Impact of MRI technique on clinical decision-making in patients with liver iron overload: comparison of FerriScan- versus R2-derived liver iron concentration

Objectives The purpose of this study was to compare clinical decision-making in iron overload patients using FerriScan and an R2*-based approach. Methods One-hundred and six patients were imaged at two consecutive timepoints (454 ± 158 days) on a 1.5-T Siemens MAGNETOM Avanto Fit scanner. For both timepoints, patients underwent the standard FerriScan MRI protocol. During the second exam, each patient additionally underwent R2*-MRI mapping. For each patient, a retrospective (simulated) decision was made to increase, decrease, or maintain chelator levels. Two different decision models were considered: The fixed threshold model assumed that chelator adjustments are based strictly on fixed liver iron concentration (LIC) thresholds. Decisions made with this model depend only on the most recent LIC value and do not require any clinician input. The second model utilized decisions made by two hematologists retrospectively based on trends between two consecutive LIC values. Agreement ( κ A ) between hematologists (i.e., interobserver variability) was compared with the agreement ( κ B ) between a single hematologist using the two different LIC techniques. Results Good agreement between R2*- and FerriScan-derived decisions was achieved for the fixed threshold model. True positive/negative rates were greater than 80%, and false positive/negative rates were less than 10%. ROC analysis yielded areas under the curve greater than 0.95. In the second model, the agreement in clinical decision-making for the two scenarios ( κ A vs. κ B ) was equal at the 95% confidence level. Conclusions Switching to R2*-based LIC estimation from FerriScan has the same level of agreement in patient management decisions as does switching from one hematologist to another. Key Points • Good agreement between R2*- and FerriScan-derived decisions in liver iron overload patient management • Switching to R2*-based LIC estimation from FerriScan has the same level of agreement in patient management decisions as does switching from one hematologist to another.