The relationship between subarachnoid hemorrhage volume and development of cerebral vasospasm

Posthemorrhagic Hydrocephalus in Patients with Subarachnoid Hemorrhage Occurs Independently of CSF Osmolality

The molecular mechanisms underlying the development of posthemorrhagic hydrocephalus (PHH) remain incompletely understood. As the disease pathogenesis often cannot be attributed to visible cerebrospinal fluid (CSF) drainage obstructions, we here aimed to elucidate whether elevated CSF osmolality following subarachnoid hemorrhage (SAH) could potentiate the formation of ventricular fluid, and thereby contribute to the pathological CSF accumulation observed in PHH. The CSF osmolality was determined in 32 patients with acute SAH after external ventricular drainage (EVD) placement and again upon EVD removal and compared with the CSF osmolality from 14 healthy control subjects undergoing vascular clipping of an unruptured aneurism. However, we found no evidence of elevated CSF osmolality or electrolyte concentration in patients with SAH when compared to that of healthy control subjects. We detected no difference in CSF osmolality and electrolyte content in patients with successful EVD weaning versus those that were shunted due to PHH. Taken together, elevated CSF osmolality does not appear to underlie the development of PHH following SAH. The pathological CSF accumulation observed in this patient group must thus instead be attributed to other pathological alterations associated with the abnormal presence of blood within the CSF compartments following SAH.

Evaluation of techniques to improve a deep learning algorithm for the automatic detection of intracranial haemorrhage on CT head imaging

BACKGROUND

Deep learning (DL) algorithms are playing an increasing role in automatic medical image analysis.

PURPOSE

To evaluate the performance of a DL model for the automatic detection of intracranial haemorrhage and its subtypes on non-contrast CT (NCCT) head studies and to compare the effects of various preprocessing and model design implementations.

METHODS

The DL algorithm was trained and externally validated on open-source, multi-centre retrospective data containing radiologist-annotated NCCT head studies. The training dataset was sourced from four research institutions across Canada, the USA and Brazil. The test dataset was sourced from a research centre in India. A convolutional neural network (CNN) was used, with its performance compared against similar models with additional implementations: (1) a recurrent neural network (RNN) attached to the CNN, (2) preprocessed CT image-windowed inputs and (3) preprocessed CT image-concatenated inputs. The area under the receiver operating characteristic curve (AUC-ROC) and microaveraged precision (mAP) score were used to evaluate and compare model performances.

RESULTS

The training and test datasets contained 21,744 and 491 NCCT head studies, respectively, with 8,882 (40.8%) and 205 (41.8%) positive for intracranial haemorrhage. Implementation of preprocessing techniques and the CNN-RNN framework increased mAP from 0.77 to 0.93 and increased AUC-ROC [95% confidence intervals] from 0.854 [0.816-0.889] to 0.966 [0.951-0.980] (p-value = 3.91 × 10^-12).

CONCLUSIONS

The deep learning model accurately detected intracranial haemorrhage and improved in performance following specific implementation techniques, demonstrating clinical potential as a decision support tool and an automated system to improve radiologist workflow efficiency.

KEY POINTS

• The deep learning model detected intracranial haemorrhages on computed tomography with high accuracy. • Image preprocessing, such as windowing, plays a large role in improving deep learning model performance. • Implementations which enable an analysis of interslice dependencies can improve deep learning model performance. • Visual saliency maps can facilitate explainable artificial intelligence systems. • Deep learning within a triage system may expedite earlier intracranial haemorrhage detection.

Automated Quantification of Compartmental Blood Volumes Enables Prediction of Delayed Cerebral Ischemia and Outcomes After Aneurysmal Subarachnoid Hemorrhage

OBJECTIVE

The role of hemorrhage volume in risk of vasospasm, delayed cerebral ischemia (DCI), and poor outcomes after aneurysmal subarachnoid hemorrhage (SAH) is well established. However, the relative contribution of blood within individual compartments is unclear. We present an automated technique for measuring not only total but also volumes of blood in each major compartment after SAH.

METHODS

We trained convolutional neural networks to identify compartmental blood (cisterns, sulci, and ventricles) from baseline computed tomography scans of patients with SAH. We compared automated blood volumes against traditional markers of bleeding (modified Fisher score [mFS], Hijdra sum score [HSS]) in 190 SAH patients for prediction of vasospasm, DCI, and functional status (modified Rankin Scale) at hospital discharge.

RESULTS

Combined cisternal and sulcal volume was better correlated with mFS and HSS than cisternal volume alone (ρ = 0.63 vs. 0.58 and 0.75 vs. 0.70, P < 0.001). Only blood volume in combined cisternal plus sulcal compartments was independently associated with DCI (OR 1.023 per mL, 95% CI 1.002-1.048), after adjusting for clinical factors while ventricular blood volume was not. Total and specifically sulcal blood volume was strongly associated with poor outcome (OR 1.03 per mL, 1.01-1.06, P = 0.006 and OR 1.04, 1.00-1.08 for sulcal) as was HSS (OR 1.06 per point, 1.00-1.12, P = 0.04), while mFS was not (P = 0.24).

CONCLUSIONS

An automated imaging algorithm can measure the volume of bleeding after SAH within individual compartments, demonstrating cisternal plus sulcal (and not ventricular) blood contributes to risk of DCI/vasospasm. Automated blood volume was independently associated with outcome, while qualitative grading was not.

Comparison of Symptomatic Vasospasm after Surgical Clipping and Endovascular Coiling

Vasospasm, initial neurological damage, rebleeding, and periprocedural complications are associated prognostic factors for clinical outcomes after aneurysmal subarachnoid hemorrhage (SAH). In this study, factors related to delayed ischemic neurological deficit (DIND) are evaluated using data from our institute for the last 18 years. Data from 2001 to 2018 of patients with aneurysmal SAH who underwent surgical clipping (SC) or endovascular coiling (EC) within 7 days of onset were retrospectively analyzed. Cases of mortality within 5 days after treatment were excluded. Multivariate analysis was used to identify the risk factors for DIND. In total, 840 cases of SAH were assessed; among these cases, 384 (45.7%) and 456 (54.3%) were treated with SC and EC, respectively. The frequency of DIND in the EC group was significantly less than that in the SC group (11.8% vs. 17.7%; p = 0.016). In the results of multivariate analysis, internal carotid artery (ICA) aneurysm and hemorrhagic complications were the risk factors for DIND. Cilostazol administration and EC were significant factors for vasospasm prevention after aneurysmal SAH (odds ratio of ICA aneurysm: 1.59, hemorrhagic complications: 1.76, SC: 1.51, and cilostazol administration: 0.51, respectively). Cilostazol administration was also a significant factor in patients who were treated with EC. ICA aneurysm, treatment strategy, hemorrhagic complications, and cilostazol administration were associated with DIND. Oral administration of cilostazol and avoiding hemorrhagic complications were effective in DIND prevention. If both treatments are available for ruptured aneurysms, clinicians should choose EC on the basis of its ability to prevent DIND.

Role of serum biomarkers and transcranial Doppler in predicting cerebral vasospasm after aneurysmal subarachnoid hemorrhage

Background

Aneurysmal subarachnoid hemorrhage (aSAH) can have serious consequences related to vasospasm and delayed cerebral ischemia (DCI). Serum biomarkers have emerged as a promising assessment tool to facilitate earlier diagnosis of cerebral vasospasm (CV) and to identify pre-clinical vessel narrowing.

Objectives

Our aim was to detect the predictive value of serum biomarkers such as von Willebrand factor (vWF), vascular endothelial growth factor (VEGF) and matrix metalloproteinase9 (MMP-9) in CV after aSAH.

Subjects and methods

Thirty five patients with recent aSAH were included. Patients were divided into two groups; 19 patients (CV group) and 16 patients (non-CV group). The CV group was further subdivided into 9 symptomatic (DCI) and 10 asymptomatic patients. All patients underwent transcranial Doppler (TCD) evaluations three times a week for 2 weeks measuring the mean flow velocities. Serum level of vWF, MMP-9, and VEGF were assessed twice (at onset and within 2 weeks).

Results

A statistically significant increase in serum biomarker levels was found in the CV group. Cutoff value for vWF, MMP-9, and VEGF were > 4985 ng/ml, > 495 ng/ml, and > 184 pg/ml, respectively. Statistically significant positive correlations were found between serum levels of biomarkers and degree of vasospasm. No difference was found in the biomarkers between symptomatic CV and asymptomatic CV.

Conclusion

Serum biomarkers are a reliable tool to predict CV following aSAH, their levels reflect the severity of vascular vasospasm, yet, they cannot predict DCI. TCD has a strong role in early detection, monitoring of post subarachnoid vasospasm and successfully capturing asymptomatic DCI.

Phase dynamics of cerebral blood flow in subarachnoid haemorrhage in response to sodium nitrite infusion

Aneurysmal subarachnoid haemorrhage (SAH) is a devastating subset of stroke. One of the major determinates of morbidity is the development of delayed cerebral ischemia (DCI). Disruption of the nitric oxide (NO) pathway and consequently the control of cerebral blood flow (CBF), known as cerebral autoregulation, is believed to play a role in its pathophysiology. Through the pharmacological manipulation of in vivo NO levels using an exogenous NO donor we sought to explore this relationship. Phase synchronisation index (PSI), an expression of the interdependence between CBF and arterial blood pressure (ABP) and thus cerebral autoregulation, was calculated before and during sodium nitrite administration in 10 high-grade SAH patients acutely post-rupture. In patients that did not develop DCI, there was a significant increase in PSI around 0.1 Hz during the administration of sodium nitrite (33%; p-value 0.006). In patients that developed DCI, PSI did not change significantly. Synchronisation between ABP and CBF at 0.1 Hz has been proposed as a mechanism by which organ perfusion is maintained, during periods of physiological stress. These findings suggest that functional NO depletion plays a role in impaired cerebral autoregulation following SAH, but the development of DCI may have a distinct pathophysiological aetiology.

The role of haptoglobin and hemopexin in the prevention of delayed cerebral ischaemia after aneurysmal subarachnoid haemorrhage: a review of current literature

Delayed cerebral ischaemia (DCI) after aneurysmal subarachnoid haemorrhage (aSAH) is a major cause of mortality and morbidity. The pathophysiology of DCI after aSAH is thought to involve toxic mediators released from lysis of red blood cells within the subarachnoid space, including free haemoglobin and haem. Haptoglobin and hemopexin are endogenously produced acute phase proteins that are involved in the clearance of these toxic mediators. The aim of this review is to investigate the pathophysiological mechanisms involved in DCI and the role of both endogenous as well as exogenously administered haptoglobin and hemopexin in the prevention of DCI.

The Nimodipine-Sparing Effect of Perioperative Dexmedetomidine Infusion During Aneurysmal Subarachnoid Hemorrhage: A Prospective, Randomized, Controlled Trial

Background: Nimodipine can block the influx of calcium into the vascular smooth muscle cell and prevent secondary ischemia in patients with aneurysmal subarachnoid hemorrhage. However, the reduction of blood pressure after long-term intravenous administration of nimodipine has been associated with neurological deterioration. Yet, no effective solutions have been suggested to address this phenomenon. The use of neuroprotective drug combinations may reduce the risk of sudden blood pressure loss. This prospective, randomized, controlled trial was performed to evaluate the nimodipine-sparing effect of perioperative dexmedetomidine infusion during aneurysmal subarachnoid hemorrhage.

Methods: One hundred nine patients who underwent aneurysm embolization were divided into three groups: group C (n = 35, infused with 0.9% sodium chloride at the same rate as other two groups), group D1 (n = 38, dexmedetomidine infusion at 0.5 µg·kg^–1 for 10 min, then adjusted to 0.2 µg·kg^–1·h^–1), and group D2 (n = 36, dexmedetomidine infusion at 0.5 µg·kg^–1 for 10 min, then adjusted to 0.4 µg·kg^–1·h^–1). Patient-controlled analgesia was given for 48 h after surgery. The primary outcome measure was the total consumption of nimodipine during the first 48 h after surgery. The secondary outcome measures were recovery time at post-anesthesia care unit (PACU), postoperative pain intensity scores, dexmedetomidine and sufentanil consumption, hemodynamic, satisfaction of patients and neurosurgeon, neurologic examination (Glasgow Coma Scale, GCS), Bruggemann comfort scale, and adverse effects. Intraoperative hemodynamics were recorded at the following time-points: arrival at the operating room (T1); before intubation (T2); intubation (T3); 5 min (T4), 10 min (T5), and 15 min (T6) after intubation; suturing of femoral artery (T7); end of surgery (T8); extubation (T9); and 5 min (T10), 10 min (T11), and 15 min (T12) after arrival at the PACU. The level of sedation was recorded at 15 min, 30 min, 1 h, and 2 h after extubation. We also recorded the incidence of symptomatic cerebral vasospasm during 7 days after surgery, Glasgow Outcome Score (GOS) at 3 months, and incidence of cerebral infarction 30 days after surgery.

Results: The consumption of nimodipine during the first 48 h after surgery was significantly lower in group D2 (P < 0.05). Compared with group C, HR and MAP were significantly decreased from T2 to T12 in group D1 and D2 (P < 0.05). Patients in group D2 showed a significantly decreased MAP from T5 to T9 compared with group D1 (P < 0.05). The consumption of sevoflurane, remifentanil, dexmedetomidine, and nimodipine were all significantly reduced in groups D1 and D2 during surgery (P < 0.05). Compared with group C, MAP was significantly decreased in groups D1 and D2 during the first 48 h after surgery (P < 0.05). Compared with group C, consumption of sufentanil and dexmedetomidine at 1 h, pain intensity at 1 h, and 8 h after surgery were significantly decreased in groups D1 and D2 (P < 0.05). FAS was significantly higher in group D2 at 8 h, 16 h, and 24 h after surgery. LOS was significantly lower only in group D2 at 0.5 h after surgery (P < 0.05). Compared with group C, BCS was significantly higher group D2 at 4 h and 8 h after surgery (P < 0.05). There were no significant differences among the three groups in consumption of propofol, cisatracurium, fentanyl, and vasoactive drugs during operation, recovery time at PACU, satisfaction of patients and neurosurgeon, and number of applied urapidil and GCS during the first 48 h after surgery. The incidence of symptomatic cerebral vasospasm during 7 days after surgery, GOS of 3 months, and cerebral infarction after 30 days were also comparable among the three groups.

Conclusions: Dexmedetomidine (infusion at 0.5 µg·kg^–1 for 10 min, then adjusted to 0.4 µg·kg^–1·h^–1 during the surgery) significantly reduced the total consumption of nimodipine during the first 48 h after surgery and promoted early rehabilitation of patients although the incidences of symptomatic cerebral vasospasm, GOS, and cerebral infarction were not reduced.

Hybrid 3D/2D Convolutional Neural Network for Hemorrhage Evaluation on Head CT

BACKGROUND AND PURPOSE

Convolutional neural networks are a powerful technology for image recognition. This study evaluates a convolutional neural network optimized for the detection and quantification of intraparenchymal, epidural/subdural, and subarachnoid hemorrhages on noncontrast CT.

MATERIALS AND METHODS

This study was performed in 2 phases. First, a training cohort of all NCCTs acquired at a single institution between January 1, 2017, and July 31, 2017, was used to develop and cross-validate a custom hybrid 3D/2D mask ROI-based convolutional neural network architecture for hemorrhage evaluation. Second, the trained network was applied prospectively to all NCCTs ordered from the emergency department between February 1, 2018, and February 28, 2018, in an automated inference pipeline. Hemorrhage-detection accuracy, area under the curve, sensitivity, specificity, positive predictive value, and negative predictive value were assessed for full and balanced datasets and were further stratified by hemorrhage type and size. Quantification was assessed by the Dice score coefficient and the Pearson correlation.

RESULTS

A 10,159-examination training cohort (512,598 images; 901/8.1% hemorrhages) and an 862-examination test cohort (23,668 images; 82/12% hemorrhages) were used in this study. Accuracy, area under the curve, sensitivity, specificity, positive predictive value, and negative-predictive value for hemorrhage detection were 0.975, 0.983, 0.971, 0.975, 0.793, and 0.997 on training cohort cross-validation and 0.970, 0.981, 0.951, 0.973, 0.829, and 0.993 for the prospective test set. Dice scores for intraparenchymal hemorrhage, epidural/subdural hemorrhage, and SAH were 0.931, 0.863, and 0.772, respectively.

CONCLUSIONS

A customized deep learning tool is accurate in the detection and quantification of hemorrhage on NCCT. Demonstrated high performance on prospective NCCTs ordered from the emergency department suggests the clinical viability of the proposed deep learning tool.

Cerebral vasospasm: current understanding

PURPOSE OF REVIEW

With recent research trying to explore the pathophysiologic mechanisms behind vasospasm, newer pharmacological and nonpharmacological treatments are being targeted at various pathways involved. This review is aimed at understanding the mechanisms and current and future therapies available to treat vasospasm.

RECENT FINDINGS

Computed tomography perfusion is a useful alternative tool to digital subtraction angiography to diagnose vasospasm. Various biomarkers have been tried to predict the onset of vasospasm but none seems to be helpful. Transcranial Doppler still remains a useful tool at the bedside to screen and follow up patients with vasospasm. Hypertension rather than hypervolemia and hemodilution in 'Triple-H' therapy has been found to be helpful in reversing the vasospasm. Hyperdynamic therapy in addition to hypertension has shown promising effects. Endovascular approaches with balloon angioplasty and intra-arterial nimodipine, nicardipine, and milrinone have shown consistent benefits. Endothelin receptor antagonists though relieved vasospasm, did not show any benefit on functional outcome.

SUMMARY

Endovascular therapy has shown consistent benefit in relieving vasospasm. An aggressive combination therapy through various routes seems to be the most useful approach to reduce the complications of vasospasm.

Predictors of Postoperative Cerebral Ischemia in Patients with Ruptured Anterior Communicating Artery Aneurysms

OBJECTIVE

Cerebral ischemia is a major contributor to poor outcome after ruptured anterior communicating artery aneurysms (ACoAs), and is not well classified. In this article, we develop a classification and identify risk factors of cerebral ischemia after ruptured ACoAs.

METHODS

Three hundred sixty patients with ruptured ACoAs undergoing microsurgical clipping were collected. Sex, age, smoking status, Hunt-Hess grade, Fisher grade, hospital stay, surgical timing, hypertension, diabetes, postoperative cerebral ischemia, and postoperative modified Rankin Scale score were collected. Postoperative ischemic changes are classified according to a novel grade (ischemic grade I-IV).

RESULTS

Predictive factors of postoperative ischemia (grade I-IV) included sex (odds ratio [OR], 1.956; 95% confidence interval [CI], 1.262-3.032; P = 0.003) and Fisher grade (OR, 1.813; 95% CI, 1.144-2.871; P = 0.011). Male sex had a tendency to develop postoperative cerebral ischemia (61.3% in the ischemia group vs. 45.7% in the nonischemia group), while surgical timing did not. However, in patients with postoperative ischemia, early surgery within 3 days (OR, 3.334; 95% CI, 1.411-7.879; P = 0.006) and advanced age greater than 55 years (OR, 2.783; 95% CI, 1.214-6.382; P = 0.016) were risk factors for postoperative neurologic deficits (grade III-IV).

CONCLUSIONS

Male sex and higher Fisher grade predict postoperative ischemia (grade I-IV), whereas surgical timing does not. However, in patients with postoperative cerebral ischemia, early surgery within 3 days and age greater than 55 years can increase the frequency of postoperative neurological deficits (grade III-IV). Older male patients undergoing early microsurgery had a tendency to develop neurologic deficits.

Non-aneurysmal spontaneous subarachnoid hemorrhage: perimesencephalic versus non-perimesencephalic

Objective

To compare the clinical evolution of perimesencephalic subarachnoid hemorrhage and non-perimesencephalic subarachnoid hemorrhage.

Methods

The study was conducted retrospectively in a tertiary hospital center in the north region of Portugal. Included patients had no identifiable cause for subarachnoid hemorrhage. Several epidemiologic, clinical and imaging aspects were statistically analyzed, taking into account the differences in perimesencephalic subarachnoid hemorrhage and non-perimesencephalic subarachnoid hemorrhage.

Results

Sixty-two patients met the inclusion criteria (46.8% - perimesencephalic subarachnoid hemorrhage; 53.2% - non-perimesencephalic subarachnoid hemorrhage). Demographic and clinical background characteristics were similar in both groups. Complications were more frequent in patients with non-perimesencephalic subarachnoid hemorrhage - 84.8% of the patients had at least one complication versus 48.3% in perimesencephalic subarachnoid hemorrhage. Vasospasm, infection and hydrocephaly were the most common complications (each was detected more frequently in the non-perimesencephalic subarachnoid hemorrhage group than in perimesencephalic subarachnoid hemorrhage group). Two patients died, both had a non-perimesencephalic subarachnoid hemorrhage. The median inpatient time was longer in the non-perimesencephalic subarachnoid hemorrhage group (21 versus 14 days). No incidents of rebleeding were reported during the follow-up period (mean time of 15 ± 10.3 months).

Conclusion

Perimesencephalic subarachnoid hemorrhage and non-perimesencephalic subarachnoid hemorrhage are two different entities that have different clinical outcomes, namely in terms of complication rate and median inpatient time. The management of these patients should respect this difference to improve treatment and optimize health care resources.

Can S100B Predict Cerebral Vasospasms in Patients Suffering from Subarachnoid Hemorrhage?

Background: Protein S100B has proven to be a useful biomarker for cerebral damages. Increased levels of serum and cerebrospinal fluid (CSF) S100B have been shown in patients suffering subarachnoid hemorrhage (SAH), severe head injury and stroke. In patients with SAH, the course of S100B levels has been correlated with neurological deficits and outcome. Cerebral vasospasm is a major contributor to morbidity and mortality. The primary aim of this study was to investigate the potential of S100B protein as a predictor of cerebral vasospasm in patients with severe SAH.

Materials and Methods: Patients with SAH, Fisher grade 3 and 4, were included in the study. Five samples of CSF and serum S100B were collected from each patient. The first sample (baseline sample) was drawn within the first 3 days following ictus and the following four samples, once a day on days 5–8, with day of ictus defined as day 1. Clinical suspicion of cerebral vasospasm confirmed by computed tomography angiography was used to diagnose cerebral vasospasm.

Results: A total of 18 patients were included. Five patients (28%) developed cerebral vasospasm, two (11%) developed ventriculitis. There were no significant differences between S100B for those with and without vasospasm. Serum S100B levels in patients with vasospasm were slightly lower within the first 5 days following ictus, compared to patients without vasospasm. Two out of five patients had elevated and increasing serum S100B prior to vasospasm. Only one showed a peak level of S100B 1 day before vasospasm could be diagnosed. Due to the low number of patients in the study, statistical significance could not be reached.

Conclusion: Neither serum nor CSF S100B can be used as predictor of cerebral vasospasm in patients suffering from SAH.