Absence of susceptibility vessel sign with cancer- associated hypercoagulability-related stroke

BACKGROUND AND PURPOSE: Susceptibility vessel sign (SVS), a hypointense signal on MR T2-weighted gradient-recalled echo images, is associated with erythrocyte-predominant thrombi, which are often present in cardioembolism (CE). In contrast, cancer-associated hypercoagulability (CAH)-related stroke, which is presumably caused by fibrin-predominant thrombi, is associated with the absence of SVS. We hypothesized that the prevalence of SVS may be of help in distinguishing CAH-related stroke from CE. This study attempted to validate this hypothesis and investigated the usefulness of SVS in differentiating CAH-related stroke from CE.

MATERIALS AND METHODS: We retrospectively studied both CAH-related stroke patients (CAH group) and CE patients (CE group), who had major cerebral artery occlusion on MR angiography that was performed within 6 hours of stroke onset. All patients visited our department from 2015 to 2021. CAH-related stroke was defined as 1) complication of active cancer, 2) pre-treatment D-dimer value >3 μg/mL, 3) multiple vascular territories infarctions, and 4) lack of any other specifically identified causes of stroke. We compared SVS positivity rates within each group. Multivariable logistic regression analysis was used to assess the association between the absence of SVS and CAH-related stroke.

RESULTS: Of 691 patients with CAH-related stroke or CE, major cerebral artery occlusion was observed in 10 patients in the CAH group and 198 patients in the CE group. The absence of SVS was identified in 55 of 208 patients and was significantly more frequent in the CAH versus the CE group (90% versus 24%, p < 0.05). For predicting CAH-related stroke, absence of SVS demonstrated a sensitivity of 90% (95% confidence interval [95%CI] 59-99), specificity of 78% (95%CI 71-83), positive predictive value of 18 (95%CI 10- 31), negative predictive value of 99% (95%CI 96-99), and a likelihood ratio of 4.06. Multivariable logistic regression analysis revealed that the absent of SVS was independently associated with CAH-related stroke (odds ratio 43, 95% [CI] 6.8-863; p < 0.01).

CONCLUSIONS: The absence of SVS was more frequent in CAH-related stroke versus that found for CE. These findings could potentially be helpful for clinical management and differentiating between CE and CAH-related stroke.

ABBREVIATIONS: CAH, cancer-associated hypercoagulability; CE, cardioembolism; SVS, susceptibility vessel sign; GRE, gradient recalled echo.