accurate understanding, description, and a more uniform nomenclature. The aim of this study was to evaluate patients to identify anatomical clusters and better classify CCJ disorders with ventral compression and guide treatment. METHODS A retrospective review of adult and pediatric patients with ventral CCJ compression from 2008 to 2022 at a single center was performed. The incidence of anatomical abnormalities and compressive etiologies was assessed. Surgical approach, radiographic data, and outcomes were recorded. Association rules analysis (ARA) was used to assess variable clustering. RESULTS Among 51 patients, the main causes of compression were either purely bony (retroflexed dens [n = 18]; basilar invagination [BI; n = 13]) or soft tissue (degenerative pannus [n = 16]; inflammatory pannus [n = 2]). The primary cluster in ARA was a retroflexed dens, platybasia, and Chiari malformation (CM), and the secondary cluster was BI, C1–2 subluxation, and reducibility. These, along with degenerative pannus, formed the three major classes. In assessing the optimal treatment strategy, reducibility was evaluated. Of the BI cases, 12 of the 13 patients had anterolisthesis of C1 that was potentially reducible, compared with 2 of the 18 patients with a retroflexed dens (both with concomitant BI), and no pannus cases. The mean C1–2 facet angle was significantly higher in BI at 32.4°, compared with −2.3° in retroflexed dens and 8.1° in degenerative pannus (p < 0.05). Endonasal decompression with posterior fixation was performed in 48 (94.0%) of the 51 patients, whereas posterior reduction/fixation alone was performed in 3 patients (6.0%). Of 16 reducible cases, open posterior reduction alone was successful in 3 (60.0%) of 5 cases, with all successes containing isolated BI. Reduction was not attempted if vertebral anatomy was unfavorable (n = 9) or the C1 lateral mass was absent (n = 5). The mean follow-up was 28 months. Symptoms improved in 88.9% of patients and were stable in the remaining 11.1%. Tracheostomy and percutaneous G-tube placement occurred in 7.8% and 11.8% of patients, respectively. Reoperation for an endonasal CSF leak repair or posterior cervical wound revision both occurred in 3.9% of patients. CONCLUSIONS In classifying, one cluster caused decreased posterior fossa volume due to an anatomical triad of retroflexed dens, platybasia, and CM. The second cluster caused pannus formation due to degenerative hypertrophy. For both, endonasal decompression with posterior fixation was ideal. The third group contained C1 anterolisthesis characterized by a steep C1–2 facet angle causing reducible BI. Posterior reduction/fixation is the first-line treatment when anatomically feasible or endonasal decompression with in situ posterior fixation when anatomical constraints exist.