Surrogate Thresholds for Iron Chelation in Patients with Hematological Malignancies

Background: The reference standard for evaluating the magnitude of body iron overload is measurement of the liver iron concentration (LIC). Until recently, the conventional clinical method for measuring LIC was through analysis of biopsy specimens considered the "gold standard" of LIC measurement. To counter the shortcomings of liver biopsy, several noninvasive techniques have been applied to estimate LIC. The spin-density projection-assisted (SDPA) R2-MRI (FerriScan) technique has been recently validated to estimate LIC. There are reports in the literature that ferritin levels in patients with hematological malignancies (HM) may not correlate sufficiently with LIC. We conducted R2-MRI Ferriscan sequencing in consecutive patients with HM followed in our center to better evaluate the diagnostic value of ferritin levels and packed red blood cell (pRBC) transfusions for iron overload.

Methods: HM patients with a history of transfusion requirement were evaluated for inclusion in the study within the Liver Iron Cutino Network (Licnet) project. Patients were excluded if they had contraindications to MRI assessment (heart pacemaker, metallic foreign body or devices, and claustrophobia). Demographics and HM disease data were collected, together with ferritin levels and transfusion requirement. R2 LIC measures by FerriScan were defined as indicating normal/marginally high iron load for values 0 to <3, mild liver iron overload for values 3 to < 7, moderate between 7 and 15 and severe if >15. Correlations and the generation of receiver operating characteristic (ROC) curves were performed to test the diagnostic power of ferritin levels and volume of transfused blood (ml/kg in the preceeding 12 months) to indicate moderate-severe iron overload.

Results: 28 patients (20 males) of mean age 62, SD 17 years with HM (myelodysplastic syndromes N=14; acute myeloid leukemia N=7, of which 5 post-transplant; acute lymphoblastic leukemia post-transplant N=2; aplastic anemia N=2; non Hodgkin lymphoma N=1; chronic lymphocytic leukemia N=1; erythroblastopenia N=1) were evaluated. Median baseline ferritin levels were 1617 (interquartile range 1185-2360). Mean R2 values were 10.9, SD 10.1 mg/kg. Three patients had normal or marginally high LIC, 12 patients had mild overload, 6 patients moderate and 7 patients severe overload. Both ferritin levels and amount of blood volume transfused in the preceding 12 months correlated well with R2 measures (Spearman rho=0.478, p=0.10 and r=0.509, p=0.006, respectively). However, neither ferritin levels nor transfusions correlated with R2 MRI in patients with normal-mild iron overload while the correlations were confirmed and highly significant in patients with moderate-severe iron overload (r=0.692, p=0.009 and r= 0.754, p=0.003, respectively). ROC curves denied the diagnostic power of ferritin levels and confirmed transfusions to indicate moderate to severe liver iron overload (Figure 1), indicating that 37 ml/kg/year (about 14 units of pRBC) has a sensitivity of 77% and a specificity of 80% to predict moderate-severe iron overload

Conclusions: We confirm that, though ferritin levels do generally correlate with LIC, they are not to be considered an appropriate surrogate marker for iron overload in HM. R2 MRI sequencing is valuable to diagnose the state of LIC. The results of this study indicate that as soon as an HM patient begins to receive regular transfusions, since the probability of developing moderate-severe iron overload is high, it is warranted not to await high ferritin levels, nor a large number of pRBC units, to initiate iron chelation.

View largeDownload slide

View largeDownload slide

Close modal