AI to measure pain?

Hello, Sumarrians!

AI is starting to pop up everywhere. Sometimes in prominent places and other times more subtly. Here is one example that I would never have imagined.

Please give me feedback on ways I can make it more useful to you.

If it is valuable to you please consider sending it on to a friend 😄 

Humerus OCD in Dogs

The study you've provided discusses the condition known as proximal humeral osteochondrosis (OC) in dogs. Proximal humeral OC is a disorder that disrupts the normal process of bone development in the shoulder joint, leading to the formation of abnormally thick cartilage on the joint surface, which can detach within the joint. This condition primarily affects large- and giant-breed dogs, and it often results in lameness and muscle atrophy. Surgical intervention, involving the removal of the abnormal cartilage and the diseased bone underneath, is a common approach to treat this condition. The study aimed to evaluate the long-term outcomes of dogs that underwent surgical treatment for proximal humeral OC. The researchers used arthroscopic examination, diagnostic imaging, and owner-completed questionnaires to assess the dogs' clinical status. The study's findings indicated that despite surgical intervention, a majority of dogs showed evidence of progressive osteoarthritis (OA) and persistent defects in the joint's cartilage. The dogs exhibited decreased range of motion, muscle atrophy, and asymmetric loading of their forelimbs. However, interestingly, the study also noted that despite these abnormal findings, the dogs' actual limb usage, as measured by kinetic data, did not significantly differ between affected and unaffected limbs. The study also highlighted a potential discrepancy between the clinical assessments and the owners' perceptions of their dogs' function. While clinical examinations revealed abnormalities, many owners reported favorable outcomes in their dogs' daily activities. The study acknowledged some limitations, including the retrospective nature of the case selection, variability in surgical techniques, and potential impact of soft tissue pathologies on the results. The study suggested that surgical intervention for proximal humeral OC in dogs may offer clinical benefits, but it doesn't necessarily prevent the progression of osteoarthritis or restore the joint's original structure. The findings also indicated the need for further research into alternative treatment options and long-term comparisons between surgical and conservative management strategies for this condition in dogs.

Zann, GJ, Jones, SC, Selmic, LS, et al. Long-term outcome of dogs treated by surgical debridement of proximal humeral osteochondrosis. Veterinary Surgery. 2023; 52(6): 810-819. doi:10.1111/vsu.13866 

Bottom line — Offers clinical benefits but doesn’t stop the progression of degenerative changes. Paradigm shift in human medicine ???

Insect Bite Hypersensitivity Hope for Horses ???

Insect Bite Hypersensitivity (IBH) is a prevalent allergic condition in horses, characterized by various symptoms such as itching, hair loss, wounds, and crusts, often exacerbated by secondary infections. Despite its common occurrence, the intricate immune system interactions underlying IBH remain incompletely understood. The disease combines features of both type I and type IVb allergies, prominently involving eosinophils and cytokine-producing TH2 cells, particularly driven by interleukin 5 (IL-5). Conventional treatments like glucocorticoids are hampered by significant side effects, prompting the search for safer and effective alternatives. To address this, researchers have developed antibodies against IL-5, a key regulator of eosinophils, using phage display technology. These antibodies, carrying equine constant domains, were synthesized in mammalian cells and rigorously assessed for efficacy and stability. An innovative in vitro assay was established to gauge their ability to hinder IL-5 from binding to its receptor. The leading antibody exhibited a promising outcome, successfully inhibiting IL-5 receptor binding with an IC50 value of 4 nM. Importantly, safety and tolerability tests in horses indicated no adverse effects. In a groundbreaking in vivo study, the new antibody demonstrated both safety and indications of therapeutic effectiveness. The integration of these in vivo findings with the previously obtained in vitro data supports the conclusion that this antibody holds substantial promise as a potential treatment for IBH. Given the disease's impact on a significant number of horses and the current lack of targeted treatments, this specific antibody targeting IL-5 presents a novel avenue for managing IBH and improving the well-being of affected animals. While the precise molecular intricacies of IBH demand further exploration, this research provides a tangible and targeted solution to address the condition's underlying mechanisms. 

Recombinant antibodies against equine interleukin 5 to treat insect bite hypersensitivity. J Vet Pharmacol Therap, 46: 47-47. https://doi.org/10.1111/jvp.13206 

Bottom line — Possible hope for these challenging cases.

AI to Help Assess Pain via Grimace Score.

In this study, the researchers aimed to employ deep neural network and machine learning models to predict both facial landmark positions and Feline Grimace Scale© (FGS) scores in cats. The FGS is a tool used to assess pain in cats based on facial expressions. The study utilized a dataset of 3447 cat face images from various studies, each annotated with 37 facial landmarks corresponding to FGS action units (AU). The images were augmented using geometric and color modifications. For FGS scores, geometric descriptors (GD) including angles, distance ratios between landmarks, and area ratios of quadrilaterals (n=35) were used. For predicting facial landmark positions, the researchers employed pre-trained models and convolutional neural networks (CNN), with ShuffleNetV2 and EfficientNetB0 showing the best performance based on normalized root mean squared error (NRMSE). These models were selected based on predictive performance, inference time, and parameter count. Accuracy for distinguishing between painful and pain-free cats was achieved at 95.5%. For predicting FGS scores, XGBoost models incorporating the geometric descriptors were trained. The best regression model achieved a mean squared error (MSE) of 0.0104 for total FGS scores. The study found that the models' prediction errors were larger for some specific FGS action units, such as muzzle tension and whiskers change, while others like ear position and orbital tightening had smaller errors. Interestingly, models that used total FGS scores performed better than those focusing on specific action units. In conclusion, the study successfully demonstrated the use of deep learning-based predictors to automate the assessment of acute pain in cats using the FGS. The models achieved high accuracy in discriminating between painful and non-painful cats and showed promise in predicting FGS scores based on facial landmarks. 

Artificial intelligence for prediction of grimace scale scores for acute pain management: A model using the Feline Grimace Scale©. J Vet Pharmacol Therap, 46: 46-46. https://doi.org/10.1111/jvp.13204 

Bottom line — Could help improve the quality of patient care.

Just putting things in perspective …