IMPROVING THE DIAGNOSIS OF ACUTE STROKE

Neurological Disorders And Stroke

Investigators

Linked publications & trials

Abstract

Patients diagnosed with acute ischemic stroke are offered interventions of proven clinical benefit (e.g., thrombolysis, in-patient stroke units, secondary prevention medicines), therefore greater diagnostic accuracy would lead to increased utilization of these treatments and improve the overall outcomes of patients with stroke compared to their misdiagnosed counterparts. Several years ago we initiated an investigation, the primary objective of which was to compare the diagnostic information contained in non-contrast CT with that of non-contrast MRI in the full range of patients with stroke-like symptoms presenting to a community hospital emergency department (Suburban Hospital, Bethesda, MD) to test the hypothesis that magnetic resonance imaging (MRI) was superior to computed tomography (CT) in acute stroke diagnosis. Free from the selection bias of other studies on the subject, our study was designed to give the definitive comparison of MRI to CT in this real world sample and scenario of the emergency diagnosis of potential stroke. In the past year we completed and published the results of the study in The Lancet. MRI is increasingly used for the diagnosis of acute stroke, but its superiority to CT in the emergency setting has not been established. This was a prospective blind comparison of non-contrast CT and MRI (with diffusion-weighted DWI and susceptibility-weighted images) in a consecutive series of patients referred for emergency evaluation of suspected acute stroke. Patients were included without regard to time from onset, symptom severity or ultimate clinical diagnosis. The scans were independently interpreted by four experts, blinded to clinical information, MRI-CT pairings, and follow-up imaging. The sample consisted of 356 patients, 217 of whom had a final clinical diagnosis of acute stroke. MRI was superior to CT for the detection of any acute stroke (ischemic or hemorrhagic), acute ischemic stroke, and chronic hemorrhage (p <0.0001, for all comparisons). MRI was comparable to CT for the detection of acute intracranial hemorrhage. MRI detected acute ischemic stroke in 164 of 356 patients, 46% (95% CI 41-51%); CT in 35 of 356 patients, 10% (7-14%). In the subset of patients scanned within 3 hours from symptom onset, MRI detected acute ischemic stroke in 41 of 90, 46% (95% CI 35-56%); CT in 6 of 90, 7% (3-14%). Relative to the final clinical diagnosis, emergency brain imaging had a sensitivity of 83% (181 of 217; 95% CI 78-88%) for MRI and 26% (56 of 217; 95% CI 20 - 32%) for CT in the diagnosis of any acute stroke. The MRI superiority was greatest for the type of strokes most difficult to accurately diagnose on purely clinical criteria the milder, smaller strokes. The superiority of MRI for the detection of acute ischemia coupled with its ability to detect acute and chronic hemorrhage make it the preferred test for accurate diagnosis of patients with suspected acute stroke. Because the patient sample studied covered the spectrum of disease likely to be encountered when evaluating emergent cases of suspected stroke, these results are directly applicable to clinical practice. [unreadable] [unreadable] In the above-mentioned study we identified clinical circumstances in which even MRI gave false negative results on DWI as early, milder and brainstem strokes, i.e., small lesions with less contrast. To further improve the diagnostic yield of DWI we performed an analysis of DWI sensitivity using a DWI sequence of higher spatial resolution. Ischemic lesion conspicuity on routine diffusion-weighted imaging (DWI 7 mm thick, 3 directions, 30 seconds acquisition) was compared with an improved sequence (high-resolution DWI 2.5 mm thick DWI-HR, 13 directions, 256 seconds, acquisition) having increased spatial resolution and signal to noise and decreased eddy current artifact in patients with acute ischemic stroke. We hypothesized that improving image quality by increasing spatial resolution and signal to noise ratio will allow for a more accurate lesion detection, and an improved lesion pattern characterization. We studied a total of 69 patients imaged within 6 hours of symptom onset, with both standard diffusion weighted imaging DWI (7mm thick, 3 gradient directions, 28s) and high resolution DWI (DWI-HR) (2.5 mm thick, 13 gradient directions, 246s). A greater number of lesions (194 vs 98) lesions were identified on DWI-HR and DWI, respectively (p<0.05). Among the lesions identified on DWI-HR, 101 (50%) were not seen on DWI, whereas only 5 lesions were unique to DWI. Lesions unique to DWI-HR were predominantly found in the cortical gray matter. Overall, the level of confidence in lesion identification rose from an average of 82% on DWI to 91% on DWI-HR. A total of 19 patients (28%) had lesions on DWI-HR in vascular territories that were not affected on DWI; 9 patients (13%) had multiple vascular territories affected while a single vascular territory was involved on DWI, and 10 patients (15%) had lesions on DWI-HR while DWI was considered negative. From this study we could conclude that increased lesion conspicuity resulted from higher resolution DWI, however an unexpected result was the identification of lesions previously missed by DWI. Those lesions were primarily located in the cortical gray matter and white matter, affected vascular territories that were seen as unaffected on DWI and caused a change in lesion pattern in a subset of patients. Higher resolution DWI may hence be a powerful tool in helping institute the most appropriate management in acute stroke patients based on the pattern and distribution of acute ischemic injury.[unreadable] [unreadable] Having discovered several clinically important features and biomarkers of ischemic stroke pathology by performing exploratory studies of repeated scans in the acute period using advanced MRI, we have adapted this strategy to the study acute intracerebral hemorrhage. In the past year, as part of our Evaluation, Pathogenesis, and Treatment of Patients with or at Risk for Cerebrovascular Disease (Protocol 01-N-0007), we began a study at our two acute stroke translational research sites (Suburban Hospital and Washington Hospital Center) of serial MRI scans in patients with acute supratentorial intracerebral hematoma at the time of presentation, and perform follow-up scans at 6 hours, 24 hours, 72 hours, and 5-7 days later. These scan time points are accompanied by collection of stroke clinical severity scales and blood biomarker samples. This ongoing study will be of several years duration and ongoing data analyses are expected to generate testable hypotheses regarding the evolution and treatment of primary brain hemorrhage.