An explainable machine learning model for prognosis prediction in sudden sensorineural hearing loss under integrated therapy
-
By
-
Xiaoxiao Ye
-
Yuxin Deng
-
Binbin Xiong
-
Min Chen
-
Gang Chen
-
Chen Huang
-
June 1, 2026
-
Clinical Scorecard: A Transparent Machine Learning Approach for Predicting Outcomes in Sudden Sensorineural Hearing Loss During Combined Treatment
At a Glance
| Category | Detail |
|---|
| Condition | Sudden Sensorineural Hearing Loss (SSNHL) |
| Key Mechanisms | Qi stagnation and blood stasis syndrome; inner ear microcirculation disturbance |
| Target Population | Patients with unilateral SSNHL receiving integrated therapy |
| Care Setting | Otolaryngology, specifically at Zhuhai Hospital of Integrated Traditional Chinese and Western Medicine |
Key Highlights
- XGBoost model achieved AUC of 0.718 in predicting outcomes for SSNHL patients.
- Key prognostic predictors identified: APTT, disease duration, platelet count, and total protein.
- Critical intervention window within 10 days of symptom onset for optimal recovery.
Guideline-Based Recommendations
Diagnosis
- Diagnosis of SSNHL based on sudden hearing loss of ≥20 dB HL affecting at least 2 consecutive frequencies within 72 hours.
Management
- Integrated therapy combining conventional Western medicine and Traditional Chinese Medicine.
Monitoring & Follow-up
- Follow-up for at least 12 weeks post-treatment.
Risks
- Increased risk of treatment failure associated with APTT values <25 s and PLT counts >300 × 10^9/L.
Patient & Prescribing Data
Patients diagnosed with unilateral SSNHL of Qi stagnation and blood stasis syndrome.
Optimal recovery associated with PLT counts of 200–250 × 10^9/L and total protein levels of 65–75 g/L.
Clinical Best Practices
- Utilize explainable machine learning models for prognostic prediction.
- Implement early risk stratification based on identified key predictors.
Related Resources & Content
