A new image-based stroke registry containing quantitative magnetic resonance imaging data.

Area of Science:

• Neurology
• Medical Imaging
• Data Science

Background:

• Conventional stroke registries lack comprehensive imaging data capture.
• There is a need for advanced systems to integrate both text and image-based clinical information for stroke patients.

Purpose of the Study:

• To develop and assess a next-generation stroke registry, DUIH_SRegI, incorporating quantitative magnetic resonance imaging (MRI) data.
• To evaluate the feasibility and clinical utility of an image-based stroke registry for research and patient care.

Main Methods:

• Designed DUIH_SRegI with the Image_QNA software package for mapping and storing MRI lesion data.
• Assessed inter- and intra-user variability for lesion mapping, comparing Image_QNA with SPM5 for lesion registration accuracy.
• Studied 47 patients with lacunar infarcts and analyzed data from 183 patients in a prospective registry.

Main Results:

• Image_QNA demonstrated low variability and superior lesion registration compared to SPM5.
• Enriched datasets revealed posterolateral brain location overlap of diffusion MR lesions in patients with higher NIH Stroke Scale scores.
• Prospective registry data collection was feasible, with an acceptable labor time of approximately 1 hour per patient.

Conclusions:

• A novel, image-based stroke registry (DUIH_SRegI) was successfully developed.
• The system facilitates intuitive, image-based hypothesis formulation and testing, advancing stroke research capabilities beyond traditional registries.