Reproducibility of quantitative susceptibility mapping in lumbar vertebra

Background: To evaluate the reliability and reproducibility of quantitative susceptibility mapping (QSM) in the lumbar vertebra. Methods: From May to September 2017, 61 subjects who underwent QSM MRI and quantitative computed tomography (QCT) were consecutively enrolled in this prospective study. QSM examination was performed two times with an interval of less than 1 week for each subject. For each data set, the QSM and QCT values on L1-L4 vertebral bodies were measured independently by two radiologists. The correlation coefficient between QSM and QCT values was calculated on L1-L4 vertebral bodies. The intraclass correlation coefficient (ICC) and Bland-Altman plots were used to evaluate the inter-observer reliability and the inter-scan reproducibility on QSM. Results: A total of 61 subjects (mean age, 55.5 +/- 13.7 years) with 244 vertebral bodies were analyzed. Overall, QSM and QCT showed good correlation in the L1-L4 vertebral body, especially in the L3 (R=-0.75). QSM value showed excellent inter-observer reliability (ICC, 0.992, 95% CI: 0.985-0.996) with a mean difference of 0.35 and 95% limits of agreements of within-22.74 to 23.45 ppb, and very good inter-scan reproducibility (ICC, 0.877, 95% CI: 20.52-28.32) with a mean difference of-7.60 ppb and 95% limits of agreements of within of -92.85 to 77.62 ppb. Conclusions: QSM in the lumbar vertebra is a reliable and reproducible technique for evaluating bone mineral density.