Bone marrow changes in beta-thalassemia major: quantitative MR imaging findings and correlation with iron stores

Abstract

The purpose of this study is to describe the MR imaging features of bone marrow in beta-thalassemia major and investigate their relation to ferritin, liver and spleen siderosis. Spinal bone marrow was prospectively assessed on abdominal MR studies of 40 transfused beta-thalassemic patients and 15 controls using T1-w, Pd, T2*-w Gradient Echo (GRE) and T1-w turbo Spin Echo (TSE) sequences. Signal intensity (SI) ratios of liver, spleen and bone marrow to paraspinous muscles (L/M, S/M, B/M respectively) and the respective T2 relaxation rates (1/T2) were calculated. Serum ferritin levels were recorded. Bone marrow hypointensity in at least T2*-w GRE sequence was noted in 29/40 (72.5%) patients. Eleven/40 patients exhibited normal B/M on all MR sequences. Five/40 patients had normal B/M and low L/M. B/M correlated with L/M in T1-w TSE sequence only (r = 0.471, p = 0.05). B/M correlated with S/M and mean ferritin values in all sequences (r > 0.489, p < 0.01 and r > - 0.496, p < 0.03 respectively). Marrow 1/T2 did not correlate with ferritin values or liver and spleen 1/T2. B/M in transfused beta-thalassemic patients is related to splenic siderosis and ferritin levels. Although marrow is usually hypointense, it may occasionally display normal SI coexisting with liver hypointensity, a pattern typical of primary hemochromatosis.

Use of fat suppression in R2 relaxometry with MRI for the quantification of tissue iron overload in beta-thalassemic patients

“…In contradistinction, we documented a weaker correlation between bone marrow R 2 without fat suppression and ferritin (r=0.49) and no correlations between bone marrow, hepatic and pancreatic R 2 values without fat suppression. A previous study of our group with a larger number of patients with β-thalassemia revealed only a moderate correlation between bone marrow R 2 values and ferritin, splenic and hepatic R 2 values, but not with pancreatic R 2 values [18], whereas other studies [20,25] did not document statistically significant correlations between bone marrow, hepatic and pancreatic siderosis, estimated with the T2, T2⁎ or SIR methods.…”

“…To the best of our knowledge, no previous study has applied fat suppression in multiecho sequences for R 2 or R 2 ⁎ relaxometry for iron overloaded tissues well known to contain significant amounts of fat [18][19][20][21][22][23][24][25]. Schwenzer et al [32] have been the only group that had applied fat suppression in a multiecho gradient sequence for T2* calculation in a large cohort of 129 healthy subjects, and they found a significant correlation between hepatic T2* and ferritin (r=0.62) but no correlation between pancreatic T2* and ferritin.…”

“…Moreover, in extrahepatic tissues with iron overloaded and large amounts of fat, such as the pancreas and vertebral bone marrow, MR determination of iron cannot be representative of the degree of iron deposition due to severe fat influences. This may explain why MR measurements for pancreas and bone marrow with any method have yielded no correlation or poor correlations with indices of iron stores such as serum ferritin or hepatic siderosis so far, even in studies with large numbers of patients with iron overload pathologic conditions [18][19][20][21][22][23][24][25][26]. Recent research studies report methods of R 2 * quantification of liver iron after correction for fat [27] or allow discrimination of fat and water by separate reconstruction on opposed-phase images [28].…”

Bone marrow MR imaging as predictors of outcome in hemopoietic stem cell transplantation

“…Previously, a few imaging studies decribed that the main findings of bone marrow in thalassemic patients were red marrow hyperplasia and marrow haemosiderosis [14,15]. Recently, Drakonaki et al [24] had prospectively assessed the spinal bone marrow signal intensity ratio on abdominal MR studies of 40 transfused beta-thalassemic patients and correlated with ferritin values. In the current study, we focused on the MR assessment of femoral marrow in beta-thalassemic patients and its capacity to predict HSCT outcome.…”

“…Another limitation is the lack of biopsies due to ethical reasons. However, MR imaging is considered accurate for iron quantification even more than biopsy, due to the examining of the entire organ [24] and sensitivity of FFE images with long TE to slight iron deposits [18]. Iron and trabecular bone and firbrotic tissue could produce low signal intensity in both the gradient echo and TSE sequences [26].…”