INTRODUCTION
Cerebral vasospasm (CV) is a major cause of mortality and morbidity in patients with subarachnoid hemorrhage (SAH).
10) Therefore after a SAH, during the period in which CV frequently occurs, accurate diagnosis is helpful in early treatment and it is ultimately possible to assist in the neurological recovery of the patient.
12)
There are several methods for diagnosis of CV. Digital subtraction angiography (DSA) has been the gold standard diagnostic test,
20) however, its total complication rate is approximately 5% with a permanent stroke rate of approximately 0.5% to 1%.
4)5)7)11)18) Thus there is a significant demand for noninvasive alternatives for accurate detection of CV.
20) Transcranial Doppler sonography (TCD) is a noninvasive method for detection of vasospasm which has shown high sensitivity and specificity compared with DSA,
13) however, its operator dependence and difficulty in accurately detecting vasospasm at sites other than the proximal middle cerebral artery (MCA) limit its use to a screening method only.
1) While magnetic resonance angiography (MRA) can be used as an alternative, there are many disadvantages to this method, including a decrease in the visualization of the distal segment.
Computed tomographic angiography (CTA) has also been developed to provide clear vascular images, and it can detect CV after SAH with high sensitivity and specificity.
14) Also with the recent advancement in maximum intensity projection (MIP) imaging, CTA can obtain a more accurate and clearer image than before in a relatively short amount of time. The purpose of this article is to perform a retrospective study on the accuracy of MIP images of CTA for the diagnosis of CV following SAH compared to that of DSA.
DISCUSSION
The presence of significant vasospasm carries the risk of stroke, thus the role of imaging in these patients is not only to reach a diagnosis of vasospasm but mainly to make an appropriate clinical decision regarding the most suitable treatment.
12) DSA is considered the standard method to confirm the diagnosis of vasospasm.
16) The major advantage of DSA lies in its accuracy and the capacity to immediately perform endovascular treatment by balloon angioplasty and/or intra-arterial injection of vasodilatory drugs.
19) However, this procedure is invasive and is not always available in critically ill patients and not without its own risk.
3)11) There has been, therefore, great demand for noninvasive alternatives for accurate depiction of cerebral vasospasm.
20) TCD is an examination with many advantages including its accuracy, relative simplicity, and it can be performed every day, etc. However, the technical limitations of TCD include examiner experience and the difficulty of accurately detecting vasospasm at sites other than the proximal middle cerebral artery.
10) In addition, approximately 10% of studies using TCD failed to detect signals of cerebral artery blood flow caused by poor insonation of the cranial window.
8)17) Thus, TCD alone is inadequate for detection of vasospasm, and other methods are required.
9) MRA can depict the cerebral vasculature less invasively, however its sensitivity and specificity for detection of vasospasm after SAH remain relatively low. In addition, MRA has limitations for detecting the severity of vasospasm and visualization of the distal arteries.
15) Therefore, MRA cannot be considered as a practical alternative to DSA.
By contrast with these modalities, although the distal segments of the cerebral arteries are poorly imaged because of the limited spatial resolution,
10) CTA is a noninvasive alternative to DSA, and can be rapidly performed immediately after a non-contrast head CT, which is often performed in patients to evaluate for hydrocephalus.
2) However, there are also disadvantages. It is important to mention that imaging for vasospasm is usually performed after clipping or coiling of an aneurysm. Streak artifacts from these metal devices could obscure adjacent arteries especially after coiling and influence CTA's evaluation. However, recently developed MIP images have overcome some of the limitations of CTA. MIP images have advantages for differentiating calcification from true lumen and determining degree of atherosclerotic disease.
6)
This study examined the accuracy of MIP images of CTA for diagnosis of CV compared to DSA. When diagnosed with SAH, initial MIP images of CTA and DSA were checked, and the same examinations were repeated one week later. We considered that the MIP images of CTA could predict a significant part of the cerebral vasospasm. Actually, in this study compared with DSA, the accuracy of the MIP image of CTA was higher overall (88-100%). Previous studies have reported higher sensitivity (up to 97.5%) and specificity (up to 98.1%) of CTA for vasospasm.
20) This is explained by the fact that the previous study only considered severe vasospasm, and considered all vessels that were not severely narrowed as normal. This dichotomization of the interpretation of the CTA images resulted in apparent increased sensitivity and specificity at the expense of not detecting mild or moderate vasospasm.
However, in this study the degrees of vasospasm were categorized according to normal, mild, moderate, and severe. As a result, in the case of moderate and severe vasospasms, it was confirmed that sensitivity and specificity were high, and in the case of normal and mild vasospasms, it was confirmed that sensitivity and specificity were relatively lower. The fact that the accuracy of the moderate and severe vasospasms was very high is considered the most important findings of this study (
Fig. 2). This signifies that a moderate and severe vasospasm in the MIP images of CTA has a high probability of also being a moderate and severe vasospasm in DSA as well, respectively. In other words, it can be concluded that when diagnosed with moderate or severe vasospasms by the MIP images of CTA, it is unnecessary to check DSA for a more accurate diagnosis. However, by correlation with clinical factors, endovascular treatment can be performed simultaneously with DSA as necessary. When comparing with a distinction between the central vessel segment (A1, M1, P1) and the peripheral segment (A2, M2, P2), both readers confirmed higher accuracy of the central vessel segment compared with that of the peripheral vessel segment. From these results it is considered that the central vessel segment is more advantageous than the peripheral vessel segment when measuring the vasospasm.
This study is distinct from other studies for the following reasons. First, this study compares the four group images starting from the group without cerebral vasospasm up to the severe vasospasm group. Second, this research compares the MIP images of CTA and DSA images when the SAH diagnosis is initially given to the MIP images of CTA and DSA images one week later. Third, this study performs a comparison based on the detailed categorization of A1, A2, M1, M2, P1, P2 without knowing whether the patients were normal or diagnosed with vasospasm. And, because of separate analysis on the MIP images of CTA and DSA with two-week duration, the bias was reduced. Fourth, a clear distinction between spasm and hypoplastic vessel was possible because both initial and one week examinations were checked for the CTA and DSA.
However this study also had a few limitations. First, although the period in between was short, the time between duration for the MIP images of CTA and DSA in the follow-up image test was an average of 15.7 hours. Because cerebral vasospasm is a dynamic process, there was a limitation in that there was a possibility of change even within a short period of time. Second, because this was a study only on the diameter of the vessel, there is a limitation that the results might not be identical to the clinical results. Therefore, there is a need for future research in the area of modalities such as perfusion CT. Third, the patients included in the research protocol were in relatively good clinical states. In other words there was a high probability of CV in patients who were not in good clinical states, resulting in the limitation that these patients were excluded from the study. Fourth, the low power of statistics due to the small study population was a major limitation of our study.