Real-time planar angiographic cerebral perfusion imaging

Area of Science:

• Medical Imaging
• Interventional Radiology
• Neurology

Background:

• Digital Subtraction Angiography (DSA) is crucial for minimally invasive procedures.
• Enhancing DSA with tomographic methods like 3D reconstruction and CT perfusion imaging improves diagnostic value.

Purpose of the Study:

• To demonstrate a hardware and software setup for expanding DSA capabilities to functional perfusion imaging.
• To assess planar cerebral perfusion in real-time during angiographic procedures.

Main Methods:

• Utilized an HDMI video splitter and frame-grabber to duplicate DSA video output.
• Applied numerical approximation and neural network inference for perfusion parameter map generation.
• Compared completion times of various perfusion estimation methods.

Main Results:

• Identified two perfusion estimation methods (SVD-based deconvolution, numerical curve feature estimation) yielding results in 1-2 seconds.
• Optimized conventional curve fitting to reduce calculation time from five minutes to 30 seconds.
• Achieved clinically feasible real-time cerebral perfusion evaluation.

Conclusions:

• Direct access to image data outside the angiography suite enables real-time cerebral perfusion evaluation.
• On-the-fly analysis of angiograms is now possible in clinical settings.
• Facilitates future clinical studies on angiographic perfusion.