Can novel CT-and MR-based neuroimaging biomarkers further improve the etiological diagnosis of lobar intra-cerebral hemorrhage?

Abstract

Lobar hematomas represent around half of all supratentorial hemorrhages and have high mortality and morbidity. Their management depends on the underlying cause. Apart from local causes such as vascular malformation, which are rare and can usually be easily excluded thanks to imaging, the vast majority of lobar hematomas equally frequently result from either hypertensive arteriolopathy (HA) or cerebral amyloid angiopathy (CAA). Distinguishing between CAA and HA is important for prognostication (risk of recurrence nearly sevenfold higher in the former), for decision-making regarding, e.g., antithrombotic therapies (for other indications) and for clinical trials of new therapies. Currently, a non-invasive diagnosis of probable CAA can be made using the MR-based modified Boston criteria, which have excellent specificity but moderate sensitivity against histopathological reference, leading to the clinically largely irrelevant diagnosis of "possible CAA". Furthermore, the Boston criteria cannot be applied when both lobar and deep MRI hemorrhagic markers are present, a not uncommon situation. Here we propose to test whether new CT and MR-based imaging biomarkers, namely finger-like projections of the hematoma and adjacent subarachnoid hemorrhage on acute-stage CT or MRI, and remote punctate diffusion-weighted imaging ischemic lesions on acute or subacute-stage MRI, have the potential to improve the performance of the Boston criteria. Furthermore, we also propose to test whether clinical-radiological biomarkers may also allow a positive diagnosis of HA to be made in lobar hematomas, which, if feasible, would not only further reduce the prevalence of "possible CAA" but also permit a diagnosis of HA and/or CAA to be made in the presence of mixed deep and lobar MRI hemorrhagic markers.