Radiology: MDCT shows promise as first-line stroke exam
 | |
| CT images in a patient with left hemiparesis. Image source: Michael H. Lev, MD, and Angelos Konstas, MD, Massachusetts General Hospital, Boston. |
Standard etiologic workup of patients admitted for acute ischemic stroke consists of a multi-modality imaging protocol, including ultrasound, MR angiography or CT angiography, transthoracic echocardiography (TTE) and transesophageal echocardiography (TEE). Common causes of ischemic stroke identified by these imaging techniques include extra- or intracranial atheroma, cardioembolic sources, microvascular disease and aortic arch atheroma. The authors evaluated the accuracy of single-session MDCT for detecting the causes of acute ischemic stroke compared with these established imaging methods.
Forty-six patients were imaged using "a one-time multidetector CT examination of the heart, aorta and extra- and intracranial vessels and a standard protocol involving combined MR angiography of the extra- and intracranial vessels, duplex ultrasound of the extracranial vessels and TTE and/or TEE of the heart and aorta," according to Loic Boussel, MD, PhD, of Louis Pradel Hospital in Bron, France, and co-authors.
Seven of the 46 patients imaged were ineligible for TEE due to swallowing abnormalities. The mean patient age was 63 years, and patients presented with an average National Institutes of Health stroke score of 8.4.
The final etiologic workup of participants, determined using the conventional modalities, attributed strokes to cardiac sources in 20 patients, major arterial atheroma in nine cases, multiple sources in four patients and cryptogenic sources in 13 patients.
CT was used to correctly classify 38 patients, or 83 percent of the sample population. CT misidentified the causes of stroke by underestimating cardiac sources and multiple sources in five and three patients, respectively, while overestimating major arterial atheroma and cryptogenic sources by three cases each.
MDCT achieved overall specificity and sensitivity of 72 and 95 percent, respectively. Relative to TTE and TEE's identification of cardiac sources of emboli, CT detected one case of intracardiac thrombus missed by the two modalities while missing seven other embolic sources detected by TTE or TEE.
The authors emphasized the positive predictive value of MDCT for stroke workup, with MDCT having "facilitated correct etiologic classification in up to 83 percent of the patients." Boussel and colleagues cautioned, however, that "[n]egative multidetector CT results should be confirmed with TEE and MR imaging."
The authors elaborated several limitations to the findings. Because all CT findings relied on the consensus of several experienced reviewers, the authors acknowledged that "the accuracy values are probably overestimated." The study design thus prevented the measurement of inter- and intraobserver variability. Boussel and colleagues also noted that even though some large-detector CT scanners might enable a more optimal imaging strategy, the research protocols resulted in "high radiation exposure," with an average dose measuring 20.8 mSv.
Despite these concerns, the authors called for larger randomized studies to investigate the cost-effectiveness of MDCT, saying that CT "may be the first-line imaging modality for identifying acute ischemic stroke causes."