Risk Factors for Unilateral Trigeminal Neuralgia Based on Machine Learning

Area of Science:

• Neurology
• Medical Imaging
• Machine Learning

Background:

• Neurovascular compression (NVC) is the primary cause of classical trigeminal neuralgia (CTN).
• A subset of idiopathic trigeminal neuralgia (ITN) may also stem from NVC (ITN-nvc).
• Understanding risk factors for unilateral CTN or ITN-nvc with bilateral NVC is crucial.

Purpose of the Study:

• To explore risk factors for unilateral CTN or ITN-nvc (UC-ITN) exhibiting bilateral NVC.
• To utilize machine learning (ML) for identifying these risk factors.
• To differentiate UC-ITN based on the presence of NVC on the unaffected side.

Main Methods:

• Prospective recruitment of 89 patients with UC-ITN.
• Magnetic resonance imaging (MRI) scans to assess bilateral trigeminal nerve cisternal segments.
• Manual delineation of nerves, identification of offending vessels (Ofv), and extraction of textural features.
• Application of ML for dimensionality reduction, feature selection, model construction, and evaluation.

Main Results:

• Distinct textural features (rad_score) showed significant differences between affected and unaffected sides (p < 0.05).
• Four and six high-weight textural features were identified for patients without and with NVC on the unaffected side, respectively.
• A nomogram model demonstrated optimal diagnostic power with an AUC of 0.76 (training) and 0.77 (validation).
• Offending vessels (Ofv) and rad_score were identified as significant risk factors for UC-ITN.

Conclusions:

• Trigeminal nerve texture features and Ofv are significant risk factors for UC-ITN, beyond NVC alone.
• These findings contribute to a deeper understanding of the microscopic etiology of UC-ITN.
• The study provides a foundation for further research into the underlying causes of trigeminal neuralgia.