Predicting Cerebral Aneurysm Recurrence after Coil Embolization: A Novel Deep Learning Approach Using Time-of-flight...

Area of Science:

• Neurosurgery
• Radiology
• Artificial Intelligence in Medicine

Background:

• Cerebral aneurysm recurrence after coil embolization presents a significant clinical challenge.
• Accurate prediction of recurrence is crucial for effective patient management and treatment planning.

Purpose of the Study:

• To develop and evaluate a novel approach combining machine learning and deep learning for predicting cerebral aneurysm recurrence.
• To assess the utility of time-of-flight magnetic resonance angiography (TOF-MRA) imaging data in conjunction with clinical information for recurrence prediction.

Main Methods:

• A retrospective multicenter study involving 154 patients with coil-embolized unruptured cerebral aneurysms.
• Development of three predictive models: logistic regression, a neural network using clinical data, and a combined deep learning model incorporating clinical and 3D TOF-MRA data.
• Evaluation using leave-one-out cross-validation, assessing area under the receiver operating characteristic curve (AUC), accuracy, sensitivity, and specificity.

Main Results:

• Recurrence was observed in 46 (29.9%) of the cases.
• The combined deep learning model using both pre- and post-operative TOF-MRA images achieved the highest performance (AUC: 0.822, sensitivity: 0.848).
• The model using only pre-operative images showed good accuracy (0.766) and specificity (0.833) but a lower AUC.

Conclusions:

• Integrating TOF-MRA data into AI-driven predictive models significantly enhances the assessment of recurrence risk after cerebral aneurysm coil embolization.
• The combined model, especially with post-operative imaging, offers potential for precise, individualized risk prediction.
• This approach may lead to improved patient management strategies for cerebral aneurysms.