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.

Retrospective evaluation of hypertrophic cardiomyopathy in 68 dogs.

BACKGROUND: There is a lack of clinical data on hypertrophic cardiomyopathy (HCM) in dogs. HYPOTHESIS/OBJECTIVES: To investigate signalment, clinical signs, diagnostic findings, and survival in dogs with HCM. ANIMALS: Sixty-eight client-owned dogs. METHODS: Retrospective multicenter study. Medical records were searched between 2003 and 2015. The diagnosis of left ventricular (LV) hypertrophy was made by echocardiographic examination. RESULTS: Three hundred and forty-five dogs with LV hypertrophy were identified, of which 277 were excluded. The remaining 68 dogs were 0.3 to 14 years old and predominantly <10 kg (85%), and without a sex predilection. Twenty-four % were Shih Tzu and 24% terrier breeds. Most (80%) had a systolic heart murmur. Owner-determined exercise intolerance (37%) and syncope (18%) were most commonly reported signs. The majority (84%) of dogs had symmetrical LV hypertrophy, whereas asymmetrical septal and LV free wall hypertrophy was observed in 9% and 6% of dogs, respectively. Isolated basal interventricular septal hypertrophy was not observed. Commonly recorded were systolic anterior motion of the mitral valve (60%) and LV diastolic dysfunction (89% of dogs where diastolic function was evaluated). Six dogs died unexpectedly, and 3 developed congestive heart failure. Known survival times were between 1 day and 114 months after diagnosis. CONCLUSIONS AND CLINICAL IMPORTANCE: Hypertrophic cardiomyopathy in dogs should be considered as a differential diagnosis if LV hypertrophy is identified. Small breed dogs are overrepresented, and it is uncommon for dogs with HCM to develop CHF although sudden death can occur.

Case report - azygos vein drainage into the left atrium in a dog with cor triatriatum dexter and a patent foramen ovale.

A 2-year-old Airedale terrier was presented with exercise intolerance since birth and newly developed chylous pleural effusion. Imaging procedures including echocardiography, cardiac magnetic resonance imaging, computed tomography, and selective angiography revealed an aberrant connection of the azygos vein and the left atrium, a membrane in the right atrium consistent with cor triatriatum dexter, and a patent foramen ovale with right-to-left shunt. Balloon dilation of the membrane in the right atrium seemed to result in transient improvement of exercise tolerance compared with the previous 2 years. When chylothorax relapsed after three months, the dog was euthanized. Necropsy confirmed the azygos vein to left atrial connection, the patent foramen ovale, and the cor triatriatum dexter.

Two-dimensional echocardiographic estimates of left atrial function in healthy dogs and dogs with myxomatous mitral valve disease.

OBJECTIVES: To provide reference intervals for 2-dimensional linear and area-based estimates of left atrial (LA) function in healthy dogs and to evaluate the ability of estimates of LA function to differentiate dogs with subclinical myxomatous mitral valve disease (MMVD) and similarly affected dogs with congestive heart failure (CHF). ANIMALS: Fifty-two healthy adult dogs, 88 dogs with MMVD of varying severity. METHODS: Linear and area measurements from 2-dimensional echocardiographs in both right parasternal long and short axis views optimized for the left atrium were used to derive estimates of LA active emptying fraction, passive emptying fraction, expansion index, and total fractional emptying. Differences for each estimate were compared between healthy and MMVD dogs (based on ACVIM classification), and between MMVD dogs with subclinical disease and CHF that had similar LA dimensions. Diagnostic utility at identifying CHF was examined for dogs with subclinical MMVD and CHF. Relationships with bodyweight were assessed. RESULTS: All estimates of LA function decreased with increasing ACVIM stage of mitral valve disease (p<0.05) and showed negative relationships with increasing LA size (all r2 values < 0.2), except for LA passive emptying fraction, which did not differ or correlate with LA size (p=0.4). However, no index of LA function identified CHF better than measurements of LA size. Total LA fractional emptying and expansion index showed modest negative correlations with bodyweight. CONCLUSIONS: Estimates of LA function worsen with worsening MMVD but fail to discriminate dogs with CHF from those with subclinical MMVD any better than simple estimates of LA size.

Comparison of left and right ventricular stroke volume of dogs calculated on the basis of morphology and blood flow determined by use of cardiac magnetic resonance imaging.

OBJECTIVE To compare stroke volume (SV) calculated on the basis of cardiac morphology determined by MRI and results of phase-contrast angiography (PCA) of ventricular inflow and outflow in dogs. ANIMALS 10 healthy Beagles. PROCEDURES Cardiac MRI was performed twice on each Beagle. Cine gradient echo sequences of both ventricles in short-axis planes were used for morphological quantification of SVs by assessment of myocardial contours. From the long-axis plane, SVs in 4-chamber and left ventricular 2-chamber views were acquired at end diastole and end systole. For calculation of SV on the basis of blood flow, PCA was performed for cardiac valves. RESULTS Mean ± SD values for SV quantified on the basis of blood flow were similar in all valves (aortic, 17.8 ± 4.1 mL; pulmonary, 17.2 ± 5.4 mL; mitral, 17.2 ± 3.9 mL; and tricuspid, 16.9 ± 5.1 mL). Morphological quantification of SV in the short-axis plane yielded significant differences between left (13.4 ± 2.7 mL) and right (8.6 ± 2.4 mL) sides. Morphological quantification of left ventricular SV in the long-axis plane (15.2 ± 3.3 mL and 20.7 ± 3.8 mL in the 4- and 2-chamber views) yielded variable results, which differed significantly from values for flow-based quantification, except for values for the morphological 4-chamber view and PCA for the atrioventricular valves, for which no significant differences were identified. CONCLUSIONS AND CLINICAL RELEVANCE In contrast to quantification based on blood flow, calculation on the basis of morphology for the short-axis plane significantly underestimated SV, probably because of through-plane motion and complex right ventricular anatomy.