OBJECTIVE To assess the ability of a contrast-enhanced magnetic resonance imaging (MRI) technique to quantitatively determine glycosaminoglycan content in canine articular cartilage. METHODS Fifty-four full-thickness cartilage discs were collected from the femorotibial and scapulohumeral joints of three adult dogs immediately following euthanasia. One set of discs from each dog was analysed for glycosaminoglycan content using a colourimetric laboratory assay. The remaining position-matched set of discs from contralateral limbs underwent pre- and post-contrast gadolinium-enhanced MRI, using repeated saturation recovery pulse sequences which were used to generate calculated T1 maps of the cartilage discs. Linear regression analysis was then performed relating delayed gadolinium-enhanced MRI T1 calculated signal intensity to the cartilage glycosaminoglycan content normalized to DNA content. Repeatability of triplicate measurements was estimated by calculating the coefficient of variation. RESULTS Mean coefficient of variation estimates for the gadolinium-enhanced MRI T1 signal intensity values for nine sampling sites from three dogs ranged from 5.9% to 7.5%. Gadolinium-enhanced MRI T1 signal intensity was significantly correlated (p <0.05) with normalized glycosaminoglycan content in two dogs (r = 0.79, p = 0.011; r = 0.78, p = 0.048), but not in the third dog (r = 0.53, p = 0.071). CONCLUSIONS Gadolinium-enhanced MRI assessment of cartilage may be predictive of glycosaminoglycan content and therefore offer an in vivo assessment of changes in cartilage characteristics over time. Additional studies appear indicated to determine the reliability and clinical applicability of gadolinium-enhanced MRI in detecting changes in cartilage over time.