Elastography is an emerging imaging modality for characterizing tendon injury in horses, but its ability to differentiate tissue deformability relative to treatment group and biochemical properties using a prospective, experimental study design remain unknown. Objectives of the current study were to (a) to investigate differences in glycosaminoglycan, DNA, and soluble collagen levels in mesenchymal stem cell (MSC) treated limbs compared to untreated control limbs utilizing a collagenase model of tendinopathy; (b) compare elastographic features between treatment groups; and (c) determine tissue-level predictive capabilities of elastography in relation to biochemical outcomes. Bone marrow was collected for MSC culture and expansion. Tendinopathy of both forelimb deep digital flexor tendons (DDFTs) was induced with collagenase under ultrasonographic guidance. One randomly assigned limb was treated with intra-lesional MSC injection with the opposite limb serving as an untreated control. Horses were placed into a controlled exercise program with elastographic evaluations performed baseline (0) and 14, 60, 90, and 214 days post-treatment. Postmortem biochemical analysis was performed. MSC-treated limbs demonstrated significantly less (42%) glycosaminoglycan (P = .006). Significant differences in elastographic region of interest (ROI) percent hardness, ROI color histogram, and subjective lesion stiffness were appreciated between treatment groups at various study time points. Elastographic outcome parameters were weak predictors of biochemical tissue analysis, with all R2 values ≤ 0.50. Within this range of differences in glycosaminoglycan content between treatment groups, elastography outcomes did not predict biochemical differences. Tissue-specific differences between DDFTs treated with MSCs compared to controls were apparent biochemically, but not predicted by elastography.