A WEb-Accessible comprehensiVE platform for automatic vestibular schwannoma segmentation and longitudinal volumetric...

Vestibular schwannomas (VS) require long-term tracking for treatment decisions and outcome assessment. This study aims to develop a WEb-Accessible comprehensiVE (WEAVE) platform that combines AI-driven segmentation with a user-friendly interface to enable longitudinal volume tracking for disease assessment, planning, and monitoring.

WEAVE was built using nnU-Net as the backbone for auto-segmentation, with 3 models trained and validated on combinations of various image modalities. Auto-segmentation performance was evaluated with multiple metrics including absolute and relative volume differences (AVD/RVD), Dice score, mean surface-to-surface distance, and 95th percentile Hausdorff distance (HD95). The platform features a central database with DICOM-RT import/export capabilities, and its interface is built using Rust and WebAssembly.

Three models demonstrated comparable performance without significant differences with mean Dice scores, AVD, and RVD ranged from (0.89 to 0.90), (0.11 to 0.13 cc), and (10.70% to 13.44%), respectively. Mean surface-to-surface distance and HD95 values were consistently low (0.14-0.19 and 0.74-0.88 mm, respectively). Average inference time was ∼60 s per case. The platform successfully enabled longitudinal tumor volume tracking and provided flexible visualization options, including single and multiple image views and a graphical representation of volume changes over time.

WEAVE is a comprehensive platform that combines automated segmentation with longitudinal tracking to support VS management. The AI models achieved Dice scores comparable to interobserver variability in manual contouring, indicating clinical adequacy. The tracking capability provides consistency in treatment planning and monitoring and opens the possibility to advance AI-driven segmentation and streamline workflows for other intracranial pathologies.