Differentiating Causes of Post-Surgical Conductive Hearing Loss with Optical Coherence Tomography Vibrometry

To evaluate whether optical coherence tomography (OCT) vibrometry can differentiate causes of conductive hearing loss (CHL) following ossiculoplasty, specifically distinguishing among effusion, soft-tissue fixation, and prosthesis disconnection.

We simulated three post-surgical CHL conditions, effusion, soft-tissue fixation, and prosthesis disconnection, in a cadaveric temporal bone model (N = 10 per condition), with a partial ossicular replacement prosthesis (PORP) and cartilage graft implanted via a retrotympanic approach. A custom-built swept-source middle-ear OCT system (λ₀ = 1550 nm, Δλ = 40 nm) was used to capture cross-sectional images and mobility measurements at the umbo and center of the cartilage graft across six stimulus frequencies (500, 750, 1000, 1500, 2000, and 3000 Hz). Mobility values (point velocity normalized to sound pressure) served as input features for a Random Forest classifier. Changes in mobility from baseline were also statistically analyzed.

The classifier achieved 90.9% accuracy (40/44 datasets) in differentiating conditions in leave-one-out cross-validation, and 100% when trained on the full dataset. Simulated effusion and soft-tissue fixation were associated with broadband mobility decreases of 17.5 dB and 8.0 dB, respectively, averaged across both locations. Prosthesis disconnection at the stapes and at the tympanic membrane led to low-frequency (500 and 750 Hz) mobility increases of 9.9 dB and 11.1 dB, respectively, averaged across both locations.

In a cadaveric model of post-surgical CHL, OCT vibrometry accurately distinguished effusion, soft-tissue fixation, and prosthesis disconnection. The ability to identify the cause of post-surgical CHL highlights OCT vibrometry's potential to assist clinical decision-making in revision surgery.