Predicting Anticoagulation Need for Otogenic Intracranial Sinus Thrombosis: A Machine Learning Approach

Area of Science:

• Neurology
• Otolaryngology
• Medical Informatics

Background:

• The role of anticoagulation (AC) in managing otogenic cerebral venous sinus thrombosis (OCVST) is debated.
• Existing literature lacks clear guidelines on AC use in OCVST cases.
• OCVST is a serious condition often associated with complications like cholesteatoma and intracranial abscess.

Purpose of the Study:

• To define optimal anticoagulation (AC) strategies for otogenic cerebral venous sinus thrombosis (OCVST).
• To identify patient subgroups that benefit most from AC in OCVST management.
• To leverage machine learning for predictive analysis in OCVST treatment decisions.

Main Methods:

• Systematic literature search of MEDLINE, EMBASE, and Cochrane Library up to February 2019.
• Risk of bias assessment using the Methodological Index for Non-Randomized Studies (MINORS) scale.
• Random effects meta-regression and random forest machine learning analysis across 16 moderator variables comparing AC and non-anticoagulated (NAC) cohorts.

Main Results:

• 92% of all treated patients were neurologically asymptomatic at long-term follow-up.
• AC was associated with lower rates of persistent neurological symptoms (4% vs. 14% for NAC).
• AC showed higher recanalization rates (81% vs. 63% for NAC) and was predicted by intracranial abscess and cholesteatoma comorbidities.

Conclusions:

• This study pioneers machine learning application in OCVST management.
• Findings support therapeutic AC for OCVST complicated by thrombophilia, intracranial abscess, and cholesteatoma.
• Combined AC and surgery may benefit patients with intracranial abscess and cholesteatoma; NAC and surgery may suffice for cholesteatoma alone.