Machine learning-assisted screening for canine Cushing's syndrome.

Cushing's syndrome (CS) is a common endocrine disorder in dogs that can significantly impair their quality of life. Diagnosis is often challenging because of its variable clinical presentation, making it difficult to identify suitable candidates for further diagnostic tests. This study employed machine learning algorithms to assist in CS diagnosis using routinely available screening diagnostics, including complete blood count, serum chemistry panel, and urinalysis parameters such as urine specific gravity and urine protein-to-creatinine ratio. Data were collected from 153 control dogs initially suspected of CS but later excluded and 152 dogs with confirmed CS. A boosted tree algorithm (gradient boosting) was trained on 80% of the collected data, with the remaining 20% reserved for testing. The developed model demonstrated an accuracy of 88.5% [95% confidence interval (CI): 80.5-96.5%], a sensitivity of 83.3% (95% CI: 70.7-96.7%), a specificity of 93.5% (95% CI: 84.9-100%), and an area under the receiver operating characteristic curve of 0.912 (95% CI: 0.835-0.988), indicating excellent discriminatory ability. A user-friendly graphical interface was also developed to facilitate clinical implementation, potentially improving diagnostic efficiency and owner satisfaction.