Tendinopathies of the superficial digital flexor tendon rank among the most common orthopaedic diseases in horses and their healing often results in chronic disease processes. Besides ultrasound, a low-field magnetic resonance imaging (MRI) system which allows examinations of standing, sedated horses has been established for the diagnosis of tendon lesions and monitoring of its healing success. In human medicine, high-field MRI systems are the gold standard for this purpose, whereas 3 T MRI systems in particular are still rarely used in horses. For the treatment of tendinopathies, the intralesional injection of mesenchymal stromal cells (MSCs) is a promising therapeutic approach. In acute lesions, positive effects, especially on the recurrence rate of lesions, have been demonstrated which are probably primarily due to the immunomodulatory effects of MSCs. However, little is known about the efficiency of MSCs in chronic lesions and their potential mechanisms of action so far. Therefore, the first aim of the present thesis was to characterise chronic tendon lesions in the low-field MRI system established for horses as well as in a 3 T high-field MRI system. For this investigation, dissected superficial digital flexor tendons were used in vertical orientation in the low-field system, as it is common for living horses. In the high-field MRI, the tendons were positioned both horizontally and, analogous to the low-field system, vertically in order to identify possible effects of the different positioning on the visualisation of tendon lesions. A total of n = 195 MRI images per sequence, representing the same anatomical areas, were evaluated qualitatively by scoring and quantitatively by measuring the signal intensities of the tendon and, where appropriate, also of the lesion. One histologic sample per tendon (n = 12) served as the gold standard for analysing the detectability of lesions. The agreement of the qualitative assessment in the different sequences was analysed based on Cohen’s kappa statistics. In addition, the agreement of the measured signal intensities was tested using the calculation of Spearman´s rank correlations and the backtransformed logarithmised limits of agreement. The tendons unambiguously identified as either healthy or chronically diseased were subsequently used as scaffolds in a cell culture study aiming to extend the knowledge on the influence of chronically altered tendon matrix on the regeneration-promoting properties of MSCs. For this purpose, in the first part of the study, the behaviour of MSCs from six different donor horses cultured on scaffolds from one healthy tendon was analysed in comparison to MSC aggregate cultures. Then, MSCs from the donor horse best reflecting the results of the first part of the experiment were used for seeding scaffolds of five diseased and five healthy tendons and incubated for up to 21 days. The focus in both parts of the study was on the tenogenesis of the MSCs, their expression of matrix components and their matrix remodelling properties. Differences between experimental groups were analysed using Mann-Whitney-U tests, while differences over time were analysed using Wilcoxon signed-rank tests. Regarding the characterisation of chronic tendon lesions on MRI, all histologically proven lesions could be detected in low-field MRI using the T1-weighted (w) gradient recalled echo (GRE) sequence. In high-field examinations, this was only possible with the T1w fast-low-angle-shot (FLASH) sequence when the tendon was positioned vertically. The agreement of the detectability of lesions in the high-field sequences with the T1w GRE low-field sequence, analysed using Cohen’s kappa statistics, was better when using the FLASH sequence or vertical orientation of the tendon than when using the T1w spin-echo (SE) sequence or horizontal orientation of the tendon. Moreover, the agreement of the signal intensity of the tendons in high-field MRI was higher when using different sequences but the same positioning (r = 0.964 and r = 0.957; p < 0.001) than when using the same sequence but different positioning (r = 0.833 and r = 0.802; p < 0.001). In the cell culture study, culturing MSCs as aggregates supported cell fitness, demonstrated by a higher expression of the matrix components tenascin-C, collagen 1A2 and collagen 3A1, and the growth factors transforming growth factor ß (TGFB)1 and -3 than in the scaffold-based culture (p < 0.01 for tenascin-C on day 6, collagen 1A2 on days 3 and 6, and TGFB1 and -3 on day 3, p < 0.05 for collagen 3A1 on day 3 and TGFB1 on day 6). In contrast, the matrix remodelling activities of MSCs cultured on scaffolds were higher, indicated by a higher expression of the matrix-metalloproteinases (MMPs) MMP9 and MMP13 (p < 0.01 for MMP9 on day 6 and MMP13 on day 3) and a lower expression of the tissue-inhibitors-of-metalloproteinases (TIMPs) TIMP1 and TIMP3 (p < 0.05 for TIMP1 on day 6 and p < 0.01 for TIMP3 on day 3). This was confirmed by a higher release of glycosaminoglycans into the medium (p < 0.01 on day 3 and 6) and a higher MMP activity (p < 0.05 on day 6) in the scaffold-based group. The alterations in the extracellular matrix of tendons with chronic lesions did not lead to significant differences in tenogenic differentiation and only minor differences in the expression of matrix components by the MSCs. However, the matrix components collagen 3A1 and decorin as well as TGFB3 were initially (day 3) expressed less by MSCs on scaffolds of chronically diseased tendons (p < 0.05 for decorin, p < 0.01 for collagen 3A1 and TGFB3). Furthermore, matrix remodelling by MSCs was reduced within the first week of culture in diseased tendon scaffolds, demonstrated by a lower expression of MMP1, MMP3, MMP13 and MMP14, as well as TIMP2 (p < 0.05 for MMP3 at day 6, p < 0.01 for MMP1, MMP13 and MMP14, and TIMP2 at day 3). With this doctoral thesis, the level of knowledge on diagnosis and treatment options for chronic tendon lesions could be expanded. The results of the first study included here support the use of the low-field MRI system for the visualisation of chronic tendon lesions. In 3 T high-field MRI, FLASH sequences were better suited for this than SE sequences. However, the vertical positioning of the tendon in the magnetic field was more crucial for the visualisation of tendon lesions, which might be due to randomly oriented individual tendon fibres causing the magic angle effect in that position. With regard to the therapy of chronic tendon lesions, injection of MSCs cannot be recommended without reservation as yet. Despite the unaffected tenogenic differentiation on scaffolds of chronic lesion areas, the lowered expression level of various genes can be interpreted as a maladaptation of the MSCs. This is, on the one hand, the reduced expression of TGFB3, but more importantly, the reduced matrix remodelling by the MSCs, which is most likely required, at least to some extent, for the therapeutic success in chronic fibrotic tendon lesions. However, the adaptation effect of the MSCs could possibly be avoided by adequate pre-conditioning, as the behaviour of the MSCs in both groups converged over time.