Explainable text-tabular models for predicting mortality risk in companion animals.

As interest in using machine learning models to support clinical decision-making increases, explainability is an unequivocal priority for clinicians, researchers and regulators to comprehend and trust their results. With many clinical datasets containing a range of modalities, from the free-text of clinician notes to structured tabular data entries, there is a need for frameworks capable of providing comprehensive explanation values across diverse modalities. Here, we present a multimodal masking framework to extend the reach of SHapley Additive exPlanations (SHAP) to text and tabular datasets to identify risk factors for companion animal mortality in first-opinion veterinary electronic health records (EHRs) from across the United Kingdom. The framework is designed to treat each modality consistently, ensuring uniform and consistent treatment of features and thereby fostering predictability in unimodal and multimodal contexts. We present five multimodality approaches, with the best-performing method utilising PetBERT, a language model pre-trained on a veterinary dataset. Utilising our framework, we shed light for the first time on the reasons each model makes its decision and identify the inclination of PetBERT towards a more pronounced engagement with free-text narratives compared to BERT-base's predominant emphasis on tabular data. The investigation also explores the important features on a more granular level, identifying distinct words and phrases that substantially influenced an animal's life status prediction. PetBERT showcased a heightened ability to grasp phrases associated with veterinary clinical nomenclature, signalling the productivity of additional pre-training of language models.

Stone decision engine accurately predicts stone removal and treatment complications for shock wave lithotripsy and laser ureterorenoscopy patients.

Kidney stones form when mineral salts crystallize in the urinary tract. While most stones exit the body in the urine stream, some can block the ureteropelvic junction or ureters, leading to severe lower back pain, blood in the urine, vomiting, and painful urination. Imaging technologies, such as X-rays or ureterorenoscopy (URS), are typically used to detect kidney stones. Subsequently, these stones are fragmented into smaller pieces using shock wave lithotripsy (SWL) or laser URS. Both treatments yield subtly different patient outcomes. To predict successful stone removal and complication outcomes, Artificial Neural Network models were trained on 15,126 SWL and 2,116 URS patient records. These records include patient metrics like Body Mass Index and age, as well as treatment outcomes obtained using various medical instruments and healthcare professionals. Due to the low number of outcome failures in the data (e.g., treatment complications), Nearest Neighbor and Synthetic Minority Oversampling Technique (SMOTE) models were implemented to improve prediction accuracies. To reduce noise in the predictions, ensemble modeling was employed. The average prediction accuracies based on Confusion Matrices for SWL stone removal and treatment complications were 84.8% and 95.0%, respectively, while those for URS were 89.0% and 92.2%, respectively. The average prediction accuracies for SWL based on Area-Under-the-Curve were 74.7% and 62.9%, respectively, while those for URS were 77.2% and 78.9%, respectively. Taken together, the approach yielded moderate to high accurate predictions, regardless of treatment or outcome. These models were incorporated into a Stone Decision Engine web application (http://peteranoble.com/webapps.html) that suggests the best interventions to healthcare providers based on individual patient metrics.

Home spirometry appears accurate and feasible for monitoring chronic respiratory disease.

Spirometry is underutilised and can be difficult to access. This study assessed the accuracy and feasibility of home spirometry compared to gold standard. Findings suggest home spirometry is accurate and feasible across many respiratory disease groups. https://bit.ly/42TLoYd.

Stiffness Mediated-Mechanosensation of Airway Smooth Muscle Cells on Linear Stiffness Gradient Hydrogels.

In obstructive airway diseases such as asthma and chronic obstructive pulmonary disease (COPD), the extracellular matrix (ECM) protein amount and composition of the airway smooth muscle (ASM) is often remodelled, likely altering tissue stiffness. The underlying mechanism of how human ASM cell (hASMC) mechanosenses the aberrant microenvironment is not well understood. Physiological stiffnesses of the ASM were measured by uniaxial compression tester using porcine ASM layers under 0, 5 and 10% longitudinal stretch above in situ length. Linear stiffness gradient hydrogels (230 kPa range) were fabricated and functionalized with ECM proteins, collagen I (ColI), fibronectin (Fn) and laminin (Ln), to recapitulate the above-measured range of stiffnesses. Overall, hASMC mechanosensation exhibited a clear correlation with the underlying hydrogel stiffness. Cell size, nuclear size and contractile marker alpha-smooth muscle actin (αSMA) expression showed a strong correlation to substrate stiffness. Mechanosensation, assessed by Lamin-A intensity and nuc/cyto YAP, exhibited stiffness-mediated behaviour only on ColI and Fn-coated hydrogels. Inhibition studies using blebbistatin or Y27632 attenuated most mechanotransduction-derived cell morphological responses, αSMA and Lamin-A expression and nuc/cyto YAP (blebbistatin only). This study highlights the interplay and complexities between stiffness and ECM protein type on hASMC mechanosensation, relevant to airway remodelling in obstructive airway diseases.

Bronchial Thermoplasty Improves Ventilation Heterogeneity Measured by Functional Respiratory Imaging in Severe Asthma.

Purpose: Bronchial thermoplasty (BT) is a bronchoscopic intervention for the treatment of severe asthma. Despite demonstrated symptomatic benefit, the underlying mechanisms by which this is achieved remain uncertain. We hypothesize that the effects of BT are driven by improvements in ventilation heterogeneity as assessed using functional respiratory imaging (FRI). Patient and Methods: Eighteen consecutive patients with severe asthma who underwent clinically indicated BT were recruited. Patients were assessed at baseline, 4-week after treatment of the left lung, and 12-month after treatment of the right lung. Data collected included short-acting beta-agonist (SABA) and oral prednisolone (OCS) use, asthma control questionnaire (ACQ-5) and exacerbation history. Patients also underwent lung function tests and chest computed tomography. Ventilation parameters including interquartile distance (IQD; measure of ventilation heterogeneity) were derived using FRI. Results: 12 months after BT, significant improvements were seen in SABA and OCS use, ACQ-5, and number of OCS-requiring exacerbations. Apart from pre-bronchodilator FEV1, no other significant changes were observed in lung function. Ventilation heterogeneity significantly improved after treatment of the left lung (0.18 ± 0.04 vs 0.20 ± 0.04, p=0.045), with treatment effect persisting up to 12 months later (0.18 ± 0.05 vs 0.20 ± 0.04, p=0.028). Ventilation heterogeneity also improved after treatment of the right lung, although this did not reach statistical significance (0.18 ± 0.05 vs 0.19 ± 0.04, p=0.06). Conclusion: Clinical benefits after BT are accompanied by improvements in ventilation heterogeneity, advancing our understanding of its mechanism of action. Beyond BT, FRI has the potential to be expanded into other clinical applications.

Adipose tissue in the small airways: How much is enough to drive functional changes?

Obesity is a contributing factor to asthma severity; while it has long been understood that obesity is related to greater asthma burden, the mechanisms though which this occurs have not been fully elucidated. One common explanation is that obesity mechanically reduces lung volume through accumulation of adipose tissue external to the thoracic cavity. However, it has been recently demonstrated that there is substantial adipose tissue within the airway wall itself, and that the presence of adipose tissue within the airway wall is related to body mass index. This suggests the possibility of an additional mechanism by which obesity may worsen asthma, namely by altering the behaviour of the airways themselves. To this end, we modify Anafi & Wilson's classic model of the bistable terminal airway to incorporate adipose tissue within the airway wall in order to answer the question of how much adipose tissue would be required in order to drive substantive functional changes. This analysis suggests that adipose tissue within the airway wall on the order of 1%-2% of total airway cross-sectional area could be sufficient to drive meaningful changes, and further that these changes may interact with volume effects to magnify the overall burden.

Unraveling the Enigma of Organismal Death: Insights, Implications, and Frontiers.

Significant knowledge gaps exist regarding the responses of cells, tissues, and organs to organismal death. Examining the survival mechanisms influenced by metabolism and environment, this research has the potential to transform regenerative medicine, redefine legal death, and provide insights into life's physiological limits, paralleling inquiries in embryogenesis.

Text mining for disease surveillance in veterinary clinical data: part one, the language of veterinary clinical records and searching for words.

The development of natural language processing techniques for deriving useful information from unstructured clinical narratives is a fast-paced and rapidly evolving area of machine learning research. Large volumes of veterinary clinical narratives now exist curated by projects such as the Small Animal Veterinary Surveillance Network (SAVSNET) and VetCompass, and the application of such techniques to these datasets is already (and will continue to) improve our understanding of disease and disease patterns within veterinary medicine. In part one of this two part article series, we discuss the importance of understanding the lexical structure of clinical records and discuss the use of basic tools for filtering records based on key words and more complex rule based pattern matching approaches. We discuss the strengths and weaknesses of these approaches highlighting the on-going potential value in using these "traditional" approaches but ultimately recognizing that these approaches constrain how effectively information retrieval can be automated. This sets the scene for the introduction of machine-learning methodologies and the plethora of opportunities for automation of information extraction these present which is discussed in part two of the series.

Risk factors for tick attachment in companion animals in Great Britain: a spatiotemporal analysis covering 2014-2021.

BACKGROUND: Ticks are an important driver of veterinary health care, causing irritation and sometimes infection to their hosts. We explored epidemiological and geo-referenced data from > 7 million electronic health records (EHRs) from cats and dogs collected by the Small Animal Veterinary Surveillance Network (SAVSNET) in Great Britain (GB) between 2014 and 2021 to assess the factors affecting tick attachment in an individual and at a spatiotemporal level. METHODS: EHRs in which ticks were mentioned were identified by text mining; domain experts confirmed those with ticks on the animal. Tick presence/absence records were overlaid with a spatiotemporal series of climate, environment, anthropogenic and host distribution factors to produce a spatiotemporal regression matrix. An ensemble machine learning spatiotemporal model was used to fine-tune hyperparameters for Random Forest, Gradient-boosted Trees and Generalized Linear Model regression algorithms, which were then used to produce a final ensemble meta-learner to predict the probability of tick attachment across GB at a monthly interval and averaged long-term through 2014-2021 at a spatial resolution of 1 km. Individual host factors associated with tick attachment were also assessed by conditional logistic regression on a matched case-control dataset. RESULTS: In total, 11,741 consultations were identified in which a tick was recorded. The frequency of tick records was low (0.16% EHRs), suggesting an underestimation of risk. That said, increased odds for tick attachment in cats and dogs were associated with younger adult ages, longer coat length, crossbreeds and unclassified breeds. In cats, males and entire animals had significantly increased odds of recorded tick attachment. The key variables controlling the spatiotemporal risk for tick attachment were climatic (precipitation and temperature) and vegetation type (Enhanced Vegetation Index). Suitable areas for tick attachment were predicted across GB, especially in forests and grassland areas, mainly during summer, particularly in June. CONCLUSIONS: Our results can inform targeted health messages to owners and veterinary practitioners, identifying those animals, seasons and areas of higher risk for tick attachment and allowing for more tailored prophylaxis to reduce tick burden, inappropriate parasiticide treatment and potentially TBDs in companion animals and humans. Sentinel networks like SAVSNET represent a novel complementary data source to improve our understanding of tick attachment risk for companion animals and as a proxy of risk to humans.

Quantification of smooth muscle in human airways by polarization-sensitive optical coherence tomography requires correction for perichondrium.

Quantifying airway smooth muscle (ASM) in patients with asthma raises the possibility of improved and personalized disease management. Endobronchial polarization-sensitive optical coherence tomography (PS-OCT) is a promising quantitative imaging approach that is in the early stages of clinical translation. To date, only animal tissues have been used to assess the accuracy of PS-OCT to quantify absolute (rather than relative) ASM in cross sections with directly matched histological cross sections as validation. We report the use of whole fresh human and pig airways to perform a detailed side-by-side qualitative and quantitative validation of PS-OCT against gold-standard histology. We matched and quantified 120 sections from five human and seven pig (small and large) airways and linked PS-OCT signatures of ASM to the tissue structural appearance in histology. Notably, we found that human cartilage perichondrium can share with ASM the properties of birefringence and circumferential alignment of fibers, making it a significant confounder for ASM detection. Measurements not corrected for perichondrium overestimated ASM content several-fold (P < 0.001, paired t test). After careful exclusion of perichondrium, we found a strong positive correlation (r = 0.96, P < 0.00001) of ASM area measured by PS-OCT and histology, supporting the method's application in human subjects. Matching human histology further indicated that PS-OCT allows conclusions on the intralayer composition and in turn potential contractile capacity of ASM bands. Together these results form a reliable basis for future clinical studies.NEW & NOTEWORTHY Polarization-sensitive optical coherence tomography (PS-OCT) may facilitate in vivo measurement of airway smooth muscle (ASM). We present a quantitative validation correlating absolute ASM area from PS-OCT to directly matched histological cross sections using human tissue. A major confounder for ASM quantification was observed and resolved: fibrous perichondrium surrounding hyaline cartilage in human airways presents a PS-OCT signature similar to ASM for birefringence and optic axis orientation. Findings impact the development of automated methods for ASM segmentation.

Postnatal steroids as lung protective and anti-inflammatory in preterm lambs exposed to antenatal inflammation.

BACKGROUND: Lung inflammation and impaired alveolarization precede bronchopulmonary dysplasia (BPD). Glucocorticoids are anti-inflammatory and reduce ventilator requirements in preterm infants. However, high-dose glucocorticoids inhibit alveolarization. The effect of glucocorticoids on lung function and structure in preterm newborns exposed to antenatal inflammation is unknown. We hypothesise that postnatal low-dose dexamethasone reduces ventilator requirements, prevents inflammation and BPD-like lung pathology, following antenatal inflammation. METHODS: Pregnant ewes received intra-amniotic LPS (E.coli, 4 mg/mL) or saline at 126 days gestation; preterm lambs were delivered 48 h later. Lambs were randomised to receive either tapered intravenous dexamethasone (LPS/Dex, n = 9) or saline (LPS/Sal, n = 10; Sal/Sal, n = 9) commencing <3 h after birth. Respiratory support was gradually de-escalated, using a standardised protocol aimed at weaning from ventilation towards unassisted respiration. Tissues were collected at day 7. RESULTS: Lung morphology and mRNA levels for inflammatory mediators were measured. Respiratory support requirements were not different between groups. Histological analyses revealed higher tissue content and unchanged alveolarization in LPS/Sal compared to other groups. LPS/Dex lambs exhibited decreased markers of pulmonary inflammation compared to LPS/Sal. CONCLUSION: Tapered low-dose dexamethasone reduces the impact of antenatal LPS on ventilation requirements throughout the first week of life and reduces inflammation and pathological thickening of the preterm lung IMPACT: We are the first to investigate the combination of antenatal inflammation and postnatal dexamethasone therapy in a pragmatic study design, akin to contemporary neonatal care. We show that antenatal inflammation with postnatal dexamethasone therapy does not reduce ventilator requirements, but has beneficial maturational impacts on the lungs of preterm lambs at 7 days of life. Appropriate tapered postnatal dexamethasone dosing should be explored for extuabtion of oxygen-dependant neonates.

Evaluating ChatGPT text mining of clinical records for companion animal obesity monitoring.

BACKGROUND: Veterinary clinical narratives remain a largely untapped resource for addressing complex diseases. Here we compare the ability of a large language model (ChatGPT) and a previously developed regular expression (RegexT) to identify overweight body condition scores (BCS) in veterinary narratives pertaining to companion animals. METHODS: BCS values were extracted from 4415 anonymised clinical narratives using either RegexT or by appending the narrative to a prompt sent to ChatGPT, prompting the model to return the BCS information. Data were manually reviewed for comparison. RESULTS: The precision of RegexT was higher (100%, 95% confidence interval [CI] 94.81%-100%) than that of ChatGPT (89.3%, 95% CI 82.75%-93.64%). However, the recall of ChatGPT (100%, 95% CI 96.18%-100%) was considerably higher than that of RegexT (72.6%, 95% CI 63.92%-79.94%). LIMITATIONS: Prior anonymisation and subtle prompt engineering are needed to improve ChatGPT output. CONCLUSIONS: Large language models create diverse opportunities and, while complex, present an intuitive interface to information. However, they require careful implementation to avoid unpredictable errors.

PetBERT: automated ICD-11 syndromic disease coding for outbreak detection in first opinion veterinary electronic health records.

Effective public health surveillance requires consistent monitoring of disease signals such that researchers and decision-makers can react dynamically to changes in disease occurrence. However, whilst surveillance initiatives exist in production animal veterinary medicine, comparable frameworks for companion animals are lacking. First-opinion veterinary electronic health records (EHRs) have the potential to reveal disease signals and often represent the initial reporting of clinical syndromes in animals presenting for medical attention, highlighting their possible significance in early disease detection. Yet despite their availability, there are limitations surrounding their free text-based nature, inhibiting the ability for national-level mortality and morbidity statistics to occur. This paper presents PetBERT, a large language model trained on over 500 million words from 5.1 million EHRs across the UK. PetBERT-ICD is the additional training of PetBERT as a multi-label classifier for the automated coding of veterinary clinical EHRs with the International Classification of Disease 11 framework, achieving F1 scores exceeding 83% across 20 disease codings with minimal annotations. PetBERT-ICD effectively identifies disease outbreaks, outperforming current clinician-assigned point-of-care labelling strategies up to 3 weeks earlier. The potential for PetBERT-ICD to enhance disease surveillance in veterinary medicine represents a promising avenue for advancing animal health and improving public health outcomes.

Airway-associated adipose tissue accumulation is increased in a kisspeptin receptor knockout mouse model.

Airway-associated adipose tissue increases with body mass index and is a local source of pro-inflammatory adipokines that may contribute to airway pathology in asthma co-existing with obesity. Genetic susceptibility to airway adiposity was considered in the present study through kisspeptin/kisspeptin receptor signalling, known to modulate systemic adiposity and potentially drive airway remodelling. Therefore, the aim of the study was to determine the effects of kisspeptin/kisspeptin receptor signalling in the lung, focusing on airway-associated adipose tissue deposition and impact on airway structure-function. Wild-type, heterozygous and kisspeptin receptor knockout mice were studied at 6 or 8 weeks of age. Lung mechanics were assessed before and after methacholine challenge and were subsequently fixed for airway morphometry. A separate group of mice underwent glucose tolerance testing and bronchoalveolar lavage. At 6 weeks of age, kisspeptin/kisspeptin receptor signalling did not affect body adiposity, airway inflammation, wall structure or function. Despite no differences in body adiposity, there was a greater accumulation of airway-associated adipose tissue in knockout mice. By 8 weeks of age, female knockout mice displayed a non-diabetic phenotype with increased body adiposity but not males. Airway-associated adipose tissue area was also increased in both knockout females and males at 8 weeks of age, but again no other respiratory abnormality was apparent. In summary, airway-associated adipose tissue is decoupled from body adiposity in prepubescent mice which supports a genetic susceptibility to fatty deposits localised to the airway wall. There was no evidence that airway-associated adipose tissue drives pathology or respiratory impairment in the absence of other environmental exposures.

Dose-related cardiac outcomes in response to postnatal dexamethasone treatment in premature lambs.

BACKGROUND: Postnatal corticosteroids are used in the critical care of preterm infants for the prevention and treatment of bronchopulmonary dysplasia. We aimed to investigate the effects of early postnatal dexamethasone therapy and dose on cardiac maturation and morphology in preterm lambs. METHODS: Lambs were delivered prematurely at ~128 days of gestational age and managed postnatally according to best clinical practice. Preterm lambs were administered dexamethasone daily at either a low-dose (n = 9) or a high-dose (n = 7), or were naïve to steroid treatment and administered saline (n = 9), over a 7-day time-course. Hearts were studied at postnatal Day 7 for gene expression and assessment of myocardial structure. RESULTS: High-dose dexamethasone treatment in the early postnatal period led to marked differences in cardiac gene expression, altered cardiomyocyte maturation and reduced cardiomyocyte endowment in the right ventricle, as well as increased inflammatory infiltrates into the left ventricle. Low-dose exposure had minimal effects on the preterm heart. CONCLUSION: Neonatal dexamethasone treatment led to adverse effects in the preterm heart in a dose-dependent manner within the first week of life. The observed cardiac changes associated with high-dose postnatal dexamethasone treatment may influence postnatal growth and remodeling of the preterm heart and subsequent long-term cardiac function.

Dogs with cropped ears in the UK: A population-based study using electronic health records.

BACKGROUND: Ear cropping for cosmetic purposes was made illegal in the UK in 2006. Despite this, a lack of import regulations and celebrity and media influences mean cropped dogs are increasingly reported. METHODS: The demographics, temporal trends and patient-level associated factors for dogs with cropped ears were evaluated in a large sentinel population of dogs visiting UK veterinary practices. RESULTS: A total of 132 dogs with cropped ears were identified, with rates peaking in 2021. In 84 cases (63.6%), there was evidence of importation, most commonly from countries where cropping is also illegal, including Romania, Hungary, Bulgaria, Serbia, Spain, Poland and Ireland. American Bulldogs, Dobermanns, Italian Mastiffs (Cane Corso), Bulldogs and Mastiffs were all significantly overrepresented. Affected dogs were more likely to be unneutered (odds ratio 11.04, 95% confidence interval 5.84-20.90). LIMITATIONS: The study likely underestimates true levels of ear cropping. Identified cases are from a sentinel network of veterinary practices, and as such may not be representative of the wider UK population. CONCLUSION: These data suggest a need to educate owners and veterinary surgeons about the welfare and legal implications of ear cropping. The data presented can inform future targeted policies in veterinary practices and at a governmental level.

Seasonality and other risk factors for fleas infestations in domestic dogs and cats.

Fleas in the genus Ctenocephalides are the most clinically important parasitic arthropods of dogs and cats worldwide yet risk factors that might increase the risk of infestation in small animals remains unclear. Here we developed a supervised text mining approach analysing key aspects of flea epidemiology using electronic health records from domestic cats and dogs seen at a sentinel network of 191 voluntary veterinary practices across Great Britain between March 2014 and July 2020. Our methods identified fleas as likely to have been present during 22,276 of 1,902,016 cat consultations (1.17%) and 12,168 of 4,844,850 dog consultations (0.25%). Multivariable logistic regression modelling found that animals originating from areas of least deprivation were associated with 50% reductions in odds of veterinary-recorded flea infestation compared to the most deprived regions in England. Age of the animal was significantly associated with flea presentation in both cats and dogs, with cases peaking before animals reached 12 months. Cases were recorded through each study years, peaking between July and October, with fluctuations between each year. Our findings can be used towards healthcare messaging for veterinary practitioners and owners.