Conference proceeding
Rapid, automated prediction of post-stroke cognitive impairment for ischemic stroke
Vol.13407, pp.134071Y-134071Y-8
Progress in Biomedical Optics and Imaging
04/04/2025
DOI: 10.1117/12.3047488
Abstract
Post-stroke cognitive impairment represents a significant clinical challenge, affecting up to 70% of survivors and standing as one of the most critical determinants of patient outcomes. While research has demonstrated that lesion location serves as a powerful predictor of cognitive outcomes, translating these insights into clinical practice faces a fundamental barrier: the current process requires neuroimaging experts to spend several hours manually analyzing each patient’s lesions, making it impractical for routine clinical use. Our automated pipeline processes standard clinical MRI sequences, including diffusion-weighted imaging (DWI), apparent diffusion coefficient (ADC) maps, and anatomical scans, to identify patients at high risk for cognitive impairment based on lesion location. The system was validated using acute clinical data from 114 ischemic stroke patients who underwent comprehensive neuropsychological testing in the chronic phase of stroke. The analysis demonstrated strong predictive performance, where automatically detected lesions in critical left hemisphere white matter regions effectively identified patients who developed significant cognitive impairment (p < 0.0001), with results remaining robust after controlling for confounding variables such as age and lesion volume. The entire analysis pipeline, from raw DICOM data to final prediction, consistently completes in under 10 minutes per patient, representing a dramatic improvement over traditional manual methods that require hours of expert analysis. This efficiency enables rapid processing of large-scale clinical datasets while maintaining accuracy comparable to manual analysis. Our system demonstrates the feasibility of integrating personalized, lesion location-based prognostic information into clinical workflows, potentially transforming acute care decision-making and rehabilitation planning across diverse healthcare settings.
Details
- Title: Subtitle
- Rapid, automated prediction of post-stroke cognitive impairment for ischemic stroke
- Creators
- Michal Brzus - University of IowaJamie Kaminski - University of IowaTrevor Viohl - The Univ. of Iowa (United States)Joel Bruss - University of IowaDaniel Tranel - University of IowaAaron D. Boes - University of IowaHans J. Johnson - University of Iowa
- Contributors
- Susan M. Astley (Editor) - University of ManchesterAxel Wismüller (Editor) - University of Rochester
- Resource Type
- Conference proceeding
- Publication Details
- Vol.13407, pp.134071Y-134071Y-8
- Publisher
- SPIE
- Series
- Progress in Biomedical Optics and Imaging
- DOI
- 10.1117/12.3047488
- ISSN
- 1605-7422
- Grant note
- NIH NINDS: R01NS114405, R01NS119896 Roy J. Carver Trust: T32GM108540
This work was supported by the NIH NINDS R01NS114405 and R01NS119896, The Roy J. Carver Trust, and BBIP T32 (T32GM108540).
- Language
- English
- Date published
- 04/04/2025
- Academic Unit
- Roy J. Carver Department of Biomedical Engineering; Neurology; Electrical and Computer Engineering; Psychiatry; Stead Family Department of Pediatrics; Psychological and Brain Sciences; Iowa Neuroscience Institute; The Iowa Institute for Biomedical Imaging; Neurology (Pediatrics); The Iowa Initiative for Artificial Intelligence; Iowa Informatics Initiative
- Record Identifier
- 9984813317702771
Metrics
8 Record Views