.jpg)
Objective: With advancements in imaging technologies, the use of neuroimaging biomarkers in characterizing vestibular schwannomas (VS), aiding treatment selection, outcomes prognostication, and post-treatment care has greatly expanded. However, an evidence-based systematic review on this topic is lacking. As such, the appropriate interpretation and utilization of strategic biomarkers remains nebulous. This review aims to synthesize the current body of knowledge on this topic, with a focus on best-supportive evidence for modalities that may alter clinical management of VS. Secondary aim is to review promising novel modalities under investigation.
Method: A systematic review of publications from 1997 to 2023 was conducted using index databases. Key phrase search parameters included imaging modalities (e.g. diffusion MRI), VS, radiomics, and morphology. A flow chart with search characteristics in various neuroradiology domains will be provided upon presentation. Included were in vivo MR studies of adults diagnosed with VS with sample sizes ≥ 6. To be considered, abstracts had to exhibit clearly defined methodology, MR modality employed, and a statistical analysis linking imaging findings and clinical presentation/outcome. Exclusions included studies of NF-2 and smaller case reports/ series (N ≤ 5). Out of 10184 initial results, 228 studies were considered in multiple domains. Roughly 75 manuscripts were ultimately included in this review.
Results: The most studied imaging biomarker is tumor size, which has been reported heterogeneously and with questionable statistical associations to outcomes measures. Linear measurement of size along one or more axes was the most often reported metric. Grading systems taking into account the relative tumor size and interactions with surrounding anatomy (Koos, Hannover classifications) were also widely reported, but were prognostically questionable. Volumetric segmentation of tumors has been more prevalent in studies published in the past 10 years and may have superior statistical value. Other biomarkers demonstrating clinical-prognostic value included pre-operative facial nerve tractography, cystic sac characteristics, tumor necrosis, and tissue calcification. Categorical variables such as fundal fluid cap were reported, but their quantification is not standardized and prognostic value remains in question. Finally, radiomics is a burgeoning discipline to study tumor features based on clinically available scans, yet its reliability and validity remain indeterminate. At the time of this review, no imaging biomarker has emerged as a reliable radiographic correlate to tumor adhesion, vascularity, aggressive phenotypic presentations/growth patterns, and/or microenvironment; although a few show promise. This presentation will review the significant trends and associations for various biomarkers that do have some evidence-based support. Machine-learning (semi-automation/artificial intelligence) is a growing area of investigation that may carry promise for enhanced accuracy, efficiency, and clinical correlation.
Conclusions: This review synthesizes a broad body of literature, with intent to underscore the transformative potential of imaging biomarkers in VS management. While some biomarkers show promise, further development and statistical validation is required. Ongoing research in this field is focused on developing imaging acquisition protocols, improving new image processing algorithms, and correlating a growing dataset with phenotypic and genotypic variables.