Multicenter validation of spin-density projection-assisted R2-MRI for the noninvasive measurement of liver iron concentration

Abstract

Purpose: Magnetic resonance imaging (MRI)-based techniques for assessing liver iron concentration (LIC) have been limited by single scanner calibration against biopsy. Here, the calibration of spin-density projection-assisted (SDPA) R2-MRI (FerriScan®) in iron-overloaded β-thalassemia patients treated with the iron chelator, deferasirox, for 12 months is validated.

Methods: SDPA R2-MRI measurements and percutaneous needle liver biopsy samples were obtained from a subgroup of patients (n = 233) from the ESCALATOR trial. Five different makes and models of scanner were used in the study.

Results: LIC, derived from mean of MRI- and biopsy-derived values, ranged from 0.7 to 50.1 mg Fe/g dry weight. Mean fractional differences between SDPA R2-MRI- and biopsy-measured LIC were not significantly different from zero. They were also not significantly different from zero when categorized for each of the Ishak stages of fibrosis and grades of necroinflammation, for subjects aged 3 to <8 versus ≥8 years, or for each scanner model. Upper and lower 95% limits of agreement between SDPA R2-MRI and biopsy LIC measurements were 74 and -71%.

Conclusion: The calibration curve appears independent of scanner type, patient age, stage of liver fibrosis, grade of necroinflammation, and use of deferasirox chelation therapy, confirming the clinical usefulness of SDPA R2-MRI for monitoring iron overload.

Keywords: ESCALATOR; biopsy; deferasirox; iron overload; β-thalassemia.

Figure 1

a: LIC measurements by SDPA R2-MRI vs. LIC measurements by biopsy (n = 233). The solid line is the line of equivalence; b: Bland–Altman plot showing the fractional differences between the two measurements of LIC for each subject. The solid line shows the mean fractional difference and the dotted lines show the 95% limits of agreement between the two measurements. Symbol shape indicates scanner used: • Philips NT Intera; ▪ Philips Gyroscan; ♦ Siemens Sonata; ▴ GE Signa Excite; ▾ Siemens Symphony.

Figure 2

Distribution of (a) Ishak fibrosis stages and (b) Ishak necroinflammatory grades for patients in the study. Three patients unknown.

Figure 3

Effect of fibrosis stage on agreement between SDPA R2-MRI and biopsy measurements of LIC.

Figure 4

Fractional differences between LIC measurements by SDPA R2-MRI and biopsy for patients with different (a) liver fibrosis stages and (b) necroinflammatory grades. Means and standard errors of the means are shown. Fibrosis stages and necroinflammatory grades were unavailable for three of the 233 patients.

Figure 5

Fractional differences between LIC measurements by SDPA R2-MRI and biopsy for 233 patients on five different MRI scanners. Means and standard errors of the means are shown. The median (range) of mean LIC measured on each scanner were: Philips NT Intera, 16.2 (4.7–50.1) mg Fe/g dw; Philips Gyroscan, 9.5 (0.7–42.6) mg Fe/g dw; Siemens Sonata, 27.4 (12.5–38.6) mg Fe/g dw; GE Signa Excite, 9.3 (1.4–39.8) mg Fe/g dw; and Siemens Symphony, 21.4 (6.4–40.2) mg Fe/g dw.

Figure 6

Fractional differences between LIC measurements by SDPA R2-MRI and biopsy for 39 subjects aged less than 8 years, and 194 patients aged 8 years and above. Means and standard errors of the means are shown. The medians (ranges) of the mean LIC measurements for the subjects less than 8 years old were 13.7 (5.1–38.0) mg Fe/g dw and for the subjects 8 years old and above were 13.4 (0.7–50.1) mg Fe/g dw.