Spectral analysis of respiratory noise in horses with upper airway disorders

BrachySound: machine learning based assessment of respiratory sounds in dogs

The early and accurate diagnosis of brachycephalic obstructive airway syndrome (BOAS) in dogs is pivotal for effective treatment and enhanced canine well-being. Owners often do underestimate the severity of BOAS in their dogs. In addition, traditional diagnostic methods, which include pharyngolaryngeal auscultation, are often compromised by subjectivity, are time-intensive and depend on the veterinary surgeon's experience. Hence, new fast, reliable assessment methods for BOAS are required. The aim of the current study was to use machine learning techniques to bridge this scientific gap. In this study, machine learning models were employed to objectively analyze 366 audio samples from 69 Pugs and 79 other brachycephalic breeds, recorded with an electronic stethoscope during a 15-min standardized exercise test. In classifying the BOAS test results as to whether the dog is affected or not, our models achieved a peak accuracy of 0.85, using subsets from the Pugs dataset. For predictions of the BOAS results from recordings at rest in Pugs and various brachycephalic breeds, accuracies of 0.68 and 0.65 were observed, respectively. Notably, the detection of laryngeal sounds achieved an F1 score of 0.80. These results highlight the potential of machine learning models to significantly streamline the examination process, offering a more objective assessment than traditional methods. This research indicates a turning point towards a data-driven, objective, and efficient approach in canine health assessment, fostering standardized and objective BOAS diagnostics.

Upper airway endoscopy in exercising horses: Findings in 164 barrel racing horses with respiratory clinical signs and/or poor performance

Barrel racing involves sprinting through a cloverleaf pattern in under 20 s. The prevalence of upper airway obstructions (UAO) in barrel racers is unknown, thus a retrospective analysis of 164 client-owned horses referred for overground endoscopy (OGE) between 2014 and 2022 was performed. Referring complaints included respiratory noise, cough, epistaxis, behaviour (owner reported stress/anxiety, refusal to enter arena, excessive head shaking), and/or poor performance. Horses performed a standardized exercise test with low and high-speed components. Videoendoscopic recordings were systematically scored by one experienced clinician. Exercising abnormalities included palatal instability (PI), intermittent dorsal displacement of the soft palate (iDDSP), nasopharyngeal collapse (NPC), recurrent laryngeal neuropathy (RLN), ventro-medial luxation of the apex of the corniculate process of the arytenoid (VLAC), medial deviation of the aryepiglottic fold (MDAF), vocal fold collapse (VFC), and cricotracheal ligament collapse. Associations between co-existing UAOs, age and sex were examined (McNemar's test; logistic regression; P < 0.05). During exercise, 24/164 (15%) horses had no UAO. In the remaining 140 horses, 56% had two or more. NPC was observed in 68/140 (49%), PI in 59/140 (42%), iDDSP in 55/140 (39%), VFC in 38/140 (27%), RLN in 27/140 (19%), VLAC in 14/140 (10%), MDAF in 8/140 (6%) and cricotracheal ligament collapse in 8/140 (6%). Nasopharyngeal collapse and RLN were primarily Grade 3 and Grade B, respectively. Associations exist between several co-existing UAOs, but not age or sex. Abnormalities were more frequent at low-speed (55%) than high-speed (22%), highlighting the importance of OGEs at different exercise intensities. Multiple UAOs appear common in barrel racers with clinical presentations.

Soft Palate Modification Using a Collagen Crosslinking Reagent for Equine Dorsal Displacement of the Soft Palate and Other Upper Airway Breathing Disorders

The mechanical properties of the soft palate can be associated with breathing abnormalities. Dorsal displacement of the soft palate (DDSP) is a naturally occurring equine soft palate disorder caused by displacement of the caudal edge of the soft palate. Snoring and a more serious, sometimes life-threatening, condition called obstructive sleep apnea (OSA) are forms of sleep-related breathing disorders in humans which may involve the soft palate. The goal of this study was to investigate the effect of injecting the protein crosslinker genipin into the soft palate to modify its mechanical properties for the treatment of equine DDSP with potential implications for the treatment of snoring and OSA in humans. Ex vivo experiments consisted of mechanical testing and a wind tunnel study to examine the effect of genipin on the mechanical properties, displacement, and vibration of equine soft palates. A pilot in vivo study was completed using DDSP and control horses to test the safety and effectiveness of injecting a genipin reagent into the soft palate. The wind tunnel testing demonstrated a greater than 50% decrease in transient deformation and a greater than 33% decrease in steady-state vibrations for all doses of genipin tested. Ultimate tensile stress, yield stress, and Young's modulus were higher in the genipin-treated distal soft palate specimens by 52%, 53%, and 63%, respectively. The pilot in vivo study showed a reduction of snoring loudness in all DDSP horses and elimination of DDSP in at least one of three horses. The difficulty of using a 1-meter-long endoscopic injection needle contributed to a consistent overinjection of the equine soft palates, causing excessive stretching (pillowing) and related degradation of the tissue. These ex vivo and in vivo results demonstrated reduced vibration amplitude and flaccidity and increased strength of genipin-treated soft palates, suggesting that genipin crosslinking could become an effective and safe treatment for soft palate related breathing abnormalities.

Factors influencing variation in locomotor-respiratory coupling in Standardbred Trotters in the field

REASONS FOR PERFORMING STUDY

A close relationship between limb and respiratory rhythms has been shown in clinically healthy galloping horses due to mechanical constraints in the thoracic region. This synchronisation leads to a 1/1 ratio between stride frequency (SF) and respiratory frequency (RF) during galloping. Very little is known about locomotor-respiratory coupling (LRC) during fast trot.

OBJECTIVES

To investigate stride and respiratory rates during a standardised exercise test on the track in Standardbred Trotters.

METHODS

Forty-four French trotters age 2-10 years performed a standardised exercise test consisting of three 3 min steps at constant speeds. Speed (V) of exercise varied between 500 and 750 m/min. Variables measured during exercise: SF, heart rate, RF using a microphone between the nostrils, blood lactate concentration. Physiological variables V4 and V200 were calculated and race performance index recorded.

RESULTS

There was no age-related difference in RF or in LRC. Two LRC strategies were found: around 1/1 ratio and 3/2 ratio between SF and RF, respectively. A relationship between SF/RF ratio and physiological parameters showed a higher LRC in good performers because of a lower RF during submaximal and maximal exercise.

CONCLUSION

RF is significantly lower and LRC significantly higher in good compared to poor performers.

POTENTIAL RELEVANCE

Investigation is required to ensure that the breathing techniques of trotters may be used in the selection process of racehorses.

Videoendoscopic evaluation of the upper respiratory tract in 93 sport horses during exercise testing on a high-speed treadmill

REASONS FOR PERFORMING STUDY

Videoendoscopy of the upper respiratory tract (URT) during high-speed treadmill exercise has proved to be invaluable in the assessment of URT dysfunction in racehorses. However, very little information exists regarding dynamic airway collapse in other sport horses used in nonracing equestrian disciplines.

OBJECTIVES

To evaluate the videoendoscopic findings at rest and during exercise in a mixed population of sport horses referred for investigation of poor athletic performance and/or abnormal respiratory noise.

METHODS

Videoendoscopy of the upper airway was performed at rest and during high-speed treadmill exercise in 93 horses.

RESULTS

Dynamic airway obstructions were diagnosed in 77% of horses and were frequently complex in nature. The most common forms of dynamic collapse included soft palate dysfunction (54%), dynamic laryngeal collapse (38%), axial deviation of the aryepiglottic folds (24%) and pharyngeal wall collapse (18%). In the majority of horses, no obvious abnormalities were identified at rest. Enforced poll flexion was found to be a contributing factor in 24% of cases.

CONCLUSIONS

Dynamic obstructions of the URT were a common cause of poor performance and/or abnormal respiratory noise in sport horses referred for investigation of performance problems.

POTENTIAL RELEVANCE

This study highlights the importance of videoendoscopic evaluation of the URT during exercise in horses utilised for equestrian sports where exercise during competition is submaximal in nature.

Dynamic obstructions of the equine upper respiratory tract. Part 2: Comparison of endoscopic findings at rest and during high-speed treadmill exercise of 600 Thoroughbred racehorses

REASONS FOR PERFORMING STUDY

The reliability of diagnoses of obstructive conditions of the upper respiratory tract (URT) based on examinations performed at rest vs. at exercise is controversial.

OBJECTIVE

To compare diagnosis of URT by endoscopy at rest with that achieved during high-speed treadmill exercise (HSTE).

HYPOTHESIS

Endoscopy of URT at rest, when performed in isolation from other simpler techniques is unreliable in the prediction of dynamic respiratory obstructions.

METHODS

Endoscopic findings of 600 Thoroughbred racehorses during quiet breathing were compared with findings during high-speed treadmill exercise. Other parameters were also assessed for their specificity in diagnosis.

RESULTS

Endoscopy of the resting horse showed low sensitivity (0.15) in the diagnosis of dorsal displacement of the soft palate (DDSP) and palatal instability (PI). When endoscopy and reported noises were taken together there was still a 35% misdiagnosis rate. Although there was significant association between resting laryngeal function score (LFS) and dynamic vocal cord and/or arytenoid cartilage collapse at exercise, 19% of horses with a grade 4/5 LFS were able to attain and maintain full abduction during exercise and 7% of those with 'normal' grades 1 or 2 LFS at rest showed dynamic laryngeal collapse when exerted. Sensitivity of the diagnostic model was greatly increased (80%) when a history of inspiratory noise and palpable intrinsic muscle atrophy were included.

CONCLUSIONS AND POTENTIAL RELEVANCE

Endoscopy of the upper respiratory tract of static horses is unreliable in the diagnosis of dynamic obstructions of the URT and should not be used in isolation in surgical decision-making or in the assessment of horses at the time of sale.

The displaced equine soft palate as a source of abnormal respiratory noise during expiration

REASONS FOR PERFORMING STUDY

To develop the technique of respiratory sound analysis for the diagnosis of dorsal displacement of the equine soft palate.

OBJECTIVES

To show that 1) the displaced soft palate can spontaneously vibrate in the presence of expiratory airflow; 2) the observed frequency of vibration can be detected in respiratory sound recordings; and 3) the frequency of vibrations measured in cadaveric specimens are similar to those obtained from audio spectra in exercising horses with DDSP.

METHODS

Palatal movements and respiratory sounds were recorded from artificially ventilated cadaveric heads. The fundamental mode of palate vibrations calculated from high-speed videoendoscopic recordings was compared with frequencies detected in the audio spectra and with data collected from exercising horses with DDSP.

RESULTS

Palate vibrations occurred when air was blown through the trachea. The fundamental mode of vibrations was present in the audio spectra and was not significantly different from those recorded in vivo.

CONCLUSIONS

Vibrations of the soft palate are the probable source of expiratory sounds recorded in horses with DDSP.

POTENTIAL RELEVANCE

This study strengthens the case for the application of respiratory sound analysis in the diagnosis of DDSP in exercising horses.

Dynamic obstructions of the equine upper respiratory tract. Part 1: observations during high-speed treadmill endoscopy of 600 Thoroughbred racehorses

REASONS FOR PERFORMING STUDY AND OBJECTIVE

To review the prevalence of single and complex forms of dynamic airway obstructions within a large group of Thoroughbred horses in training referred for investigation of poor performance.

METHODS

Video-endoscopic recordings of the upper respiratory tract made during a standardised treadmill exercise test of 600 Thoroughbred racehorses were reviewed and analysed in real time and slow motion to identify dynamic collapse by the tissues bordering onto the pharyngeal and laryngeal airways.

RESULTS

Dynamic collapse within the nasopharynx or larynx was confirmed in 471 of the 600 horses. Dorsal displacement of the soft palate (DDSP; 50%) and palatal instability (33%) were the disorders most frequently identified. It was concluded that deglutition is not a significant event in the triggering of DDSP. Complex forms of dynamic collapse were present in 30% of the horses with upper respiratory tract obstructions. A significant influence of age on the prevalence of DDSP and dynamic laryngeal collapse was identified. There was an increased risk of DDSP in younger horses, and of laryngeal collapse in older horses. No association with gender or format of racing was identified.

CONCLUSIONS AND POTENTIAL RELEVANCE

Palatal instability and DDSP comprised the most frequently encountered forms of dynamic collapse within the upper respiratory tract of the Thoroughbred racehorses in this study and are probably expressions of the same nasopharyngeal malfunction. Complex obstructions, i.e. where more than one structure collapses into the airway, occur frequently and therefore treatments that address solitary disorders may often be unsuccessful. Younger horses were found to be at greater risk of sustaining DDSP while older horses seemed more at risk to vocal cord collapse but not to collapse of the arytenoid cartilage itself.

Laser Vocal Cordectomy Fails to Effectively Reduce Respiratory Noise in Horses with Laryngeal Hemiplegia

OBJECTIVE

To report the effect of unilateral laser vocal cordectomy on respiratory noise and airway function in horses with experimentally induced laryngeal hemiplegia (LH).

STUDY DESIGN

Experimental study.

ANIMALS

Six Standardbred horses without upper airway abnormalities at rest or during high-speed treadmill exercise.

METHODS

Respiratory sounds and inspiratory trans-upper airway pressure (P(Ui)) were measured before (baseline) and 14 days after induction of LH by left recurrent laryngeal neurectomy, and again 30, 60, 90, and 120 days after endoscopically assisted laser cordectomy of the left vocal cord. Data were collected with the horses exercising on a treadmill at a speed producing maximum heart rate (HR(max)).

RESULTS

In horses exercising at HR(max), induction of LH caused a significant increase in P(Ui), sound level (SL), and the sound intensity of formant 2 (F(2)) and 3 (F(3)). The sound intensity of formant 1 (F(1)) was unaffected by induction of LH. Laser vocal cordectomy had no effect on SL, or on the sound intensity of F(1) and F(3). At 30, 60, 90, and 120 days after surgery, P(Ui) and the sound intensity of F(2) were significantly reduced, but these variables remained significantly different from baseline values.

CONCLUSIONS

Unilateral laser vocal cordectomy did not effectively improve upper airway noise in horses with LH. The procedure decreased upper airway obstruction to the same degree as bilateral ventriculocordectomy.

CLINICAL RELEVANCE

Currently, laser vocal cordectomy cannot be recommended for the treatment of upper airway noise in horses with LH.