Urinary metabolite profiles differ by tumor size and malignancy in dogs with mammary tumors.

Background: Urine metabolomics may offer a non-invasive way to detect systemic metabolic alterations associated with canine mammary tumors, but previous studies have been few and have not accounted for diet as a potential confounder. Methods: In this observational case-control study, urine from 118 client-owned female dogs in Finland was analyzed by untargeted liquid chromatography-mass spectrometry, including 71 dogs with mammary tumors and 47 tumor-free controls. Case dogs were further evaluated by tumor malignancy (28 benign, 27 malignant, 16 with both benign and malignant tumors), tumor size (53 small, 12 medium, 6 large), and tumor number (43 single, 28 multiple). Statistical models were adjusted for diet, age, body size, and sterilization status. Results: After covariate adjustment, no features remained significant in the overall case-control comparison or in analyses of tumor number, whereas tumor size showed the clearest signal. Three features remained significant in the covariate-adjusted analysis of variance for tumor size, and per-contrast linear models identified four false discovery rate-significant urinary features associated with tumor size and four associated with tumor malignancy. All showed lower intensities in tumor-bearing dogs than in controls. Hippurate was the most robust and consistently supported finding, with lower levels in dogs with large tumors. Several unknown aromatic and sulfur-containing features covaried with hippurate and tumor size, suggesting a broader host metabolic response. Exploratory pathway analysis indicated glucose homeostasis as the strongest enrichment signal in adjusted models, although this did not remain significant after false discovery rate correction. Adjustment for diet and other background variables substantially reduced the number of significant findings, highlighting the importance of accounting for physiological and environmental variation in veterinary metabolomics. Conclusion: Urine metabolomics detected a limited but biologically plausible metabolic signature associated primarily with tumor size rather than with overall case-control status or tumor number. Hippurate and a small set of unknown features emerged as candidate urinary markers of tumor-associated systemic metabolic change, supporting the value of urine as a non-invasive matrix for studying canine mammary tumors while underscoring the need for careful covariate control.