Bone is a composite material consisting of mineral, collagen and water in a complex hierarchical structure ¹. Bone water exists in the form of free water in the Haversian canals or lacunar-canalicular system, as well as bound water either loosely bound to collagen or more tightly bound to mineral ². Pore water content provides a surrogate measure of cortical porosity, which is difficult to measure directly in vivo due to the limited spatial resolution of most imaging techniques ³. Loosely bound water content provides a surrogate measure of collagen content, which is inaccessible with x-ray based techniques ⁴. Ultrashort echo time (UTE) sequences with TEs as short as 8 µs can potentially detect signal from pore water and loosely bound water ^5-7. In this study, we introduce an approach for fast volumetric mapping of bound and pore water content in vivo using a clinical 3T MR scanner.A 3D UTE Cones sequence (Figure 1) was implemented on a GE 3T Signa TwinSpeed MR scanner. The sequence was further combined with an adiabatic inversion preparation pulse (IR-Cones) for selective imaging of bound water ⁷. T₁ of pore water (T₁^PW) can be measured with dual-echo 3D Cones acquisitions with a series of TRs (TRs = 7.4, 10, 15, 20 ms) based on Fig.2 Eq. [1]. The T₁ of bound water (T₁^BW) can be measured with IR-Cones acquisitions with a series of TR/TI combinations (e.g., TR/TI = 50/24; 100/45; 200/81; 300/110 ms) based on Fig 2. Eq. [2], on the condition that each TR/TI combination satisfies the inversion and nulling of pore water requirement (so that only bound water is detected) ⁸. Total water concentration (WC^Total) can be measured by comparing Cones signal intensity of bone with that of a water phantom (water doped with MnCl₂ with T₂* similar to that of bone) based on Fig 2. Eq. [3]. Bound water concentration (WC^Bound) can be measured by comparing the IR-Cones signal intensity of bone with that of the water phantom based on Fig 2. Eq. [4]. Pore water concentration (WC^Pore) can be calculated as the difference between total and bound water concentration based on Fig 2. Eq. [5]. η is a coil sensitivity correction. Typical imaging parameters included a flip angle of 20°, a bandwidth of 250kHz, a FOV of 15cm, matrix of 192×192, slice thickness of 8cm, 10 slices, a total scan time of 20min including 7min for T₁^CW measurement, 5min for T₁^PW measurement, 4 min for total water content measurement and 4 min for bound water content measurement. Five healthy volunteers were recruited for this feasibility study.

Figure 2 IR-UTE imaging of the tibia-midshaft of a 27y healthy male volunteer. Exponential fitting of the IR-Cones signal with variable TR/TI combinations shows a short T₁^BW of 137±7 ms for bound water.

Figure 3 shows volumetric imaging and mapping of total, collagen-bound and pore water in the tibial midshaft of the same volunteer using a clinical 3T scanner. Both Cones and IR-Cones sequences show high signal for cortical bone and a water phantom. Further analysis shows single component signal decay for the 3D IR-UTE signal (results not shown), confirming that only collagen-bound water is detected. A mean bound water concentration of ~18% and pore water concentration of ~4% was found in the tibial mid-shaft of the healthy volunteers.

Figure 4 shows volumetric total and bound water mapping in ulna and radius of the forearm. Chemical shift effect leads to increased total water content, especially for the thin cortical cortex in the ulna. Higher spatial resolution and sampling bandwidth are expected to minimize this error.

We have demonstrated that collagen-bound water and pore water in cortical bone can be measured with 3D Cones and IR-Cones sequences using a clinical whole-body scanner, with correction of T₁^PW measured with dual-echo Cones acquisitions with variable TRs, and T₁^CW measured with a series of IR-Cones acquisitions with different TR/TI combinations. Volumetric mapping of total, collagen-bound and pore water concentration in vivo, including T₁^CW measurement for collagen-bound water, can be achieved in less than 15 minutes. Further clinical studies will be necessary to evaluate the diagnostic power of this method, including studies of patients with osteoporosis, osteomalacia, osteopenia, Paget disease, insufficiency fractures in the setting of biophosphonate therapy, raloxifene treatment, etc.