Head computed tomography for subarachnoid hemorrhage diagnosis
A good quality (e.g. no motion artifact) non-contrast high-resolution Head computed tomography will detect SAH in ≥ 95% of cases if scanned within 48 hrs of SAH. Blood appears as high density (white) within subarachnoid spaces.
For subtle SAH findings, look in the occipital horns of the lateral ventricles and the dependent portions of the Sylvian fissures.
Acute headache may be the primary symptom of subarachnoid hemorrhage (SAH). Guidelines suggest that non-contrast computed tomography (CT) is adequate to exclude aneurysmal SAH if performed within 6 h after symptom onset. However, most studies of acute headache including CT, lumbar puncture and SAH are multicenter studies from referral hospitals with highly selected patient populations. In acute headache, a CT scan taken within 6 h is practically 100% sensitive for detecting any SAH 1)
CT also assesses:
1. ventricular size: hydrocephalus occurs acutely in 21% of aneurysmal ruptures.
2. hematoma: intracerebral hemorrhage or large amount of subdural blood with mass effect may need emergent evacuation
3. infarct: not sensitive in first 24 hours after infarct
4. amount of blood in cisterns and fissures: important prognosticator for vasospasm and can identify pretruncal nonaneurysmal hemorrhage
5. CT can predict aneurysm location based on the pattern of blood in ≈ 78% of cases (but mostly for MCA and A-comm aneurysms)
a) blood predominantly in anterior interhemispheric fissure (±blood in lateral ventricles)or within the gyrus rectus suggests a-comm aneurysm
b) blood predominantly in Sylvian fissure is compatible with p-commor MCAaneurysm on that side
c) blood predominantly in the prepontine or peduncular cistern suggests a basilar apex or SCA aneurysm
d) blood predominantly within ventricles
● blood primarily in 4th and third ventricle: suggests lower posterior fossa source, such as PICA aneurysm or VA dissection
● blood primarily in the 3rd ventricle suggests a basilar apex aneurysm
6. with multiple aneurysms, CT may help identify which one bled by the location of blood
Extensive subarachnoid hemorrhage distributed throughout perimesencephalic cisterns (predominantly in the chiasmatic cistern), prepontine cistern, both Sylvian fissures, and interhemispheric fissure, with a component of bifrontal parasagittal intraparenchymal hematoma with mild adjacent vasogenic edema. The hemorrhage also extends into the ventricular system (third ventricle and lateral ventricles). Findings consistent with modified Fisher grade IV subarachnoid hemorrhage.
This extension is compatible with an aneurysmal subarachnoid hemorrhage pattern.
Erasure of cerebral sulci, predominantly frontotemporal, in relation to diffuse cerebral edema.
Increased ventricular size for the patient's age.
In the early phase, during the first 24 hours, Head computed tomography, combined with intracranial CT angiography is recommended to make a positive diagnosis of subarachnoid hemorrhage (SAH), to identify the cause and to investigate for an intracranial aneurysm. Cerebral MRI may be proposed if the patient's clinical condition allows it. FLAIR imaging is more sensitive than CT to demonstrate a subarachnoid hemorrhage and offers greater degrees of sensitivity for the diagnosis of restricted subarachnoid hemorrhage in cortical sulcus. A lumbar puncture should be performed if these investigations are normal while clinical suspicion is high 2).
Noncontrast computed tomography is highly sensitive in detecting subarachnoid blood, especially within 6 hours of haemorrhage.
The distribution of the subarachnoid blood as shown on the first CT scan after aneurysm rupture barely allows to predict the symptomatic aneurysm site. Thus, neurosurgical decision making (identification of the ruptured aneurysm in patients with multiple aneurysms; surgical exploration in patients with non-perimesencephal SAH, but negative angiography) should not rely on the first CT scan after SAH 3).
Rogers et al. report significant differences in the diagnostic approach of Australasian emergency physicians and trainees to this condition, in particular the utility of CT within 6 h for exclusion of SAH 4).
Volume
The amount of blood in cisterns and fissures: important prognosticator for vasospasm.
Accurate volumetric assessment of spontaneous aneurysmal subarachnoid hemorrhage (aSAH) is a labor-intensive task performed with current manual and semiautomatic methods that might be relevant for its clinical and prognostic implications. In a research, García-García et al. sought to develop and validate an artificial intelligence-driven, fully automated blood segmentation tool for SAH patients via non-contrast computed tomography (NCCT) scans employing a transformer-based Swin-UNETR architecture.
They retrospectively analyzed NCCT scans from patients with confirmed aSAH utilizing the Swin-UNETR for segmentation. The performance of the proposed method was evaluated against manually segmented ground truth data using metrics such as Dice score, Intersection over Union (IoU), Volumetric Similarity Index(VSI), Symmetric Average Surface Distance(SASD), Sensitivity and Specificity. A validation cohort from an external institution was included to test the generalizability of the model.
The model demonstrated high accuracy with robust performance metrics across the internal and external validation cohorts. Notably, it achieved high Dice coefficient (0.873±0.097), IoU (0.810±0.092), VSI (0.840±0.131), Sensitivity (0.821±0.217) and Specificity (0.996±0.004) values and a low SASD (1.866±2.910), suggesting proficiency in segmenting blood in SAH patients. The model's efficiency was reflected in its processing speed, indicating potential for real-time applications.
The Swin UNETR-based model offers significant advances in the automated segmentation of blood in SAH patients on NCCT images. Despite the computational demands, the model operates effectively on standard hardware with a user-friendly interface, facilitating broader clinical adoption. Further validation across diverse datasets is warranted to confirm its clinical reliability 5)