Histological classification of canine and feline lymphoma using a modular approach based on deep learning and advanced image processing.

Histopathological examination of tissue samples is essential for identifying tumor malignancy and the diagnosis of different types of tumor. In the case of lymphoma classification, nuclear size of the neoplastic lymphocytes is one of the key features to differentiate the different subtypes. Based on the combination of artificial intelligence and advanced image processing, we provide a workflow for the classification of lymphoma with regards to their nuclear size (small, intermediate, and large). As the baseline for our workflow testing, we use a Unet++ model trained on histological images of canine lymphoma with individually labeled nuclei. As an alternative to the Unet++, we also used a publicly available pre-trained and unmodified instance segmentation model called Stardist to demonstrate that our modular classification workflow can be combined with different types of segmentation models if they can provide proper nuclei segmentation. Subsequent to nuclear segmentation, we optimize algorithmic parameters for accurate classification of nuclear size using a newly derived reference size and final image classification based on a pathologists-derived ground truth. Our image classification module achieves a classification accuracy of up to 92% on canine lymphoma data. Compared to the accuracy ranging from 66.67 to 84% achieved using measurements provided by three individual pathologists, our algorithm provides a higher accuracy level and reproducible results. Our workflow also demonstrates a high transferability to feline lymphoma, as shown by its accuracy of up to 84.21%, even though our workflow was not optimized for feline lymphoma images. By determining the nuclear size distribution in tumor areas, our workflow can assist pathologists in subtyping lymphoma based on the nuclei size and potentially improve reproducibility. Our proposed approach is modular and comprehensible, thus allowing adaptation for specific tasks and increasing the users' trust in computer-assisted image classification.

Automated diagnosis of 7 canine skin tumors using machine learning on H&E-stained whole slide images.

Microscopic evaluation of hematoxylin and eosin-stained slides is still the diagnostic gold standard for a variety of diseases, including neoplasms. Nevertheless, intra- and interrater variability are well documented among pathologists. So far, computer assistance via automated image analysis has shown potential to support pathologists in improving accuracy and reproducibility of quantitative tasks. In this proof of principle study, we describe a machine-learning-based algorithm for the automated diagnosis of 7 of the most common canine skin tumors: trichoblastoma, squamous cell carcinoma, peripheral nerve sheath tumor, melanoma, histiocytoma, mast cell tumor, and plasmacytoma. We selected, digitized, and annotated 350 hematoxylin and eosin-stained slides (50 per tumor type) to create a database divided into training, n = 245 whole-slide images (WSIs), validation (n = 35 WSIs), and test sets (n = 70 WSIs). Full annotations included the 7 tumor classes and 6 normal skin structures. The data set was used to train a convolutional neural network (CNN) for the automatic segmentation of tumor and nontumor classes. Subsequently, the detected tumor regions were classified patch-wise into 1 of the 7 tumor classes. A majority of patches-approach led to a tumor classification accuracy of the network on the slide-level of 95% (133/140 WSIs), with a patch-level precision of 85%. The same 140 WSIs were provided to 6 experienced pathologists for diagnosis, who achieved a similar slide-level accuracy of 98% (137/140 correct majority votes). Our results highlight the feasibility of artificial intelligence-based methods as a support tool in diagnostic oncologic pathology with future applications in other species and tumor types.

Isolation and Characterization of Novel Canine Osteosarcoma Cell Lines from Chemotherapy-Naïve Patients.

The present study aimed to establish novel canine osteosarcoma cell lines (COS3600, COS3600B, COS4074) and characterize the recently described COS4288 cells. The established D-17 cell line served as a reference. Analyzed cell lines differed notably in their biological characteristics. Calculated doubling times were between 22 h for COS3600B and 426 h for COS4074 cells. COS3600B and COS4288 cells produced visible colonies after anchorage-independent growth in soft agar. COS4288 cells were identified as cells with the highest migratory capacity. All cells displayed the ability to invade through an artificial basement membrane matrix. Immunohistochemical analyses revealed the mesenchymal origin of all COS cell lines as well as positive staining for the osteosarcoma-relevant proteins alkaline phosphatase and karyopherin α2. Expression of p53 was confirmed in all tested cell lines. Gene expression analyses of selected genes linked to cellular immune checkpoints (CD270, CD274, CD276), kinase activity (MET, ERBB2), and metastatic potential (MMP-2, MMP-9) as well as selected long non-coding RNA (MALAT1) and microRNAs (miR-9, miR-34a, miR-93) are provided. All tested cell lines were able to grow as multicellular spheroids. In all spheroids except COS4288, calcium deposition was detected by von Kossa staining. We believe that these new cell lines serve as useful biological models for future studies.

Validation of digital microscopy: Review of validation methods and sources of bias.

Digital microscopy (DM) is increasingly replacing traditional light microscopy (LM) for performing routine diagnostic and research work in human and veterinary pathology. The DM workflow encompasses specimen preparation, whole-slide image acquisition, slide retrieval, and the workstation, each of which has the potential (depending on the technical parameters) to introduce limitations and artifacts into microscopic examination by pathologists. Performing validation studies according to guidelines established in human pathology ensures that the best-practice approaches for patient care are not deteriorated by implementing DM. Whereas current publications on validation studies suggest an overall high reliability of DM, each laboratory is encouraged to perform an individual validation study to ensure that the DM workflow performs as expected in the respective clinical or research environment. With the exception of validation guidelines developed by the College of American Pathologists in 2013 and its update in 2021, there is no current review of the application of methods fundamental to validation. We highlight that there is high methodological variation between published validation studies, each having advantages and limitations. The diagnostic concordance rate between DM and LM is the most relevant outcome measure, which is influenced (regardless of the viewing modality used) by different sources of bias including complexity of the cases examined, diagnostic experience of the study pathologists, and case recall. Here, we review 3 general study designs used for previous publications on DM validation as well as different approaches for avoiding bias.

Composition of lymphocyte subpopulations in normal and mildly reactive peripheral lymph nodes in cats.

OBJECTIVES: Flow cytometric (FCM) immunophenotyping of lymphoid tissue aspirates is an available adjunct for feline lymphoma diagnostics. Reference data have only been established for feline peripheral blood. Studies investigating the composition of normal and mildly reactive feline lymph nodes (LNs) are lacking. The aim of this prospective study was to establish reference data for lymphocyte subpopulations in normal and mildly reactive feline peripheral LNs using a standardised multicolour panel of antibodies. METHODS: Macroscopically inconspicuous mandibular and/or popliteal LNs from 31 adult cats, which were euthanased for reasons other than haematological diseases, were excised and processed within 5 h after death. Multicolour flow cytometry using eight different feline-specific, anti-canine and human cross-reactive monoclonal antibodies used in current diagnostic marker panels was performed after cytological exclusion of pathological states and complemented by lymphocyte clonality testing, histopathology and immunohistochemistry (IHC) to ensure the absence of lymphoid disease. RESULTS: Of 31 cats, the immunophenotyping data of 24 individuals could be included as histopathology and clonality testing excluded a pathological condition. Lymphocyte populations showed the following positive antibody reactions: CD18+ 86.3% ± 13.86%, CD3+ 54.81% ± 11.10%, CD5+ 57.39% ± 12.66%, CD21+ 40.42% ± 12.40%, CD79alphacy+ (CD79αcy) 30.41% ± 13.49% and CD14+ 0.75% ± 1.35%. There were 30.88% ± 13.48% CD4+ and 12.91% ± 6.68% CD8+ cells. Cytology revealed a mixed population of mostly lymphoid cells in all samples. The absence of a monoclonal/oligoclonal neoplastic population was confirmed by lymphocyte clonality testing. Histopathology and IHC showed a normal or mildly reactive pattern in all cases. CONCLUSIONS AND RELEVANCE: This study establishes FCM immunophenotyping data of lymphocyte populations of normal and mildly reactive feline peripheral LNs. For the first time, anti-CD5, CD4, CD8 and CD21 reference data in normal and mildly reactive feline peripheral LNs are presented. CD18, CD3, CD14 and CD79αcy have been used to establish reference data for the first time in any feline material.

Computer-assisted mitotic count using a deep learning-based algorithm improves interobserver reproducibility and accuracy.

The mitotic count (MC) is an important histological parameter for prognostication of malignant neoplasms. However, it has inter- and intraobserver discrepancies due to difficulties in selecting the region of interest (MC-ROI) and in identifying or classifying mitotic figures (MFs). Recent progress in the field of artificial intelligence has allowed the development of high-performance algorithms that may improve standardization of the MC. As algorithmic predictions are not flawless, computer-assisted review by pathologists may ensure reliability. In the present study, we compared partial (MC-ROI preselection) and full (additional visualization of MF candidates and display of algorithmic confidence values) computer-assisted MC analysis to the routine (unaided) MC analysis by 23 pathologists for whole-slide images of 50 canine cutaneous mast cell tumors (ccMCTs). Algorithmic predictions aimed to assist pathologists in detecting mitotic hotspot locations, reducing omission of MFs, and improving classification against imposters. The interobserver consistency for the MC significantly increased with computer assistance (interobserver correlation coefficient, ICC = 0.92) compared to the unaided approach (ICC = 0.70). Classification into prognostic stratifications had a higher accuracy with computer assistance. The algorithmically preselected hotspot MC-ROIs had a consistently higher MCs than the manually selected MC-ROIs. Compared to a ground truth (developed with immunohistochemistry for phosphohistone H3), pathologist performance in detecting individual MF was augmented when using computer assistance (F1-score of 0.68 increased to 0.79) with a reduction in false negatives by 38%. The results of this study demonstrate that computer assistance may lead to more reproducible and accurate MCs in ccMCTs.

Proliferation Activity in Canine Gastrointestinal Lymphoma.

Gastrointestinal lymphomas are uncommon in dogs and little is known about their distinct subtypes or proliferation rate. The aim of this study was to stratify 33 canine gastrointestinal lymphoma samples according to the latest World Health Organization classification and to determine the Ki67 proliferation index by manual counting, digital image analysis and visual estimation. The Ki67 index was then correlated with subtype, immunophenotype, mitotic index, grade and tumour location. The mitotic index correlated positively with the Ki67 index. A significantly higher number of Ki67-positive cells was found in enteropathy-associated T-cell lymphoma type I and in diffuse large B-cell lymphoma compared with enteropathy-associated T-cell lymphoma type II. There was also a significant difference in Ki67 immunolabelled cells between grade 1 and grade 2 lymphomas. Moderate agreement was found between the Ki67 index as obtained by manual counting and visual estimation, but there was strong agreement between manual counting and digital image analysis. The user-friendly digital imaging system used in this study could have potential for future determination of the Ki67 index in lymphoid neoplasms.

A Recurrent STAT5BN642H Driver Mutation in Feline Alimentary T Cell Lymphoma.

Alimentary lymphomas arising from T cells are rare and aggressive malignancies in humans. In comparison, they represent the most common anatomical form of lymphoma in cats. Due to the low prevalence in humans, the underlying pathomechanism for these diseases is poorly characterised, limiting experimental analysis and therapeutic exploration. To date, activating mutations of the JAK/STAT core cancer pathway and particularly the STAT5B oncoprotein have been identified in human enteropathy-associated T cell lymphoma. Here, we describe a high homology of human and feline STAT3 and STAT5B proteins and strong conservation at the genomic level. Analysis of 42 samples of feline T cell alimentary lymphoma reveals broad activation of STAT3 and STAT5B. Screening for known activating mutations in STAT3 or STAT5B identifies the presence of the STAT5BN642H driver mutation in feline enteropathy-associated T cell lymphoma in 7 out of 42 (16.67%) samples in total. Regarding lymphoma subtypes, the majority of mutations with 5 out of 17 (29.41%) cases were found in feline enteropathy-associated lymphoma type II (EATL II). This identification of an oncogenic STAT5B driver mutation in felines recapitulates the genetic situation in the corresponding human disease, thereby establishing the cat as a potential new model for a rare and incurable human T cell disease.

Revisiting lymphocyte clonality testing in feline B-cell lymphoma.

Differentiation between resident mature lymphocyte populations and small-cell lymphoma cannot be made by cytological review alone and remains challenging in histopathological review. These cases warrant application of complementary tools like PCR-based immunoglobulin (IG) and T-cell receptor (TCR) clonality testing for confirmation. Although primer coverage of potential IG gene rearrangements in feline B-cell neoplasms constantly improves, the possibility of false negative and false positive test results still poses a problem. In this retrospective study, we assessed diagnostic sensitivity and specificity of a novel developed multiplex PCR assay for routine diagnosis of B-cell clonality. Therefore, 24 feline patients were subjected to comparative clonality testing by using different primer sets. Feline lymphoma cell lines and confirmed patient material served as positive control. Compared to previous studies, this novel developed multiplex PCR assay showed positive effects on diagnostic sensitivity, specificity, accuracy, and positive predictive value accompanied by a slight impairment of negative predictive value. Notably, none of the primer sets was superior; hence, we recommend the combined application of the herein tested primer sets in routine diagnostics. However, a more in-depth-evaluation of the dynamic of assay specific parameters in dependency on primer set usage requires prospective studies on larger cohorts of feline patients.

Flow Cytometric Assessment of Ki-67 Expression in Lymphocytes From Physiologic Lymph Nodes, Lymphoma Cell Populations and Remnant Normal Cell Populations From Lymphomatous Lymph Nodes.

Recent literature suggests conventional flow cytometric (FCM) immunophenotyping complemented by Ki-67 FCM assessment as a reliable tool to classify canine lymphomas. Ki-67 expression assessed by FCM is rarely reported in canine lymphoma cases and reference data for normal canine lymph nodes are missing. Moreover, nothing is known about the Ki-67 expression within the occasionally observed remnant cell population within the gates of normal lymphocytes in lymphoma cases. Aim of this study was to compare flow cytometric Ki-67 expression of lymphocyte populations from normal canine lymph nodes, lymphoma cells from World-Health-Organisation (WHO) classified lymphoma patient samples and their neighboring normal remnant cell population. Cryopreserved lymphocyte cell suspensions from normal lymph nodes from eight dogs free of lymphoma served as reference material. Fourteen cases diagnosed by cytology, FCM, clonality testing, histopathology including immunohistochemistry consisting of 10 DLBCL, 1 MZL, 1 PTCL and 2 TZL showed a residual small lymphocyte population and were investigated. The Ki-67 expression in normal canine lymphoid tissue was 3.19 ± 2.17%. Mean Ki-67 expression in the malignant cell populations was 41 ± 24.36%. Ki-67 positivity was 12.34 ± 10.68% in the residual physiologic lymphocyte population, which otherwise exhibited a physiologic immunophenotype pattern. This ratio was equivalent (n = 3) or lower (n = 11) than the Ki-67 expression of the malignant cell population within the sample. This is the first report of FCM derived Ki-67 expression combined with immunophenotype patterns in normal canine lymph nodes, compared with lymphoma cell populations and residual normal cell populations of lymphoma cases diagnosed by state of the art technology.

Ki-67 assessment-agreeability between immunohistochemistry and flow cytometry in canine lymphoma.

Recent literature suggests a combination of flow cytometric determination of Ki-67 and immunophenotype as a reliable tool to classify canine lymphomas. Immunohistochemistry (IHC) on histological samples is the gold standard technique assessing Ki-67 index. Agreement between IHC and FCM derived Ki-67 indices has never been investigated. The aim of this study was to investigate the agreement between IHC and FCM in the assessment of Ki-67 expression/index, in order to evaluate whether FCM may serve as a non-invasive alternative method for the estimation of proliferative activity in canine lymphoma. Dogs with previously untreated canine lymphoma undergoing diagnostic lymphadenectomy were prospectively enrolled. Ki-67 expression/index was assessed by FCM and IHC and expressed as percentage of positive cells. 39 dogs classified by histopathology matched the inclusion criteria. With both methods, Ki-67 expression/index was higher in intermediate/high-grade lymphomas. Spearman's coefficient of correlation was ρ = 0.57; (95% CI0.33-0.75) suggesting a moderate correlation. A Bland-Altman plot revealed a negative constant bias of -3.55 (95% CI: -10.52 to 3.42) with limits of agreement from -45.71 to 38.61. The study confirmed agreement albeit with wide confidence intervals between the values of Ki-67 expression/index assessed with FCM and IHC. Discrepancies were observed in a subset of cases. Possible explanation could be that Ki-67 index in IHC is determined in the most proliferative areas of the slide, which could introduce kind of sampling bias, whereas FCM evaluates many more cells in cell suspension. Further studies are warranted to investigate this phenomenon.

Are B-symptoms more reliable prognostic indicators than substage in canine nodal diffuse large B-cell lymphoma.

In humans B-symptoms refer to systemic symptoms of lymphoma such as fever, weight loss, and night sweats and influence the prognosis of patients. In canine lymphoma, substage B is used to describe any clinical sign observed. Aim of the retrospective study was to compare the prognostic value of substage B with B-symptoms to predict treatment response and survival in canine nodal diffuse large B-cell lymphoma. Affected dogs treated with CHOP chemotherapy between 2008 and 2019 were included. B-symptoms were defined by weight loss greater than 10% of normal weight, fever and the occurrence of unexplained resting tachypnoea substituted human night sweats. Substage B was defined as any symptoms but lymphadenopathy. Fifty-five cases were included. B-symptoms were present in 20/55 (36%) and substage B in 40/55 (74%) patients. No significant associations between B-symptoms or substage B and weight, sex, breed, WHO stage and lymphoma grade were found. Treatment response was negatively associated with both substage B (P = .02) and B-symptoms (P = .001). B-symptoms significantly decreased progression free survival (PFS) (95 vs 330 days, P = .001) and lymphoma specific survival (LSS) (160 vs 462 days, P = .001). Data showed that B-symptoms might be a more reliable prognostic indicator than substage B in canine nodal diffuse large B-cell lymphoma. Prospective studies assessing B-symptoms in a larger cohort of patients and in other common lymphoma types are warranted. The abstract was presented at the fourth meeting of the European Canine Lymphoma Network Group in Lugano, 22 June 2019 and published in the proceeding of the meeting on the page 26.

Diagnostic Yield of Micro-Computed Tomography (micro-CT) Versus Histopathology of a Canine Oral Fibrosarcoma.

Micro-computed tomography (micro-CT) imaging currently gains increased interest in human as well as veterinary medicine. The ability to image 3-dimensional (3D) biopsy specimens nondestructively down to 1 µm spatial resolution makes it a promising tool for microscopic tissue evaluation in addition to histopathology. Visualizing tumor margins and calculating tumor load on 3D reconstructions may also enhance oncological therapies. The objective of this study was to describe the workflow from tumor resection to histopathological diagnosis, using both routine hematoxylin-eosin (HE)-stained sections and micro-CT tomograms on a stage II oral fibrosarcoma in a 7-year-old Hovawart dog. The maxillectomy specimen was fixed with formalin and stained with an X-ray dense soft tissue contrast agent. Micro-CT imaging was done using an ex vivo specimen micro-CT device. Tumor margins could not be exactly determined on micro-CT tomograms due to limited image resolution and contrast. Histopathology was performed after washing out the contrast agent. It showed neoplastic cells infiltrating the surrounding tissue further than assumed from micro-CT images. A total tumor volume of 10.3 cm3 could be calculated based on correlating micro-CT tomograms with HE-stained sections. This correlative approach may be of particular interest for oncological therapy. More than that, micro-CT imaging technology supported histopathology by means of 3D orientation and selection of slices to be cut on determining tumor margins. In this clinical case report, micro-CT imaging did not provide unambiguous clinical evidence for oncological decision-making, but it showed potential to support histopathology and calculate tumor volume for further clinical use.

Pneumocystosis in dogs: meta-analysis of 43 published cases including clinical signs, diagnostic procedures, and treatment.

We evaluated 43 published cases of dogs with confirmed Pneumocystis infection regarding the value of clinical parameters indicating the presence of the disease as well as tools for the detection of the pathogen. The assessed parameters included clinical signs, laboratory findings, results of thoracic radiography, autopsy, histopathology, methods for the detection of Pneumocystis, as well as medical therapy. Pneumocystosis was diagnosed most often in certain breeds (Cavalier King Charles Spaniel, Miniature Dachshund) with a predisposition for impaired immunity. The median age of the dogs was 1 y. Chronic therapy-resistant respiratory signs, such as tachypnea, dyspnea, and cough, along with leukocytosis, neutrophilia, and hypogammaglobulinemia, were the most frequently described clinical and clinicopathologic abnormalities. Pneumocystosis can be masked by coinfections with other respiratory pathogens, and the successful detection of Pneumocystis organisms is of major relevance. Several detection methods have been used in the past, but only a few provide reliable results. In 2017, the cytologic evaluation of Giemsa-stained bronchoalveolar lavage samples is generally used, even if sensitivity is only moderate. More reliable results can be achieved using special stains or sensitive molecular techniques. Fast and reliable detection of Pneumocystis is the essential basis for appropriate treatment and higher survival chances for dogs.

Pneumocystis carinii infection with severe pneumomediastinum and lymph node involvement in a Whippet mixed-breed dog.

A 3.5-y-old Whippet mixed-breed dog was presented with a history of respiratory distress, exercise intolerance, and generalized demodicosis. Hematologic alterations included marked leukocytosis and neutrophilia. Radiographic examination showed a diffuse interstitial and mild peripheral alveolar lung pattern and pneumomediastinum. Because the cytologic examination of the bronchoalveolar aspirate was not diagnostic and a persistent perforation of the upper respiratory tract could not be ruled out, the dog was submitted to thoracoscopy, and subsequently the left cranial lung lobe as well as mediastinal and sternal lymph nodes were resected. Pulmonary pneumocystosis with spread to the thoracic lymph nodes was suspected after histologic investigation of lung and lymph nodes, which was confirmed by in situ hybridization, PCR, and subsequent Sanger sequencing. We document a rare, simultaneous occurrence of severe pulmonary and thoracic lymph node pneumocystosis with spontaneous pneumomediastinum in a dog. Definitive diagnosis was achieved through the use of Grocott methenamine silver staining, in situ hybridization, and PCR.

Does categorisation of lymphoma subtypes according to the World Health Organization classification predict clinical outcome in cats?

Objectives The purpose of this study was to specify lymphoma subtypes according to the World Health Organization (WHO) classification in a group of cats and to investigate their potential prognostic value. Methods Records of cats from the University of Veterinary Medicine Vienna suffering from lymphoma were reviewed in this retrospective study. To diagnose various subtypes specified in the WHO classification, histopathological and immunohistochemical examinations, as well as clonality assays in some cases, were performed. Results Of the 30 cats included in this study and classified according to the WHO guidelines, peripheral T-cell lymphoma was the most prevalent lymphoma subtype (37% of cases; n = 11), followed by diffuse large B-cell (23%; n = 7), intestinal T-cell (10%; n = 3), T-cell-rich B-cell (10%; n = 3), large granular lymphocytic (7%; n = 2), anaplastic large T-cell (7%; n = 2), B-cell small lymphocytic (3%; n = 1) and T-cell angiotropic lymphoma (3%; n = 1). The median survival time (MST) was 5.4 months (range 6 days to 2.2 years), with two cats still alive after 1.7 and 2.0 years, respectively. Treating cats prior to chemotherapy with glucocorticoids did not worsen their prognosis. Adding to chemotherapy, radiotherapy or surgery did not improve the clinical outcome. We observed that patients with intestinal T-cell lymphoma lived significantly longer (MST 1.7 years) than those with a diffuse large B-cell (MST 4.5 months) or peripheral T-cell lymphoma (MST 6.1 months). Cats with T-cell-rich B-cell lymphoma survived significantly longer (MST 1.2 years) than those with a diffuse large B-cell lymphoma. Conclusions and relevance A detailed diagnosis of feline lymphoma can be obtained by allocating different subtypes according to the WHO classification. From the eight detected lymphoma subtypes, two, intestinal T-cell lymphoma and T-cell-rich B-cell lymphoma, showed promising survival times in cats.

Ezrin and moesin expression in canine and feline osteosarcoma.

Biological features of canine osteosarcomas (OS) differ markedly from those found in feline and resemble more human osteosarcomas, in particular for their high rate of metastasis and poor prognosis. Ezrin, radixin and moesin are members of the ERM protein family and link the actin cytoskeleton with the cell membrane. Ezrin and moesin have been shown to be of prognostic significance in tumor progression due to their role in the metastatic process. The objective of this study was to analyze ezrin and moesin protein expression in a series of dog (n = 16) and cat (n = 8) osteosarcoma samples using immunohistochemistry and western blot techniques. We found that cat OS have a higher moesin expression compared to dog OS, however, the active phosphorylated forms of moesin and ezrin Tyr353 were more abundant in the dog samples. A statistically significant difference was found for the low and high immunohistochemical scores of ezrin and pan-phospho-ERM proteins between cat and dog. Although phospho-ezrin Thr567 was higher in feline OS, the membranous localization in dog OS samples indicates the presence of the biologically active form. Therefore, the observed differences in phosphorylated forms of ezrin and moesin status should be further studied to demonstrate if they are relevant for different biological behavior between dog and cat OS.

Stem cell growth factor receptor in canine vs. feline osteosarcomas.

Osteosarcoma is considered the most common bone cancer in cats and dogs, with cats having a much better prognosis than dogs, since the great majority of dogs with osteosarcoma develop distant metastases. In search of a factor possibly contributing to this disparity, the stem cell growth factor receptor KIT was targeted, and the messenger (m)RNA and protein expression levels of KIT were compared in canine vs. feline osteosarcomas, as well as in normal bone. The mRNA expression of KIT was quantified by reverse transcription-quantitative polymerase chain reaction, and was observed to be significantly higher in canine (n=14) than in feline (n=5) osteosarcoma samples (P<0.001). KIT protein expression was evaluated by immunohistochemistry, which revealed that 21% of canine osteosarcoma samples did not exhibit KIT staining in their neoplastic cells, while in 14% of samples, a score of 1 (<10% positive tumour cells) was observed, and in 50% and 14% of samples, a score of 2 (10-50% positivity) and 3 (>50% positivity), respectively, was observed. By contrast, the cancer cells of all the feline bone tumour samples analysed were entirely negative for KIT. Notably, canine and feline osteocytes of healthy bone tissue lacked any KIT expression. These results could be the first evidence that KIT may be involved in the higher aggressiveness of canine osteosarcoma compared with feline osteosarcoma.

Characterization of the T-cell receptor gamma chain gene rearrangements as an adjunct tool in the diagnosis of T-cell lymphomas in the gastrointestinal tract of cats.

Feline alimentary lymphoma is the most common hematopoietic neoplasia in cats. It affects mainly the small intestines and is most frequently of T-cell origin. Evaluation of a fine needle aspirate is often the first step in the diagnostic work-up. Differentiation between a resident mature lymphocyte population as encountered in inflammatory bowel disease and small cell lymphoma cannot be achieved by cytology alone. Even full thickness biopsies evaluated by histopathology can be inconclusive. These cases warrant the application of complementary tools like PCR-based T-cell receptor (TCR) clonality testing for confirmation. The aim of this study was to optimize the DNA extraction protocol for formalin fixed and paraffin embedded tissues (FFPE) and to establish a heteroduplex analysis to enhance resolution of the PCR fragments of the T-cell receptor gamma (TCRG) V-J gene. The new protocols resulted in improved quantity and quality of the extracted DNA. Heteroduplex analysis of the samples improved the resolution of the electrophoresis results so that rules for interpretation of the different patterns could be established. Application of this improved setup detected clonal rearrangements in at least one TCRG primer reaction in 31 of 36 of our feline intestinal lymphoma samples after DNA quality testing.

Distribution and activity levels of matrix metalloproteinase 2 and 9 in canine and feline osteosarcoma.

Overexpression of matrix metalloproteinases (MMPs) has been associated with increased tumor aggressiveness and metastasis dissemination. We investigated whether the contrasting metastatic behavior of feline and canine osteosarcoma is related to levels and activities of MMP2 and MMP9. Zymography and immunohistochemistry were used to determine expression levels of MMP2 and MMP9 in canine and feline osteosarcoma. Using immunohistochemistry, increased MMP9 levels were identified in most canine osteosarcomas, whereas cat samples more often displayed moderate levels. High levels of pro-MMP9, pro-MMP2, and active MMP2 were detected by gelatin zymography in both species, with significantly higher values for active MMP2 in canine osteosarcoma. These findings indicate that MMP2 is probably involved in canine and feline osteosarcoma and their expression and activity could be associated with the different metastatic behavior of canine and feline osteosarcoma.

La surexpression de métalloprotéases de matrice (MPMs) a été associée avec une augmentation de l'agressivité des tumeurs et de la dissémination métastasique. Nous avons cherché à savoir si le comportement métastasique contrastant d'ostéosarcomes félin et canin est relié aux quantités et à l'activité de MPM2 et MPM9. La zymographie et l'immunohistochimie ont été utilisées afin de déterminer les niveaux d'expression de MPM2 et de MPM9 dans des ostéosarcomes canins et félins. En utilisant l'immunohistochimie, des quantités augmentées de MPM9 ont été identifiées dans la plupart des ostéosarcomes canins, alors que les échantillons félins montraient plus souvent des quantités modérées. Des niveaux élevés de pro-MPM9, pro-MPM2, et de la MPM2 active ont été détectés par zymographie sur gélatine chez les deux espèces, avec des valeurs significativement plus élevées pour de la MPM2 active dans les ostéosarcomes canins. Ces données indiquent que MPM2 est probablement impliquée dans les ostéosarcomes canins et félins et que leur expression et activité pourraient être associé avec le comportement métastasique différent des ostéosarcomes canins et félins.(Traduit par Docteur Serge Messier).