International Session 6　　SNHL 1

Objective : To develop a multidimensional classification of Menière's disease(MD)integrating audio-vestibular tests and vertigo burden.Methods : We studied 101 unilateral MD patients undergoing pure tone audiometry(PTA) , vestibularevoked myogenic potentials(VEMPs) , video head impulse testing(vHIT) , and rotary chair testing(RCT) .Machine-learning models identified variables associated with DHI scores.Cluster analysis using PTA, number of affected vestibular organs (NAVO) , and vertigo attacks over 6 months(NVA)derived MD subtypes.Results : RCT showed the highest abnormal rate (~69． 31％)and was the strongest predictor of DHI, followed by anterior-canal vHIT and cervical VEMP.A GA-NN indicated that NAVO alone predicted DHI severity with accuracy comparable to the full vestibular battery.Using bi-categorized NAVO plus PTA and NVA, five clusters were identified, from Cluster I(27． 72％)to Cluster V, differing in 2-year management and distributions of affected organs.Conclusions : Classification based on PTA, NVA, and NAVO yields a clinically meaningful staging system for monitoring progression and guiding individualized treatment.