Noninvasive Targeted Temperature Management of the Inner... : Otology & Neurotology (2025)

INTRODUCTION

About 1.5 billion individuals worldwide experience some form of hearing decline in their lifetime (¹). Hearing loss (HL) is a significant global health issue, affecting over 6% of the world's population and projected to increase to nearly 10% by 2050 (²). It is the fourth leading contributor to years lived with disability worldwide, with an estimated annual economic impact exceeding $750 billion (³). HL is associated with lower educational attainment, reduced income, higher unemployment rates (⁴), and recent studies show an association between hearing loss and cognitive decline (⁵). HL may also contribute to social isolation, loneliness, and depression in older adults (⁵). Importantly, hearing loss is a modifiable risk factor for dementia, potentially accounting for up to 9% of the overall modifiable risk (⁵). Hearing loss can result from inner ear injuries, which include exposure to excessive noise, trauma associated with blast explosions, and electrode insertion trauma (^6,7). Mechanisms underlying cochlear damage involve the generation of reactive oxygen species (^8,9), production of free radical (¹⁰), activation of inflammatory pathways (^8,9,11), and glutamate excitotoxicity (¹²). All of these primarily lead to programed hair cell death at the organ of Corti (^13,14), swelling of afferent dendrites (^15–17), and spiral ganglion degeneration (^18–20). The latter has been shown in both animal models and in histological assessments of human temporal bones (^21,22). Functionally, these injuries manifest as a decline in auditory thresholds both after noise exposure (²³) and cochlear implantation (^6,7,24).

Therapeutic hypothermia has long been recognized as a therapeutic option against cochlear injury (^25–28) as it reduces concentrations of free radicals (²⁹), decreases glutamate toxicity (²⁵), reduces lipid peroxidation (^30,31), and attenuates the expression of inflammatory response mediators (³²) (³³). Systemic application of hypothermia resulted in reduced cochlear hair loss caused by noise exposure in rodent models (³⁴), as well as decreased auditory brainstem responses (ABR) and distortion product otoacoustic emissions (DPOAEs) threshold shifts (³⁵) in preclinical models with cochlear implantation. However, the application of systemic therapeutic hypothermia for inner ear pathologies is limited due to significant risks including adverse effects on heart rate, coagulation pathways, drug metabolism, and infections (^36,37). Invasive, localized techniques, such as administration of cooled fluid in the cochlea (^38–41) and ear canals (⁴²) of rodent and temporal bone models, have also been studied as alternatives for delivering hypothermia (^38–41). These approaches have demonstrated efficacy in preserving residual hearing and maintain hair cell density post-electrode insertion (⁴³) and drug-induced hearing loss (⁴⁴). Nevertheless, they pose a risk of inner ear structural damage and carry clinical challenges, which underscores the need for noninvasive delivery methods. Recently, data from our group have demonstrated the feasibility, safety, and efficacy of cooling inner ear structures in rodents by utilizing an external, localized, and noninvasive technique (⁴⁵). This approach has not yet been employed in human cadaveric models, which account for unique physical properties of human tissues, which could influence temperature transfer including blood perfusion and fluid movement. Further evaluation in human cadaveric models is necessary to assess the effectiveness and safety of this strategy in a more clinically relevant context.

The aim of this study was to examine the viability of achieving temperature alterations in the inner ear structures of human cadaveric specimens using a targeted and non-intrusive method. Findings were replicated in 18 separate experiments and further validated through a computational model that simulated the distribution of temperature within the three-dimensional geometry of the inner ears. Results from this study will provide new insights that could lead to the development of a more practical hearing protection method for clinical use.

MATERIALS AND METHODS

Specimen Preparation

A total of five human adult half cadaver heads (two males and three females) from donors were acquired for the study (Science Care Body Donation). Before experimental use, each cadaver sample was scanned with MRI and CT scans for use in numerical modeling. CT and MRI protocols followed clinical standards (⁴⁶). CT scan of the head was as follows: axial 3 × 3 cerebrum window, axial 3 × 3 bone window, coronal 2 × 2 cerebrum window, sagittal 2 × 2 cerebrum window, and axial 0.75 × 0.75 cerebrum window. The CT of the temporal bone was as follows: bilateral, axial soft tissue 0.6 × 0.6; unilateral, axial bone window 0.6 × 0.6; bilateral, coronal soft tissue 0.6 × 0.6; unilateral, coronal bone windows 0.6 × 0.6; and SCC views, bone windows 0.4 × 0.4. MRI sequences were 3D MPRAGE, 3D FLAIR, and 3D T2 weighted. The relevant methods are described below.

The frozen cadaver heads were thawed overnight at room temperature to facilitate the experimental procedures. Each specimen was then prewarmed to a target temperature of 36.9°C ± 3.2°C by immersing it in an aluminum container filled with heated metallic beads (Lab Armor Beads, cat no. A1254302). The beads were warmed to 44°C using a heating pad (RestorEar Devices LLC) placed beneath the container, which was connected to a self-designed heater (RestorEar Devices LLC) to ensure a constant temperature throughout the experiment. The specimens remained submerged in the bead bath throughout the experiment to ensure a consistent temperature was maintained.

For each experimental trial, the precise anatomical location and permeability of the eustachian tube (ET) were carefully assessed. Three micro-thermistors (Omega Engineering Inc., 5SC-TT-K-40-36-ROHS) were inserted approximately 3 cm into the ET, starting at the nasopharynx, to reach the middle ear and cochlear region. To ensure stable probe placement throughout the experiment, superglue was applied at the tube's opening, and sutures were used to secure the cadaver tissue. A single thermistor was positioned above the temporal bone, where the cooling device would later be placed, whereas another served as a reference point on the scalp, away from the other probes. Additional thermistors were placed near the eardrum and within the bead bath. An overview of the experimental setup is shown in Figure 1.

Hypothermia Experiments

Experiments were conducted following two protocols: 1) specimens that underwent a 60-minute cooling period with a total of 11 repetitions, and 2) specimens that underwent a 30-minute cooling period with a total of 7 repetitions.

A custom-designed cooling device, consisting of a horseshoe-shaped gel pack filled with aqua pearl beads (ReBound; RestorEar Devices LLC), was tailored to fit around the ear and cover the temporal bone. The device was stored at −20°C before the start of the experiments. After stable temperature readings were recorded from the middle ear and scalp thermistors, the cooling device was carefully placed over the temporal bone for either 60 or 30 minutes, depending on the protocol. Once the cooling phase concluded, the device was removed, allowing the specimens to naturally rewarm over a 30-minute period, which was also monitored. Temperature changes from the micro-thermistors were continuously measured and logged every 2 seconds using a Graphtec Datalogger (Omega RDXL 12SD).

Imaging and Computational Fluid Dynamics

3D digital head models were reconstructed from MRI and CT scanned data corresponding to the same cadaver heads utilized in the experimental set-up. Axial CT scans of cadaveric heads were acquired using a Siemens SOMATOM Perspective scanner (130 kVp, 75 mAs, 512 × 512 acquisition matrix, 0.7-mm slice thickness, H31s kernel, 240-mm field of view). CT images were segmented by a single operator and reconstructed in Seg3D (v2.5.1; Scientific Computing Institute, University of Utah). External soft tissues (ear and surrounding skin), bone (skull and mandible), inner ear, and the cranial cavity were isolated in a region of interest (ROI) centered on the ear canal. Three-dimensional isosurface reconstructions were generated from the respective segmentations and exported in .stl format. Surface decimation, smoothing, and conversion to from .stl to solid body (.sldprt) were conducted in Solidworks (Dassault Systèmes, Vélizy-Villacoublay, France) using the Mesh2Surface plugin application (KVS Ltd., Cambridge, UK). Boolean operations were utilized to ensure congruent interfaces between reconstructed solid bodies before finite element mesh generation. CFdesign, a computational fluid dynamics (CFD) software that utilizes finite element analysis (FEA), was utilized for this purpose. Tissue types were then distinguished and meshed to form 3D geometry with appropriate material properties as detailed in Figure 2. Subsequently, transient analyses were conducted to determine both the time duration and temperature fluctuations within the inner ear structures. The analysis was conducted using a model comprising 3,472,108 elements and extended for the equivalent of 30 minutes (180-time steps). To validate our results, we repeated the analysis with a 15% increase and decrease in the number of elements, ensuring the convergence of the solution.

To ensure the model's stability and the consistency of results, we performed a mesh-independent study, systematically varying the number of elements to confirm convergence. Several key assumptions were made during the analysis: 1) conductive materials, including bone and blood/tissue, were treated as having homogeneous properties; 2) blood was modeled as noncirculating and represented as a “solid” with approximate thermal properties; 3) the sole mode of heat transfer considered was conduction; 4) all external surfaces were treated as adiabatic, implying no heat transfer to or from the environment; 5) the model incorporated symmetry about the dorsal plane; and 6) the initial temperature of the head model was set to 37°C. Notably, we deemed cold pack material properties irrelevant, as the entire volume maintained a steady-state temperature boundary condition of 0°C, simulating the effect of a cold treatment.

Statistical Analysis

The mean and standard deviation of the maximum change in temperature achieved at the middle ear and the scalp were compared between the 60- and 30-minute cooling protocol using a Student t test. Sex differences were also evaluated using Student t test. All statistical analyses were performed using GraphPad PRISM software.

RESULTS

Thermal Measurements From Cadaveric Specimens

Figure 3A provides a detailed visualization of temperature changes from a single repetition of the 60-minute cooling protocol. Figure 3B presents the mean and standard deviations of middle ear and scalp temperature changes across 12 repetitions of the same protocol. The results demonstrate a gradual decline in middle ear temperatures, which were sustained at hypothermic levels throughout the experiment. Thermistors placed in the ear canal and beneath the gel pack recorded a sharp temperature drop during cooling, followed by a rapid return to baseline after the gel pack was removed. In contrast, the scalp temperature, used as the control, displayed fluctuations that closely matched those of the bead bath rather than the middle ear. No obvious changes in skin characteristics were observed in the auricle or skin of the temporal bone after device removal.

Figure 4 shows the summarized results from thermal measurements across all cadaver samples. Temperature shifts were calculated as the total change in degrees from the baseline to the lowest achieved temperature. Over a 60-minute cooling period (n = 11), the average temperature drop was 4.655°C ± 1.22°C in the middle ear and 2.79°C ± 1.38°C in the scalp (Fig. 4A). Statistical analysis revealed a significant difference between the measurements (p = 0.0017). During a 30-minute cooling period (n = 7), the average temperature drop was 2.87°C ± 1.80°C in the middle ear and 0.77°C ± 1.31°C in the scalp (Fig. 4A). These results did not show statistical significance (p = 0.1888).

Figure 4B depicts the differences in temperature shifts calculated from experiments conducted separately in cadaver heads of women and men during a 60-minute cooling period. Male cadaver heads (n = 4) exhibited an average temperature drop of 3.9°C ± 0.7°C in the middle ear and 2.4°C ± 0.6°C in the scalp. Female cadaver heads (n = 7) showed an average temperature drop of 5.08°C ± 1.27°C in the middle ear and 2.98°C ± 1.69°C in the scalp. Statistical analysis revealed significant differences between middle ear and scalp temperatures in both male and female analyses (men, p = 0.0267; women, p = 0.0239). Figure 4C displays the differences in female and male temperature changes during a 30-minute cooling protocol. The average temperature shift in males (n = 2) was 0.595°C ± 0.154°C in the middle ear and 0.9°C ± 0.42°C in the scalp. For female cadaver heads (n = 5), the shift was 3.78°C ± 1.13°C in the middle ear and 1.44°C ± 0.76°C in the scalp. Statistical significance was found in females (p = 0.005) but not in the male analysis (p = 0.333).

Thermal Measurement Using a Finite Element Modeling

Figure 5A provides a visual representation of temperature transfer obtained from a single modeled head during a 30-minute cooling protocol. Each frame corresponds to 10-minute increments. Figure 5B illustrates temperature shifts at key locations within the cochlea, including the apex, middle turn, promontory, and round window. These shifts are calculated as the change from the baseline (in degrees Celsius) to the lowest achieved temperature after 30 minutes. The average temperature drop in these regions was 2.82°C ± 1.02°C in the cochlear apex, 2.53°C ± 1.01°C in the middle turn, 1.99°C ± 0.68°C at the promontory, and 2.13°C ± 0.89°C at the round window.

DISCUSSION

Despite the extensively studied benefits of mild therapeutic hypothermia (MTH) in mitigating cellular and synaptic damage caused by cochlear trauma, its clinical application remains limited due to challenges in achieving rapid and controlled hypothermia, as well as the associated side effects. Systemically administered MTH has been shown to reduce hearing threshold shifts and hair cell loss in rodent models following various injuries, including cochlear blood flow interruption (^47,48), cochlear electrode-insertion trauma (³⁵), and noise overexposure (^34,49–51). In contrast, locally delivered hypothermia may offer a more effective approach. For instance, Smith and Mills (⁵²) demonstrated that irrigating cold fluid directly into a rat's cochlear bulla can reduce the temperature of the cochlear basal turn by 3°–4°C within minutes. Similarly, Tamames et al. introduced a novel experimental design that delivers localized, invasive hypothermia in a rodent model. Their device, consisting of a fluorocarbon-filled probe attached to a custom thermoelectric Peltier unit, significantly reduced cochlear temperatures by 4°C–6°C and preserved residual hearing and outer hair cells following cochlear implantation. Additionally, this custom cooling device was able to deliver local MTH to human cadaveric temporal bone cochleae without compromising surgical visualization (³⁹). However, localized MTH delivery often requires invasive techniques that pose risks to delicate cochlear structures, such as the facial nerve, and are not well suited for rapid application in outpatient settings—where speed is critical given the limited window of opportunity to treat conditions like noise- and blast-induced hearing loss. Moreover, many studies exploring MTH in the context of hearing loss rely on animal models or isolated human temporal bones, which, although informative, may not fully reflect clinical outcomes due to differences between experimental models and human biology.

In this study, we developed and tested a novel, noninvasive cooling technique for reducing inner ear temperatures using a custom-designed device. The device, composed of an aqua pearl-filled gel pack positioned over the mastoid region of human cadaver specimens, enabled gradual and controlled cochlear cooling. During a 60-minute protocol, we observed a significant temperature reduction in the middle ear compared with the scalp (p = 0.0017), with the middle ear showing an average drop of 4.655°C ± 1.22°C and the scalp showing a drop of 2.79°C ± 1.38°C. In contrast, during the 30-minute protocol, temperature changes in the middle ear (2.87°C ± 1.80°C) and scalp (0.77°C ± 1.31°C) were not statistically different (p = 0.1888). These results indicate the feasibility of external cooling and suggest that the duration of the cooling period may influence the extent of temperature change. Additionally, we analyzed gender-based differences in temperature shifts. Over 60 minutes, female cadaver heads exhibited greater temperature drops in both the middle ear (5.08°C ± 1.27°C) and scalp (2.98°C ± 1.69°C) compared with males (middle ear: 3.9°C ± 0.7°C; scalp: 2.4°C ± 0.6°C). In the 30-minute protocol, only female cadavers showed significant middle ear cooling. These findings imply that anatomical and physiological differences between genders may affect temperature regulation and response to cooling.

In the majority of repetitions conducted, we observed temperature fluctuations in both the scalp and bead bath, likely due to variations in thermistor positioning and movement of heated beads near the scalp thermistor tip. Additionally, a sharp temperature decrease occurred in thermistors located in the ear canal and near the gel pack, though unlike middle ear temperatures, these drops rapidly recovered once the cooling device was removed. This suggests that superficial temperature declines are quickly reversible and may not significantly impact nearby structures or functions. The inner ear is highly vascularized, and the blood flow may alter temperature distribution by heat transfer. However, given the gradual and sustained delivery of hypothermia, it is unlikely to significantly impact the cooling of the target endorgans. Studies in human subjects combined with sensitive electrophysiological measures will be needed to further characterize the cooling efficiency of these devices.

Additionally, 3D digital reconstructions of the cadaver heads used in this study were generated from MRI and CT scans to provide a comprehensive numerical model for analysis. These reconstructions allowed us to apply finite element analysis (FEA) to accurately simulate the transfer of temperature from the cooling device to the cochlea and surrounding structures of the inner ear. The numerical simulations from the FEA revealed that, during the 30-minute cooling protocol, the average temperature reduction was 2.82°C ± 1.02°C at the cochlear apex, 2.53°C ± 1.01°C in the middle turn, 1.99°C ± 0.68°C at the promontory, and 2.13°C ± 0.89°C at the round window. These computational results closely mirrored the experimental data obtained from the cadaveric specimens, thereby validating the accuracy and reliability of the experimental approach.

It is important to note that this study does not aim to establish the exact duration of MTH required for clinical effectiveness, as this may vary with the type of cochlear injury among other factors. However, previous auditory research has shown that MTH delivered to preclinical models for durations between 20 minutes and 7 hours yields beneficial effects (^{43,45,47,48,53}). Combined with our findings that MTH temperatures can be achieved within 30 minutes, we hypothesize that this relatively straightforward approach could be applicable in real-life scenarios and likely achieve clinical benefits. Further clinical studies will be required to determine the specific duration for different types of cochlear injuries. We also note that frostbite and other cold-induced skin effects are typically associated with prolonged exposure to temperatures below 0°C. In our protocol, the gel pack is initially maintained at 5°C–6°C; however, it gradually warms over time as its cooling capacity diminishes. This controlled temperature progression significantly reduces the likelihood of sustaining cold-related injuries at low temperatures.

In this study, we did not characterize temperature changes in the vestibular system. The numerical model suggested ~2°C change in the lateral semicircular canal with the cooling protocol. However, we hypothesize this change combined with bilaterally delivered therapeutic hypothermia is unlikely to induce caloric responses. The gradual cooling protocol prevents the rapid endolymphatic flow required for cupular deflection (⁵⁴). Additionally, the final design of our product incorporates bilateral simultaneous cooling, removing the unilateral temperature differences needed to trigger caloric responses.

Several limitations arose within the scope of this study. First, the sample size was relatively small, which may limit the generalizability of the findings. Increasing the sample size in future research would enhance the robustness of the conclusions and offer more definitive evidence. Although we made meticulous efforts to ensure the correct placement of thermistors within the middle ear, we were unable to visually confirm their precise final positioning. This decision was made to avoid compromising the structural integrity of the inner ear and surrounding anatomy, which could have affected the outcome of the experiment. Nonetheless, future studies could improve upon this methodology by utilizing micro-thermistors placed directly at the cochlea, providing more localized and accurate temperature measurements.

In conclusion, our study introduces a promising, noninvasive technique for inducing localized hypothermia within the human middle ear. This approach demonstrates the potential for applying mild therapeutic hypothermia (MTH) in clinical settings to address hearing-related conditions in a safe and effective manner. The ability to induce hypothermia without the need for invasive procedures could greatly expand the accessibility and practicality of MTH, allowing for its use in outpatient settings while minimizing risks to patient safety. Moving forward, further research and larger-scale studies will be essential to fully explore the clinical potential of this technique and to optimize its implementation in therapeutic applications.

REFERENCES