A TRI-MODALITY COMPARISON OF VOLUMETRIC BONE MEASURE QUANTIFICATION USING 1.0 TESLA PERIPHERAL MAGNETIC RESONANCE IMAGING, PERIPHERAL QUANTITATIVE COMPUTED TOMOGRAPHY AND HIGH-RESOLUTION-PERIPHERAL QUANTITATIVE COMPUTED TOMOGRAPHY IMAGES

Abstract

<p>This comparative study of peripheral (p) QCT, high-resolution pQCT (hr-pQCT) and 1.0 Tesla pMRI technologies quantified short-term test-retest reproducibility, validity, one-year detection limit and clinical sensitivity of each modality’s derived bone measures. Select bone outcomes were evaluated from scans performed on Hamiltonian women above 50 years old and externally validated in a population-based cohort. In the local cohort (age: 74 ± 9 years and BMI 27.65 ± 5.74 kg/m²), Tb.Sp measured on pMRI, Ct.Th and vBMD from pQCT showed significant correlations (r²=0.52-0.85) with hr-pQCT, yielding slopes near unity. Bland-Altman analyses revealed significant relationships between pQCT and pMRI bone outcome values (Tb.Th(-), Tb.N(+), BV/TV(-)) and agreement with hr-pQCT. Short-term reproducibility was < 5% for pQCT but only BV/TV was < 5% for pMRI. Co-registration and excluding individuals with fractures mildly reduced precision error and one-year change. In the local cohort, only Ct.Th and cortical vBMD associated with fractures (OR: 1.09-3.28) using hr-pQCT, which was externally validated in the national cohort. Certain trabecular measures on pMRI and pQCT erred towards increased odds for fractures locally. For pQCT, these became significant in the national cohort (OR:1.04-3.81). This Canadian reference dataset for hr-pQCT showed larger Tb.Sp and smaller Tb.N compared to Americans but age-related decline in Ct.Th and BV/TV was faster in Europeans. This study demonstrated validity of pMRI and pQCT-derived volumetric bone outcomes and reasonable short- and long-term precision error for pQCT but not 1.0T pMRI. A single CT slice from pQCT was comparable to 110 slices from hr-pQCT in associations with fractures.</p>