RADAR - Radiomics on aSDH: predicting outcome with surface area
Antonia Richter1, Johannes Wach2, Alim Basaran2
1Department of Neurosurgery, University Hospital Leipzig, 04103, Leipzig, Germany. Antonia.Richter@medizin.uni-leipzig.de.
Acta Neurochirurgica
|January 19, 2025
Summary
Hematoma surface area predicts outcomes in acute subdural hematoma (aSDH). Smaller surface areas correlate with better survival and neurological prognosis, especially after craniotomy surgery.
Area of Science:
- Neurosurgery
- Radiology
- Medical Imaging Analysis
Background:
- Acute subdural hematoma (aSDH) presents a critical neurosurgical challenge with significant morbidity and mortality.
- Limited clinical improvement post-evacuation necessitates better decision-making tools for surgical intervention.
- Radiomics can extract complex imaging patterns to reflect pathophysiology and aid in decision support.
Purpose of the Study:
- To analyze radiomic parameters of aSDH to guide decisions between urgent surgery and conservative management.
- To investigate the predictive value of radiomic features, specifically surface area, for neurological outcomes and survival rates.
Main Methods:
- Retrospective radiomic analysis of aSDH patients.
- Measurement of radiomic parameters using 3D Slicer software.
- Statistical analyses included correlation, AUC-analysis, and Kaplan-Meier survival analysis.
Main Results:
- Significant correlations were found between hematoma surface area and poorer neurological prognosis.
- Surface area significantly predicted poorer outcomes in patients undergoing craniotomy (p < 0.05).
- Smaller surface areas (< 339.50 cm² ) were associated with advantageous survival across all management subgroups.
Conclusions:
- Hematoma surface area is a potential predictor of neurological outcome and survival in aSDH patients.
- This finding is particularly relevant for guiding treatment decisions in patients undergoing craniotomy.
More Related Videos
07:15Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model
Published on: August 16, 2020
6.7K
04:09Predicting Treatment Response to Image-Guided Therapies Using Machine Learning: An Example for Trans-Arterial Treatment of Hepatocellular Carcinoma
Published on: October 10, 2018
8.2K
