Artificial intelligence in veterinary diagnostic imaging: Perspectives and limitations.

The field of veterinary diagnostic imaging is undergoing significant transformation with the integration of artificial intelligence (AI) tools. This manuscript provides an overview of the current state and future prospects of AI in veterinary diagnostic imaging. The manuscript delves into various applications of AI across different imaging modalities, such as radiology, ultrasound, computed tomography, and magnetic resonance imaging. Examples of AI applications in each modality are provided, ranging from orthopaedics to internal medicine, cardiology, and more. Notable studies are discussed, demonstrating AI's potential for improved accuracy in detecting and classifying various abnormalities. The ethical considerations of using AI in veterinary diagnostics are also explored, highlighting the need for transparent AI development, accurate training data, awareness of the limitations of AI models, and the importance of maintaining human expertise in the decision-making process. The manuscript underscores the significance of AI as a decision support tool rather than a replacement for human judgement. In conclusion, this comprehensive manuscript offers an assessment of the current landscape and future potential of AI in veterinary diagnostic imaging. It provides insights into the benefits and challenges of integrating AI into clinical practice while emphasizing the critical role of ethics and human expertise in ensuring the wellbeing of veterinary patients.

Non-contact detection method of pregnant sows backfat thickness based on two-dimensional images.

Since sow backfat thickness (BFT) is highly correlated with its service life and reproductive effectiveness, dynamic monitoring of BFT is a critical component of large-scale sow farm productivity. Existing contact measures of sow BFT have their problems including, high measurement intensity and sows' stress reaction, low biological safety, and difficulty in meeting the requirements for multiple measurements. This article presents a two-dimensional (2D) image-based approach for determining the BFT of pregnant sows when combined with the backfat growth rate (BGR). The 2D image features of sows extracted by convolutional neural networks (CNN) and the artificially defined phenotypic features of sows such as hip width, hip height, body length, hip height-width ratio, length-width ratio, and waist-hip ratio, were used respectively, combined with BGR, to construct a prediction model for sow BFT using support vector regression (SVR). Following testing and comparison, it was shown that using CNN to extract features from images could effectively replace artificially defined features, BGR contributed to the model's accuracy improvement. The CNN-BGR-SVR model performed the best, with R2 of 0.72 and mean absolute error of 1.21 mm, and root mean square error of 1.50 mm, and mean absolute percentage error of 7.57%. The results demonstrated that the CNN-BGR-SVR model based on 2D images was capable of detecting sow BFT, establishing a new reference for non-contact sow BFT detection technology.

A novel rabbit fixator made of a thermoplastic mask for awake imaging experiments.

This study aimed to develop and validate a novel rabbit fixator made from a thermoplastic mask for awake imaging experiments. When heated in a hot-water bath at 65-70 °C for 2-5 min, the thermoplastic mask became soft and could be molded to fit over the entire body of an anesthetized rabbit (4 ml of 3% pentobarbital sodium solution by intramuscular injection). Twenty rabbits were randomly divided into fixator (n = 10) and anesthesia (n = 10) groups. The animals' vital signs, stress hormones (cortisol and adrenaline), and subjective image quality scores for the computed tomography (CT), positron emission tomography (PET), and magnetic resonance imaging (MRI) scanning were measured and compared. Phantom CT, MRI and PET studies were performed to assess the performance with and without the thermoplastic mask by using image agents at different concentrations or with different radioactivity. The respiration rate (RR), systolic blood pressure (SBP), diastolic blood pressure (DBP), peripheral capillary oxygen saturation (SpO2) and body temperature (T) decreased after anesthesia (all P < 0.05) but did not significantly decrease after fixation (all P > 0.05). The heart rate (HR), cortisol and adrenaline did not significantly decrease after either anesthesia or fixation (all P > 0.05). The subjective image quality scores for the CT and MRI images of the head, thorax, liver, kidney, intestines and pelvis and the subjective image quality scores for the PET images did not significantly differ between the two groups (all P > 0.05). For all examined organs except the muscle, 18F-FDG metabolism was lower after fixation than after anesthesia, and was almost identical of liver between two groups. The phantom study showed that the CT values, standard uptake values and MR T2 signal values did not differ significantly with or without the mask (all P > 0.05). A novel rabbit fixator created using a thermoplastic mask could be used to obtain high-quality images for different imaging modalities in an awake and near-physiological state.

Influence of radiographic examination findings on recommendations made during routine clinical re-evaluation of dogs with uncomplicated tibial plateau leveling osteotomy.

OBJECTIVE: To determine the influence of follow-up radiographic examination on recommendations made during routine clinical re-evaluation of dogs that had undergone uncomplicated tibial plateau leveling osteotomy (TPLO). STUDY DESIGN: Retrospective multi-institutional case series. ANIMALS: Client-owned dogs (N = 1010) that underwent uncomplicated TPLO. METHODS: Records from 11 institutions were searched for dogs that had been treated with unilateral TPLO and had no history of postoperative complications before their routine follow-up examination. The frequency of change in further clinical recommendations resulting from client- or clinician-voiced concerns or radiographic abnormalities was investigated. RESULTS: Follow-up evaluation was performed at a median of 6 (range, 4-15) weeks after TPLO. Radiographic examination findings contributed to a change in recommendations in 4.15% (38/915) of dogs presented without client concerns and without abnormalities at orthopedic examination. Abnormal radiographic findings alone influenced the management of 3.76% (38/1010) of dogs. An association was detected between clinical features and radiological findings leading to a change in recommendations (P < .0001). Administration of analgesia at the time of follow-up was associated with radiographic abnormalities (P = .017) and change in postoperative plans (P = .0007). CONCLUSION: Radiographic examination findings at follow-up did not influence the management of most dogs with uncomplicated TPLO. CLINICAL SIGNIFICANCE: Radiographic examination findings are unlikely to influence the treatment of dogs that seem to be recovering uneventfully from an uncomplicated TPLO without concerns from clients, analgesia, or abnormal findings on thorough orthopedic examination by a surgical specialist, at the time of the planned clinical re-evaluation.

Technological Advances in Diagnostic Imaging in Exotic Pet Medicine.

Diagnostic imaging relies on interpretation of interactions between the body tissue and various energies, such as x-rays, ultrasound, and magnetic or nuclear energies, to differentiate normal from abnormal tissues. Major technological improvements regarding emission and detection of the energetic waves, as well as reconstruction and interpretation of the images, have occurred. These advances made possible visualization of smaller structures, quantitative evaluation of functional processes, and development of unique imaging-guided procedures. This article reviews the technological advances that allowed development of cone beam computed tomography, dual-energy x-ray absorptiometry, and contrast-enhanced ultrasonography, which all could have applications in exotic pet medicine.

Twitter in the Veterinary Diagnostic Imaging Classroom: Examination Outcomes and Student Views.

Radiographic lesion identification and differential diagnosis list generation can be difficult for veterinary students; thus, a novel means of distributing cases for study could improve students' engagement and learning. The goal of this study was to determine whether using Twitter as an adjunct means of studying diagnostic imaging would improve student outcomes on the final exam for a radiology course. A secondary goal was to determine students' preferred means of accessing additional cases for study. Twitter was used in a third-year veterinary radiology course to provide additional optional radiographic cases that were relevant to the topics covered in the course. At the end of the semester, students completed a survey to report their prior and current use of Twitter and to give preferences as to further distribution of optional cases. Mean final examination scores were compared between students who used Twitter in their studies and those who did not. No significant difference was found between the mean final examination score for each group (22.2; p = .98). Only 3% of respondents ( n = 2/79) preferred Twitter as a means of receiving additional radiographic cases; Moodle (the Web platform for classwork used at this institution) and Facebook were the most preferred platforms for further cases, receiving 41% ( n = 32/79) and 23% ( n = 18/79) of votes, respectively. Educational use of Twitter did not improve student examination performance in diagnostic imaging, and other media platforms may be more beneficial than Twitter for encouraging student use of additional resources.

Peripherally acting α-adrenoceptor antagonist MK-467 with intramuscular medetomidine and butorphanol in dogs: A prospective, randomised, clinical trial.

The aim of this study was to investigate the clinical usefulness of MK-467 (vatinoxan; L-659'066) in dogs sedated for diagnostic imaging with medetomidine-butorphanol. It was hypothesised that MK-467 would alleviate bradycardia, hasten drug absorption and thus intensify the early-stage sedation. In a prospective, randomised, blinded clinical trial, 56 client-owned dogs received one of two IM treatments: (1) 0.5mg/m2 medetomidine+0.1mg/kg butorphanol (MB, n=29); or (2) 0.5mg/m2 medetomidine+0.1mg/kg butorphanol+10mg/m2 MK-467 (MB-MK, n=27). Heart rates and visual sedation scores were recorded at intervals. Plasma drug concentrations were determined in venous samples obtained approximately 14min after injection. Additional sedation (50% of original dose of medetomidine IM) and/or IM atipamezole for reversal were given when needed. The area under the sedation score-time curve for visual analogue scale (AUCVAS30) was calculated for the first 30min after treatment using the trapezoidal method. Repeated ANOVA, Mann-Whitney U test and Fisher's exact test were used for parametric, non-parametric and dichotomous data. Heart rate was significantly higher from 10 to 40min with MB-MK than with MB. AUCVAS30 was significantly higher after MB-MK. More dogs treated with MB-MK required additional sedation after 30min, but fewer needed atipamezole for reversal compared with MB. Plasma concentrations of both medetomidine and butorphanol were higher after MB-MK. All procedures were successfully completed. MK-467 alleviated the bradycardia, intensified the early stage sedation and shortened its duration in healthy dogs that received IM medetomidine-butorphanol.

Randomised clinical trial comparing clinically relevant sedation outcome measures in dogs after intramuscular administration of medetomidine in combination with midazolam or butorphanol for routine diagnostic imaging procedures.

The aim of this study was to investigate the sedative effects of medetomidine in combination with midazolam or butorphanol for routine imaging procedures in dogs. Eighty client owned dogs were recruited in a prospective, randomised, blinded clinical study and randomly assigned to receive one of four treatments intramuscularly (IM): (1) 30µg/kg medetomidine (Med30); (2) 20µg/kg medetomidine combined with 0.3mg/kg butorphanol (Med20But0.3); (3) 20µg/kg medetomidine combined with 0.3mg/kg midazolam (Med20Mid0.3); and (4) 10µg/kg medetomidine combined with 0.3mg/kg midazolam (Med10Mid0.3). The level of sedation was evaluated using a composite sedation scale assessed by one investigator (0=no sedation, 15=profound sedation). The number of dogs deemed to be adequately clinically sedated and the dose of propofol administered as rescue sedation were recorded. Mean±standard deviation sedation scores at 30min after the commencement of treatment in the groups that received Med20But0.3 (9.8±4) and Med20Mid0.3 (8.9±4.4) were not statistically significantly different from each other, but were significantly different from the group receiving Med10Mid0.3 (5.6±3.6). Only Med20But0.3 was significantly associated with adequate clinical sedation, while Med10Mid0.3 was associated with 85% sedation failure. The rescue sedation dose of propofol (1.5±1mg/kg) for the Med10Mid0.3 group was significantly higher than for other treatments. A sedation score≥10 out of 15 was a satisfactory cut-off to predict adequate clinical sedation. In healthy dogs, the combination of medetomidine with midazolam did not provide comparable sedation to the same dose of medetomidine in combination with butorphanol in a clinical setting.

A Diagnostic Approach for Rodent Progressive Cardiomyopathy and Like Lesions in Toxicology Studies up to 28 Days in the Sprague Dawley Rat (Part 2 of 2).

To test the diagnostic approach described in part 1 of this article, 2 exercises were completed by pathologists from multiple companies/agencies. Pathologist's examination of whole slide image (WSI) heart sections from rats using personal diagnostic approaches (exercise #1) corroborated conclusions from study #1. Using the diagnostic approach described in part 1, these pathologists examined the same WSI heart sections (exercise #2) to determine whether that approach increased consistency of diagnosis of rodent progressive cardiomyopathy (PCM) lesions. In exercise #2, there was improved consistency of categorization of small borderline morphologies and mild lesions, but a decrement in consistency of categorizing minimal lesions. Exercises 1 and 2 suggest the described diagnostic approach is representative of that in use by the majority of toxicologic pathologists across companies/agencies and that application by all may improve diagnostic consistency of PCM/like lesions. Additionally, a criterion of approximately 5% heart section involvement is suggested for separating mild from moderate or greater severity. While evidence is not absolute, until further investigation shows otherwise, microscopic changes resembling PCM, but located in the epicardial and subepicardial region of the right ventricle, may be considered as part of the spectrum of PCM.

A Diagnostic Approach for Rodent Progressive Cardiomyopathy and Like Lesions in Toxicology Studies up to 28 Days in the Sprague Dawley Rat (Part 1 of 2).

Spontaneous rodent progressive cardiomyopathy (PCM) in the Sprague Dawley rat may confound identification and/or interpretation of potential test article (TA)-related cardiotoxicity. Pathologists apply diagnostic term(s) and thresholds for diagnosing and assigning severity grades for PCM and/or PCM-like (PCM/like) lesions consistently within a study, which is necessary to identify and interpret TA-related findings. Due to differences in training and/or experiences, diagnostic terms and thresholds may vary between pathologists. Harmonized terminology and thresholds across studies will generate better historical control data, will likely enhance interpretation of study data, and may further enhance our understanding of the spontaneous change. An assessment of the diagnostic approaches of a group of 37 pathologists identified an approach that is relatively easily applied; and if adopted, it could enhance diagnostic consistency across studies. This approach uses the single "slash" term "necrosis/inflammatory cell infiltrate (NICI)" as the diagnosis for the spectrum of lesions seen in younger rats, uses no threshold for diagnosis (e.g., diagnose all lesions clearly identifiable as PCM/like), and uses aggregate lesion size of approximately ≥45% of the field of view (FOV) using a 10×/22 eyepiece and the 40× objective or approximately ≥100% of the FOV using the 60× objective as the criterion separating minimal from mild severities.

RADIOGRAPHY IN THE FIELD: ASSESSING A LIGHTWEIGHT, HANDHELD, BATTERY-POWERED DENTISTRY UNIT FOR FIELD DIAGNOSTIC APPLICATIONS.

Radiography units are not used commonly in wildlife medicine field settings, primarily because of their weight and requirement for a power supply. In this study, a portable, battery-powered, and lightweight radiography unit, originally developed for dentistry, was assessed for its potential field applications. Radiographs of various animal species (ranging in weight from 14 g to 1,000 kg) were imaged using varying source image distance (SID) and exposure time. The quality of these images was evaluated for their resolution, image noise, and motion blur. When required, image resolutions were further enhanced using computed radiography postprocessing. Other parameters evaluated were the freehand use of the device, its battery durability, the maximum obtainable image size, and multiple use of a single computed radiography cassette. Using an SID of 60 cm, radiographs delivered adequate image quality. The quality, however, was found deteriorated in images of larger animals (>50 kg) or thicker tissues (>15 cm). The use of a tripod proved unnecessary in most cases, and its exclusion greatly facilitated equipment handling. Under field conditions, the battery was depleted after a total running time of 1.6 hr or 36 radiographs. The maximum size of a radiographic image reached a diameter of 40 cm, and radiation shielding allowed the multiple use of a single computed radiography cassette. Taken together, the radiography unit evaluated in this study presented a balanced compromise between portability and radiograph quality for field use. However, the unit image resolution cannot replace those of the fixed standard radiography units commonly used in veterinary medicine.

FORENSIC RADIOLOGY AND IMAGING FOR VETERINARY RADIOLOGISTS.

Imaging studies are often of evidentiary value in medicolegal investigations involving animals and the role of the veterinary radiologist is to interpret those images for courts as an expert or opinion witness. With progressing interest in prosecuting animal crimes and strengthening of penalties for crimes against animals, the participation of veterinary radiologists in medicolegal investigations is expected to increase. Veterinary radiologists who are aware of radiographic and imaging signs that result in animal suffering, abuse, or neglect; knowledgeable in ways radiology and imaging may support cause of death determinations; conversant in postmortem imaging; comfortable discussing mechanisms and timing of blunt or sharp force and projectile trauma in imaging; and prepared to identify mimics of abuse can assist court participants in understanding imaging evidence. The goal of this commentary review is to familiarize veterinary radiologists with the forensic radiology and imaging literature and with the advantages and disadvantages of various imaging modalities utilized in forensic investigations. Another goal is to provide background information for future research studies in veterinary forensic radiology and imaging.