Related Experiment Video
Updated: Mar 14, 2026
Automated Midline Shift and Intracranial Pressure Estimation based on Brain CT Images
Published on: April 13, 2013
Development and Validation of an Automatic Segmentation Algorithm for Quantification of Intracerebral Hemorrhage
Moritz Scherer1, Jonas Cordes2, Alexander Younsi2
1From the Department of Neurosurgery (M.S., A.Y., Y.-A.S., A.U., B.O.), Institute of Medical Biometry and Informatics (IMBI) (C.S.), and Department of Neurology (J.B.), University Hospital Heidelberg, Germany; Junior Group Medical Image Computing, German Cancer Research Center (DKFZ), Heidelberg, Germany (J.C., M.G., K.M.-H.); and Division of Neuroradiology, Heidelberg University Hospital, Germany (M.M.). moritz.scherer@med.uni-heidelberg.de.
A new automatic segmentation algorithm accurately quantifies spontaneous intracerebral hemorrhage (ICH) volume, outperforming the ABC/2 method. This tool improves ICH volume measurement for outcome prediction and surgical decisions.
Area of Science:
- Medical Imaging
- Neurology
- Radiology
Background:
- The ABC/2 method is standard for estimating spontaneous intracerebral hemorrhage (ICH) volume but has limitations.
- These limitations may contribute to inconsistent results in outcome studies.
Purpose of the Study:
- To develop and validate an automatic segmentation algorithm for precise ICH volume quantification.
- To compare the algorithm's accuracy against manual segmentation and the ABC/2 method.
Main Methods:
- A random-forest algorithm was trained using manual ICH segmentations and incorporating texture and threshold features.
- Algorithm performance was assessed for agreement with manual segmentations in two independent patient cohorts.
Main Results:
- The automatic algorithm demonstrated strong agreement with manual ICH volume measurements (concordance correlation coefficient 0.95).
- The ABC/2 method significantly overestimated ICH volumes compared to manual and algorithmic measurements.
- The algorithm's accuracy was validated in an independent cohort, showing excellent agreement (concordance correlation coefficient 0.99).
Conclusions:
- A novel automatic segmentation algorithm provides accurate and efficient volumetric analysis of spontaneous ICH.
- The algorithm overcomes the overestimation limitations of the ABC/2 method.
- This tool can enhance the evaluation of ICH volume as a prognostic factor and guide treatment decisions.
