Shortwave infrared otoscopy for diagnosis of middle ear effusions: a machine-learning-based approach

Abstract

Otitis media, a common disease marked by the presence of fluid within the middle ear space, imparts a significant global health and economic burden. Identifying an effusion through the tympanic membrane is critical to diagnostic success but remains challenging due to the inherent limitations of visible light otoscopy and user interpretation. Here we describe a powerful diagnostic approach to otitis media utilizing advancements in otoscopy and machine learning. We developed an otoscope that visualizes middle ear structures and fluid in the shortwave infrared region, holding several advantages over traditional approaches. Images were captured in vivo and then processed by a novel machine learning based algorithm. The model predicts the presence of effusions with greater accuracy than current techniques, offering specificity and sensitivity over 90%. This platform has the potential to reduce costs and resources associated with otitis media, especially as improvements are made in shortwave imaging and machine learning.

Color Dependence Analysis in a CNN-Based Computer-Aided Diagnosis System for Middle and External Ear Diseases

“…Multispectral imaging analysis has several advantages over standard otoscopy, including increased image contrast, clear visualization of middle ear elements, better assessment of tympanic membrane vascularity, and improved demarcation of critical morphological structures (e.g., the malleus and the promontory). Although there are no commercial otoscopes available that acquire images in different spectral bands in addition to RGB, the research results of the few works [9,14,15] in state of the art encourage the development of new technologies considering a multispectral approach. In this work, we explored the dependence of three different color wavelengths, including red, green, and blue channels, in the performance of a CNN-based model to predict the diagnosis of four ear conditions.…”

“…Due to the advent of advanced processing techniques and the large volume of digital data available today, new technology to assist medical diagnosis has been generated as an increasing tendency to reduce diagnostic error in the last few years [6]. Some studies [7][8][9] have demonstrated the benefit of using artificial intelligence (AI) to improve and diagnose safety for ENT pathologies of various types by achieving a diagnostic performance of over 90%.…”

“…In contrast, the red channel has a high penetrance to the tympanic membrane, allowing a better visualization of the structures inside the tympanic cavity. Another study [9] identified the presence of effusions in the middle ear using machine learning algorithms. The most important contribution was developing an otoscope that visualizes middle ear structures and fluid in the shortwave infrared region.…”

Development of a shortwave infrared sinuscope for the detection of cerebrospinal fluid leaks

“…Fluid in the middle ear, for example, shows strong light absorption between 1400 and 1550 nm. As a result, straightforward middle ear fluid detection in a model using a SWIR otoscope is possible and can be used to diagnose otitis media with greater accuracy than is currently possible with a standard otoscope 11 , 12 . SWIR technology can also be used to improve the characterization of allergic contact dermatitis using the different absorption properties of different SWIR channels 13 …”

Development of a shortwave infrared (SWIR) sinuscope for the detection of cerebrospinal fluid (CSF) leaks

“…Fluid in the middle ear, for example, shows strong light absorption between 1,400 and 1,550 nm. As a result, straightforward middle ear fluid detection in a model using a SWIR otoscope is possible and can be used to diagnose otitis media with greater accuracy than is currently possible with a standard otoscope (Carr et al, 2018) (Kashani et al, 2021).…”