Diagnostic accuracy of high-resolution T2-weighted MRI vs contrast-enhanced T1-weighted MRI to screen for cerebellopontine angle lesions in symptomatic patients.

OBJECTIVE: To evaluate diagnostic accuracy of high-resolution T2-weighted MRI (T2w) for detecting cerebellopontine angle (CPA) lesions compared to a combined protocol including gadolinium enhanced T1-weighted MRI (GdT1w). SETTING: Department of Radiology & Nuclear Medicine, Radboud University Medical Centre, Nijmegen, The Netherlands. PARTICIPANTS: A random sample of MRIs from 350 patients (700 CPAs) with asymmetrical audiovestibular complaints was used, acquired between 2013 and 2016. MAIN OUTCOME MEASURES: Sensitivity, specificity, positive and negative predictive values of T2w results compared to GdT1w and, in patients with any suggestion of CPA pathology, to the complete examination (T1w, GdT1w and T2w). Inter-rater agreement between an experienced neuroradiologist and a less experienced observer was calculated. RESULTS: Results of 678 CPAs in 340 patients were analysed. On T2w, the neuroradiologist identified all 27 lesions >2 mm in size out of a total of 30 CPA lesions (sensitivity: 90% [95% CI: 73.5%-97.9%]). Negative predictive value reached 99.5% (95% CI: 98.7-99.9). One missed lesion of 2 mm would have been detected in clinical practice, as this was one of 14 patients for which additional GdT1w would have been ordered based on T2w alone, increasing sensitivity to 93% (95% CI: 77.9%-99.2%) and negative predictive value to 99.7% (95% CI: 98.9%-100%). Inter-rater agreement for T2w was 98% (95% CI: 96.4-98.8). CONCLUSION: T2w has a very high diagnostic accuracy for the presence of CPA lesions in patients with asymmetrical audiovestibular complaints. However, in a screening protocol with T2w only, smallest vestibular schwannomas as well as rare differential diagnoses that probably only would be detected on GdT1w may remain unnoticed.

Potential savings in the diagnosis of vestibular schwannoma.

INTRODUCTION: Magnetic resonance imaging (MRI) is used to screen patients at risk for vestibular schwannoma (VS). These MRIs are costly and have an extremely low yield; only 3% of patients in the screening population has an actual VS. It might be worthwhile to develop a test to predict VS and refer only a subset of all patients for MRI. OBJECTIVE: To examine the potential savings of such a hypothetical diagnostic test before MRI. DESIGN: We built a decision analytical model of the diagnostic strategy of VS. Input was derived from literature and key opinion leaders. The current strategy was compared to hypothetical new strategies, assigning MRI to the following: (i) all patients with pathology, (ii) all patients with important pathology and (iii) only patients with VS. This resulted in potential cost savings for each strategy. We conducted a budget impact analysis to predict nationwide savings for the Netherlands and the United Kingdom (UK), and a probabilistic sensitivity analysis to address uncertainty. RESULTS: Mean savings ranged from €256 (95%CI €250 - €262) or approximately US$284 (95%CI US$277 - US$291) per patient for strategy 1 to €293 (95%CI €290 - €296) or approximately US$325 (95%CI US$322 - US$328) per patient for strategy 3. Future diagnostic strategies can cost up to these amounts per patient and still be cost saving. Annually, for the Netherlands, €2.8 to €3.2 million could be saved and €10.8 to €12.3 million for the UK. CONCLUSIONS: The model shows that substantial savings could be generated if it is possible to further optimise the diagnosis of VS.