Identifying perfusion deficits on CT perfusion images using temporal similarity perfusion (TSP) mapping

Area of Science:

• Neuroradiology
• Medical Imaging Analysis
• Cerebrovascular Disease Research

Background:

• Cerebral Perfusion CT (CTP) imaging is crucial for acute stroke assessment.
• Deconvolution methods for CTP map generation can be affected by signal transit delays.
• Developing transit-delay-insensitive CTP analysis is essential for accurate perfusion deficit detection.

Purpose of the Study:

• To introduce Temporal Similarity Perfusion (TSP) analysis as a novel method for CTP data.
• To evaluate TSP's ability to generate perfusion maps less dependent on transit times.
• To assess TSP's sensitivity and interpretability in detecting perfusion deficits in acute stroke patients.

Main Methods:

• CTP data from 65 acute stroke patients within 9 hours of symptom onset were analyzed.
• Temporal Similarity Perfusion (TSP) analysis was performed, comparing pixel time-series curves to whole-brain averages.
• TSP maps were compared qualitatively and quantitatively to deconvolution-derived maps (MTT, CBV, TTP) and follow-up CT findings.

Main Results:

• TSP analysis identified perfusion deficits in 84% of patients, comparable to deconvolution methods.
• TSP map interpretation showed higher intra-rater agreement for inexperienced readers compared to CBV/MTT maps.
• TSP showed a strong correlation with final infarct volume (R²=0.63), second only to CBV (R²=0.77).

Conclusions:

• Temporal Similarity Perfusion (TSP) mapping provides a viable alternative for CTP analysis.
• TSP maps demonstrate reduced dependence on transit delays and improved interpretability for less experienced clinicians.
• Further research is warranted to fully elucidate the clinical utility of TSP in stroke management.