In vivo calibration of the T2* cardiovascular magnetic resonance method at 1.5 T for estimation of cardiac iron in a minipig model of transfusional iron overload

Figures

• 

  Fig. 1. Individual pig timelines of dextran-iron dosing and timepoints of investigations. a Illustrates for each pig in group A the weekly intramuscular dosing of dextran iron, given in mg Fe/kg BW/week, the duration of iron loading in months and the timepoints of cardiovascular magnetic resonance (CMR) and of endomyocardial biopsies (EMB), performed at the end of the sessions. The dose of dextran iron in mg Fe/kg/ week was fixed in all pigs during the whole follow-up, but the total weekly dose injected was adjusted to the body weight (BW) immediately after each CMR/biopsy session. Pig 2, 4, 6 and 7 died of biopsy complications, indicated by * (for details see the Results). b Illustrates the dextran-iron dosing and duration of iron loading of the pigs in group B, comprising 16 pigs, who had taken a full-wall biopsy from the cardiac septum immediately after last CMR session followed by euthanasia. Group B also includes pig 8 to 11 from group A, who also had these biopsies taken after euthanasia. Principles of iron loading are identical in both groups, except in pig 13 and 14: The dextran-iron dose was increased from 5 and 10 mg Fe/g, respectively, to 35 mg/kg in both pigs after about 12…

• 

  Fig. 2. Investigation of segmental and transmural myocardial iron distribution. The photograph shows a mid-myocardial cut slice from the formalin-fixed heart of pig 8. The slice was cut into 6 segments: Anteroseptal (1), inferolateral (2), inferior (3), inferoseptal (4), anterolateral (5) and anterior (6). Each segment was divided into an epicardial-, mesocardial and endocardial layer (as indicated for segment 3), leading to 18 tissue samples from each heart for chemical cardiac iron concentration (CIC) assessment. Epicardial layers of the right ventricular septum (segment 1 and 2) represent endocardium

• 

  Fig. 3. Color microphotograph of iron stains of sections from cardiac septum during iron loading. Light photomicrographs of endomyocardial right-ventricular catheter-biopsy sections taken during follow up of iron-dextran loading in pig 8, having received weekly, intramuscular dextran-iron injections, corresponding to 100 mg Fe/kg, for 15 months. Sections were stained using Perls’ Prussian blue reaction to demonstrate intracellular iron. Upper left corner in each panel indicates the month of iron loading when the biopsy was taken (left value) and contemporary chemical cardiac iron concentration (CIC) in mg/g (right value). Scale bar is 80 µm

• 

  Fig. 4. Simultaneous follow up of cardiac R2* and chemical CIC by EMBs during iron loading. Cardiac T2* (a) and chemical CIC (b) were followed up with 1–3 months intervals during dextran-iron loading in 11 minipigs for up to 15 months. Pigs 2, 4, 6, 7 died due to biopsy-related complications after 6.5, 5, 2 and 2 months of iron loading, respectively. In pig 4, parallel 5-months results of CIC and R2* were not available. Chemical CIC was assessed by ICP in EMBs, taken from cardiac septum in connection to the CMR sessions. For each pig, the weekly applied dextran-iron dose is indicated within the figure label as mg/kg/BW

• 

  Fig. 5. Calibration of cardiac R2* with CIC assessed in EMBs. The figure displays the relationship of cardiac R2* (= 1/T2*) to CIC. Non-parametric regression analysis (Passing and Bablok) led to the calibration equation [CIC] = (R2* − 17.16/41.12 with CIC in mg/g DW and R2* in Hz. The total number of parallel cardiac R2* and chemical CIC measurements was 55. Solid line is the regression line and dotted lines display its confidence interval. ULN for cardiac R2* is 28.18 Hz and ULN for CIC 0.31 mg Fe/g DW

• 

  Fig. 6. Calibration of cardiac T2* with full-wall biopsies. Full-wall biopsies were taken from cardiac septum (anteroseptal region) from 16 minipigs for assessment of chemical CIC by ICP. Biopsies were taken immediately after final CMR session, performed after completed iron loading. Fourteen of the 16 pigs had been dextran-iron loaded from 10.5 to 22 months by weekly intramuscular dextran-iron injections, and 2 had not (control pig). Weekly dextran-iron dose was from 5 to 140 mg Fe/kg BW. a Displays the relationship of cardiac R2* to chemical CIC, investigated by classical linear regression analysis, leading to the calibration equation [CIC] = (R2* − 17.19)/31.82 with CIC in mg/g DW and R2* in Hz. In b non-parametric regression analysis was used (Passing and Bablok), leading to the calibration equation [CIC] = (R2* − 16.71)/31.33. Solid lines represent the regression lines and dotted lines display their confidence interval. c Displays the same data as in panel A and B, but the individual cardiac R2* measurements are given with error bars (mean of triplicate measurements ± SD)

• 

  Fig. 7. Comparison of chemical CIC measurements with T2*-based CIC estimates using 3 different calibration curves. a–c Display the relationship between CIC values, assessed chemically by use of post-mortem full-wall biopsies from cardiac septum (independent variable) and the corresponding CIC values predicted by the T2* method (dependent variable) using different calibration curves. In a, CIC values were predicted by the calibration curve obtained by use of EMBs for chemical assessment of CIC. In b, the predicted CIC values were calculated by use of the calibration curve from Carpenter et al. [12], based on a study on formalin-fixed hearts from TM patients with transfusion iron overload (TIO). In c, the calibration curve from Wood et al. [11] was used, based on dextran-iron loaded Mongolian gerbils. For details see the Results. Fat, dotted lines represent lines of equality. d–f display the agreement between CIC, assessed by use of post-mortem full-wall biopsies, and by CIC values predicted by use of the calibration curve coming from EMBs (d), from Carpenter et al. [12] (e) and from Wood et al. [11] (f). Agreement is assessed by Bland–Altman plots of the mean of 2 methods (x-axis),…

• 

  Fig. 8. Segmental and transmural distribution of myocardial iron. One mid-myocardial slice was cut from formalin-fixed dextran iron-loaded hearts. Formalin-fixed hearts were available from 11 pigs (7, 8, 12, 13, 14, 16, 17, 18, 19, 20, and 21). The slice was cut into 6 segments: Anteroseptal (1), inferolateral (2), inferior (3), inferoseptal (4), anterolateral (5) and anterior (6). Each section was divided into an epicardial-, mesocardial and endocardial layer, leading to 18 samples from each heart for assessment of chemical CIC (for details see Fig. 2 and the Results). a to f display segmental and transmural CIC values for each pig. g to l display the corresponding mean CIC levels of all 11 pigs for each of the 18 different samples given in mg Fe/g. Significant differences are indicated by horizontal bars and p-values. Label panel gives weekly dextran-iron dose for each pig in mg Fe/kg BW

• 

  Fig. 9. Individual segmental iron distribution investigated in 3 myocardial layers. The individual CIC values are displayed in cell line graphs for the endocardial (a), mesocardial (b) and the epicardial layer (c) for 6 different segmental myocardial regions: Anteroseptal (1), inferolateral (2), inferior (3), inferoseptal (4), anterolateral (5) and anterior (6). CIC values are given as mg Fe/g DW. Label panel gives weekly dextran-iron dose for each pig in mg Fe/g BW