Automated echocardiographic left ventricular dimension assessment in dogs using artificial intelligence: Development and validation.

BACKGROUND: Artificial intelligence (AI) could improve accuracy and reproducibility of echocardiographic measurements in dogs. HYPOTHESIS: A neural network can be trained to measure echocardiographic left ventricular (LV) linear dimensions in dogs. ANIMALS: Training dataset: 1398 frames from 461 canine echocardiograms from a single specialist center. VALIDATION: 50 additional echocardiograms from the same center. METHODS: Training dataset: a right parasternal 4-chamber long axis frame from each study, labeled by 1 of 18 echocardiographers, marking anterior and posterior points of the septum and free wall. VALIDATION DATASET: End-diastolic and end-systolic frames from 50 studies, annotated twice (blindly) by 13 experts, producing 26 measurements of each site from each frame. The neural network also made these measurements. We quantified its accuracy as the deviation from the expert consensus, using the individual-expert deviation from consensus as context for acceptable variation. The deviation of the AI measurement away from the expert consensus was assessed on each individual frame and compared with the root-mean-square-variation of the individual expert opinions away from that consensus. RESULTS: For the septum in end-diastole, individual expert opinions deviated by 0.12 cm from the consensus, while the AI deviated by 0.11 cm (P = .61). For LVD, the corresponding values were 0.20 cm for experts and 0.13 cm for AI (P = .65); for the free wall, experts 0.20 cm, AI 0.13 cm (P < .01). In end-systole, there were no differences between individual expert and AI performances. CONCLUSIONS AND CLINICAL IMPORTANCE: An artificial intelligence network can be trained to adequately measure linear LV dimensions, with performance indistinguishable from that of experts.

Cardiac auscultation skills in final year veterinary students and recent veterinary graduates, referral hospital veterinary surgeons and veterinary cardiologists or cardiology residents.

BACKGROUND: Cardiac auscultation is an important part of the physical examination. This study evaluated cardiac auscultation skills in veterinary students and compared their abilities to recent veterinary graduates, referral hospital veterinary surgeons and veterinary cardiologists or cardiology residents. In addition it compared their self-predicted quiz scores to their actual scores, evaluating if they could accurately predict their own performance level. METHODS: A digital recording device was used to record auscultation sounds from 12 different patients with a diagnosis confirmed by a board-certified veterinary cardiologist. The sound files and associated phonocardiograms were uploaded to a video sharing website. A cloud-based online multiple-choice quiz was generated and shared with final year veterinary students, recent veterinary graduates, referral hospital veterinary surgeons and veterinary cardiologists or cardiology residents. RESULTS: There were 128 participants: 51 final year veterinary students, 62 recent veterinary graduates, and 10 referral hospital veterinary surgeons and five veterinary cardiologists or cardiology residents. No difference was found between the cardiac auscultation skills of recent veterinary graduates and final year veterinary students. Veterinary students' self-predicted scores were lower than actual scores. CONCLUSIONS: Recent veterinary graduates did not perform better than final year veterinary students in this study, suggesting that auscultation skills do not continue to improve in the first few years after graduation. Efforts should be made to maximise students' learning in cardiac auscultation skills. Veterinary students show a lack of confidence in cardiac auscultation skills.

Left heart dimensions in anemic cats and dogs before and after blood transfusion.

BACKGROUND: Whether anemic cats and dogs with increased left heart dimensions are at higher risk of transfusion-associated circulatory overload, and the effect of blood transfusion on left heart dimensions in naturally occurring anemia is unknown. HYPOTHESIS/OBJECTIVES: To evaluate the effect of blood transfusion on left heart dimensions in clinically relevant anemia. ANIMALS: Twenty dogs and 20 cats presenting to a university veterinary teaching hospital. METHODS: In this prospective observational study, anemic dogs and cats requiring blood transfusion were included. Packed cell volume (PCV), total solids, and echocardiography were performed before and within 24 hours of blood transfusion. Signalment, bodyweight, disease process, transfusion duration and volume, and prior treatments were recorded. Nonparametric statistics were reported as median [range]. Post hoc Bonferroni correction set significance at P < .006. RESULTS: After transfusion, PCV increased in cats (12% [6-16] to 18% [10-33], P = .001) and dogs (14% [7-24] to 25% [9-37], P = .001), heart rate decreased in dogs (104 bpm [86-166] to 87 bpm [56-138], P < .001), and fractional shortening decreased in cats (57.1% [36.0-84.7] to 41.0% [28.1-69.6], P = .002) and dogs (33.79% [19.33-62.79] to 31.89% [19.06-51.47], P = .006). Left ventricular internal diameter in systole increased in cats (6.5 mm [2.7-9.8] to 7.9 mm [5.3-11.1], P = .001). Normalized left ventricular internal diameter in diastole (1.48 [1.25-1.79] to 1.57 [1.33-2.00], P = .001) and systole (0.87 [0.58-1.19] to 1.00 [0.74-1.36], P = .001) increased in dogs. Incidence of volume overload did not differ before (14/20 cats, 70%; 9/20 dogs, 45%) or after (12/20 cats, 60%; 11/20 dogs, 55%) transfusion (P = .64). CONCLUSIONS AND CLINICAL IMPORTANCE: Blood transfusion is well tolerated when signs of volume overload are present before transfusion.

Investigation of optimal sample preparation conditions with potassium triiodide and optimal imaging settings for microfocus computed tomography of excised cat hearts.

OBJECTIVE: To determine optimal sample preparation conditions with potassium triiodide (I2KI) and optimal imaging settings for microfocus CT (micro-CT) of excised cat hearts. SAMPLE: 7 excised hearts (weight range, 10 to 17.6 g) obtained from healthy adult cats after euthanasia by IV injection of pentobarbital sodium. PROCEDURES: Following excision, the hearts were preserved in 10% formaldehyde solution. Six hearts were immersed in 1.25% I2KI solution (n = 3) or 2.5% I2KI solution (3) for a 12-day period. Micro-CT images were acquired at time 0 (prior to iodination) then approximately every 24 and 48 hours thereafter to determine optimal sample preparation conditions (ie, immersion time and concentration of I2KI solution). Identified optimal conditions were then used to prepare the seventh heart for imaging; changes in voltage, current, exposure time, and gain on image quality were evaluated to determine optimal settings (ie, maximal signal-to-noise and contrast-to-noise ratios). Images were obtained at a voxel resolution of 30 µm. A detailed morphological assessment of the main cardiac structures of the seventh heart was then performed. RESULTS: Immersion in 2.5% I2KI solution for 48 hours was optimal for sample preparation. The optimal imaging conditions included a tube voltage of 100 kV, current of 150 µA, and exposure time of 354 milliseconds; scan duration was 12 minutes. CONCLUSIONS AND CLINICAL RELEVANCE: Results provided an optimal micro-CT imaging protocol for excised cat hearts prepared with I2KI solution that could serve as a basis for future studies of micro-CT for high resolution 3-D imaging of cat hearts.

ACVIM consensus guidelines for the diagnosis and treatment of myxomatous mitral valve disease in dogs.

This report, issued by the ACVIM Specialty of Cardiology consensus panel, revises guidelines for the diagnosis and treatment of myxomatous mitral valve disease (MMVD, also known as endocardiosis and degenerative or chronic valvular heart disease) in dogs, originally published in 2009. Updates were made to diagnostic, as well as medical, surgical, and dietary treatment recommendations. The strength of these recommendations was based on both the quantity and quality of available evidence supporting diagnostic and therapeutic decisions. Management of MMVD before the onset of clinical signs of heart failure has changed substantially compared with the 2009 guidelines, and new strategies to diagnose and treat advanced heart failure and pulmonary hypertension are reviewed.

Arrhythmogenic right ventricular cardiomyopathy in Boxer dogs: the diagnosis as a link to the human disease.

BACKGROUND: Arrhythmogenic right ventricular cardiomyopathy (ARVC) is a myocardial disease with an increased risk for ventricular arrhythmias. The condition, which occurs in Boxer dogs, shares phenotypic features with the human disease arrhythmogenic cardiomyopathy (ACM) suggesting its potential as a natural animal model. However, there are currently no universally accepted clinical criteria to diagnose ARVC in Boxer dogs. We aimed to identify diagnostic criteria for ARVC in Boxer dogs defining a more uniform and consistent phenotype. METHODS AND RESULTS: Clinical records from 264 Boxer dogs from a referral veterinary hospital were retrospectively analysed. ARVC was initially diagnosed according to the number of ventricular premature complexes (VPCs) in the 24-hour-Holter-ECG in the absence of another obvious cause. Dogs diagnosed this way had more VPCs, polymorphic VPCs, couplets, triplets, VTs and R-on-T-phenomenon and syncope, decreased right ventricular function and dilatation in comparison to a control group of all other Boxer dogs seen by the Cardiology Service over the same period. Presence of couplets and R-on-T-phenomenon on a 24h-ECG were identified as independent predictors of the diagnosis. A diagnosis based on ≥100 VPCs in 24 hours, presence of couplets and R-on-T phenomenon on a 24h-ECG was able to select Boxer dogs with a phenotype most similar to human ACM. CONCLUSION: We suggest the diagnosis of ARVC in Boxer dogs requires two out of the three following criteria: presence of ≥ 100 VPCs, presence of couplets or R-on-T-phenomenon on a 24 h-ECG. This results in a uniform phenotype similar to that described in human ACM and may result in the adoption of the term ACM for this analogous condition in Boxer dogs.

Echocardiographic predictors of survival in dogs with myxomatous mitral valve disease.

OBJECTIVES: To evaluate vena contracta and other echocardiographic measures of myxomatous mitral valve disease (MMVD) severity in a multivariable analysis of survival in dogs. ANIMALS: 70 dogs diagnosed with MMVD from stored echocardiographic images that met study inclusion criteria. METHODS: Left heart dimensions were measured as well as mitral regurgitant jet area/left atrial area (JAR), early mitral filling velocity (Evel), extent of mitral valve prolapse in right and left views (ProlR, ProlL), Prol indexed to aortic diameter (ProlR:Ao, ProlL:Ao), presence of a flail leaflet (FlailR, FlailL), and mitral regurgitation vena contracta diameter (VCR, VCL) indexed to aortic diameter (VCR:Ao, VCL:Ao). Follow-up from referring veterinarians was obtained by questionnaire or telephone to determine survival times. Inter- and intra-observer agreement was evaluated with Bland-Altman plots and weighted Kappa analysis. Survival was analyzed using Kaplan-Meier curves, logrank tests and Cox's proportional hazards. RESULTS: Logrank analysis showed VCL:Ao, VCR:Ao, FlailL, ProlR:Ao, ProlL:Ao, left ventricular internal dimension in diastole indexed to aortic diameter (LVIDD:Ao) >2.87, left atrium to aorta ratio (LA/Ao) >1.6, and Evel >1.4 m/s were predictors of cardiac mortality. In a multivariable analysis, the independent predictors of cardiac mortality were Evel >1.4 m/s [hazard ratio (HR) 5.0, 95% confidence interval (CI) 2.5-10.3], FlailL (HR 3.1, 95% CI 1.3-7.9), and ProlR:Ao (HR 2.8, 95% CI 1.3-6.3). CONCLUSIONS: Echocardiographic measures of mitral regurgitation severity and mitral valve pathology provide valuable prognostic information independent of chamber enlargement in dogs with MMVD.

The influence of clinical and genetic factors on left ventricular wall thickness in Ragdoll cats.

OBJECTIVES: To investigate the effect of various genetic and environmental modifiers on left ventricular (LV) wall thickness in a cohort of cats genotyped for the myosin binding protein C3 mutation (MYBPC3). ANIMALS: Sixty-four Ragdoll cats. METHODS: All cats were screened for HCM with echocardiography and genotyping for the HCM-associated MYBPC3:R820W mutation. Cats were also genotyped for previously identified variant polymorphisms of the angiotensin-converting enzyme (ACE) and cardiac beta-adrenergic receptor (ADRB1) genes. Plasma N-terminal pro-B-type natriuretic peptide and cardiac troponin I were also measured. Associations were evaluated between genotype (MYBPC3 negative/positive, and ACE and ADRB1 negative/heterozygous/homozygous), patient factors (body weight, age and sex) and echocardiographic measurements of LV wall thickness. RESULTS: Male cats had greater maximum wall thickness (LVmax; 5.8 mm, IQR 5.1-6.4 mm) than females (4.7 mm, IQR 4.4-5.3 mm, p = 0.002). Body weight positively correlated with LVmax (ρ = 0.604, p < 0.001). The MYBPC3:R820W-positive cats had a greater LVmax (5.44 mm, IQR 4.83-6.28 mm) than the negative cats (4.76 mm, IQR 4.36-5.32 mm, p = 0.001). Also, the ACE polymorphism genotype was associated with LVmax: the homozygous cats (5.37 mm, IQR 5.14-6.4 mm) had greater LVmax than the heterozygous cats (4.73 mm, IQR 4.41-5.55 mm, p = 0.014). Only the MYBPC3 genotype and body weight were independently associated with wall thickness in multivariable analysis. CONCLUSIONS: This study provides evidence that the MYBPC3:R820W mutation is independently associated with LV wall thickness in Ragdoll cats. Body weight is also independently associated with maximum LV wall thickness, but is not currently accounted for in HCM screening. In addition, other genetic modifiers may be associated with variation in LV wall thickness in Ragdolls.

Management of incidentally detected heart murmurs in dogs and cats.

A dog or a cat has an incidentally detected heart murmur if the murmuris an unexpected discovery during a veterinary consultation that was not initially focused on the cardiovascular system. This document presents approaches for managing dogs and cats that have incidentally-detected heart murmurs, with an emphasis on murmur characteristics, signalment profiling, and multifactorial decision-making to choose an optimal course for a given patient.

Effect of treatment with atenolol on 5-year survival in cats with preclinical (asymptomatic) hypertrophic cardiomyopathy.

OBJECTIVES: To investigate the effect of treatment with atenolol on 5-year survival in cats with preclinical hypertrophic cardiomyopathy (HCM). ANIMALS: 63 Client-owned cats with preclinical HCM and 31 healthy control cats. METHODS: Prospective, observational, open-label, clinical cohort study. Cats with HCM were diagnosed by echocardiography, treated with atenolol (6.25-12.5 mg q12h, PO; n = 42) or untreated (n = 21), and were observed for 5 years after enrollment. The study end point was death from any cause. Cats of similar body weight, age, gender, and breed without evidence of heart disease were studied concurrently and served as controls. RESULTS: During the observational period, 27 cats with HCM died; 14 (22%) due to cardiac disease and 13 (21%) due to non-cardiac disease. Ten control cats (32%) died of non-cardiac disease. There was no significant difference (P = 0.307) in all-cause mortality between control and HCM. Cardiac mortality was higher in cats with HCM compared to control cats (P = 0.005). There was no significant difference in all-cause mortality (P = 0.729) and cardiac mortality (P = 0.897) between cats with HCM treated or untreated with atenolol. Age and left atrial size at diagnosis were the only predictors of 5-year outcome. CONCLUSIONS: Our study failed to demonstrate an effect of atenolol on 5-year survival in cats with preclinical HCM.

Evaluation of a technique for measurement of flow-mediated vasodilation in healthy ponies.

OBJECTIVE: To determine between-pony and within-pony variations and interobserver and intraobserver agreements of a technique for measurement of flow-mediated vasodilation (FMD) in healthy ponies. ANIMALS: 6 healthy pony mares (weight range, 236 to 406 kg; body condition score range, 3/9 to 7/9; age range, 14 to 25 years). PROCEDURES: In each pony, the left median artery was occluded with a blood pressure cuff (inflated to > 300 mm Hg for 5 minutes). Two-dimensional ultrasonographic images of the artery were recorded for 30 seconds before cuff inflation and for 2 minutes after cuff deflation. Maximum luminal diameters of arteries were compared with their baseline diameters to calculate FMD (relative percentage increase in luminal size). Images were obtained from 6 ponies 1 time and from 1 pony 6 times. Independent analysis of images was performed by 2 investigators, 1 of whom analyzed images on 2 occasions. RESULTS: Mean ± SD FMD in 6 ponies (1 time) was 12.57 ± 4.28% and in 1 pony (6 times) was 7.30 ± 2.11%. Between-pony and within-pony coefficients of variation were 34.09% and 28.84%, respectively. Interobserver agreement was fair (intraclass correlation coefficient, 0.47); intraobserver agreement was poor (intraclass correlation coefficient, 0.30). CONCLUSIONS AND CLINICAL RELEVANCE: FMD was identified and measured in ponies. Measurement of FMD is used to assess endothelial function in humans and has been investigated in dogs. Measurement of FMD in ponies appeared to be feasible and could be used to assess endothelial function (to determine predisposition for development of laminitis or cardiovascular diseases).

Effects of sedation on echocardiographic variables of left atrial and left ventricular function in healthy cats.

Although sedation is frequently used to facilitate patient compliance in feline echocardiography, the effects of sedative drugs on echocardiographic variables have been poorly documented. This study investigated the effects of two sedation protocols on echocardiographic indices in healthy cats, with special emphasis on the assessment of left atrial size and function, as well as left ventricular diastolic performance. Seven cats underwent echocardiography (transthoracic two-dimensional, spectral Doppler, color flow Doppler and tissue Doppler imaging) before and after sedation with both acepromazine (0.1 mg/kg IM) and butorphanol (0.25 mg/kg IM), or acepromazine (0.1 mg/kg IM), butorphanol (0.25 mg/kg IM) and ketamine (1.5 mg/kg IV). Heart rate increased significantly following acepromazine/butorphanol/ketamine (mean±SD of increase, 40±26 beats/min) and non-invasive systolic blood pressure decreased significantly following acepromazine/butorphanol (mean±SD of decrease, 12±19 mmHg). The majority of echocardiographic variables were not significantly different after sedation compared with baseline values. Both sedation protocols resulted in mildly decreased left ventricular end-diastolic dimension and mildly increased left ventricular end-diastolic wall thickness. This study therefore failed to demonstrate clinically meaningful effects of these sedation protocols on echocardiographic measurements, suggesting that sedation with acepromazine, butorphanol and/or ketamine can be used to facilitate echocardiography in healthy cats.

Ultrasonographic measurement of flow-mediated vasodilation in dogs with chronic valvular disease.

OBJECTIVES: To measure flow-mediated vasodilation (FMD) in healthy dogs and in client-owned dogs with chronic valvular disease (CVD) and to investigate possible correlations between markers of CVD severity and FMD. ANIMALS: Twelve dogs with CVD and 11 healthy weight-matched dogs. METHODS: Brachial artery FMD following 5 min inflation of a cuff around the antebrachium was measured in 12 dogs with CVD and 11 healthy weight-matched dogs. Measurements were also obtained in the healthy dogs 5 min after cuff placement but without inflation ('sham cuff placement'). Dogs with CVD underwent echocardiography to confirm and characterize their disease. RESULTS: In healthy dogs (median age 4 [2-6] years), median FMD was 7.7% versus 3.4% with sham cuff placement (P = 0.003). In dogs with CVD (median age 8 [4-16] years) median FMD was 5.5% versus 7.7% in healthy dogs (P = 0.131). FMD showed an inverse correlation with left ventricular end-diastolic diameter normalized for body weight (r = -0.76, P = 0.0043). CONCLUSIONS: Brachial FMD in dogs with early CVD inversely correlates with severity of left ventricular remodelling.

Comparison of flow-mediated vasodilation in femoral and brachial arteries in healthy dogs.

OBJECTIVE: To compare flow-mediated vasodilation (FMD) measurements in brachial and femoral arteries of healthy dogs habituated to the assessment method, evaluate repeatability of these measurements, and investigate effects of blood pressure cuff inflation time on femoral artery FMD measurements. ANIMALS: 11 healthy adult Miniature Schnauzers. PROCEDURES: Arterial luminal diameter and blood flow velocity integral (FVI) were measured before and after cuff inflation of 5 minutes' (brachial and femoral arteries) or 3 minutes' duration (femoral artery) in separate experiments. A blood pressure cuff was inflated to > 200 mm Hg distal to each imaging site to increase local blood flow to induce reactive hyperemia. Changes in FVI after cuff deflation, FMD, and between-dog and within-dog coefficients of variation (CVs) were determined. RESULTS: After cuff inflation of 5 minutes' duration, greater changes were detected in median change in FVI and FMD of brachial arteries (174.0% and 8.0%, respectively), compared with values determined for femoral arteries (32.0% and 2.1%, respectively). Between-dog CV for brachial artery FMD was 34.0%, compared with 89.6% for femoral arteries, and within-dog CV was 32.5% for brachial arteries versus 51.6% for femoral arteries after cuff inflation of 5 minutes' duration. CONCLUSIONS AND CLINICAL RELEVANCE: In healthy Miniature Schnauzers, FMD was greater and more repeatable in brachial arteries than in femoral arteries. Reactive hyperemia was inconsistently induced in femoral arteries following 3- or 5-minute cuff inflation times. Brachial, but not femoral, artery FMD measurement is a potentially useful research technique for measurement of endothelial function in dogs.

Gerbode defect associated with blunt trauma in a dog.

A 6 year-old Labrador retriever was presented after being struck by a car. A ventricular arrhythmia, attributed to myocardial trauma, developed 12 h post-trauma. Echocardiography revealed lesions consistent with a subaortic paramembranous ventricular septal defect (VSD) with shunting of blood from the left ventricle to the right atrium (Gerbode defect). A right-to-left shunting atrial septal defect (ASD) was visualised. Pleural and peritoneal effusions developed within 48 h. Fifteen days post-trauma flow across the ASD was left-to-right while left-to-right shunting across the VSD persisted. No cavitary effusions were detected at 15 days post-trauma or subsequently.

The effect of protease inhibition on the temporal stability of NT-proBNP in feline plasma at room temperature.

OBJECTIVES: To evaluate whether the addition of a protease inhibitor (PI) to feline plasma improves the temporal stability of N-terminal pro-B-type natriuretic peptide (NT-proBNP). ANIMALS, MATERIALS AND METHODS: Forty-four EDTA blood samples were collected from 42 cats with cardiac disease or hyperthyroidism. Samples were separated within 15 min of collection and the plasma stored at -80 °C. Samples were thawed and each separated into 2 aliquots, one of which was mixed with PI. Each was subdivided into 5 aliquots, which remained at room temperature (RT) for zero, 24, 48, 96 or 120 h before being returned to storage at -80 °C. NT-proBNP was measured using a commercially-available ELISA. RESULTS: There was no difference in NT-proBNP measurements between plain and PI samples at time zero (P = 0.836) or 24 h (P = 0.293). At subsequent time-points NT-proBNP was higher in the PI than the plain samples (P < 0.05). An overall effect of time at RT was detected for plain (P < 0.001) and PI samples (P < 0.001). CONCLUSIONS: The addition of PI to feline plasma significantly reduces, but does not eliminate, degradation of NT-proBNP at RT. This degradation occurs over a time course comparable to postal transportation of samples for laboratory analysis and may alter the clinical interpretation of results.

Comparison of auscultatory and echocardiographic findings in healthy adult cats.

OBJECTIVES: This pilot study was performed to investigate murmur prevalence and to explore the association between auscultatory and echocardiographic findings in apparently healthy cats in order to design a larger study. ANIMALS, MATERIALS AND METHODS: Adult cats in 4 rehoming centres were screened by auscultation and echocardiography (echo) over 2 periods of 2 weeks each. In the first period, echo was attempted only in cats with murmurs. In the second period, all cats underwent auscultation by 2 observers and echo. LVH was defined in 5 ways: maximal diastolic left ventricular (LV) wall thickness ≥ 6 mm or ≥ 5.5 mm with 2D (LVH(6 2D), LVH(5.5 2D), respectively) or M-Mode echo (LVH(6 MM) or LVH(5.5 MM) respectively), or LV wall thickness ≥ 6 mm (2D) for >50% of a wall segment (LVH(50%)). RESULTS: 67/199 (34%) cats had a murmur. Interobserver agreement on murmur presence was moderate (κ 0.47). 61 cats with a murmur and 31 cats without underwent both auscultation and echo. Depending on the criteria, LVH was present in 31 (LVH(6 2D)), 21 (LVH(50%)) and 11 (LVH(6 MM)) scanned cats. 18-62% of cats with murmurs had LVH, depending on the echo criteria used. Agreement was best between observers in identifying LVH using LVH(6 2D) and LVH(50%) (κ = 1.0). CONCLUSIONS: Heart murmurs are common in apparently healthy cats. The prevalence of LVH varies depending on the criteria used.