The hyperdense middle cerebral artery (MCA) sign, as illustrated in Figure 1, has been observed to be associated with embolic large vessel occlusion (LVO) (9, 10). The appropriate threshold for detecting thrombus has been established as a Hounsfield Unit (HU) value of 51, which exceeds the standard value of 45 HU (11). Intracranial arterial wall calcification is a hallmark of atherosclerosis and has been correlated with lower rates of reperfusion following thrombectomy (12). A substudy within the MR CLEAN trial enrolled 344 patients, with 156 in the endovascular treatment (EVT) group and 188 in the control group, excluding individuals who were difficult to evaluate due to 3-mm ultraslice imaging or body motion. The results indicated significant differences in reperfusion rates and outcomes based on the type of intracranial carotid artery calcification. Patients with medial calcification, i.e., calcification within the occluded vessel, had better outcomes in the EVT group compared to the control group [adjusted common odds ratio (OR), 2.32; 95% confidence interval (CI), 1.23–4.39], while endovascular therapy did not have a significant impact on patients with intimal calcification (adjusted common OR, 0.82; 95% CI, 0.40–1.68) (12). In addition, a study on quantitative HU values measured the HU of intra-arterial radiation in 102 consecutive Chinese stroke patients who underwent multiphase CT angiography and EVT within 6 h of onset and examined the HU distal/proximal ratio to predict emboli The optimal cutoff was an HU ratio < 0.6 measured at 2 mm from the embolization site (area under the maximum curve = 0.87, sensitivity 96%, specificity 81%) (13).

The concept of SVS has been classically defined as “a manifestation of low signal intensity in occluded thrombi on T2*-weighted Gradient Echo (GRE) images, caused by the magnetic susceptibility effect of deoxyhemoglobin in red blood cells” (14). The diagnostic value of SVS in ischemic stroke is widely acknowledged, with a prevalence of 77.5% in cases of cardioembolism (15). This is particularly evident in thrombi with high concentrations of red blood cells (known as “red thrombi”) (16). Moreover, the presence of GRE-SVS has been associated with cardioembolism and spontaneous recanalization of occluded vessels (17, 18). The diameter of SVS has been independently linked to the likelihood of cardioembolism (adjusted odds ratio [OR], 1.97; 95% confidence interval [CI], 1.34–2.90; p < 0.01). However, it is important to note that a diagnosis of SVS does not necessarily indicate cardioembolism, as 25.5% of patients without cardioembolism also exhibit SVS (14). In cases of the middle cerebral artery (MCA) diagnosed via cerebral angiography, the sensitivity of SVS was higher than that of the hyperdense MCA sign NCCT (82% vs. 54%) (14). The concept of “2-layer SVS” (as shown in Figure 2) on 3-T MR T2*-weighted GRE has been reported in Japan, where 47.7% of 132 patients (72 men, mean age 74.5 years) were diagnosed with cardioembolism. The sensitivity of SVS for cardioembolism and large-artery atherosclerosis was not statistically significant (74.6% vs. 58.0%), but the sensitivity of 2-layer SVS alone was found to be significantly higher for cardioembolism (42.9%) than for large-artery atherosclerosis (2.9%; p < 0.001) (19). The specificity of 2-layer SVS for cardioembolism and the diagnostic ratios were 97.1 and 25.1%, respectively (42.0 and 2.1% for SVS). These findings may be attributed to the magnetic heterogeneity within the thrombus, suggesting a significant correlation between 2-layer SVS and higher thrombus weight and red blood cell components (19–21).

The EPITHET-DEFUSE study disclosed a correlation between cerebral perfusion imaging attributes and atrial fibrillation. Of the 124 patients who underwent perfusion imaging and were enrolled in DEFUSE (22) or EPITHET (23), 28 patients were designated as the “definite AF group,” having been identified as having atrial fibrillation (AF) upon admission, while the remaining 96 were classified as the “NO AF group.” The comparison of perfusion imaging profiles revealed that the definite AF group displayed elevated profiles relative to the NO AF group, as indicated by higher time to maximum concentration (T_max) values (T_max > 4 s: 136 mL vs. 81 mL, p < 0.01; T_max > 6 s: 83 mL vs. 50 mL, p < 0.01; T_max > 8 s: 48 mL vs. 29 mL, p = 0.02) (24). The DEFUSE 2 (25) trial quantified hypoperfusion intensity rate (HIR) as T_max > 10 s divided by T_max > 6 s, with a median HIR value of 0.4 being linked to the extent of collateral bleeding (26). In a recent single-center observational study, HIR ≤0.22 (OR, 22.5; 95% CI, 2.9–177.0; p = 0.003) and cerebral blood volume index ≥0.9 (OR, 75.7; 95% CI, 5.8–994.0; p < 0.001) were found to be associated with ICAS-related LVO and to potentially predict underlying ICAS prior to EVT (27); (Figure 3). Moreover, cortical collateral vessels are also helpful in the preoperative differentiation of stroke etiology. Cortical vessels are prominent on prominent cortical vessels on susceptibility-weighted imaging (PCV-SWI) in 30.3% of ICAS-related LVO patients, whereas PCV-SWI is positive in only 13.4% of ICAS-related LVO patients, and PCV-SWI helps diagnose stroke etiology (28). Previous studies have also reported that prominent cortical and/or medullary veins on SWI can indicate to neuro-interventionists that the cause of LVO is more likely cardioembolism rather than ICAS-related LVO (29). A summary of previous reports of useful PWI diagnostic markers in the diagnosis of ICAS is shown in Table 1.

“Fixed focal stenosis of substantial magnitude” has been proffered as a diagnostic hallmark of ICAS-related LVO (Figure 4) (34). This criterion is articulated as “a concentration of significant stenosis circumscribed to the site of occlusion,” as observed on postoperative or definitive imaging of the MCA and was initially a term employed to characterize ICAS-related LVO. Although few systematic investigations have focused on fixed focal stenosis, a Korean registry utilizing stent retrievers noted an incidence of approximately 15–20% (35).

A multicenter observational study from Korea determined that occlusions of a truncal type were correlated with a lack of responsiveness to stent retrievers and were the fundamental cause of strokes (36). The subjects were patients undergoing MT for intracranial occlusions of the internal carotid artery, MCA, proximal MCA, intracranial vertebral artery, or basilar cerebral artery. The occlusions were classified as either branching-site occlusion or truncal-type occlusion. Branching-site occlusion was defined as at least one of the following three conditions (Figure 5); (1) anterior communicating artery collateral flow that could not proceed to the contralateral ICA or MCA because it involved the internal cerebral artery bifurcation site (T occlusion); (2) direct visualization of a Y-or T-shaped filling defect involving a bifurcation site (Y-or T-shaped clot); and (3) another branch could not be visualized or was only partially visualized when the retriever was deployed to a branch across the occlusion site. Truncal-type occlusion, on the other hand, was defined as all branches and bifurcations visible beyond the occluded vessel, including those observed at recanalization. After a comprehensive evaluation involving chest electrocardiogram, echocardiography, cardiac CT, and cervical vascular echocardiography, the patients were classified as having embolic or non-embolic LVO. Of the 259 patients (mean age 70.3 years; male/female ratio 132:127), 83.4% had embolic LVO. Multivariate analysis revealed that younger age, prior coronary artery disease, and truncal-type occlusion were independently linked to the absence of embolic LVO (OR, 9.07; 95% CI, 3.74–22.0). Furthermore, truncal-type occlusion was associated with a higher frequency of reocclusion and a longer time to recanalization during stent retriever treatment. In a subanalysis of this study, truncal-type occlusion was associated with 93% of ICAS-related LVO and 10% of embolic LVO (p < 0.01) (35), whereas branching-site occlusion was associated with 7% of ICAS-related LVO and 90% of embolic LVO. In a separate study among 115 LVO patients in China, truncal-type occlusion was present in 93% of ICAS-related LVO and 10% of embolic LVO, while branching-site occlusion was observed in 7% of ICAS-related LVO and 90% of embolic LVO, yielding a significant difference between the two LVO types (p < 0.01 for each). The area under the curve of ICAS-related LVO in truncal-type occlusion was 0.916, with the sensitivity of 92.86% and specificity of 90.41% (36). CT angiography can also assess truncal-type occlusion, and although it is not a direct predictor of pathogenesis, branching-site occlusion as determined by CT angiography has been reported to independently predict the success of recanalization with stent retrievers (OR, 8.20; 95% CI, 3.45–19.5) (37). Representative cases of truncal-type occlusion/branching-site occlusion are shown in Figure 6.

Jet-like appearance on cerebral angiography is characterized by a tapered end of the occluded vessel (Figure 7). In a Chinese observational study of 164 cases of LVO, 20.7% presented with this distinctive trait. Patients with the jet-like appearance were determined to be younger (mean age 68 years compared with 62.7 years) and had fewer severe symptoms (as indicated by a lower National Institutes of Health Stroke Scale [NIHSS] score of 16.6 compared with 12.4) than those without this feature. Multivariate logistic regression analysis revealed that a jet-like appearance was independently correlated with ICAS-related LVO (OR, 180.813; 95% CI, 17.966–1819.733; p < 0.001). The diagnostic performance of the jet-like appearance for identifying ICAS-related LVO was determined to have the sensitivity of 96%, specificity of 78%, and accuracy of 83% (38).

A Canadian study delineated two patient groups based on the occlusion location at the initial angiography and compared their demographic characteristics. Of 131 participants, 31 (23.6%) were classified as having a tapered presentation, while 100 (76.3%) were in the non-tapered group (Figure 8). The tapered group delivered a lower NIHSS score (10 vs. 16, with a significance level of p < 0.001), higher Alberta Stroke Program Early CT Score (9 vs. 7, with a significance level of p = 0.003), higher immediate reocclusion rate (26.7% vs. 8.2%, with a significance level of p = 0.025), and a lower rate of complete recanalization (45.2% vs. 71.0%, with a significance level of p = 0.028). The tapered group was also more likely to have LVO associated with ICAS (54.8% vs. 18.0%, with a significance level of p < 0.001) and to present with truncal-type occlusions (76.9% vs. 31.1%, with a significance level of p < 0.001) (39).

The benefits of identifying such occlusion margins through digital subtraction angiography are considerable, as they remain unaltered by therapeutic intervention and are thus useful in determining the causative mechanism of occlusion, even if complete recanalization of the affected vessel is not realized. Representative cases of “tapered” or “non-tapered” types are shown in Figure 8.