Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2025 Aug 14;25(1):329.
doi: 10.1186/s12880-025-01864-1.

AI post-intervention operational and functional outcomes prediction in ischemic stroke patients using MRIs

Emily Wittrup  1 , John Reavey-Cantwell  2 , Aditya S Pandey  3 , Dennis J Rivet Ii  2 , Kayvan Najarian  4   5   6   7
Affiliations

AI post-intervention operational and functional outcomes prediction in ischemic stroke patients using MRIs

Emily Wittrup et al. BMC Med Imaging. .

Abstract

Background: Despite the potential clinical utility for acute ischemic stroke patients, predicting short-term operational outcomes like length of stay (LOS) and long-term functional outcomes such as the 90-day Modified Rankin Scale (mRS) remain a challenge, with limited current clinical guidance on expected patient trajectories. Machine learning approaches have increasingly aimed to bridge this gap, often utilizing admission-based clinical features; yet, the integration of imaging biomarkers remains underexplored, especially regarding whole 2.5D image fusion using advanced deep learning techniques.

Methods: This study introduces a novel method leveraging autoencoders to integrate 2.5D diffusion weighted imaging (DWI) with clinical features for refined outcome prediction.

Results: Results on a comprehensive dataset of AIS patients demonstrate that our autoencoder-based method has comparable performance to traditional convolutional neural networks image fusion methods and clinical data alone (LOS > 8 days: AUC 0.817, AUPRC 0.573, F1-Score 0.552; 90-day mRS > 2: AUC 0.754, AUPRC 0.685, F1-Score 0.626).

Conclusion: This novel integration of imaging and clinical data for post-intervention stroke prognosis has numerous computational and operational advantages over traditional image fusion methods. While further validation of the presented models is necessary before adoption, this approach aims to enhance personalized patient management and operational decision-making in healthcare settings.

Clinical trial number: Not applicable.

Keywords: Acute ischemic stroke; Artificial intelligence; Autoencoder; Deep learning; Length of stay; MRI; Modified rankin scale; Resnet.

PubMed Disclaimer

Conflict of interest statement

Declarations. Ethics approval and consent to participate: This study was approved by the Institutional Review Board at University of Michigan (protocol number: HUM00234328), ensuring ethical conduct in accordance with all relevant guidelines for research involving human subjects including the Helsinki Declaration. Written informed consent from patients in this published dataset was waived by the Johns Hopkins Internal Review Board (IRB00228775) since this study involved no more than minimal risk to the involved subjects. Consent for publication: Not applicable. Competing interests: The authors are applying for a patent on the methodologies and applications discussed in this manuscript.

Figures

Fig. 1
Fig. 1
Distribution of outcome variables for each cohort presented as density histograms. Vertical gray dotted lines show the selected dichotomization thresholds. Three patients had a length of stay (LOS) formula image days. SD: standard deviation; IQR: interquartile range; mRS: modified Rankin Scale
Fig. 2
Fig. 2
Architecture diagram of A. Resnet10, B. Autoencoder encoder, and C. Classifier. MRI: magnetic resonance imaging; EHR: electronic health records; IDF: image-derived features; ReLU: Rectified; avg: average
Fig. 3
Fig. 3
Comparison of AUC and AUPRC test performance averaged across folds for each outcomes across all methods tested. AUC: area under the receiver operating characteristic curve; AUPRC: area under the precision recall curve; EHR: electronic health records; IDF: image-derived features; AE: autoencoder; DAFT: dynamic affine feature map transform
See this image and copyright information in PMC

References

    1. Campbell BC, De Silva DA, Macleod MR, Coutts SB, Schwamm LH, Davis SM, Donnan GA. Ischaemic stroke. Nat Rev Dis Primers. 2019;5(1):70. - PubMed
    1. Mendelson SJ, Prabhakaran S. Diagnosis and management of transient ischemic attack and acute ischemic stroke: A review. JAMA. 2021;325(11):1088–98. - PubMed
    1. Tadi P, Lui F. Acute stroke. In: StatPearls [Internet]. StatPearls Publishing, Treasure Island (FL); 2023.
    1. Saver JL. Penumbral salvage and thrombolysis outcome: A drop of brain, a week of life. Brain. 2017;140(3):519–22. - PubMed
    1. Bacchi S, Oakden-Rayner L, Menon DK, Moey A, Jannes J, Kleinig T, Koblar S. Prospective and external validation of stroke discharge planning machine learning models. J Clin Neurosci. 2022;96:80–84. - PubMed

LinkOut - more resources