Optimizing Ischemic Stroke Treatment: Brain Region-Based Analysis of DSA via CTA-DSA Registration

The primary objective of this thesis is to investigate whether early post-endovascular thrombectomy (EVT) perfusion in critical brain regions can predict short-term neurological recovery, as measured by NIHSS scores at 24 hours. While overall reperfusion, quantified by TICI scores, has been established as a predictor of long-term functional outcomes, a more targeted analysis of regional perfusion may provide deeper insights into early neurological improvement. To achieve this, we employ a multi-step methodology, including: (1) Supervised registration of 3D CTA to 2D DSA, (2) Unsupervised registration of an ATLAS to 3D CTA, and (3) Unsupervised registration of both onto DSA, aligned through the supervised transformation. Perfusion analysis is conducted by computing TDT perfusion masks and comparing them against anatomical prediction maps to derive regional perfusion percentages. Our study focuses on assessing perfusion in functionally critical regions related to motor control, memory, speech and auditory processing, as well as tactile and sensory-motor integration. By correlating localized perfusion patterns with follow-up NIHSS scores, we aim to determine whether specific regions play a key role in early post-stroke recovery. Establishing this relationship could refine stroke management strategies, enabling personalized rehabilitation planning, optimized secondary prevention approaches, and improved patient monitoring. Ultimately, these findings may contribute to more targeted and effective post-stroke care.