To construct and validate an explainable machine learning model to predict prognostic outcomes of SSNHL patients receiving integrated therapy.
Key Findings:
The XGBoost model showed superior performance in the validation set (AUC: 0.718; 95% CI: 0.590–0.846).
Key predictors identified: activated partial thromboplastin time (APTT), disease duration, platelet count (PLT), and total protein (TP).
Critical intervention window identified within 10 days of symptom onset.
Optimal recovery associated with PLT counts of 200–250 × 10^9/L and APTT values of 25–30 s.
TP exhibited a tri-phasic association with peak prognostic probability at 65–75 g/L.
Interpretation:
The study developed an explainable XGBoost-based model for prognostic prediction in SSNHL patients, integrating machine learning with TCM-informed therapy.
Limitations:
Lack of external validation.
Results should be interpreted with caution.
Conclusion:
Further multicenter prospective studies are crucial to confirm the model’s generalizability and validate the identified biological predictors.
