The fuzzy MAD stroke conjecture, using Fuzzy C Means to classify multimodal apparent diffusion for ischemic stroke...

Area of Science:

• Neuroimaging
• Radiology
• Biomedical Engineering

Background:

• Current acute ischemic stroke imaging classifies lesions into core and penumbra, which may be too rigid.
• Advancements suggest a need for more detailed, patient-specific approaches to stroke assessment.
• This study addresses the limitations of current radiological practice in discerning stroke lesion variability.

Purpose of the Study:

• To demonstrate the noninvasive capability of providing insights into the in vivo stroke lesion cascade.
• To develop a more fine-grained depiction of acute focal ischemic stroke.
• To offer a patient-tailored approach to stroke imaging.

Main Methods:

• Retrospective application of a multimodal apparent diffusion (MAD) method.
• Utilized multi-b-value diffusion-weighted imaging (DWI) up to b=10,000 s/mm² in 34 acute ischemic stroke patients.
• Employed Fuzzy C Means clustering for MAD parameter analysis.

Main Results:

• Identified 18 normal-appearing tissue (NAT) clusters and 14 potential ischemic lesion types.
• Provided insights into lesion progression variability and aggressiveness.
• Achieved 92% autonomous identification of previously identified lesions, with 87% rated as efficacious or better than radiologists.

Conclusions:

• Findings suggest the MAD method can differentiate various tissue and lesion types in acute ischemic stroke.
• Highlights potential clinical imaging considerations, including white matter fluid accumulation and blood-brain barrier compromise.
• Suggests that the b=1000 core may not represent dead tissue and explains DWI (pseudo) normalization.