Vasospasm after Subarachnoid Hemorrhage: Utility of Perfusion CT and CT Angiography on Diagnosis and Management

Design

At our institution, all patients with an acute aneurysmal SAH are monitored for vasospasm with serial neurologic examinations and daily TCD studies. When the clinical findings are suspicious or TCD FVs are elevated, conventional DSA is performed. In some cases, CTA and PCT are performed before DSA. Angiographic vasospasm is treated with intra-arterial verapamil or endovascular balloon angioplasty when appropriate. In addition, standard medical treatment for SAH patients includes oral nimodipine26 and hypertensive therapy, with the goal of increasing the mean arterial blood pressure to 100–120 mmHg.27, 28 External ventricular drainage is used to treat symptomatic hydrocephalus.

Between January and September 2003, all patients with aneurysmal SAH who underwent DSA because of a suspicion of vasospasm were retrospectively identified. Patients were included in the present study if they underwent CTA/PCT and DSA (with or without endovascular treatment) within a 12-hour interval. The TCD examinations performed the same day as the CTA/PCT were also included in the analysis.

Our institutional review board approved this retrospective review of patient studies obtained as part of standard clinical care.

Imaging Protocol and Data Processing

The CT protocol for patients with suspected vasospasm includes a noncontrast CT (NCT) of the brain, CTA, 2-level PCT series, and contrast-enhanced CT of the brain.

The CTA data acquisition was performed according to the following protocol: spiral mode, 0.6-second sections; collimation, 8 or 16 × 1.25 mm; pitch, 1.75; section thickness, 1.25 mm; reconstruction interval, 1.00 mm; acquisition parameters, 120 kVp/240 mA. A caudocranial scanning direction was selected, covering the whole brain down to 1 cm below the foramen magnum, to encompass the posteroinferior cerebellar arteries in the volume analysis. Seventy milliliters of iohexol (Omnipaque, Amersham Health, Princeton, NJ; 300 mg/mL of iodine) were administered into an antecubital vein by using a power injector at an injection rate of 4 mL/s. Adequate timing of the CTA acquisition was achieved according to a bolus-chasing technique. CTA raw data were reformatted in axial, sagittal, and coronal 3-mm-thick maximal intensity projection (MIP) images.

PCT was performed 5 minutes after the CTA. PCT data acquisition consisted of a 45-second series, with 45 gantry rotations performed in cine mode during intravenous administration of iodinated contrast material. The temporal sampling rate was one rotation of the gantry per second on acquisition. Multidetector raw CT technology permitted the assessment of 2 adjacent 10-mm-thick sections for each series. Two 10-mm-thick sections were preferred to 4 adjacent 5-mm-thick sections per location, to maximize signal intensity-to-noise ratio for the same acquisition parameters. Two successive PCT data acquisitions were performed, separated by 3–5 minutes. The 2 sections of the first PCT level were at the level of the third ventricle and the basal ganglia, positioned above the orbits to protect the lenses. The second PCT series was selected at a level of 3.5 cm rostral to the first section of the first series. For each series of PCT, a bolus of 40 mL of iohexol was administered into an antecubital vein (right arm preferred) by using a power injector at an injection rate of 4 mL per second. The acquisition parameters for both PCT series included 80 kVp and 120 mAs. CT scanning was initiated 7 seconds after initiation of contrast bolus injection. The delay before contrast material reached the brain parenchyma allowed the acquisition of baseline images without contrast enhancement.

TCD FVs were recorded with a Neuroguard ultrasonographic system (Medasonics, Fremont, Calif) in the internal carotid artery bifurcations (ICABs), anterior cerebral (ACA), middle cerebral (MCA), posterior cerebral (PCA), vertebral (VA) and basilar (BA) arteries, as well as in the external carotid arteries (ECA). In addition to the absolute FV values, extracranial-to-intracranial (EICA) ratios were calculated as the ratio between the FV in one specific artery and the FV in the ECA.29, 30 The concept of EICA ratio does not apply to the PCA, because it is most often supplied by the vertebrobasilar system. Instead, a BA-to-VA ratio was calculated according to the following formula: BA FV/[(right VA FV + left VA FV)/2].

Three- or 4-vessel DSA was performed via a transfemoral approach under monitored sedation or general anesthesia. Conventional angiographic views (frontal, lateral, oblique) were obtained, as were dedicated magnified and focused views on the treated aneurysmal sites.

Data Analysis

The NCT examinations were reviewed by an experienced neuroradiologist for the presence or absence of hypoattenuation.

The PCT data were analyzed by using software developed by Philips Medical Systems (Cleveland, Ohio). This software relies on the central volume principle, which is the most accurate for low injection rates of iodinated contrast material.31 The central volume principle uses mathematical deconvolution to calculate the mean transit time (MTT).32 The regional cerebral blood volume (rCBV) map is calculated from the areas under the time-enhancement curves.33 Finally, a simple equation combining rCBV and MTT values allows the calculation of regional cerebral blood flow (rCBF): rCBF = rCBV/MTT.15–17 Time-to-peak (TTP) values were also calculated. rCBV, MTT, rCBF, and TTP values were recorded in the gray matter of the arterial territories listed in Table 1.

CTA MIP images were reviewed by a neuroradiologist for vasospasm, which was qualified as absent (0), present and mild/moderate (1), or present and severe (2). When the reviewer could not decide between a hypoplastic and a vasospastic segment of the circle of Willis, the considered segment was qualified as “indeterminate” and considered as negative with respect to vasospasm. The reviewer also evaluated the quality of the CTA images (metallic artifacts, quality of the contrast material bolus).

All TCD examinations were performed by the same technologist. The following thresholds were used in the interpretation of absolute values and EICA ratios11, 13:

For the vertebrobasilar system, the following thresholds were used11, 13:

The DSA examinations were performed and interpreted by an interventional neuroradiologist for vasospasm. Vasospasm was graded as absent (0), present but not requiring endovascular treatment (1), or present and requiring endovascular treatment (2).

Because the goal of the present study was to compare CTA/PCT, TCD, and DSA techniques for the evaluation of vasospasm, we developed and used the anatomic chart reported in Table 1 to compare the values in the different arterial territories as assessed by the different imaging techniques. For each patient, 18 arterial territories were assessed and compared across techniques.

Statistical Analysis

The patients and assessed arterial territories were distributed into the following groups:

1 + 2. No angiographic vasospasm

3. Angiographic vasospasm but no endovascular treatment,

4. Angiographic vasospasm and endovascular treatment.

To compare PCT and TCD, the group of patients with no vasospasm (1 + 2) diagnosed on DSA was further subdivided into 2 subgroups:

1. No vasospasm diagnosed on DSA or on TCD

2. Vasospasm diagnosed on TCD but not on DSA (TCD false-positives), leading to a final 4-group distribution.

PCT values were evaluated for statistically significant differences between the 4 groups by using Kruskall-Wallis tests. When the Kruskal-Wallis test was positive and significant (P < .001), the statistical significance was checked by comparing each pair of groups by using Wilcoxon (Mann-Whitney) tests. For both the Kruskal-Wallis and the Wilcoxon tests, the significance level was set at 0.001 to account for the multiple testing (Bonferroni).

For each PCT parameter, a “diagnosis” threshold was calculated as the mid-value between the average PCT values in the group of patients with no angiographic vasospasm and in the group of patients with angiographic vasospasm (with or without endovascular treatment). Similarly, a “treatment” threshold was calculated as the mid-value between the average PCT values in the group of patients with angiographic vasospasm but no endovascular treatment and in the group of patients with angiographic vasospasm and endovascular treatment.

The PCT “diagnosis” thresholds for each parameter were evaluated for accuracy, sensitivity, and specificity, as well as negative (NPVs) and positive predictive values (PPVs) in the prediction of territories with angiographic vasospasm. As a reminder, the accuracy of a test designates the number of true results (true positives + true negatives) divided by the total number of subjects. Selected combinations of PCT parameters and CTA were also considered. Similarly, the accuracy, sensitivity, specificity, NPV, and PPV of NCT hypodensities, CTA findings, and TCD-absolute FV (120 cm/s), and EICA ratio thresholds (3.00) were also assessed for the diagnosis of present or absent angiographic vasospasm.

The PCT “treatment” thresholds for each parameter, but also CTA (severe vasospasm) and TCD absolute FV (180 cm/s) and EICA ratio thresholds (6.00), were evaluated for accuracy, sensitivity, specificity, NPV, and PPV in predicting whether an endovascular treatment was performed or not during DSA.

Patients

Between January and September 2003, 112 patients were admitted to our institution with aneurysmal SAHs. A total of 27 patients with CTA/PCT and DSA within 12 hours were identified. Seven were men and 20 were women, with a median age of 56 years (range, 18–82 years). The presenting clinical World Federation of Neurosurgical Societies grades were as follows: grade 1, 2 patients; grade 2, 9 patients; grade 3, 8 patients; grade 4, 5 patients; grade 5, 3 patients. A total of 29 aneurysms were identified in these patients (3 ICABs, one superior hypophyseal artery, one carotid ophthalmic artery, 8 anterior communicating arteries [AComs], 4 MCA bifurcations, 10 posterior communicating arteries [PComs], one BA tip, one anterior inferior cerebellar artery), including 27 ruptured aneurysms, treated either by neurosurgical clipping18 or endovascular coiling.9

Seven patients were found to have no vasospasm on both DSA and TCD. Nine patients were suspected of vasospasm on TCD but turned out to be vasospasm-negative on DSA. Eleven patients were found to have angiographic vasospasm. Six of these 11 patients underwent endovascular therapy with intra-arterial verapamil and/or balloon angioplasty. It is interesting to note that, among the patients with posterior circulation aneurysms, the patient with the right anterior inferior cerebellar artery did not show vasospasm; the patient with the basilar tip aneurysm demonstrated vasospasm involving bilateral PCAs and was treated with intra-arterial verapamil.

Imaging Examinations

Because of a clinical suspicion of vasospasm, the 27 patients in our series underwent 35 NCT/CTA/PCT, TCD, and DSA examinations within 12 hours of each other (the imaging examinations separated by >12 hours from each other, especially the daily TCD examinations, were not considered). Eight patients underwent 3 successive imaging surveys (NCT/CTA/PCT, TCD, and DSA).

The NCT/CTA/PCT examinations were performed after a median delay of 7 days after the aneurysmal rupture (interquartile range, 6–9 days; range, 5–14 days). The TCD examinations were performed after a median delay of 7 days after the aneurysmal rupture (interquartile range, 6–8.5 days; range, 5–14 days). The DSA examinations were performed after a median delay of 7 days after the aneurysmal rupture (interquartile range, 6.5–9 days; range, 5–14.5 days).

With regard to the anatomic chart represented in Table 1, the results from the different imaging techniques were compared in each of the 630 (27 × 18) arterial territories in 27 patients. In this study, 402 arterial territories were DSA-negative and TCD-negative for vasospasm; 105 arterial territories were TCD-positive but DSA-negative (TCD false-positives); and 123 arterial territories were DSA-positive for vasospasm, 60 of which were endovascularly treated (Fig 1).

• Download figure
• Open in new tab
• Download powerpoint

Fig 1.

Patient transferred at day 8 to our neurovascular ICU from an outside institution after coiling of a ruptured ACom aneurysm. NCT obtained at the admission of the patient in our neurovascular ICU demonstrated extensive residual SAH and suspicious loss of gray-white matter contrast in the left superior frontal gyrus (white arrows). The tip of a right ventricular drain catheter is also visible. On PCT, significantly abnormal brain perfusion in the distribution of the anterior and inferior branches of the left (and also, to a lesser extent, right) ACA (arrowheads) and of the right posterior MCA branches is seen primarily on MTT and TTP maps. The rCBF was also slightly decreased in the same territories, whereas rCBV was mainly preserved (it is lowered only in the left superior frontal gyrus [star]). CTA confirmed the suspicion of moderate vasospasm of both A2 and A3 segments of the ACA (arrows), ultimately verified by gold-standard DSA. No abnormality of the right posterior MCA branches was identified. The artifacts created by the coils on the CTA images, obscuring the A1 segments bilaterally and interfering with their evaluation, are noteworthy. Endovascular therapy (intra-arterial verapamil) was performed in the ACA territories during the DSA.

• Download figure
• Open in new tab
• Download powerpoint

Fig 2.

Patient transferred to our neurovascular ICU from an outside institution at day 4 after rupture of an ACom aneurysm (arrow). Admission TCD ultrasonography demonstrated increased absolute FV of 135 cm/s in ACAs. A CTA and PCT survey were obtained to rule out vasospasm; the PCT was completely normal and the CTA demonstrated the ACom aneurysm (arrow), but no vasospasm. DSA was performed determine the configuration of the aneurysm and to determine the best therapeutic strategy (endovascular coiling vs neurosurgical clipping). It confirmed the absence of vasospasm. TCD false-positive results were most likely related to the triple H (hypertensive, hypervolemic, hemodilutional) therapy undergone by the patient at the time of the TCD examination.

Twenty-five CTA examinations were considered of good quality. Two showed poor contrast bolus but were still interpretable. Eight showed significant metallic artifacts interfering with the interpretation of some segments of the circle of Willis, especially A1 segments, M1 segments, and PComs. These segments were considered negative, depending on the final DSA interpretation either CTA true-negative or false-negative.

Imaging Examinations Results

The measured values of the parameters in the 4 categories of arterial territories are reported in Table 2. Statistically significant differences could be identified between the 4 categories of arterial territories for all the PCT parameters (Kruskal-Wallis and Wilcoxon, P < .001) except rCBV.

With regard to the diagnosis of angiographic vasospasm, the PCT thresholds were set at the mid-value between the average PCT values in the group of DSA-negative arterial territories and in the group of DSA-positive arterial territories. The accuracy, sensitivity, specificity, NPV, and PPV of the different imaging techniques for the diagnosis of angiographic vasospasm were calculated. The thresholds and results are reported in Table 3.

MTT with a threshold of 6.4 seconds was the most accurate parameter for the diagnosis of angiographic vasospasm, with a very high NPV of 98.7%.

In light of the high sensitivity of MTT and the high specificity of CTA, we evaluated the accuracy, sensitivity, specificity, NPV, and PPV of a combined interpretation involving first the assessment of MTT as a screening test and then the evaluation of CTA to confirm the diagnosis of vasospasm. The accuracy, sensitivity, specificity, NPV, and PPV of the MTT/CTA combination were 93.0%, 72.4%, 98.0%, 93.6, and 89.9%, respectively.

Regarding the endovascular treatment of vasospasm, the PCT thresholds were set at the mid-value between the average PCT values in the group of DSA-positive arterial territories without endovascular treatment and in the group of DSA-positive arterial territories with endovascular treatment. The accuracy, sensitivity, specificity, NPV, and PPV of the different imaging techniques for the diagnosis of vasospasm requiring endovascular treatment were calculated, with regard to DSA endovascular treatment as the gold standard. The thresholds and results are reported in Table 4.

MTT was again the parameter with the highest NPV, but with this time a threshold at 7.6 seconds, whereas PCT cortical rCBF with a threshold at 39.3 mL × 100 g⁻¹ × min⁻¹ was the most accurate indicator for endovascular therapy.