Bone marrow fat distribution in patients with beta-thalassemia: A study using chemical shift-based water-fat MRI

• Main Authors: Ismail, Umi Nabilah, Azlan, Che Ahmad, Khairullah, Shasha, Raja Aman, Raja Rizal Azman, Lee, Kuan Jin, Yeong, Chai Hong, Omar, Nur Farhayu, Ramli, Norlisah, Shah, Mohammad Nazri Md, Jackson, Nicholas, Ng, Kwan Hoong
• Format: Article
• Published: Elsevier Science Inc 2022
• Subjects: Medical technology
• Online Access: http://eprints.um.edu.my/42341/
• Institution: Universiti Malaya

Rationale and objectives: Molecular studies have shown the changes in bone marrow fat in relation to altered hematopoiesis. This study aims to investigate the changes in the bone marrow fat in patients affected by beta-thalassemia by using chemical shift-encoded (CSE)-MRI. Materials and methods: Twenty-three subjects, comprising of six healthy (17-31 years old) and 17 beta-thalassemia subjects (19-39 years old), were scanned using a multiecho fast low angle shot sequence (0.94 x 0.94 x 3.00 mm(3)) and a stimulated echo acquisition mode sequence using 3T MRI. Bone marrow proton density fat fraction (PDFF) was quantified in the left femur of each subject. Regression and Bland-Altman analysis were used to analyze agreement between CSE-MRI and 1H-MRS. PDFF distribution was analyzed using Hartigan's dip test and the computed Wasserstein distances. Jonckheere-Terpstra trend analysis was performed to evaluate the effect of disease severity on PDFF distribution. Results: An excellent agreement was found between PDFF measured using CSE-MRI with H-1-MRS (R-2 = 0.91; bias = -1.41%). Healthy subjects showed left-skewed or bimodal PDFF distribution while beta-thalassemia subjects showed bimodal, normal or right-skewed distribution. Jonckheere-Terpstra test shows that PDFF distribution was increasingly different from the norm as disease severity increased (TJT = 166.0, z = 3.806, p < 0.05). Increase in variability of PDFF distribution within each subject group was also seen with increasing disease severity (TJT = 169.0, z = 3.971, p < 0.05). Conclusion: CSE-MRI is a promising tool to demonstrate spatial changes and variability in marrow fat distribution, resulting from ineffective erythropoiesis.