1
|
Tanioka S, Aydin OU, Hilbert A, Ishida F, Tsuda K, Araki T, Nakatsuka Y, Yago T, Kishimoto T, Ikezawa M, Suzuki H, Frey D. Prediction of hematoma expansion in spontaneous intracerebral hemorrhage using a multimodal neural network. Sci Rep 2024; 14:16465. [PMID: 39013990 PMCID: PMC11252350 DOI: 10.1038/s41598-024-67365-3] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 04/09/2024] [Accepted: 07/10/2024] [Indexed: 07/18/2024] Open
Abstract
Hematoma expansion occasionally occurs in patients with intracerebral hemorrhage (ICH), associating with poor outcome. Multimodal neural networks incorporating convolutional neural network (CNN) analysis of images and neural network analysis of tabular data are known to show promising results in prediction and classification tasks. We aimed to develop a reliable multimodal neural network model that comprehensively analyzes CT images and clinical variables to predict hematoma expansion. We retrospectively enrolled ICH patients at four hospitals between 2017 and 2021, assigning patients from three hospitals to the training and validation dataset and patients from one hospital to the test dataset. Admission CT images and clinical variables were collected. CT findings were evaluated by experts. Three types of models were developed and trained: (1) a CNN model analyzing CT images, (2) a multimodal CNN model analyzing CT images and clinical variables, and (3) a non-CNN model analyzing CT findings and clinical variables with machine learning. The models were evaluated on the test dataset, focusing first on sensitivity and second on area under the receiver operating curve (AUC). Two hundred seventy-three patients (median age, 71 years [59-79]; 159 men) in the training and validation dataset and 106 patients (median age, 70 years [62-82]; 63 men) in the test dataset were included. Sensitivity and AUC of a CNN model were 1.000 (95% confidence interval [CI] 0.768-1.000) and 0.755 (95% CI 0.704-0.807); those of a multimodal CNN model were 1.000 (95% CI 0.768-1.000) and 0.799 (95% CI 0.749-0.849); and those of a non-CNN model were 0.857 (95% CI 0.572-0.982) and 0.733 (95% CI 0.625-0.840). We developed a multimodal neural network model incorporating CNN analysis of CT images and neural network analysis of clinical variables to predict hematoma expansion in ICH. The model was externally validated and showed the best performance of all the models.
Collapse
Affiliation(s)
- Satoru Tanioka
- Charité Lab for Artificial Intelligence in Medicine, Charité Universitätsmedizin Berlin, Charitéplatz 1, 101117, Berlin, Germany.
- Department of Neurosurgery, Mie University Graduate School of Medicine, 2-174 Edobashi, Tsu, Mie, 5148507, Japan.
| | - Orhun Utku Aydin
- Charité Lab for Artificial Intelligence in Medicine, Charité Universitätsmedizin Berlin, Charitéplatz 1, 101117, Berlin, Germany
| | - Adam Hilbert
- Charité Lab for Artificial Intelligence in Medicine, Charité Universitätsmedizin Berlin, Charitéplatz 1, 101117, Berlin, Germany
| | - Fujimaro Ishida
- Department of Neurosurgery, Mie Chuo Medical Center, 2158-5 Myojin-Cho, Hisai, Tsu, Mie, 5141101, Japan
| | - Kazuhiko Tsuda
- Department of Neurosurgery, Matsusaka Chuo General Hospital, 102 Kobo, Matsusaka, Mie, 5158566, Japan
| | - Tomohiro Araki
- Department of Neurosurgery, Suzuka Kaisei Hospital, 112-1 Ko-Cho, Suzuka, Mie, 5138505, Japan
| | - Yoshinari Nakatsuka
- Department of Neurosurgery, Suzuka Kaisei Hospital, 112-1 Ko-Cho, Suzuka, Mie, 5138505, Japan
| | - Tetsushi Yago
- Department of Neurosurgery, Mie Chuo Medical Center, 2158-5 Myojin-Cho, Hisai, Tsu, Mie, 5141101, Japan
| | - Tomoyuki Kishimoto
- Department of Neurosurgery, Matsusaka Chuo General Hospital, 102 Kobo, Matsusaka, Mie, 5158566, Japan
| | - Munenari Ikezawa
- Department of Neurosurgery, Suzuka Kaisei Hospital, 112-1 Ko-Cho, Suzuka, Mie, 5138505, Japan
| | - Hidenori Suzuki
- Department of Neurosurgery, Mie University Graduate School of Medicine, 2-174 Edobashi, Tsu, Mie, 5148507, Japan
| | - Dietmar Frey
- Charité Lab for Artificial Intelligence in Medicine, Charité Universitätsmedizin Berlin, Charitéplatz 1, 101117, Berlin, Germany
- Department of Neurosurgery, Charité Universitätsmedizin Berlin, Charitéplatz 1, 10117, Berlin, Germany
| |
Collapse
|
2
|
Mak A, Matouk CC, Avery EW, Behland J, Haider SP, Frey D, Madai VI, Vajkoczy P, Griessenauer CJ, Zand R, Hendrix P, Abedi V, Sanelli PC, Falcone GJ, Petersen N, Sansing LH, Sheth KN, Payabvash S, Malhotra A. Automated detection of early signs of irreversible ischemic change on CTA source images in patients with large vessel occlusion. PLoS One 2024; 19:e0304962. [PMID: 38870240 PMCID: PMC11175522 DOI: 10.1371/journal.pone.0304962] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 09/25/2023] [Accepted: 05/21/2024] [Indexed: 06/15/2024] Open
Abstract
PURPOSE To create and validate an automated pipeline for detection of early signs of irreversible ischemic change from admission CTA in patients with large vessel occlusion (LVO) stroke. METHODS We retrospectively included 368 patients for training and 143 for external validation. All patients had anterior circulation LVO stroke, endovascular therapy with successful reperfusion, and follow-up diffusion-weighted imaging (DWI). We devised a pipeline to automatically segment Alberta Stroke Program Early CT Score (ASPECTS) regions and extracted their relative Hounsfield unit (rHU) values. We determined the optimal rHU cut points for prediction of final infarction in each ASPECT region, performed 10-fold cross-validation in the training set, and measured the performance via external validation in patients from another institute. We compared the model with an expert neuroradiologist for prediction of final infarct volume and poor functional outcome. RESULTS We achieved a mean area under the receiver operating characteristic curve (AUC), accuracy, sensitivity, and specificity of 0.69±0.13, 0.69±0.09, 0.61±0.23, and 0.72±0.11 across all regions and folds in cross-validation. In the external validation cohort, we achieved a median [interquartile] AUC, accuracy, sensitivity, and specificity of 0.71 [0.68-0.72], 0.70 [0.68-0.73], 0.55 [0.50-0.63], and 0.74 [0.73-0.77], respectively. The rHU-based ASPECTS showed significant correlation with DWI-based ASPECTS (rS = 0.39, p<0.001) and final infarct volume (rS = -0.36, p<0.001). The AUC for predicting poor functional outcome was 0.66 (95%CI: 0.57-0.75). The predictive capabilities of rHU-based ASPECTS were not significantly different from the neuroradiologist's visual ASPECTS for either final infarct volume or functional outcome. CONCLUSIONS Our study demonstrates the feasibility of an automated pipeline and predictive model based on relative HU attenuation of ASPECTS regions on baseline CTA and its non-inferior performance in predicting final infarction on post-stroke DWI compared to an expert human reader.
Collapse
Affiliation(s)
- Adrian Mak
- Department of Radiology and Biomedical Imaging, Section of Neuroradiology, Yale School of Medicine, New Haven, CT, United States of America
- CLAIM—Charité Lab for Artificial Intelligence in Medicine, Charité Universitätsmedizin Berlin, Berlin, Germany
| | - Charles C. Matouk
- Department of Neurosurgery, Division of Neurovascular Surgery, Yale University School of Medicine, New Haven, CT, United States of America
| | - Emily W. Avery
- Department of Radiology and Biomedical Imaging, Section of Neuroradiology, Yale School of Medicine, New Haven, CT, United States of America
| | - Jonas Behland
- Department of Radiology and Biomedical Imaging, Section of Neuroradiology, Yale School of Medicine, New Haven, CT, United States of America
- CLAIM—Charité Lab for Artificial Intelligence in Medicine, Charité Universitätsmedizin Berlin, Berlin, Germany
| | - Stefan P. Haider
- Department of Radiology and Biomedical Imaging, Section of Neuroradiology, Yale School of Medicine, New Haven, CT, United States of America
- Department of Otorhinolaryngology, LMU Clinic of Ludwig-Maximilians-University of Munich, Munich, Germany
| | - Dietmar Frey
- CLAIM—Charité Lab for Artificial Intelligence in Medicine, Charité Universitätsmedizin Berlin, Berlin, Germany
| | - Vince I. Madai
- QUEST Center for Responsible Research, Berlin Institute of Health (BIH), Charité Universitätsmedizin Berlin, Berlin, Germany
- School of Computing and Digital Technology, Faculty of Computing, Engineering and the Built Environment, Birmingham City University, Birmingham, United Kingdom
| | - Peter Vajkoczy
- Department of Neurosurgery, Charité Universitätsmedizin Berlin, Berlin, Germany
| | - Christoph J. Griessenauer
- Research Institute of Neurointervention, Paracelsus Medical University, Salzburg, Austria
- Department of Neurosurgery, Paracelsus Medical University, Salzburg, Austria
| | - Ramin Zand
- Department of Neurology, Geisinger Medical Center, Danville, PA, United States of America
- Department of Neurology, Pennsylvania State University, State College, PA, United States of America
| | - Philipp Hendrix
- Department of Neurosurgery, Geisinger Medical Center, Danville, PA, United States of America
- Department of Neurosurgery, Saarland University Medical Center, Homburg, Germany
| | - Vida Abedi
- Department of Public Health Sciences, The Pennsylvania State University, Hershey, PA, United States of America
- Department of Molecular and Functional Genomics, Geisinger Medical Center, Danville, PA, United States of America
| | - Pina C. Sanelli
- Department of Radiology, Northwell Health Feinstein Institutes for Medical Research, Manhasset, New York, United States of America
| | - Guido J. Falcone
- Division of Neurocritical Care and Emergency Neurology, Department of Neurology, Yale University School of Medicine, New Haven, CT, United States of America
| | - Nils Petersen
- Division of Neurocritical Care and Emergency Neurology, Department of Neurology, Yale University School of Medicine, New Haven, CT, United States of America
| | - Lauren H. Sansing
- Division of Stroke and Vascular Neurology, Department of Neurology, Yale University School of Medicine, New Haven, CT, United States of America
| | - Kevin N. Sheth
- Division of Stroke and Vascular Neurology, Department of Neurology, Yale University School of Medicine, New Haven, CT, United States of America
| | - Seyedmehdi Payabvash
- Department of Radiology and Biomedical Imaging, Section of Neuroradiology, Yale School of Medicine, New Haven, CT, United States of America
| | - Ajay Malhotra
- Department of Radiology and Biomedical Imaging, Section of Neuroradiology, Yale School of Medicine, New Haven, CT, United States of America
| |
Collapse
|
3
|
Oura D, Gekka M, Sugimori H. The montage method improves the classification of suspected acute ischemic stroke using the convolution neural network and brain MRI. Radiol Phys Technol 2024; 17:297-305. [PMID: 37934345 DOI: 10.1007/s12194-023-00754-x] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 05/29/2023] [Revised: 10/15/2023] [Accepted: 10/17/2023] [Indexed: 11/08/2023]
Abstract
This study investigated the usefulness of the montage method that combines four different magnetic resonance images into one images for automatic acute ischemic stroke (AIS) diagnosis with deep learning method. The montage image was consisted from diffusion weighted image (DWI), fluid attenuated inversion recovery (FLAIR), arterial spin labeling (ASL), and apparent diffusion coefficient (ASL). The montage method was compared with pseudo color map (pCM) which was consisted from FLAIR, ASL and ADC. 473 AIS patients were classified into four categories: mechanical thrombectomy, conservative therapy, hemorrhage, and other diseases. The results showed that the montage image significantly outperformed pCM in terms of accuracy (montage image = 0.76 ± 0.01, pCM = 0.54 ± 0.05) and the area under the curve (AUC) (montage image = 0.94 ± 0.01, pCM = 0.76 ± 0.01). This study demonstrates the usefulness of the montage method and its potential for overcoming the limitations of pCM.
Collapse
Affiliation(s)
- Daisuke Oura
- Department of Radiology, Otaru General Hospital, Otaru, 047-0152, Japan
- Graduate School of Health Sciences, Hokkaido University, Sapporo, 060-0812, Japan
| | - Masayuki Gekka
- Department of Neurosurgery, Otaru General Hospital, Otaru, 047-0152, Japan
| | - Hiroyuki Sugimori
- Faculty of Health Sciences, Hokkaido University, Sapporo, 060-0812, Japan.
| |
Collapse
|
4
|
Khosravi M, Zare Z, Mojtabaeian SM, Izadi R. Artificial Intelligence and Decision-Making in Healthcare: A Thematic Analysis of a Systematic Review of Reviews. Health Serv Res Manag Epidemiol 2024; 11:23333928241234863. [PMID: 38449840 PMCID: PMC10916499 DOI: 10.1177/23333928241234863] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 11/05/2023] [Revised: 02/04/2024] [Accepted: 02/07/2024] [Indexed: 03/08/2024] Open
Abstract
Introduction The use of artificial intelligence (AI), which can emulate human intelligence and enhance clinical results, has grown in healthcare decision-making due to the digitalization effects and the COVID-19 pandemic. The purpose of this study was to determine the scope of applications of AI tools in the decision-making process in healthcare service delivery networks. Materials and methods This study used a qualitative method to conduct a systematic review of the existing reviews. Review articles published between 2000 and 2024 in English-language were searched in PubMed, Scopus, ProQuest, and Cochrane databases. The CASP (Critical Appraisal Skills Programme) Checklist for Systematic Reviews was used to evaluate the quality of the articles. Based on the eligibility criteria, the final articles were selected and the data extraction was done independently by 2 authors. Finally, the thematic analysis approach was used to analyze the data extracted from the selected articles. Results Of the 14 219 identified records, 18 review articles were eligible and included in the analysis, which covered the findings of 669 other articles. The quality assessment score of all reviewed articles was high. And, the thematic analysis of the data identified 3 main themes including clinical decision-making, organizational decision-making, and shared decision-making; which originated from 8 subthemes. Conclusions This study revealed that AI tools have been applied in various aspects of healthcare decision-making. The use of AI can improve the quality, efficiency, and effectiveness of healthcare services by providing accurate, timely, and personalized information to support decision-making. Further research is needed to explore the best practices and standards for implementing AI in healthcare decision-making.
Collapse
Affiliation(s)
- Mohsen Khosravi
- Department of Health Care Management, School of Management and Information Sciences, Shiraz University of Medical Sciences, Shiraz, Iran
| | - Zahra Zare
- Department of Health Care Management, School of Management and Information Sciences, Shiraz University of Medical Sciences, Shiraz, Iran
| | - Seyyed Morteza Mojtabaeian
- Department of Healthcare Economics, School of Management and Medical Informatics, Shiraz University of Medical Sciences, Shiraz, Iran
| | - Reyhane Izadi
- Department of Health Care Management, School of Management and Information Sciences, Shiraz University of Medical Sciences, Shiraz, Iran
| |
Collapse
|
5
|
Bravata DM, Ranta A. Artificial Intelligence in Clinical Decisions Support for Stroke: Balancing Opportunity With Caution. Stroke 2023; 54:1517-1518. [PMID: 37216447 DOI: 10.1161/strokeaha.123.043004] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 05/24/2023]
Affiliation(s)
- Dawn M Bravata
- Department of Veterans Affairs (VA), Health Services Research and Development (HSR&D), Precision Monitoring to Transform Care, Quality Enhancement Research Initiative (QUERI) (D.M.B.), Indianapolis, IN
- VA HSR&D Center for Health Information and Communication, Richard L. Roudebush VA Medical Center, Indianapolis, IN (D.M.B.)
- Departments of Internal Medicine and Neurology, Indiana University School of Medicine, Indianapolis (D.M.B.)
- Regenstrief Institute, Indianapolis (D.M.B.)
| | - Anna Ranta
- Department of Medicine, University of Otago - Wellington (A.R.)
- Department of Neurology, Wellington Hospital, New Zealand and Te What Ora - Health NZ, National Hyper-Acute Stroke Programme, Wellington, New Zealand (A.R.)
| |
Collapse
|