Correlation between MRI-derived and biopsy-confirmed liver iron concentration in patients with chronic liver disease

Objectives Elevated liver iron concentration (LIC) promotes liver inflammation and fibrosis. LIC can be invasively measured in liver biopsies (Biopsy-LIC) but also non-invasively quantified by MRI. This study compared Biopsy-LIC to whole-liver MRI-LIC in chronic liver disease (CLD) patients using two relaxometry techniques: 3D multi-echo gradient echo processed with Liver-Lab (LL-LIC) and 2D multi-echo gradient echo processed with MRQuantif (MRQ-LIC). Materials and methods Retrospectively, 64 CLD patients (35 females; mean age 50.9 ± 13.5 years) underwent both liver biopsy and 3 T MRI within a median interval of 19 (6–76) days. Two independent readers compared Biopsy-LIC with LL-LIC and MRQ-LIC. Group differences were assessed via regression models adjusted for age, gender, BMI, fat fraction, liver stiffness, and iron panel values. Statistical agreement between measurements was analyzed using Pearson correlation, ICC, and Bland–Altman methodology. Results Mean Biopsy-LIC, LL-LIC, and MRQ-LIC were 0.73 ± 0.80, 0.93 ± 0.59, and 1.06 ± 0.54 mg /g, respectively. Biopsy-LIC strongly correlated with both LL-LIC (r = 0.83) and MRQ-LIC (r = 0.84). Correlation between LL-LIC and MRQ-LIC was even stronger (r = 0.95, p < 0.001). Agreement was excellent between LL-LIC and MRQ-LIC (ICC = 0.946, CI 0.912–0.967) and good between Biopsy-LIC and MR-LICs (ICC = 0.821–0.824). Median biases were −0.29 mg /g (Biopsy-LIC vs. LL-LIC), −0.43 mg /g (Biopsy-LIC vs. MRQ-LIC), and −0.14 mg /g (LL-LIC vs. MRQ-LIC). Conclusion In CLD patients, MRI-based LIC estimates showed strong correlation and excellent agreement with biopsy-derived LIC, regardless of MRI technique.