Optimizing Automated Hematoma Expansion Classification from Baseline and Follow-Up Head Computed Tomography
Hematoma expansion (HE) is an independent predictor of poor outcomes and a modifiable treatment target in intracerebral hemorrhage (ICH). Evaluating HE in large datasets requires segmentation of hematomas on admission and follow-up CT scans, a process that is time-consuming and labor-intensive in la...
Saved in:
| Main Authors: | , , , , , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2024-12-01
|
| Series: | Applied Sciences |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2076-3417/15/1/111 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1841549429889302528 |
|---|---|
| author | Anh T. Tran Dmitriy Desser Tal Zeevi Gaby Abou Karam Julia Zietz Andrea Dell’Orco Min-Chiun Chen Ajay Malhotra Adnan I. Qureshi Santosh B. Murthy Shahram Majidi Guido J. Falcone Kevin N. Sheth Jawed Nawabi Seyedmehdi Payabvash |
| author_facet | Anh T. Tran Dmitriy Desser Tal Zeevi Gaby Abou Karam Julia Zietz Andrea Dell’Orco Min-Chiun Chen Ajay Malhotra Adnan I. Qureshi Santosh B. Murthy Shahram Majidi Guido J. Falcone Kevin N. Sheth Jawed Nawabi Seyedmehdi Payabvash |
| author_sort | Anh T. Tran |
| collection | DOAJ |
| description | Hematoma expansion (HE) is an independent predictor of poor outcomes and a modifiable treatment target in intracerebral hemorrhage (ICH). Evaluating HE in large datasets requires segmentation of hematomas on admission and follow-up CT scans, a process that is time-consuming and labor-intensive in large-scale studies. Automated segmentation of hematomas can expedite this process; however, cumulative errors from segmentation on admission and follow-up scans can hamper accurate HE classification. In this study, we combined a tandem deep-learning classification model with automated segmentation to generate probability measures for false HE classifications. With this strategy, we can limit expert review of automated hematoma segmentations to a subset of the dataset, tailored to the research team’s preferred sensitivity or specificity thresholds and their tolerance for false-positive versus false-negative results. We utilized three separate multicentric cohorts for cross-validation/training, internal testing, and external validation (<i>n</i> = 2261) to develop and test a pipeline for automated hematoma segmentation and to generate ground truth binary HE annotations (≥3, ≥6, ≥9, and ≥12.5 mL). Applying a 95% sensitivity threshold for HE classification showed a practical and efficient strategy for HE annotation in large ICH datasets. This threshold excluded 47–88% of test-negative predictions from expert review of automated segmentations for different HE definitions, with less than 2% false-negative misclassification in both internal and external validation cohorts. Our pipeline offers a time-efficient and optimizable method for generating ground truth HE classifications in large ICH datasets, reducing the burden of expert review of automated hematoma segmentations while minimizing misclassification rate. |
| format | Article |
| id | doaj-art-262bae82b05647f5ba80b81862835161 |
| institution | Kabale University |
| issn | 2076-3417 |
| language | English |
| publishDate | 2024-12-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Applied Sciences |
| spelling | doaj-art-262bae82b05647f5ba80b818628351612025-01-10T13:14:28ZengMDPI AGApplied Sciences2076-34172024-12-0115111110.3390/app15010111Optimizing Automated Hematoma Expansion Classification from Baseline and Follow-Up Head Computed TomographyAnh T. Tran0Dmitriy Desser1Tal Zeevi2Gaby Abou Karam3Julia Zietz4Andrea Dell’Orco5Min-Chiun Chen6Ajay Malhotra7Adnan I. Qureshi8Santosh B. Murthy9Shahram Majidi10Guido J. Falcone11Kevin N. Sheth12Jawed Nawabi13Seyedmehdi Payabvash14Department of Radiology and Biomedical Imaging, Yale School of Medicine, New Haven, CT 06519, USADepartment of Neuroradiology, Charité—Universitätsmedizin Berlin, Humboldt-Universität Zu Berlin, Freie Universität Berlin, Berlin Institute of Health, 10117 Berlin, GermanyDepartment of Radiology and Biomedical Imaging, Yale School of Medicine, New Haven, CT 06519, USADepartment of Radiology and Biomedical Imaging, Yale School of Medicine, New Haven, CT 06519, USADepartment of Neuroradiology, Charité—Universitätsmedizin Berlin, Humboldt-Universität Zu Berlin, Freie Universität Berlin, Berlin Institute of Health, 10117 Berlin, GermanyDepartment of Neuroradiology, Charité—Universitätsmedizin Berlin, Humboldt-Universität Zu Berlin, Freie Universität Berlin, Berlin Institute of Health, 10117 Berlin, GermanyDepartment of Radiology and Biomedical Imaging, Yale School of Medicine, New Haven, CT 06519, USADepartment of Radiology and Biomedical Imaging, Yale School of Medicine, New Haven, CT 06519, USAZeenat Qureshi Stroke Institute and Department of Neurology, University of Missouri, Columbia, MO 65201, USADepartment of Neurology, Weill Cornell School of Medicine, New York, NY 10065, USADepartment of Neurosurgery, Mount Sinai School of Medicine, New York, NY 10029, USADepartment of Neurology, Yale School of Medicine, New Haven, CT 06510, USADepartment of Neurology, Yale School of Medicine, New Haven, CT 06510, USADepartment of Neuroradiology, Charité—Universitätsmedizin Berlin, Humboldt-Universität Zu Berlin, Freie Universität Berlin, Berlin Institute of Health, 10117 Berlin, GermanyDepartment of Radiology and Biomedical Imaging, Yale School of Medicine, New Haven, CT 06519, USAHematoma expansion (HE) is an independent predictor of poor outcomes and a modifiable treatment target in intracerebral hemorrhage (ICH). Evaluating HE in large datasets requires segmentation of hematomas on admission and follow-up CT scans, a process that is time-consuming and labor-intensive in large-scale studies. Automated segmentation of hematomas can expedite this process; however, cumulative errors from segmentation on admission and follow-up scans can hamper accurate HE classification. In this study, we combined a tandem deep-learning classification model with automated segmentation to generate probability measures for false HE classifications. With this strategy, we can limit expert review of automated hematoma segmentations to a subset of the dataset, tailored to the research team’s preferred sensitivity or specificity thresholds and their tolerance for false-positive versus false-negative results. We utilized three separate multicentric cohorts for cross-validation/training, internal testing, and external validation (<i>n</i> = 2261) to develop and test a pipeline for automated hematoma segmentation and to generate ground truth binary HE annotations (≥3, ≥6, ≥9, and ≥12.5 mL). Applying a 95% sensitivity threshold for HE classification showed a practical and efficient strategy for HE annotation in large ICH datasets. This threshold excluded 47–88% of test-negative predictions from expert review of automated segmentations for different HE definitions, with less than 2% false-negative misclassification in both internal and external validation cohorts. Our pipeline offers a time-efficient and optimizable method for generating ground truth HE classifications in large ICH datasets, reducing the burden of expert review of automated hematoma segmentations while minimizing misclassification rate.https://www.mdpi.com/2076-3417/15/1/111intracerebral hemorrhagehematoma expansionconvolution neural networksegmentationclassificationground truth generation |
| spellingShingle | Anh T. Tran Dmitriy Desser Tal Zeevi Gaby Abou Karam Julia Zietz Andrea Dell’Orco Min-Chiun Chen Ajay Malhotra Adnan I. Qureshi Santosh B. Murthy Shahram Majidi Guido J. Falcone Kevin N. Sheth Jawed Nawabi Seyedmehdi Payabvash Optimizing Automated Hematoma Expansion Classification from Baseline and Follow-Up Head Computed Tomography Applied Sciences intracerebral hemorrhage hematoma expansion convolution neural network segmentation classification ground truth generation |
| title | Optimizing Automated Hematoma Expansion Classification from Baseline and Follow-Up Head Computed Tomography |
| title_full | Optimizing Automated Hematoma Expansion Classification from Baseline and Follow-Up Head Computed Tomography |
| title_fullStr | Optimizing Automated Hematoma Expansion Classification from Baseline and Follow-Up Head Computed Tomography |
| title_full_unstemmed | Optimizing Automated Hematoma Expansion Classification from Baseline and Follow-Up Head Computed Tomography |
| title_short | Optimizing Automated Hematoma Expansion Classification from Baseline and Follow-Up Head Computed Tomography |
| title_sort | optimizing automated hematoma expansion classification from baseline and follow up head computed tomography |
| topic | intracerebral hemorrhage hematoma expansion convolution neural network segmentation classification ground truth generation |
| url | https://www.mdpi.com/2076-3417/15/1/111 |
| work_keys_str_mv | AT anhttran optimizingautomatedhematomaexpansionclassificationfrombaselineandfollowupheadcomputedtomography AT dmitriydesser optimizingautomatedhematomaexpansionclassificationfrombaselineandfollowupheadcomputedtomography AT talzeevi optimizingautomatedhematomaexpansionclassificationfrombaselineandfollowupheadcomputedtomography AT gabyaboukaram optimizingautomatedhematomaexpansionclassificationfrombaselineandfollowupheadcomputedtomography AT juliazietz optimizingautomatedhematomaexpansionclassificationfrombaselineandfollowupheadcomputedtomography AT andreadellorco optimizingautomatedhematomaexpansionclassificationfrombaselineandfollowupheadcomputedtomography AT minchiunchen optimizingautomatedhematomaexpansionclassificationfrombaselineandfollowupheadcomputedtomography AT ajaymalhotra optimizingautomatedhematomaexpansionclassificationfrombaselineandfollowupheadcomputedtomography AT adnaniqureshi optimizingautomatedhematomaexpansionclassificationfrombaselineandfollowupheadcomputedtomography AT santoshbmurthy optimizingautomatedhematomaexpansionclassificationfrombaselineandfollowupheadcomputedtomography AT shahrammajidi optimizingautomatedhematomaexpansionclassificationfrombaselineandfollowupheadcomputedtomography AT guidojfalcone optimizingautomatedhematomaexpansionclassificationfrombaselineandfollowupheadcomputedtomography AT kevinnsheth optimizingautomatedhematomaexpansionclassificationfrombaselineandfollowupheadcomputedtomography AT jawednawabi optimizingautomatedhematomaexpansionclassificationfrombaselineandfollowupheadcomputedtomography AT seyedmehdipayabvash optimizingautomatedhematomaexpansionclassificationfrombaselineandfollowupheadcomputedtomography |