A clinical neuroimaging platform for rapid, automated lesion detection and personalized post-stroke outcome prediction

AI Summary

Fully automated neuroimaging platform forecasts individual cognitive outcomes after ischemic stroke using deep-learning lesion segmentation and location and network-based features.
Processes raw DICOM MRI from heterogeneous scanners and generates text-based, personalised prognoses, with an LLM producing interpretable reports in approximately three minutes.
Validated on a 604-lesion training cohort and independent 153-patient cohort; multiple cognitive predictions showed reasonable accuracy and 96% concordance with manual methods.

NPJ Digit Med. 2026 May 27. doi: 10.1038/s41746-026-02803-2. Online ahead of print.

ABSTRACT

Accurately predicting long-term outcomes after stroke remains a key challenge in personalized medicine. Here, we present a neuroimaging platform that forecasts individualized cognitive outcomes in patients with ischemic stroke using deep learning-based lesion segmentation and location-/network-based features. This novel, fully automated system is capable of processing raw DICOM MRI data from heterogeneous scanners and generating text-based, personalized outcome information. To demonstrate this pipeline, we trained cognitive outcome-prediction models using a large lesion cohort (N = 604) and applied them to an independent stroke cohort (N = 153). Multiple cognitive outcome predictions achieved reasonable accuracy, with 96% concordance with manual methods. A report generated by a large language model provides interpretable, patient-specific prognoses within ~3 min. This demonstrates the potential for imaging-informed prognostication to inform stroke care and guide rehabilitation strategies.

PMID:42204350 | DOI:10.1038/s41746-026-02803-2

Document this CPD