Multicenter Validation of an AI-Based Snoring Model vs. Wearables for Diagnosing Sleep Apnea

Haoyan Guan,1 Junyu Huang,2 Danling Liu,1 Long Yang,3,4 Hao Zeng,1 Jingjing Huang,4 Beiping Miao1 1Department of Otolaryngology Head and Neck Surgery, The Second People’s Hospital of Shenzhen (the First Affiliated Hospital of Shenzhen University), Shenzhen, Guangdong, People’s Republic of China; 2Department of Otolaryngology Head and Neck Surgery, The First Affiliated Hospital of Xiamen University, Xiamen, Fujian, People’s Republic of China; 3State Key Laboratory of Medical Neurobiology and MOE Frontiers Center for Brain Science, Fudan University, Shanghai, People’s Republic of China; 4Sleep Medicine Center, ENT Institute and Otorhinolaryngology Department, Eye & ENT Hospital, Fudan University, Shanghai, People’s Republic of ChinaCorrespondence: Beiping Miao, Department of Otolaryngology Head and Neck Surgery, The Second People’s Hospital of Shenzhen (The First Affiliated Hospital of Shenzhen University), Shenzhen, Guangdong, People’s Republic of China, Email miaobeiping@163.com Jingjing Huang, Sleep Medicine Center, ENT Institute and Otorhinolaryngology Department, Eye & ENT Hospital, Fudan University, Shanghai, People’s Republic of China, Email gennie_xuan@163.comBackground: Sleep apnea–hypopnea syndrome (SAHS) has high prevalence and cardiovascular/cerebrovascular comorbidities. Polysomnography (PSG) is the diagnostic gold standard but limited by complexity, cost, and accessibility; smart wearables are convenient but less accurate. This study validated an AI-based snoring-sound model and compared its diagnostic performance with a wearable, using PSG as reference.Methods: Adults with suspected SAHS (n = 134) were included and underwent overnight PSG, with simultaneous snoring sound and wearable signal recording. The snoring algorithm used short-time Fourier transform (STFT) spectrograms and a multiscale encoder–attention–decoder neural network for classification. Model performance was assessed using correlation analysis, intraclass correlation coefficient (ICC), Bland–Altman plots, and receiver operating characteristic (ROC) analyses. Accuracy, sensitivity, specificity, and area under the curve (AUC) values were computed for both models.Results: Disease severity was associated with variation in BMI (p < 0.001) and decreasing oxygen saturation (p < 0.001). The snoring model showed stronger correlation and higher agreement with PSG-derived AHI (r = 0.79, ICC = 0.753) than the wearable (r = 0.68). Although the snoring model yielded consistently higher AUCs, particularly in mild SAHS (AUC = 0.83 vs. 0.72, p = 0.07), the differences between methods were not statistically significant (all p > 0.05). Both models achieved excellent discriminative performance in severe SAHS (AUC ≥ 0.90).Conclusion: The AI-driven snoring-sound model demonstrated comparable overall performance and numerically better performance in certain settings, particularly in identifying mild SAHS.Plain Language Summary: Sleep apnea-hypopnea syndrome (SAHS) is a common disorder with serious health implications, including cardiovascular and cerebrovascular risks. PSG is the gold standard for diagnosing SAHS but is limited by cost and accessibility. This has led to the development of wearable devices, though their accuracy is compromised by the inability to measure airflow directly and detect subtle arousals.This study introduces an AI-based snoring sound analysis model as a non-invasive and scalable diagnostic tool for SAHS. Compared to PSG, the snoring model shows comparable performance, particularly in detecting mild SAHS, and demonstrates higher AUC values than wearable devices (AUC 0.83 vs. 0.72).