Introduction
Idiopathic sudden sensorineural hearing loss is one of the most common causes of hearing loss, carrying an uncertain prognosis upon initial presentation. Intratympanic steroid injection is widely applied as either initial or salvage treatment for idiopathic sudden sensorineural hearing loss. Nevertheless, no standard protocol for intratympanic steroid injection has been established.
We retrospectively reviewed more than 800 patients from two tertiary referral centers who fulfilled the diagnosis criteria of idiopathic sudden sensorineural hearing loss. We applied several machine learning models for predicting hearing outcome. The models were validated by a train-test split of 4:1. Random forest classifier achieved the highest area under the ROC curve of 0.836 with 16 selected features via feature selection. Personalized treatment planning including number of injections and treatment frequency can be achieved by implementing the model. The model enabled individualized prediction for prognosis, thereby facilitated personalized treatment for
patients with idiopathic sudden sensorineural hearing loss.
Features / strengths
The artificial intelligence analysis unit trains an AI model, and uses the AI model to rank the importance of sudden sensorineural hearing loss (SSNHL) characteristics in patient information or to assess post-treatment, generating predictive results. The processing unit then creates a predictive analysis report based on the received prediction results, which are visualized and explained in charts, enabling the rapid formulation of personalized treatment strategies.
Specification in detail
研發中
Seek of partners for business cooperation