The frontal skull Hounsfield unit value can predict ventricular enlargement in patients with subarachnoid haemorrhage.
Yu Deok WonMin Kyun NaChoong Hyun KimJae Min KimJin Hwan CheongJe Il RyuMyung-Hoon HanPublished in: Scientific reports (2018)
Hydrocephalus is a common complication following subarachnoid haemorrhage (SAH) arising from spontaneous aneurysm rupture. The Hounsfield unit (HU) value from computed tomography scans may reflect bone mineral density, which correlates with body mass index, which in turn is related to post-SAH ventricle size changes. We herein investigated potential associations between frontal skull HU values and ventricle size changes after SAH. HU values from four different areas in the frontal bone were averaged to minimize measurement errors. The bicaudate index and Evans ratio were measured using both baseline and follow-up CT images. CT images with bicaudate index >0.2 and Evans ratio >0.3 simultaneously were defined as indicating ventriculomegaly. We included 232 consecutive patients with SAH due to primary spontaneous aneurysm rupture, who underwent clipping over almost a 9-year period at a single institution. The first tertile of frontal skull HU values in older patients (≥55 years) was an independent predictor of ventriculomegaly after SAH, as compared to the third tertile in younger patients (hazard ratio, 4.01; 95% confidence interval 1.21-13.30; p = 0.023). The lower frontal skull HU value independently predicted ventricular enlargement post-SAH, due to the potential weak integrity of subarachnoid trabecular structures in younger patients.
Keyphrases
- bone mineral density
- computed tomography
- end stage renal disease
- body mass index
- working memory
- postmenopausal women
- ejection fraction
- functional connectivity
- chronic kidney disease
- newly diagnosed
- dual energy
- coronary artery
- heart failure
- contrast enhanced
- body composition
- deep learning
- prognostic factors
- emergency department
- convolutional neural network
- patient safety
- atrial fibrillation
- atomic force microscopy
- machine learning
- adverse drug
- pet ct
- electronic health record