Purpose: Evaluation of osteoarthritis (OA) is hampered by a lack of non-invasive tools to effectively detect early and reversible changes associated with the disease. Early OA is characterized by cartilage proteoglycan depletion related to an increase in enzymatic degradation. GagCEST is a promising quantitative MR technique for assessing content of Glycosaminoglycan (GAG) side chains attached to proteoglycans at 7T and is of particular interest due to its specificity to GAG content without the use of exogenous contrast agents or specialized hardware. While gagCEST has been shown to be a useful biomarker for cartilage composition at 7T, there is limited evidence of its utility at 3T. In this study, we assess the utility of an optimized gagCEST protocol at 3T by comparing gagCEST asymmetry in subjects with established OA to healthy controls. Methods: Ten healthy subjects (5 female and 5 male, ages 20-30) and twenty-four subjects with medial compartment OA were scanned on a 3T MRI system. The OA cohort (ages 60-70) was grouped by Kellgren-Lawrence (KL) grade (assessed using radiography) with five subjects with a KL grade of 1 (4 female and 1 male), twelve with a KL grade of 2 (8 female and 4 male), and five with a KL grade of 3 (4 female and 3 male). Other inclusion criteria included no previous ACL reconstruction surgery and BMI < 35. GagCEST images were acquired with a magnetization prepared 3D spoiled gradient-echo sequence (imaging parameters in Table 1). After correcting for magnetic field inhomogeneities, gagCEST asymmetry (%) was calculated using the B0-field corrected signal intensity at ±1.0 ppm, and the chemical shift of GAG hydroxyl protons, using the equation: gagCESTasym= [S(-1.0 ppm)-S(+1.0 ppm)]/ S0]. Femoral cartilage was manually segmented and subdivided into medial and lateral subregions. Statistical comparisons between cartilage compartments and healthy vs OA subjects were made using an ANOVA 2-way test with a Tukey HSD Test for multiple comparisons. Results: Subjects with OA showed a reduced gagCEST asymmetry compared to healthy controls. Further, there was a trend of an inverse relationship between GagCEST asymmetry and KL grade. There was a statistically significant difference in gagCEST asymmetry among KL grade values (p = 0.019). Further analysis comparing healthy subjects to the different OA KL grade subjects, showed a significant difference between healthy and OA KL-2 subjects and OA KL-3 subjects (p = 0.022 and p = 0.046 respectively). The comparison between healthy and KL-1 subjects did not show a significant difference (p = 0.536), which may be due to insufficient power in this cohort. There was no statistically significant difference between medial and lateral compartments in the femoral cartilage (p = 0.977). Additionally, a negative correlation of GagCEST with KL grade was observed (r= -0.404, p=0.01). The results show that optimized gagCEST methods at 3T were able to detect a statistically significant reduction in gagCEST asymmetry in osteoarthritic subjects, known to have cartilage degeneration and lower biochemical GAG composition, compared to healthy controls. Further analysis showed a significant difference in gagCEST asymmetry between the healthy and OA KL-2 and KL-3 subjects. This decrease can be seen in the unrolled gagCEST asymmetry maps, especially in the medial compartment of the knee, which corresponds well with the diagnosis in our cohort of medial compartment OA. Further, at a total acquisition time of 8:76 minutes for whole joint coverage, these methods offer translational potential. Conclusions: Our results demonstrate promise for the application of gagCEST at 3T to detect differences between healthy cartilage and cartilage in subjects with OA. Taken together, our results show that GagCEST could be a viable tool to assess early degenerative changes in cartilage in clinical trials. A strength of MRI over X-ray is that it is more sensitive to early changes in cartilage composition that occur due to OA. GagCEST has the potential for distinguishing these changes that are unseen in X-rays because it may be sensitive to signs of changes in GAG content that radiographs are not able to visualize.View Large Image Figure ViewerDownload Hi-res image Download (PPT)View Large Image Figure ViewerDownload Hi-res image Download (PPT)