In vivo microCT arthrography (µCTa) can be used to measure both quantity (volumetric) and quality (glycosaminoglycan content) of cartilage. This study investigated the accuracy of four segmentation techniques to isolate cartilage from µCTa datasets and then used the most accurate one to investigate if the µCTa method could show osteoarthritic changes in rat models during longitudinal follow-up. Volumetric measurements and glycosaminoglycan contents of patellar cartilage from in vivo µCTa-scans were compared with an ex vivo gold standard µCT-scan. Cartilage was segmented with three global thresholds and one local threshold algorithm. Comparisons were made for healthy and osteoarthritic cartilage. Next, three rat models were investigated for 24 weeks using µCTa. Osteoarthritis was induced by injection with a chemical (mono-iodoacetate), a surgical intervention (grooves applied in articular cartilage), and via exercise (strenuous running). After euthanasia, all knee joints were isolated for histology. Local thresholds accurately segmented cartilage from in vivo µCTa scans and best measured cartilage quantity and glycosaminoglycan content. Each of the three osteoarthritic rat models showed a specific pattern of osteoarthritis progression. All µCTa results were comparable to histology. In vivo µCTa is a sensitive technique for imaging cartilage degradation. Local thresholds enhanced the sensitivity of this method and will probably more accurately detect disease-modulating effects from interventional strategies. The data from rat models may serve as a reference for the time sequence of cartilage degeneration during in vivo testing of new strategies in osteoarthritis treatment.
Copyright © 2011 Orthopaedic Research Society.