Quantitative imaging biomarker for stroke - collateral score


The treatment outcome of a patient with the ischemic stroke is affected by many factors. One of the potential factors is the collateral score. The collateral score is being used by the clinician to determine in the early stage, whether the patient may benefit from intervention. However, the collateral score is a 4 score system and it is subjectively evaluated by the radiologist. The conventional collateral evaluation systems compare the differences of middle cerebral arteries (MCA) between left and right hemisphere. In addition to MCA, the anterior cerebral artery (ACA) and posterior cerebral artery might also be occluded.

Our focus is on the quantitative image analysis. We would like to develop quantitative imaging biomarkers for collateral scores (and other vessel-based image-derived markers that are relevant to treatment decision making). For this purpose, the whole vessel tree is essential. We want to employ deep learning methods to obtain the vessel tree, followed by a post-processing wherein the position of the occlusion is computed.


Project number


Main applicant

Ir. J. Su

Affiliated with

Technische Universiteit Delft, Faculteit Werktuigbouwkunde, Maritieme Techniek en Technische Materiaalws. (3ME), Maritieme en Transport Technologie


24/01/2018 to 24/01/2019