Clinical Value of a Novel Magnetic Resonance Imaging Protocol and Prognostic Model Establishment for Sudden Sensorineural Hearing Loss: A Prospective Study.

Abstract

Introduction: Sudden sensorineural hearing loss (SSNHL) is one of the most common acute symptoms in the otolaryngology department. Etiological diagnosis is the premise of effective treatment of SSNHL, and prognostic evaluation is the key. However, most of the patients are diagnosed as idiopathic due to a lack of overall assessment, while prognostic factors of SSNHL are numerous and controversial. Our purpose was to validate the potential value of a novel three-dimensional fluid-attenuated inversion recovery (3D-FLAIR) MR protocol in SSNHL and to establish a clinical-image prognostic model for unilateral SSNHL.

Methods: This prospective study included consecutive patients from May 2019 to November 2021. Pathogenic diagnosis relied on expertise-based estimation and the associations of MR findings with clinical features of unilateral SSNHL were assessed. The prognostic evaluation of unilateral SSNHL was adopted for recovery and no recovery groups and complete and incomplete recovery groups. Significant clinical and MR features were compared and screened out by single-factor analyses. The primary clinical-image prognosis assessment model was built by multifactor logistic regression analyses.

Results: A total of 101 patients were enrolled in our study who acquired the correct etiological diagnosis based on the novel 3D-FLAIR MR combined with clinical examination. Among the 93 patients with unilateral SSNHL, 30.1% (28/93) showed labyrinthine abnormalities on 3D-FLAIR images. The severity of initial hearing loss in the MR+ group was worse than that in the MR- group (p < 0.05), and patients with positive MR findings tended to have poor recovery. An excellent prognostic model was built for hearing complete recovery and no recovery. The combination of three independent risk factors, including abnormal distortion products otoacoustic emission and transient evoked otoacoustic emission, the period from onset to treatment, and PTA at the onset, was adopted for hearing recovery/no recovery (accuracy = 90.2%, AUC = 0.820). Furthermore, adding the factor of positive MRI findings could improve the confidence for the judgment of hearing no recovery. The only independent risk factor, PTA at the onset, was adopted for complete/incomplete hearing recovery (accuracy = 86.1%, AUC = 0.874).

Conclusion: The novel MR protocol had a good advantage in pathogenic diagnosis. Labyrinthine MR 3D-FLAIR signal abnormalities were related to the severity of an initial hearing loss and had a greater tendency to be found in patients with no recovery. A prognostic model with two main steps of unilateral SSNHL, mainly for SSNHL with no recovery and complete recovery, was built successfully and needed further verification by larger series of patients.