Transferrin Saturation and Serum Ferritin Are Main Predictors of Liver Iron Content in Subjects With...

Hyperferritinemia is widespread and is caused by or associated with a variety of diseases requiring complex diagnostic workup. Magnetic resonance imaging-based liver iron quantification (LIC^MRI) can differentiate subjects with iron overload requiring treatment but is still limited to specialized centers and costly. We aimed to determine whether parameters belonging to the diagnostic setting of hyperferritinemia could lead to a more selective use of LIC^MRI.

We enrolled a cohort of 570 subjects with hyperferritinemia from a tertiary hospital center. They underwent full clinical and laboratory evaluation, and LIC^MRI. We applied a partitioning tree to explore which features were relevant in classifying LIC^MRIseverity from grade 1 to 3.

A total of 66.1% had LIC^MRI≤3 mg/g (grade 1), while 11.2% had LIC^MRI above 7 mg/g (grade 3). The partitioning tree model showed a global accuracy of 78% (95% confidence interval, 74%-81%) in predicting the severity of LIC^MRI in the entire series. Sensitivity and specificity were high to distinguish patients with grade 1 from those with grade 3 but not from patients with grade 2. Focusing on the LIC^MRI threshold of 7 mg/g, we found that transferrin saturation (TSAT) ≥60% and TSAT <60% with serum ferritin ≥963 μg/L correctly classified 61 (95.3%) subjects with grade 3 LIC^MRI to class 3. Based on these cutoffs, 193 subjects with LIC^MRI grade 1 and 2 were also selected for LIC^MRI assessment, while 316 patients (55.4%) would be addressed to follow-up.

TSAT and serum ferritin allow us to identify more than 95% of patients with a LIC^MRI grade 3 and reduce MRI requirements by more than 50%.