MRI evaluation of neurofibromatosis 2 patients: A standardized approach for accuracy in interpretation

Objective: To determine the level of agreement between local radiologists' and an experienced neuroradiologist's measurements of vestibular schwannomas. Study Design: Prospective study with uniform magnetic resonance acquisition protocol parameters and reporting instructions across 30 magnetic resonance imaging facilities worldwide. Setting: Multicenter natural history study of neurofibromatosis Type 2. Subjects: One hundred fifteen magnetic resonance imaging examinations of 57 neurofibromatosis Type 2 patients older than 5 years of age. Interventions: Thin-slice, postcontrast cranial magnetic resonance imaging. Main Outcome Measures: Spearman's rho interobserver association coefficient of vestibular schwannoma linear measurements. Results: The local and experienced radiologist measurements and identification of tumors agreement was fair (K = 0.77). Discordant interpretations were adjudicated by another experienced neuroradiologist. Conclusion: The least interobserver variability was found in measurements of thin-slice postcontrast magnetic resonance imaging scans obtained at neurofibromatosis Type 2 centers in patients without previous operations and moderately sized tumors. If the schwannoma was difficult to assess, because of magnetic resonance imaging acquisition protocol, postoperative changes, or tumors smaller than 5 mm in greatest diameter, the neuroradiologist provided a more thorough assessment. The authors suggest uniform reporting criteria for vestibular schwannoma assessments to ensure clinically relevant information is communicated regarding vestibular schwannoma size.