As a pathological condition primarily afflicting the elderly, osteoporosis is becoming more prevalent with today’s lengthened life expectancy. Osteoporosis-associated fractures are dangerous, often associated with lethal complications, and impose a huge economic burden on society. Unfortunately, current techniques used to diagnose the disease cannot provide reliable fracture risk prediction, as they rely upon apparent bone mineral density (BMD) only. Recent advances in short echo time (TE) imaging have made it possible to examine different properties of bone with magnetic resonance imaging (MRI). Water bound to the collagen matrix (bound water, BW), water residing in the pore space (pore water, PW), and phosphorus (31P) in bone mineral crystals have previously been proposed as surrogate markers for bone matrix density, porosity and mineral density, to be quantitatively imaged with either 1H or 31P ultra-short TE (UTE) or zero TE (ZTE) sequence. In this dissertation, an integrated MRI protocol measuring both bone matrix and mineral properties in vivo was first designed and tested with clinical hardware. MRI-derived PW was negatively correlated with high resolution peripheral quantitative computed tomography (HR-pQCT) derived BMD, while bone mineral content based on MRI-derived phosphorus density was positively correlated with that based on HR-pQCT BMD. Second, in vivo bone phosphorus relaxation times were studied in a small cohort of healthy volunteers (aged 29 to 65). The relative invariability of relaxation properties obviated the need for individual measurements in this healthy cohort. Third, gradient imperfections were found to introduce errors in UTE image-based bone water quantification and to undermine measurement agreement across scanners, therefore requiring correction. Finally, the suggested protocol was applied in an ongoing osteoporosis treatment study. Expected observations –– elevated pore water, lowered bound water and phosphorus densities –– were observed in the patient group relative to healthy controls in limited baseline data acquired thus far. In conclusion, this dissertation proves the feasibility of measuring bone matrix and mineral surrogates in a clinical setting, and may aid in better predicting osteoporotic fractures.