Accuracy of deep learning to differentiate the histopathological grading of meningiomas on MR images: A preliminary study.

BACKGROUND: Grading of meningiomas is important in the choice of the most effective treatment for each patient. PURPOSE: To determine the diagnostic accuracy of a deep convolutional neural network (DCNN) in the differentiation of the histopathological grading of meningiomas from MR images. STUDY TYPE: Retrospective. POPULATION: In all, 117 meningioma-affected patients, 79 World Health Organization [WHO] Grade I, 32 WHO Grade II, and 6 WHO Grade III. FIELD STRENGTH/SEQUENCE: 1.5 T, 3.0 T postcontrast enhanced T1 W (PCT1 W), apparent diffusion coefficient (ADC) maps (b values of 0, 500, and 1000 s/mm2 ). ASSESSMENT: WHO Grade II and WHO Grade III meningiomas were considered a single category. The diagnostic accuracy of the pretrained Inception-V3 and AlexNet DCNNs was tested on ADC maps and PCT1 W images separately. Receiver operating characteristic curves (ROC) and area under the curve (AUC) were used to asses DCNN performance. STATISTICAL TEST: Leave-one-out cross-validation. RESULTS: The application of the Inception-V3 DCNN on ADC maps provided the best diagnostic accuracy results, with an AUC of 0.94 (95% confidence interval [CI], 0.88-0.98). Remarkably, only 1/38 WHO Grade II-III and 7/79 WHO Grade I lesions were misclassified by this model. The application of AlexNet on ADC maps had a low discriminating accuracy, with an AUC of 0.68 (95% CI, 0.59-0.76) and a high misclassification rate on both WHO Grade I and WHO Grade II-III cases. The discriminating accuracy of both DCNNs on postcontrast T1 W images was low, with Inception-V3 displaying an AUC of 0.68 (95% CI, 0.59-0.76) and AlexNet displaying an AUC of 0.55 (95% CI, 0.45-0.64). DATA CONCLUSION: DCNNs can accurately discriminate between benign and atypical/anaplastic meningiomas from ADC maps but not from PCT1 W images. LEVEL OF EVIDENCE: 2 Technical Efficacy: Stage 2 J. Magn. Reson. Imaging 2019;50:1152-1159.

A methodological approach for deep learning to distinguish between meningiomas and gliomas on canine MR-images.

BACKGROUND: Distinguishing between meningeal-based and intra-axial lesions by means of magnetic resonance (MR) imaging findings may occasionally be challenging. Meningiomas and gliomas account for most of the total primary brain neoplasms in dogs, and differentiating between these two forms is mandatory in choosing the correct therapy. The aims of the present study are: 1) to determine the accuracy of a deep convolutional neural network (CNN, GoogleNet) in discriminating between meningiomas and gliomas in pre- and post-contrast T1 images and T2 images; 2) to develop an image classifier, based on the combination of CNN and MRI sequence displaying the highest accuracy, to predict whether a lesion is a meningioma or a glioma. RESULTS: Eighty cases with a final diagnosis of meningioma (n = 56) and glioma (n = 24) from two different institutions were included in the study. A pre-trained CNN was retrained on our data through a process called transfer learning. To evaluate CNN accuracy in the different imaging sequences, the dataset was divided into a training, a validation and a test set. The accuracy of the CNN was calculated on the test set. The combination between post-contrast T1 images and CNN was chosen in developing the image classifier (trCNN). Ten images from challenging cases were excluded from the database in order to test trCNN accuracy; the trCNN was trained on the remainder of the dataset of post-contrast T1 images, and correctly classified all the selected images. To compensate for the imbalance between meningiomas and gliomas in the dataset, the Matthews correlation coefficient (MCC) was also calculated. The trCNN showed an accuracy of 94% (MCC = 0.88) on post-contrast T1 images, 91% (MCC = 0.81) on pre-contrast T1-images and 90% (MCC = 0.8) on T2 images. CONCLUSIONS: The developed trCNN could be a reliable tool in distinguishing between different meningiomas and gliomas from MR images.

Relationship of diagnostic accuracy of renal cortical echogenicity with renal histopathology in dogs and cats, a quantitative study.

BACKGROUND: Renal cortical echogenicity is routinely evaluated during ultrasonographic investigation of the kidneys. Both in dog and cat previous ex-vivo studies have revealed a poor correlation between renal echogenicity and corresponding lesions. The aim of this study was to establish the in-vivo relationship between renal cortical echogenicity and renal histopathology. RESULTS: Thirty-eight dogs and fifteen cats euthanized for critical medical conditions were included in the study. Ultrasonographic images of both kidneys were acquired ante mortem at standardized ultrasonographic settings. The echogenicity was quantified by means of Mean Gray Value (MGV) of the renal cortex measured with ImageJ. A complete histopathological examination of both kidneys was performed. Five kidneys were excluded because histopathology revealed neoplastic lesions. Only samples affected by tubular atrophy showed statistically different values in dog, and histopathology explained 13% of the total variance. MGV was not correlated neither to the degeneration nor to the inflammation scores. However, significant differences were identified between mildly and severely degenerated samples. Overall, the classification efficiency of MGV to detect renal lesions was poor with a sensitivity of 39% and a specificity of 86%. In cats, samples affected by both tubular vacuolar degeneration and interstitial nephritis were statistically different and histopathology explained 44% of the total variance. A linear correlation was evident between degeneration and MGV, whereas no correlation with inflammation was found. Statistically significant differences were evident only between normal and severely degenerated samples with a sensitivity of 54.17% and a specificity of 83.3% and MGV resulted scarce to discriminate renal lesions in this species. CONCLUSIONS: Renal cortical echogenicity shows low relevance in detecting chronic renal disease in dog whereas it results worth to identify severe renal damage in cat.

Normal computed tomographic features and reference values for the coelomic cavity in pet parrots.

BACKGROUND: The increasing popularity gained by pet birds over recent decades has highlighted the role of avian medicine and surgery in the global veterinary scenario; such a need for speciality avian medical practice reflects the rising expectation for high-standard diagnostic imaging procedures. The aim of this study is to provide an atlas of matched anatomical cross-sections and contrast-enhanced CT images of the coelomic cavity in three highly diffused psittacine species. RESULTS: Contrast-enhanced computed tomographic studies of the coelomic cavity were performed in 5 blue-and-gold macaws, 4 African grey parrots and 6 monk parakeets by means of a 4-multidetector-row CT scanner. Both pre- and post-contrast scans were acquired. Anatomical reference cross-sections were obtained from 5 blue-and-gold macaw, 7 African grey parrot, and 9 monk parakeet cadavers. The specimens were stored in a -20 °C freezer until completely frozen and then sliced at 5-mm intervals by means of a band saw. All the slices were photographed on both sides. Individual anatomical structures were identified by means of the available literature. Pre- and post-contrast attenuation reference values for the main coelomic organs are reported in Hounsfield units (HU). CONCLUSIONS: The results provide an atlas of matched anatomical cross-sections and contrast-enhanced CT images of the coelomic cavity in three highly diffused psittacine species.

Correlation of renal histopathology with renal echogenicity in dogs and cats: an ex-vivo quantitative study.

BACKGROUND: Increased cortical or cortical and medullary echogenicity is one of the most common signs of chronic or acute kidney disease in dogs and cats. Subjective evaluation of the echogenicity is reported to be unreliable. Patient and technical-related factors affect in-vivo quantitative evaluation of the echogenicity of parenchymal organs. The aim of the present study is to investigate the relationship between histopathology and ex-vivo renal cortical echogenicity in dogs and cats devoid of any patient and technical-related biases. RESULTS: Kidney samples were collected from 68 dog and 32 cat cadavers donated by the owners to the Veterinary Teaching Hospital of the University of Padua and standardized ultrasonographic images of each sample were collected. The echogenicity of the renal cortex was quantitatively assessed by means of mean gray value (MGV), and then histopathological analysis was performed. Statistical analysis to evaluate the influence of histological lesions on MGV was performed. The differentiation efficiency of MGV to detect pathological changes in the kidneys was calculated for dogs and cats. Statistical analysis revealed that only glomerulosclerosis was an independent determinant of echogenicity in dogs whereas interstitial nephritis, interstitial necrosis and fibrosis were independent determinants of echogenicity in cats. The global influence of histological lesions on renal echogenicity was higher in cats (23%) than in dogs (12%). CONCLUSIONS: Different histopathological lesions influence the echogenicity of the kidneys in dogs and cats. Moreover, MGV is a poor test for distinguishing between normal and pathological kidneys in the dog with a sensitivity of 58.3% and specificity of 59.8%. Instead, it seems to perform globally better in the cat, resulting in a fair test, with a sensitivity of 80.6% and a specificity of 56%.

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.

An AI-based algorithm for the automatic evaluation of image quality in canine thoracic radiographs.

The aim of this study was to develop and test an artificial intelligence (AI)-based algorithm for detecting common technical errors in canine thoracic radiography. The algorithm was trained using a database of thoracic radiographs from three veterinary clinics in Italy, which were evaluated for image quality by three experienced veterinary diagnostic imagers. The algorithm was designed to classify the images as correct or having one or more of the following errors: rotation, underexposure, overexposure, incorrect limb positioning, incorrect neck positioning, blurriness, cut-off, or the presence of foreign objects, or medical devices. The algorithm was able to correctly identify errors in thoracic radiographs with an overall accuracy of 81.5% in latero-lateral and 75.7% in sagittal images. The most accurately identified errors were limb mispositioning and underexposure both in latero-lateral and sagittal images. The accuracy of the developed model in the classification of technically correct radiographs was fair in latero-lateral and good in sagittal images. The authors conclude that their AI-based algorithm is a promising tool for improving the accuracy of radiographic interpretation by identifying technical errors in canine thoracic radiographs.

Preliminary Evaluation of Cortical and Medullary Echogenicity in Normal Canine Fetal Kidneys during the Last 10 Days of Pregnancy.

The objective of this study was to assess changes in the echogenicity of the cortex and medulla of canine fetal kidneys in relation to days before parturition (dbp), maternal size and litter size. Monitoring of 10 healthy pregnant bitches (2-8 years old, 8.8-40.3 kg bw) was conducted from -10 to 0 dbp using ultrasound. A single renal sonogram was obtained by scanning in a longitudinal section the three most caudal fetuses. The mean gray level (MGL) and SD of a manually drawn region of interest (ROI) in the renal cortex and medulla were measured using the Fiji Image J software (Image J 1.51h, Java 1.6 0_24 64 bit). A linear mixed model taking into account the maternal size as a fixed effect, dbp and litter size as covariates and the bitch as a random and repeated effect was used. The regression coefficients (b) were estimated. Cortical SD (C-SD) and cortico-medullary SD (C/M-SD) were influenced by dbp, with a significant decrease at the approaching day of parturition (b = 0.23 ± 0.06, p < 0.001 and b = 0.5 ± 0.02, p = 0.038, respectively). Maternal size had a significant impact on C/M-MGL with differences observed in large-sized (1.95 ± 0.13) compared to small- (1.41 ± 0.10, p = 0.027) and medium-sized bitches (1.51 ± 0.09, p = 0.016). The C/M-MGL was influenced by litter size, showing a decrease as the number of pups increased (b = -0.08 ± 0.03, p = 0.018). C-SD and C/M-SD were exclusively affected by dbp, and not by maternal and litter size. This suggests their potential as valuable parameters, warranting further investigations in future studies.

Improving the classification of veterinary thoracic radiographs through inter-species and inter-pathology self-supervised pre-training of deep learning models.

The analysis of veterinary radiographic imaging data is an essential step in the diagnosis of many thoracic lesions. Given the limited time that physicians can devote to a single patient, it would be valuable to implement an automated system to help clinicians make faster but still accurate diagnoses. Currently, most of such systems are based on supervised deep learning approaches. However, the problem with these solutions is that they need a large database of labeled data. Access to such data is often limited, as it requires a great investment of both time and money. Therefore, in this work we present a solution that allows higher classification scores to be obtained using knowledge transfer from inter-species and inter-pathology self-supervised learning methods. Before training the network for classification, pretraining of the model was performed using self-supervised learning approaches on publicly available unlabeled radiographic data of human and dog images, which allowed substantially increasing the number of images for this phase. The self-supervised learning approaches included the Beta Variational Autoencoder, the Soft-Introspective Variational Autoencoder, and a Simple Framework for Contrastive Learning of Visual Representations. After the initial pretraining, fine-tuning was performed for the collected veterinary dataset using 20% of the available data. Next, a latent space exploration was performed for each model after which the encoding part of the model was fine-tuned again, this time in a supervised manner for classification. Simple Framework for Contrastive Learning of Visual Representations proved to be the most beneficial pretraining method. Therefore, it was for this method that experiments with various fine-tuning methods were carried out. We achieved a mean ROC AUC score of 0.77 and 0.66, respectively, for the laterolateral and dorsoventral projection datasets. The results show significant improvement compared to using the model without any pretraining approach.

Development of an artificial intelligence-based method for the diagnosis of the severity of myxomatous mitral valve disease from canine chest radiographs.

An algorithm based on artificial intelligence (AI) was developed and tested to classify different stages of myxomatous mitral valve disease (MMVD) from canine thoracic radiographs. The radiographs were selected from the medical databases of two different institutions, considering dogs over 6 years of age that had undergone chest X-ray and echocardiographic examination. Only radiographs clearly showing the cardiac silhouette were considered. The convolutional neural network (CNN) was trained on both the right and left lateral and/or ventro-dorsal or dorso-ventral views. Each dog was classified according to the American College of Veterinary Internal Medicine (ACVIM) guidelines as stage B1, B2 or C + D. ResNet18 CNN was used as a classification network, and the results were evaluated using confusion matrices, receiver operating characteristic curves, and t-SNE and UMAP projections. The area under the curve (AUC) showed good heart-CNN performance in determining the MMVD stage from the lateral views with an AUC of 0.87, 0.77, and 0.88 for stages B1, B2, and C + D, respectively. The high accuracy of the algorithm in predicting the MMVD stage suggests that it could stand as a useful support tool in the interpretation of canine thoracic radiographs.

Contrast-enhanced ultrasound features of adrenal lesions in dogs.

BACKGROUND: The contrast-enhanced ultrasound (CEUS) features of adrenal lesions are poorly reported in veterinary literature. METHODS: Qualitative and quantitative B-mode ultrasound and CEUS features of 186 benign (adenoma) and malignant (adenocarcinoma and pheochromocytoma) adrenal lesions were evaluated. RESULTS: Adenocarcinomas (n = 72) and pheochromocytomas (n = 32) had mixed echogenicity with B-mode, and a non-homogeneous aspect with a diffused or peripheral enhancement pattern, hypoperfused areas, intralesional microcirculation and non-homogeneous wash-out with CEUS. Adenomas (n = 82) had mixed echogenicity, isoechogenicity or hypoechogenicity with B-mode, and a homogeneous or non-homogeneous aspect with a diffused enhancement pattern, hypoperfused areas, intralesional microcirculation and homogeneous wash-out with CEUS. With CEUS, a non-homogeneous aspect and the presence of hypoperfused areas and intralesional microcirculation can be used to distinguish between malignant (adenocarcinoma and pheochromocytoma) and benign (adenoma) adrenal lesions. LIMITATIONS: Lesions were characterised only by means of cytology. CONCLUSIONS: CEUS examination is a valuable tool for distinction between benign and malignant adrenal lesions and can potentially differentiate pheochromocytomas from adenocarcinomas and adenomas. However, cytology and histology are necessary to obtain the final diagnosis.

Comparative evaluation of the cadaveric, radiographic and computed tomographic anatomy of the heads of green iguana (Iguana iguana), common tegu (Tupinambis merianae) and bearded dragon (Pogona vitticeps).

BACKGROUND: Radiology and computed tomography are the most commonly available diagnostic tools for the diagnosis of pathologies affecting the head and skull in veterinary practice. Nevertheless, accurate interpretation of radiographic and CT studies requires a thorough knowledge of the gross and the cross-sectional anatomy. Despite the increasing success of reptiles as pets, only a few reports over their normal imaging features are currently available. The aim of this study is to describe the normal cadaveric, radiographic and computed tomographic features of the heads of the green iguana, tegu and bearded dragon. RESULTS: 6 adult green iguanas, 4 tegus, 3 bearded dragons, and, the adult cadavers of: 4 green iguana, 4 tegu, 4 bearded dragon were included in the study. 2 cadavers were dissected following a stratigraphic approach and 2 cadavers were cross-sectioned for each species. These latter specimens were stored in a freezer (-20°C) until completely frozen. Transversal sections at 5 mm intervals were obtained by means of an electric band-saw. Each section was cleaned and photographed on both sides. Radiographs of the head of each subject were obtained. Pre- and post- contrast computed tomographic studies of the head were performed on all the live animals. CT images were displayed in both bone and soft tissue windows. Individual anatomic structures were first recognised and labelled on the anatomic images and then matched on radiographs and CT images. Radiographic and CT images of the skull provided good detail of the bony structures in all species. In CT contrast medium injection enabled good detail of the soft tissues to be obtained in the iguana whereas only the eye was clearly distinguishable from the remaining soft tissues in both the tegu and the bearded dragon. CONCLUSIONS: The results provide an atlas of the normal anatomical and in vivo radiographic and computed tomographic features of the heads of lizards, and this may be useful in interpreting any imaging modality involving these species.

A Machine Learning-Based Approach for Classification of Focal Splenic Lesions Based on Their CT Features.

The aim of the study was to describe the CT features of focal splenic lesions (FSLs) in dogs in order to predict lesion histotype. Dogs that underwent a CT scan and had a FSL diagnosis by cytology or histopathology were retrospectively included in the study. For the statistical analysis the cases were divided into four groups, based on the results of cytopatholoy or hystopathology, namely: nodular hyperplasia (NH), other benign lesions (OBLs), sarcoma (SA), round cell tumour (RCT). Several qualitative and quantitative CT features were described for each case. The relationship occurring between each individual CT feature and the histopathological groups was explred by means of c chi-square test for the count data and by means of Kruskal-Wallis or ANOVA for the continuous data. Furthermore, the main features of each group were described using factorial discriminant analysis, and a decision tree for lesion classification was then developed. Sarcomas were characterised by large dimensions, a cystic appearance and an overall low post contrast-enhancement. NH and OBLs were characterised by small dimensions, a solid appearance and a high post-contrast enhancement. OBLs showed higher post-contrast values than NH. Lastly, RCTs did not exhibit any distinctive CT features. The proposed decision tree had a high accuracy for the classification of SA (0.89) and a moderate accuracy for the classification of OBLs and NH (0.79), whereas it was unable to classify RCTs. The results of the factorial analysis and the proposed decision tree could help the clinician in classifying FSLs based on their CT features. A definitive FSL diagnosis can only be obtained by microscopic examination of the spleen.

Contrast-enhanced ultrasound features of focal pancreatic lesions in dogs.

BACKGROUND: Contrast-enhanced ultrasound (CEUS) features of pancreatic lesions are poorly reported in veterinary literature. METHODS: Qualitative and quantitative features of pancreatic benign (nodular hyperplasia [NH], cyst and abscess) and malignant (adenocarcinoma and insulinoma) lesions during B-mode and CEUS examinations are described in 75 dogs. RESULTS: Adenocarcinomas (n = 23) had mixed echogenicity at B-mode, and they were hypoenhancing or non-enhancing at CEUS, with a non-homogeneous and cystic enhancement pattern. Insulinomas (n = 23) appeared as hypoechoic lesions at B-mode, and as hyperenhancing, homogeneous and solid lesions at CEUS. NH (n = 17) had an constant appearance, being hypoechoic at ultrasound (US) and isoenhancing at CEUS. Cysts (n = 7) were all anechoic, with acoustic enhancement clearly detectable at US, but were non-enhancing at CEUS. Lastly, abscesses (n = 5) had mixed echogenicity, and they showed both hyperenhancement and non-enhancement at CEUS. Hypoenhancement and non-homogeneous appearance had a moderate diagnostic accuracy in the detection of adenocarcinomas. In particular, hyperenhancement was evident only in malignant lesions (adenocarcinomas and insulinomas). CONCLUSION: CEUS, in combination with B-mode US features, is a valuable tool for distinction of benign and malignant abnormalities of the pancreas and can potentially differentiate insulinomas from adenocarcinomas.

Contrast-enhanced ultrasound features of focal pancreatic lesions in cats.

A large overlap in the ultrasound (US) features of focal pancreatic lesions (FPLs) in cats is reported. Furthermore, only a small number of studies describing the contrast-enhanced ultrasound (CEUS) features of FPLs in cats have been conducted today. The aim of this study is to describe the B-mode US and CEUS features of FPLs in cats. Ninety-eight cats cytologically diagnosed with FPL were included. The lesions were classified as adenocarcinoma (n = 40), lymphoma (n = 11), nodular hyperplasia (n = 17), other benign lesion (OBL) (n = 20), cyst (n = 4) or other malignant lesion (OML) (n = 6). Several qualitative and quantitative B-mode and CEUS features were described in each case. OMLs and cysts were not included in the statistical analysis. A decision tree to classify the lesions based on their B-mode and CEUS features was developed. The overall accuracy of the cross-validation of the decision tree was 0.74 (95% CI: 0.63-0.83). The developed decision tree had a very high sensitivity and specificity for nodular hyperplasia (1 and 0.94, respectively) as well as good sensitivity and specificity for both adenocarcinomas (0.85 and 0.77, respectively) and OBLs also (0.70 and 0.93, respectively). The algorithm was unable to detect any specific feature for classifying lymphomas, and almost all the lymphomas were classified as adenocarcinomas. The combination between CEUS and B-mode US is very accurate in the classification of some FPLs, especially nodular hyperplasia and adenocarcinomas. Cytopathology and or histopathology is still a fundamental step FPL diagnostic workflow.

Diagnostic Accuracy of Delayed Phase Post Contrast Computed Tomographic Images in the Diagnosis of Focal Liver Lesions in Dogs: 69 Cases.

To describe the computed tomographic (CT) features of focal liver lesions (FLLs) in dogs, that could enable predicting lesion histotype. Dogs diagnosed with FLLs through both CT and cytopathology and/or histopathology were retrospectively collected. Ten qualitative and 6 quantitative CT features have been described for each case. Lastly, a machine learning-based decision tree was developed to predict the lesion histotype. Four categories of FLLs - hepatocellular carcinoma (HCC, n = 13), nodular hyperplasia (NH, n = 19), other benign lesions (OBL, n = 18), and other malignant lesions (OML, n = 19) - were evaluated in 69 dogs. Five of the observed qualitative CT features resulted to be statistically significant in the distinction between the 4 categories: surface, appearance, lymph-node appearance, capsule formation, and homogeneity of contrast medium distribution. Three of the observed quantitative CT features were significantly different between the 4 categories: the Hounsfield Units (HU) of the radiologically normal liver parenchyma during the pre-contrast scan, the maximum dimension, and the ellipsoid volume of the lesion. Using the machine learning-based decision tree, it was possible to correctly classify NHs, OBLs, HCCs, and OMLs with an accuracy of 0.74, 0.88, 0.87, and 0.75, respectively. The developed decision tree could be an easy-to-use tool to predict the histotype of different FLLs in dogs. Cytology and histology are necessary to obtain the final diagnosis of the lesions.

Corrigendum: Diagnostic Accuracy of Delayed Phase Post Contrast Computed Tomographic Images in the Diagnosis of Focal Liver Lesions in Dogs: 69 Cases.

[This corrects the article DOI: 10.3389/fvets.2021.611556.].

Automatic classification of canine thoracic radiographs using deep learning.

The interpretation of thoracic radiographs is a challenging and error-prone task for veterinarians. Despite recent advancements in machine learning and computer vision, the development of computer-aided diagnostic systems for radiographs remains a challenging and unsolved problem, particularly in the context of veterinary medicine. In this study, a novel method, based on multi-label deep convolutional neural network (CNN), for the classification of thoracic radiographs in dogs was developed. All the thoracic radiographs of dogs performed between 2010 and 2020 in the institution were retrospectively collected. Radiographs were taken with two different radiograph acquisition systems and were divided into two data sets accordingly. One data set (Data Set 1) was used for training and testing and another data set (Data Set 2) was used to test the generalization ability of the CNNs. Radiographic findings used as non mutually exclusive labels to train the CNNs were: unremarkable, cardiomegaly, alveolar pattern, bronchial pattern, interstitial pattern, mass, pleural effusion, pneumothorax, and megaesophagus. Two different CNNs, based on ResNet-50 and DenseNet-121 architectures respectively, were developed and tested. The CNN based on ResNet-50 had an Area Under the Receive-Operator Curve (AUC) above 0.8 for all the included radiographic findings except for bronchial and interstitial patterns both on Data Set 1 and Data Set 2. The CNN based on DenseNet-121 had a lower overall performance. Statistically significant differences in the generalization ability between the two CNNs were evident, with the CNN based on ResNet-50 showing better performance for alveolar pattern, interstitial pattern, megaesophagus, and pneumothorax.

An AI-Based Algorithm for the Automatic Classification of Thoracic Radiographs in Cats.

An artificial intelligence (AI)-based computer-aided detection (CAD) algorithm to detect some of the most common radiographic findings in the feline thorax was developed and tested. The database used for training comprised radiographs acquired at two different institutions. Only correctly exposed and positioned radiographs were included in the database used for training. The presence of several radiographic findings was recorded. Consequenly, the radiographic findings included for training were: no findings, bronchial pattern, pleural effusion, mass, alveolar pattern, pneumothorax, cardiomegaly. Multi-label convolutional neural networks (CNNs) were used to develop the CAD algorithm, and the performance of two different CNN architectures, ResNet 50 and Inception V3, was compared. Both architectures had an area under the receiver operating characteristic curve (AUC) above 0.9 for alveolar pattern, bronchial pattern and pleural effusion, an AUC above 0.8 for no findings and pneumothorax, and an AUC above 0.7 for cardiomegaly. The AUC for mass was low (above 0.5) for both architectures. No significant differences were evident in the diagnostic accuracy of either architecture.

Undifferentiated laryngeal carcinoma with hyaline bodies in a cat.

BACKGROUND: Primary laryngeal neoplasms are rare in cats, with lymphoma and squamous cell carcinoma being the most commonly diagnosed tumour types. These tumours are usually highly aggressive, difficult to treat, and have a poor prognosis. Here an undifferentiated laryngeal carcinoma with hyaline bodies in a cat is reported. CASE PRESENTATION: A 13-year-old cat was presented for progressive respiratory signs. Diagnostic procedures revealed a partially obstructive laryngeal mass. Cytology was compatible with a poorly differentiated malignant tumour, with neoplastic cells frequently containing large intracytoplasmic hyaline bodies. After 1 month the patient was euthanised due to a worsening clinical condition and submitted for post-mortem examination, which confirmed the presence of two laryngeal masses. Histopathology confirmed the presence of an undifferentiated neoplasm with marked features of malignancy. Strong immunolabelling for pancytokeratin led to a diagnosis of undifferentiated carcinoma, however, histochemical and immunohistochemical investigations could not elucidate the origin of the large intracytoplasmic hyaline bodies observed in tumour cells, which appeared as non-membrane bound deposits of electron-dense material on transmission electron microscopy. CONCLUSION: This is the first report of primary undifferentiated laryngeal carcinoma in a cat. Our case confirms the clinical features and the short survival that have been reported in other studies describing feline laryngeal tumours. Moreover, for the first time in feline literature, we describe the presence of intracytoplasmic hyaline bodies in neoplastic cells that were compatible with the so-called hyaline granules reported in different human cancers and also in the dog.