Vol. 6, Issue 6 : Tissue Iron Deposition : MRI Quantitation in Clinical Trials

San Antonio (TX), Bolton (MA)

Multiple disease processes cause abnormal accumulation of iron in body tissues. These include severe hemoglobinopathies such as thalassemia and sickle cell disease requiring repeated transfusions, hemochromatosis, and other anemias and liver cirrhosis. In some hemoglobinopathies and hemochromatosis, excess iron deposition can lead to organ dysfunction and significant morbidity. For example, children with thalassemia often receive a significant iron load from repeated transfusions, saturating and exceeding their mechanisms for iron storage, leading to toxic accumulation of iron in multiple tissues, especially liver, heart, pancreas, and pituitary gland. Before modern therapies were introduced, cardiac dysfunction was the leading cause of death in these children. Parenteral iron chelation agents were developed 30-40 years ago, allowing removal of excess iron from body tissues and maintenance of safer iron levels, profoundly improving life expectancy for thalassemia patients. Traditionally, iron levels were measured by liver needle biopsy, an invasive procedure that allows direct iron measurements but suffers from limited repeatability and potential sampling error. Blood tests such as ferritin levels are also used but are of limited specificity.

During the last decade, MRI techniques have been developed which allow noninvasive detection and quantitation of iron in body tissues such as liver, heart, and pancreas. Quantitation of cardiac iron is especially important, as this is the key predictor of cardiac dysfunction and these measurements inform decisions about the effectiveness of iron chelation therapy. As newer oral iron chelation agents are being developed, serial MRI exams are used to evaluate their efficacy and their potential equivalency or superiority to traditional parenteral agents.

MRI exams for iron quantitation measure the paramagnetic effects of some iron-containing complexes, which accelerate the rate of decline of radio signals induced in body tissues. Several different methods have been developed, and researchers have shown that these methods yield comparable results if appropriate controls are implemented and appropriate image analysis is performed. Researchers have provided increasingly standardized mechanisms for image analysis, some of which are now commercially available. Clinical trials incorporating MRI exams for iron quantitation require careful standardization of methods for image acquisition and image analysis.

Interpreting MRI data about tissue iron levels requires some understanding of iron metabolism. Adults require approximately 20 mg of iron each day, primarily for the production of hemoglobin in new red blood cells; most of this iron is actually recycled from old blood cells, with comparatively small dietary requirements. Apart from hemoglobin, most iron is stored bound to a protein called ferritin within cells of the liver, spleen, and bone marrow. Iron balance is normally controlled in the intestine, where uptake of dietary iron is carefully regulated. This regulation is defective in hemochromatosis, in which uncontrolled iron absorption occurs. Once absorbed, there is no physiologic mechanism for clearance of excess iron. When circulating iron levels persistently exceed the binding ability of storage proteins such as ferritin, free iron circulates and accumulates within tissues. Free iron is highly toxic due to its ability to catalyze production of free radicals and lead to oxidative tissue damage, especially in the heart. When cells degrade iron-containing proteins, an insoluble iron deposit called hemosiderin is generated. Hematologists have come to understand that these intracellular and extracellular forms of iron in body tissues exist in dynamic equilibrium, with different kinetics in different tissues. For example, it is possible for chelation therapy to rapidly decrease a patient’s liver iron while more slowly decreasing toxic iron in the heart.

Toxic free iron is not paramagnetic and therefore is not detected by MRI. However different protein-bound forms of iron are variably paramagnetic and can be detected. Hemosiderin is strongly paramagnetic and more easily detected, whereas ferritin-bound iron is more weakly paramagnetic, depending on its degree of aggregation. Together, these paramagnetic forms of iron accelerate the spin-spin decay time (or dephasing rate) of hydrogen protons in local water molecules, and this dephasing effect can be directly measured by MRI. The time constant for this signal decay is designated T2, measured in milliseconds. The rate of signal decay (also called relaxivity) is designated R2, which is the inverse of the time constant, i.e. R2 is proportional to 1/T2. Increased concentration and aggregation of paramagnetic iron complexes in body tissues cause more rapid signal decay, meaning decreased T2 values (shorter half time of decay) and increased R2 values (higher rate of decay). By obtaining serial measurements of MRI signal intensity at multiple progressively longer time intervals after the signal is induced, it is possible to plot the signal decay and directly measure T2 and R2 values of body tissues.

Importantly, MRI can use two refocusing mechanisms to detect signal intensity in body tissues: spin-echo or gradient-echo. Initially, a short pulse of radio waves is used to induce signal within tissues. Then, one of the two refocusing mechanisms is employed to create “echoes” of radio waves from the tissue. These echoes are detected by the scanner and used to reconstruct images showing signal intensity (i.e. strength of received echo) from specific locations within the body. Spin-echo refocusing techniques yield signal measurements which decay according to the T2 time constant and R2 rate constant, as described above. Alternatively, gradient-echo refocusing techniques yield signal measurements which decay more quickly, influenced by local inhomogeneities in the magnetic field (influences negated by the spin-echo technique). The time and rate constants for signal decay measured using gradient-echo techniques are designated T2* and R2*, pronounced “T2-star” and “R2-star”. Gradient-echo methods can be performed rapidly and allow signal measurements from the moving heart, using ECG-gated techniques to obtain “snapshots” of heart muscle in less than a tenth of a second. Spin-echo methods often require longer acquisition times and are more sensitive to motion artifacts.

Fortunately, increased iron levels cause generally similar effects on T2 and T2* decay times, with higher iron levels causing increased R2 and R2* rates of decay. During the past decade, researchers have shown direct correspondence of MRI-derived R2 and R2* measurements with actual iron concentrations measured in patients’ liver biopsy samples. When liver R2 and R2* measures are plotted against liver iron concentration, there is a curvilinear relationship of R2 to iron, versus an approximately linear relationship R2* to iron. These calibrated curves allow physicians to use T2 and R2 or T2* and R2* measurements from individual patients to predict iron concentration in their tissues, influencing their therapeutic decisions related to chelation therapy and offering prognostic information about potential organ damage and dysfunction. With regard to the heart, while cardiac MRI measures have not been correlated with biopsies, there is a compelling correlation between cardiac T2* measurements (reported in milliseconds) and risk of cardiac failure. Clinicians may use these cardiac measurements to increase aggressiveness of chelation therapy in higher risk patients. Similarly, researchers can use MRI measurements in evaluating the efficacy of new therapeutic agents on maintaining or decreasing iron levels in the liver and heart.

Researchers have validated MRI techniques for measuring T2* in heart muscle tissue and both T2 and T2* in liver, pancreas, pituitary, and other tissues. These researchers have shown that such methods can be implemented on varied MRI equipment from different vendors. Indeed, there is an FDA-approved and widely used commercial service called FerriScan which provides (1) MRI protocols for use on their clients’ MRI equipment at sites worldwide and (2) image analysis and iron concentration estimates performed at their centralized facility. This service demonstrates the feasibility of establishing reliable acquisition protocols at numerous MRI facilities and then centralizing and standardizing the image analysis process.  During MRI acquisition, technologists must be trained to recognize and correct potentially significant artifacts and repeat acquisitions if necessary.

Appropriate image analysis procedures are critical in achieving accurate and repeatable iron measurements from MRI data. For this purpose, data received from the MRI facility includes image “slices” through body tissues, with each slice obtained at multiple echo times. Signal intensity from each region of tissue decays (decreases) at progressively longer echo times. Measuring the rate of this decrease yields the R2 or R2* values from which iron concentrations are estimated. Higher concentration of tissue iron complexes such as hemosiderin and ferritin cause faster rates of signal decrease, appearing as progressive darkening of the tissue at longer echo times (MRI images show signal intensity as brightness). Importantly, to accurately measure this decrease in signal intensity, it is necessary to “register” the images, accounting for any patient motion between acquisitions at each echo time. Gradient-echo methods often allow the entire series of images to be obtained within 20-30 milliseconds, minimizing any motion, whereas spin-echo methods may require longer acquisitions and separate breath holds by the patient, which can introduce significant motion and misregistration of tissues (i.e. position of liver or heart varies between images at different echo times). Once images are registered, signal intensity measurements can be performed from each individual pixel or from larger regions of interest (ROI) in each organ. These measurements are made within each image, from a series of 5 to 6 separate images obtained at progressively longer echo times. These numerical measurements are then plotted and fit to a curve, from which the rate of decrease in signal per unit of time is calculated as R2 (if spin-echo method was employed) or R2* (for gradient echo methods). There are various models for fitting these curves, yielding generally comparable results at clinically relevant iron concentrations. If calculated from individual pixels, the R2 or R2* values can then be displayed as a colorized map, correlated with the original images to show distribution of estimated iron concentration within the organs of interest. This pixel-wise analysis can be useful in detecting focal areas of higher iron deposition, but more importantly it allows detection of technical artifacts and allows measurements to be made from regions that are relatively free from artifacts. At some stage in analysis, averaged R2 or R2* values from larger regions of interest are used to make clinical decisions.

In the liver, there are usually multiple regions of hepatic parenchyma permitting ROI measurements away from vascular and extrahepatic structures, allowing accurate estimation of liver iron concentration in multiple separate areas. Compared to liver biopsy, this more global (or at least multisegmental) assessment is less susceptible to sampling error. In making these measurements, it is important to assess liver morphology, avoiding regions of fibrosis or focal hepatic lesions. In the heart, the interventricular septum is chosen as the region of myocardium most free from artifact, measured from short axis projections obtained obliquely through axis of the left ventricle. Measurements from smaller and more irregular organs such as pituitary gland or pancreas are limited by spatial resolution and by artifacts caused by nearby gas-containing structures. For most clinical and therapeutic research purposes, liver and cardiac studies remain the standard iron measurements. Internal controls include body tissues with low rates of iron deposition such as skeletal muscle, and potentially external objects included in the image. For example, the FerriScan commercial service uses bags of saline solution as internal controls, placed next to patient and included in the field of view.

Standardization of MRI acquisition techniques (spin or gradient-echo techniques yielding R2 or R2* measurements respectively, screening for artifacts, inclusion of internal controls, sometimes calibration using phantoms) and image analysis procedures (image registration, pixel-wise or ROI measurements, fitting measurements to decay curves and estimating R2 or R2* values, and correlation with established data from prior studies) are both critical for clinicians who titrate patients’ chelation regimens to maintain their iron balance within safe ranges.  Standardized MRI acquisition and image analysis procedures are also critical for researchers designing clinical trials to evaluate new therapies, such as the emerging class of oral iron chelation agents. These agents may demonstrate different effects on iron clearance from specific tissues. Improved MRI techniques for iron quantitation will serve a key role in current and future research studies and clinical trials.

To download a PDF of this article, please click here.

About the Author

Dr. Tubb is a Board Certified Radiologist at Intrinsic Imaging. He obtained his radiology residency training at Johns Hopkins Hospital after completing the combined MD-PhD program at Baylor College of Medicine in Houston, TX. His PhD research was in Molecular and Cellular Biology, with multiple peer-reviewed publications and experience in experiment design and data analysis. During his radiology residency, Dr. Tubb received a one year RSNA Resident Research Award and designed a research project involving targeted MRI contrast agents and high resolution pancreatic imaging. He served as chief resident at Johns Hopkins and also served on the admissions committee. Following residency, Dr. Tubb completed a one year MRI fellowship at the University of Pennsylvania, with broad experience in body/oncologic, breast, spine, and musculoskeletal imaging. Dr. Tubb is actively involved with the design and implementation of the imaging components of clinical trials, as well as application of new imaging technology such as MRI elastography for evaluation of liver disease.

About Intrinsic Imaging LLC

Located in Bolton, Massachusetts and San Antonio, Texas, Intrinsic Imaging is an FDA audited, ISO 9001:2008 and ISO 13485:2003 certified, GAMP® 5 compliant medical imaging core lab specializing in providing imaging core lab services for clinical trials. Its comprehensive medical imaging core lab services include, but are not limited to, expert radiologist consultation, protocol and charter development, site qualification, training and management, as well as image acquisition, processing and detailed radiologic analysis.

Intrinsic Imaging has more than sixty full-time, board-certified diagnostic radiologists on staff that have sub-specialization in all therapeutic areas including, but not limited to, Cardiovascular, Central Nervous System, Gastrointestinal & Genitourinary, Medical Device, Musculoskeletal and Oncology.


1. Wood, John and Noetzli, Leila. Cardiovascular MRI in thalassemia major. Annals of the New York Academy of Sciences 1202 (2010), pages 173-9.


2. Wood et al. MRI R2 and R2* mapping accurately estimates hepatic iron concentration in transfusion-dependent thalassemia and sickle cell disease patients. Blood 106:4 (2005), pages 1460-65.


3. Neufeld, Ellis. Oral chelators deferasirox and deferiprone for transfusional iron overload in thalassemia major: new data, new questions. Blood 107:9 (2006), pages 3436-41.


4. Ghugre et al. Improved R2* measurements in myocardial iron overload. Journal of Magnetic Resonance Imaging 23 (2006), pages 9-16.


5. Wood, John and Ghugre, Nilesh. Magnetic resonance imaging assessment of excess iron in thalassemia, sickle cell disease, and other iron overload diseases. Hemoglobin 32:1-2 (2008), pages 85-96.