CTP-Free Method for Automated Lesion Water Uptake in Acute Ischemic Stroke.

Area of Science:

• Neuroimaging
• Radiology
• Stroke Imaging

Background:

• Net water uptake (NWU) in ischemic stroke correlates with outcomes and lesion age.
• Traditional NWU measurement relies on CT perfusion (CTP), limiting its clinical use.
• Developing an automated NWU method using NCCT and CTA is crucial for wider applicability.

Purpose of the Study:

• To develop and validate an automated method for measuring NWU using only NCCT and CTA.
• To compare the performance of the automated CTA-NCCT approach with the traditional CTP-NCCT method.

Main Methods:

• Deep learning was used to automatically segment infarct core and hypoperfused areas from NCCT and CTA images in 90 stroke patients.
• NWU was calculated as the density difference between affected and contralateral regions.
• Agreement (ICC, Bland-Altman) and accuracy (ROC analysis) were assessed against CTP-NCCT-based NWU.

Main Results:

• The automated CTA-NCCT approach showed good agreement with CTP-NCCT for core NWU (ICC 0.81) and excellent agreement for hypoperfused area NWU (ICC 0.93).
• Both methods accurately identified patients within 4.5 hours of stroke onset based on core NWU (AUC 0.87).
• Hypoperfused area NWU did not significantly differentiate early vs. late treatment windows with either method.

Conclusions:

• Automated NWU measurement using CTA and NCCT is a viable alternative to CTP.
• This method demonstrates good agreement and comparable accuracy to traditional approaches.
• External validation is recommended to confirm the generalizability of these findings.