Thalassemia is the most common genetic disorder worldwide [1]. Non-transfusion dependent thalassemia (NTT) includes a variety of phenotypes, most commonly investigated forms are: β-thalassemia intermedia, β-thalassemia minor and α-thalassemia intermedia (hemoglobin H disease). Unlike β-thalassemia major, regular transfusion therapy is not mandatory for patients' survival [2].

For these patients, iron overload resulting from increased intestinal iron absorption represents an important clinical problem which progresses with age [3]. Serum ferritin is a well-known marker for indirect measurement of iron overload; however its concentration is affected by other factors, e.g. inflammation, chronic disorders and liver diseases [4]. Liver biopsy is another essential diagnostic tool for the estimation of the extent of iron deposition in the hepatic lobules; however, being an invasive technique, it carries the risk of postinterventional complications e.g. bleeding. Also, it cannot assess the overall iron burden within the whole organ [5].

Iron chelation therapy is used for management of non-transfusion dependent thalassemia after the age of ten (more than fifteen years in deletional hemoglobin H phenotype). It is indicated when iron concentration reaches 5 mg Fe/g or over dry weight; or when serum ferritin level exceeds 800 ng/mL because these thresholds carry the risk of iron-overload related morbidity [6].

Early diagnosis of cardiac iron overload is mandatory, as it requires intensified chelation treatment; otherwise irreversible heart failure could be inevitable [7]. Yet, usually the diagnosis is delayed because cardiac iron overload is unpredictable, symptoms and abnormal echocardiographic results appear late [1].

MRI is now established to be a reliable non-invasive modality for the diagnosis of liver iron overload [8,9]. This is owing to the high sensitivity of multi-echo gradient T2* sequences to the excess tissue iron. Quantitative assessment is achieved by fitting the appropriate decay model to the average signal intensity at various echo times, where tissue iron overload has magnetic susceptibility effect that cases the signal decay due to shortening of T2 time constant [10]. Recently, it has also been used to estimate myocardial iron overload in adult patients with acquired anemia. Gradient-echo T2* can rapidly assess cardiac iron overload in the same session of liver iron evaluation [11].

Therefore, the objective of this study was to assess the value of gradient-echo T2* in monitoring and screening of both liver and cardiac iron overload in non-transfusion dependent thalassemia. Measurements were obtained in the same 1.5 MRI examination, followed by calculation with the reference spreadsheet and then, the results were compared to the standard serum ferritin levels.

Patients

This observational cross-sectional prospective study was carried out in the period between April 2015 and August 2016 in the University Children’s Hospital and Magnetic resonance imaging (MRI) unit of Radiodiagnosis department. The study included children with mean age 11 years (ranged from 4 to 18 years), who suffered from nontransfusion dependent thalassemia (NTDT) attending or on follow-up in hematology outpatient clinic. The study was conducted after IRB approval as well as informed consent taken from legal guardians and assent from the older children (more than 12 year).

Detailed history and data including age, sex, and duration of illness, history of blood transfusion, affected parents or other affected siblings were recorded. Physical examination was carried out with special emphasis on complexion, growth parameters, and stage of puberty, size of the liver and spleen and presence of characteristic facies, splenectomy or cholecystectomy.

Serological study

Basic laboratory investigations were documented and confirmed including CBC, reticulocytic count, serum bilirubin (total and direct), serum ferritin (at the time of study), liver function tests, serum calcium, phosphorus and alkaline phosphatase, chest x-ray and abdominal ultrasound (at the time of study).

Imaging

MR imaging was obtained by using a 1.5-T, whole-body scanner (Acheiva; Philips Medical Systems, Best, Netherlands). Initial morphological T1 weighted sequences were carried out followed by axial gradient echo breath-hold multiecho T2* sequences; acquired with electrocardiogram synchronization. The parameters were as follows: TR/TE1/delta TE, 19/2.1/1.4, followed by echo train of 12 echo times from 2.2 to 19.3 milliseconds; flip angle 20 degrees; slice thickness, 10 mm; interslice gap, 2 mm; matrix size, 256 × 128; field of view, 350 × 350 × 10 mm; number of signals acquired, 1; acquisition time, 13 seconds.

Post processing and data analysis

Regions of interest (ROIs) of fixed area were placed in the liver parenchyma, clear of the hepatic veins and biliary tree, to measure the signal intensity of the affected hepatic segments (Figure 1). For cardiac T2* assessment, ROI was placed in the interventricular septum in all images, encircling both epicardial and endocardial regions, but away from the right ventricular insertion points (to avoid artifacts) (Figure 2).

Figure 1: Calculation of liver T2* value in miliseconds plotted against signal intensity obtained from axial images at 12 increasing TEs, with fitting of monoexponential decay curve. In this case TE was 10.2 ms and LIC was 3.02 mg/g; denoting moderate iron overload.

Figure 2: Calculation of myocardial T2* values obtained from shortaxis images at 16 increasing TEs. In this case TE was 29.3 ms and MIC was 0.73 mg/g; denoting normal T2* value.

To create exponential decay curves TE and SI values were manually inserted into the Excel sheet where SI (y-axis) is plotted against TE (xaxis) to make a graph (Figures 1 and 2). Truncation of data was achieved by excluding points where SI was very low (flat).

Statistical data analysis

Results are presented as mean ± SD or median (minimummaximum). Spearman correlation, age and gender-related analysis were done. The statistical procedures were performed with SPSS statistical software v20 (SPSS, Inc., Chicago, IL, USA).

The study included 31 patients with non-transfusion dependent thalassemia (NTT); 19 with thalassemia intermedia, 6 with thalassemia minor and 6 with hemoglobin H. Mean serum ferritin was 201.8 ng/mL (median 112.6; 9 ng/mL min and 663 ng/mL max). Mean liver T2* was measured at 23.7 ms (median 25; 3.2 ms min and 44 ms max). Mean myocardial T2* was measured at 30 ms (median 29 ms; 19.9 ms min and 67 ms max). These values are corresponding to mean liver iron concentration (LIC) 2.6 mg/g ± 3.8 SD; 0.69 min, 20.4 max and mean myocardial iron concentration (MIC) 0.75 mg/g ± 0.23 SD; 0.27 min, 1.17 max.

The mean, median values and ranges of serum ferritin and T2* values of the liver and cardiac iron concentrations for the 3 phenotypes of NTDT are shown in Tables 1 and 2.

Table 1: Serum ferritin in the 3 NTDT traits.

Table 2: T2* values of Liver and myocardial iron concentration in the 3 NTDT traits.

Four cases of thalassemia intermedia and one case with thalassemia minor trait showed increase in liver T2* values (8.1, 10.2, 10.4, 5.4 ms and 8.5 ms) corresponding to LIC 3.85, 3.02, 3, 5.7 and 3.63 mg/g, respectively. This was reported as mild liver iron overload, its threshold is T2* 3.8-1.14 ms, LIC 2-7; according to Garbowski et al. [12].

One case with thalassemia intermedia and one case with Hb-H showed increase in liver T2* values (3.2 and 4.4 ms) corresponding to LIC 9.8 and 11.7 mg/g, respectively. Moderate liver iron overload threshold is T2* 1.8-3.8 ms, LIC 7-15 mg/g [12].

Three cases with thalassemia intermedia were recorded as high normal T2* values 19.9, 19.9 and 20.1 ms, corresponding to 1.17, 1.17 and 1.08 mg/g, respectively. Mild myocardial iron overload threshold is 15-20 ms, MIC 1.16-1.65 mg/g; according to Carpenter et al. [13]. Meanwhile, none of the recorded myocardial T2* values was beyond the pathological threshold.

Correlation analysis revealed significant negative correlation between hepatic T2* and serum ferritin. In contrast, a week positive correlation was found for cardiac T2* and a week negative correlation was between hepatic and cardiac T2* values (Table 3).

Table 3: Correlation analysis for T2* values.

A statistically significant negative correlation with age was detected for hepatic T2* but not for cardiac T2* and serum ferritin (Table 4). Gender related analysis revealed no significant statistical correlation for serum ferritin and T2* values with sex.

Table 4: Correlation of T2* values and serum ferritin with age.

In thalassemia, iron overload results from both elevated gastrointestinal absorption and transfusion therapy [14]. In nontransfusion dependent thalassemia (NTDT), iron overload is primarily due to gut absorption. Despite the fact that transfusion dependence may develop with age, organ failure and death are far less common as compared with severe forms of thalassemia [15].

Although numerous studies have confirmed MRI as the method of choice for tissue iron assessment in thalassemia major and transfusion dependent traits, only a few studies were conducted on NTT traits [16-18].

In our study 5 cases (16%) showed mild liver iron overload and 2 cases (6.5%) showed moderately elevated liver iron concentration. All other cases (77.5%) had T2* values within the normal range. Hepatic T2* values significantly correlated with serum ferritin and with patients' ages (Tables 3 and 4). These findings are in agreement with previous studies which reported that although serum ferritin correlated with LIC, it was found to be a poor predictor of LIC [17,18]. They explained this to be probably related to a portal first pass effect, less use of iron chelation therapy and older age of patients as compared with severe forms of transfusion dependent thalassemia.

Our results showed that all T2* values were within the normal range (Table 2). No significant correlation was found between myocardial T2* and serum ferritin and week correlation was found between hepatic and cardiac T2* values (Table 3). A previous study stated that marked discordance was found between hepatic and cardiac iron overload as well as myocardial T2* and serum ferritin in a large study cohort comprising 106 patients with thalassemia major [19]. Similarly, studies conducted on thalassemia intermedia trait reported the absence of cardiac iron overload despite hepatic iron accumulation [16,20,21].

These findings are in agreement with previous studies which reported that thalassemia intermedia trait is less prone to develop cardiac overload than thalassemia major (15,18). A study conducted on animal models for iron overload in thalassemia intermedia also indicated the early hepatic iron accumulation and delayed cardiac iron accumulation over longer periods of time [22].

Meanwhile, other studies reported cardiac T2* values above the pathologic threshold in thalassemia intermedia patients who were regularly transfused and on iron chelation therapy [16,17]. Accordingly the existence of factors like transfusion history, chelation therapy or co-pathology has an impact on the location and extent of cardiac iron accumulation [23,24]. Therefore, it is recommended that the decision for the start and regimen of chelation therapy should be spared for older age group patients with regular transfusion, cardiac hemosiderosis or liver fibrosis [17].

In conclusion, liver and cardiac T2* measurement is the non-ivasive modality of choice for monitoring and screening of both liver and cardiac iron overload in NTDT. From our experience, cardiac iron overload is uncommon in this disease population even in cases with mild and moderate hepatic overload.

We recommend further investigation on NTDT on a wider scale and the inclusion of functional cardiac MRI assessment. A threshold value of LIC that could be associated with cardiac overload may be appreciated and should be sought.