Gene therapies
In May 2024, British toddler Opal Sandy was announced to have successfully had her hearing restored through a new gene therapy targeting low levels or absence of a protein caused by variants in the OTOF gene. Born deaf as a result of auditory neuropathy, eighteen month old Opal was administered the treatment as part of the CHORD study. The therapy works by boosting otoferlin protein levels, restoring the signalling capacity of inner ear hair cells and enabling auditory signal transmission to the brain.
Using adeno-associated virus (AAV) vectors to deliver the therapy directly into Opal’s ear, researchers were able to administer the treatment in an operation which took less than 20 minutes. Following a successful procedure, Opal is reported to have almost perfect hearing.
These early results have strengthened the rationale for developing gene therapies targeting other genetic hearing loss subtypes, and raise important questions about durability of effect, patient stratification, and the translational pathway from proof-of-concept to broader clinical application.
However, further research needs to be done to understand the mid- and long-term impacts of such treatments, how long improvements last and which hearing loss subtypes stand to benefit most from this approach.
Beyond viral vector delivery, researchers are also exploring whether the cochlea's own cells can be prompted to regenerate. Work led by researchers at Massachusetts Eye and Ear and Harvard Medical School has demonstrated that hair cells can be regenerated in an adult mammalian ear by inhibiting a protein called Notch on the surface of supporting cells. When Notch signaling was blocked using a gamma-secretase inhibitor, supporting cells in the cochlea converted into new hair cells, and partial hearing recovery was observed in noise-damaged mice. This represents a fundamentally different approach: rather than delivering a missing gene, it activates a regenerative pathway that mammals were previously thought to lack. While still preclinical, the finding expands the potential treatment landscape for sensorineural hearing loss beyond single-gene conditions.
Pharmacogenomics and hearing loss
Genetic screening can also serve a preventive function, identifying patients who are at risk of drug-induced hearing loss before treatment decisions are made.
Gentamicin is an antibiotic commonly used to treat infections in newborn babies. However, despite it being used to safely treat 100,000 babies per year in the UK, one in 500 babies carry a genetic variant which can lead to permanent hearing loss if they are given the drug. In babies with this variant, gentamicin enters inner ear cells, causing damage and resulting in hearing loss.
However, a simple genetic test can be used to establish if a baby is vulnerable to this type of hearing loss. Using a non-invasive cheek swab, the test can be run in under half an hour to establish if the risk variant is present and if the baby should be given a different type of antibiotic in order to protect their hearing while still effectively treating infection. Researchers estimate that this rapid test could save the hearing of 200 babies in England alone every year.
Technical innovations
Alongside therapeutic advances, AI and machine learning tools are beginning to reshape how hearing loss is assessed and diagnosed — with direct implications for screening efficiency, data interpretation, and clinical decision support.
Machine learning
Machine learning (ML) approaches are being applied to hearing loss screening, diagnostics, and risk identification. As demand for hearing assessment grows, ML algorithms offer a way to process high volumes of patient data at speed, potentially producing more consistent hearing models than manual audiological review. Early research has shown that ML offers significant advancements in the accuracy and efficiency of hearing assessments. Capable of analyzing large datasets, ML methods can enhance diagnosis using pattern recognition to analyze audiograms and, potentially, speech.
Integration of electronic health records
In tandem with ML approaches, the integration and analysis of aggregated, anonymized electronic health records is enabling precision medicine approaches to genetic hearing loss. By combining audiometric data with broader clinical records, these systems can support predictive diagnosis and surface intervention recommendations that reflect a patient's full clinical context. For example, a deep learning (DL) model called “Auto-Audio,” which utilizes audiogram databases from the Department of Otolaryngology electronic medical record system at the Sunnybrook Health Sciences Center in Toronto, Canada, has been able to interpret audiograms and give a diagnosis with up to 97.5% accuracy, reducing the risk of missed or delayed diagnoses.
Smart devices
AI technologies also have the potential to help develop more versatile and user-friendly hearing aids, enabling personalization of features (such as preferences around background noise filtering). Integration of hearing devices with connected health platforms is enabling longitudinal monitoring of hearing function, generating real-world data that could support both personalised device optimisation and ongoing clinical surveillance. Such innovations have the potential to enable enhanced monitoring and management of hearing health as well as create a highly personalized experience.
Personalized hearing care
Advances in gene therapy, pharmacogenomics, and AI-driven technology are changing the way hearing loss is diagnosed, treated, and managed. Across each of these domains, the direction is consistent: earlier identification, more precise intervention, and more personalized care.
This matters because timing is critical. Research consistently shows that earlier detection and intervention lead to better developmental outcomes, particularly in children. Genetic screening, whether for otoferlin-linked deafness or pharmacogenomic risk variants like gentamicin sensitivity, enables clinicians to act before irreversible damage occurs. Combined with AI-driven diagnostics and longitudinal health record analysis, these tools create the foundation for a hearing care model that is proactive rather than reactive.
As more therapies target specific genetic subtypes, the ability to identify, characterize, and engage the right patients early will define how quickly these innovations reach the people who need them.