What question did this study set out to answer?

The study aims to enhance the accuracy of intracranial hemorrhage volume measurement using AI and clinician adjustments.

February 2, 2026

Abstract WP227: Integration of Open Source Artificial Intelligence Based Intracranial Hemorrhage Volume Measurement and Prediction Tools into Clinical Workflows.

Key Points

The study aims to enhance the accuracy of intracranial hemorrhage volume measurement using AI and clinician adjustments.
Developed a deep learning network (SegResNet) for segmenting ICrH on non-contrast CT scans.
Used a dataset of 305 NCCT scans initially for training and 120 independent scans for testing.
Integrated the model with clinical workflow via dicom connectivity for real-time assessment by clinicians.
Created a multimodal model incorporating various clinical metrics for predicting ICrH volume change.
Initial median DICE score for total ICrH segmentation was 0.82, improving to 0.88 after clinician adjustments.
The ICH segmentation accuracy improved significantly, while EAH showed less improvement.
Multimodal prediction model achieved a Mean Absolute Error of 11.6 ml for ICrH volume change at 24h.

Abstract

Background and Aims: Accurate hematoma volume measurements facilitate prognostication and translation of Intracranial Hemorrhage (ICrH) research into clinical practice. We built a deep learning network for automated segmentation/measurement of intracerebral (ICH), intraventricular (IVH) and extra-axial (EAH) components on non-contrast CT (NCCT) and integrated it into the clinical workflow of a statewide TeleStroke Service (TSS). The primary aim was to demonstrate improved model accuracy with prospective clinician adjustment of segmentations and retraining. A second aim was to develop a model for prediction of change in ICrH volume at 24h. Methods: A 3D convolutional neural network (SegResNet) was trained using the Medical Open Networks for Artificial Intelligence (MONAI) framework and 305 NCCT scans from a recently completed trial that included ICH patients within 6 hours of onset. Initial ICrH scans (120) from the TSS were used as an independent test set. The model was deployed with dicom connectivity to the TSS PACS, allowing inference segmentation and volume measurements to be assessed by clinicians assessing ICrH patients. A multimodal model including baseline NCCT, blood pressure, anticoagulant use, GCS and NIHSS scores was developed to predict ICrH expansion. Results: Initial median 95% CI DICE scores for total ICrH were 0.82 0.81,0.84. ICH segmentation (0.83 0.80,0.85 was more accurate than that for IVH (0.25 0.14,0.38) or EAH (0.23 0.14,0.34). Bland-Altman analysis indicated a mean bias of 4.3 ml (95% limits of agreement -18.6,15.9 ml). The model was used in 230 TSS patients over a 2-year period. Inference segmentations were assessed for accuracy, edited by clinicians and used to retrain the model. Median DICE scores improved to 0.88 0.85,0.88, p<0.001 cf. the initial model, due to more accurate ICH (0.90 0.88,0.91, p<0.001) and IVH (0.67 0.56,0.71, p=0.047), but not EAH segmentation (0.31 0.21,0.47, p=0.089). The mean bias was 3.5 ml (-12.4,12.3 ml). The multimodal prediction model predicted ICrH volume change at 24h with a Mean Absolute Error of 11.6 ml (Root Mean Squared Error 18.7 ml, r=0.72). Conclusion: Fully integrated automated ICrH volume measurement tools can be continuously improved with clinician assessment as part of the clinical workflow. Multimodal prediction models require refinement and independent validation, but may be useful in improving patient selection for future trials and guide acute ICrH management.

Read Full Paperexternally

Mark Helpful

Bookmark

Relay

View Full Paper