Automated Extraction of Intracranial Aneurysm Morphometrics: A Novel System for Large-scale Analysis

Introduction: While multiple studies have established the power of morphometric parameters extracted from catheter angiography in predicting aneurysm rupture risk (Etminan et al., 2015) and treatment outcomes (Abbas et al., 2024), comprehensive morphometric analysis remains critical for understanding rupture risk and guiding treatment planning. The Siemens Syngo platform provides advanced morphometric parameters from catheter angiography, but manual extraction of measurements is time-consuming and error-prone, limiting large-scale studies. We present an automated system for extracting aneurysm morphometrics and report our preliminary implementation results.

Methods: We developed a cloud-based optical character recognition (OCR) system that extracts morphometric parameters from Syngo software, including standard measurements (volume, surface area) and advanced analytics (ostium geometry, surface areas). The system employs multiple OCR providers with a quorum-based validation approach. We applied this system retrospectively to analyze aneurysm morphometrics in patients treated at North Shore University Hospital (Manhasset, NY) between 2010-2024. Manual verification of OCR accuracy was performed on a sample of cases.

Results: We have completed clinical data collection for a cohort of 295 subjects with intracranial aneurysms treated with stent-assisted coil embolization. Validation testing of our automated system demonstrates accurate extraction of 19 morphometric parameters per case. The quorum-based approach eliminates transcription errors while reducing processing time from 10-15 minutes to seconds. The system’s robust performance in pilot testing positions us to efficiently collect over 5,600 individual morphometric measurements from our cohort, enabling detailed structural analytics at a scale previously impractical with manual methods.

Conclusion : We demonstrate the successful development of an automated system for extracting aneurysm morphometrics from the Syngo platform. This innovation enables rapid, accurate collection of morphometric data, facilitating large-scale studies of aneurysm characteristics and their relationship to clinical outcomes. Preliminary results show robust performance in data extraction, suggesting this system could significantly advance our ability to conduct advanced, large-scale morphometric analyses of intracranial aneurysms.