Predicting Mechanical Thrombectomy Outcome and Time Limit through ADC Value Analysis: A Comprehensive Clinical and...

Area of Science:

• Neurology
• Radiology
• Biomedical Engineering

Background:

• Predicting outcomes in acute ischemic stroke (AIS) patients after mechanical thrombectomy (MT) is complex.
• Apparent diffusion coefficient (ADC) analysis offers potential insights into tissue viability.

Purpose of the Study:

• To evaluate machine learning (ML) using detailed ADC analysis for predicting patient outcomes after MT.
• To simulate the time limit for MT in AIS based on ADC parameters.

Main Methods:

• Utilized ADC analysis with varying thresholds to extract quantitative imaging features.
• Employed an extra tree classifier model on 75 AIS patients with complete reperfusion.
• Simulated MT time limits using prediction scores and reperfusion times.

Main Results:

• The extra tree classifier achieved high accuracy (93.3%) and AUC (0.833) in outcome prediction.
• Simulation data showed decreased prediction scores with longer reperfusion times.
• Younger patients demonstrated a longer simulated time limit for MT.

Conclusions:

• ML combined with detailed ADC analysis is effective for predicting AIS patient outcomes post-MT.
• This approach can simulate the time tolerance for MT, aiding clinical decision-making.