Non-invasive assessment of hepatic iron overload: are we finally there?

Abstract

Iron excess in tissues can lead to toxicity and organ disease. The liver, as the primary site of body iron deposition, is also a principal target for its toxicity. Heavy iron overload in humans is usually associated with a genetically determined disturbance of iron homeostasis (i.e. hereditary hemochromatosis, HH) or with intensive transfusion regimens associated with hereditary anemia (i.e. β-thalassemia). Liver biopsy is the most accurate method for assessing hepatic iron burden and provides important information on underlying liver damage and diseases. However, it entails the risk for potential complications while accurate iron measurements may be biased by sampling variability in advanced chronic disease. Therefore, at least within a diagnostic workup, costs and benefits of liver biopsy should be carefully considered. In hereditary hemochromatosis, a positive genetic test for HFE or other hemochromatosis gene mutations (transferrin receptor 2, hemojuvelin, hepcidin) may allow to avoid liver biopsy in most symptomatic patients, whereas histology remains essential in homozygotes with abnormal transaminases, hepatomegaly or serum ferritin higher than 1000 ng/ml [[1]Pietrangelo A. Hereditary hemochromatosis—a new look at an old disease.N Engl J Med. 2004; 350: 2383-2397Crossref PubMed Scopus (823) Google Scholar].In recent years, the potential role of mild iron overload as factor of comorbidity in the progression of several hepatic disorders (e.g. viral hepatitis, alcoholic liver disease, non-alcoholic fatty liver disease, porphyria cutanea tarda) has been increasingly recognized. In fact, accurate evaluation of hepatic iron content is now regarded as an important clinical manoeuvre in the management of chronic non-HH liver disease. Due to safety concerns of liver biopsy, non-invasive methods to measure hepatic iron content have been, therefore, devised. Serum ferritin may be a good surrogate marker for evaluating iron stores but suffers from low specificity. High sensitivity has been reported for the superconducting quantum interference device (SQUID) that measures magnetic susceptibilities [[2]Fischer R. Longo F. Nielsen P. Engelhardt R. Hider R.C. Piga A. Monitoring long-term efficacy of iron chelation therapy by deferiprone and desferrioxamine in patients with beta-thalassaemia major: application of SQUID biomagnetic liver susceptometry.Br J Haematol. 2003; 121: 938-948Crossref PubMed Scopus (97) Google Scholar]. SQUID, however, is not widely available and requires experienced operators. Computed tomography shows attenuation of the liver in the presence of excess iron, but seems to suffer of low sensitivity and specificity particular at mild iron overload grades and in the presence of fat [[3]Howard J.M. Ghent C.N. Carey L.S. Flanagan P.R. Valberg L.S. Diagnostic efficacy of hepatic computed tomography in the detection of body iron overload.Gastroenterology. 1983; 84: 209-215Abstract Full Text PDF PubMed Scopus (65) Google Scholar].Magnetic resonance has been intensively exploited, until recently, as a sensitive and specific tool for assessing hepatic iron overload [4Gandon Y. Guyader D. Heautot J.F. Reda M.I. Yaouanq J. Buhe T. et al.Hemochromatosis: diagnosis and quantification of liver iron with gradient-echo MR imaging.Radiology. 1994; 193: 533-538PubMed Google Scholar, 5Ernst O. Sergent G. Bonvarlet P. Canva-Delcambre V. Paris J.C. L'Hermine C. Hepatic iron overload: diagnosis and quantification with MR imaging.AJR Am J Roentgenol. 1997; 168: 1205-1208Crossref PubMed Scopus (114) Google Scholar, 6Bonkovsky H.L. Rubin R.B. Cable E.E. Davidoff A. Rijcken T.H. Stark D.D. Hepatic iron concentration: non-invasive estimation by means of MR imaging techniques.Radiology. 1999; 212: 227-234PubMed Google Scholar, 7Gandon Y. Olivie D. Guyader D. Aube C. Oberti F. Sebille V. et al.Non-invasive assessment of hepatic iron stores by MRI.Lancet. 2004; 363: 357-362Abstract Full Text Full Text PDF PubMed Scopus (524) Google Scholar, 8St Pierre T.G. Clark P.R. Chua-Anusorn W. Fleming A.J. Jeffrey G.P. Olynyk J.K. et al.Non-invasive measurement and imaging of liver iron concentrations using proton magnetic resonance.Blood. 2004; (doi: 10.1182/blood-2004-01-017710)PubMed Google Scholar, 9Alustiza J.M. Artetxe J. Castiella A. Agirre C. Emparanza J.I. Otazua P. et al.MR quantification of hepatic iron concentration.Radiology. 2004; 230: 479-484Crossref PubMed Scopus (173) Google Scholar]. Due to the paramagnetic properties of iron, its increased hepatic content decreases both the T2 relaxation time and the liver signal's intensity (SI): this gives the typical dark liver at MRI. Gradient-recalled-echo (GRE) techniques have been shown to be more accurate in quantifying mild iron overload states than spin-echo sequences due to higher sensitivity to field inhomogeneities induced by paramagnetic substances. Most published studies agree on the fact that MRI is accurate in assessing a moderate-severe iron overload but a validated and cost effective method for taking advantage of MRI in quantifying also low grades of iron stores was still missing. In addition, an ideal method should also be simple and easy to be standardized in order to be transferred and applied to different machines.Gandon and coworkers seem to have taken a fundamental step toward this end [[7]Gandon Y. Olivie D. Guyader D. Aube C. Oberti F. Sebille V. et al.Non-invasive assessment of hepatic iron stores by MRI.Lancet. 2004; 363: 357-362Abstract Full Text Full Text PDF PubMed Scopus (524) Google Scholar]. They studied 174 patients by both percutaneous liver biopsy with biochemical assessment of hepatic iron concentration and MRI of the liver with various GRE sequences on a 1.5 T magnet. They calculated liver iron concentration by evaluating the correlation between liver to muscle (L/M) signal intensity ratio and developed an algorithm to calculate magnetic resonance hepatic iron concentration. Data obtained in a study group (n=139) that included patients with suspected iron overload or patients managed for hepatitis C, were then applied to a validation group (n=35). A highly T2-weighted GRE sequence was most sensitive, with 89% sensitivity and 80% specificity in the validation group, with an L/M ratio below 0.88. This threshold allowed detecting all clinically relevant liver iron overload from 60 μmol/g to about ten times the upper limit of normal (375 μmol/g) (normal value <36 μmol/g). The magnetic resonance signal is intrinsically dependent on multiple acquizition variables, and cannot be directly correlated with hepatic iron concentration. Thus, comparative quantitative data need to be extracted, for instance to calculate ratios of signal intensities such as liver to noise, liver to fat, or liver to muscle. Gandon et al. [[7]Gandon Y. Olivie D. Guyader D. Aube C. Oberti F. Sebille V. et al.Non-invasive assessment of hepatic iron stores by MRI.Lancet. 2004; 363: 357-362Abstract Full Text Full Text PDF PubMed Scopus (524) Google Scholar] have found more reliable, accurate and reproducible to choose muscle to compare with liver other than, for instance, using liver to noise ratio. They claim that L/M ratio would minimize inter-machine variability and increase reproducibility. In addition, since liver signal is usually more intense than muscle, a slight decline in liver signal would be easily detected by simple visual comparison. Correlation between MRI and biochemical assessment of iron overload was good. Liver fibrosis did not affect the accuracy of MRI, neither did steatosis.MRI is obviously safer and, overall, less costly than liver biopsy. A further potential advantage over liver biopsy could also relate to more accurate analysis of patients with cirrhosis: iron deposition is usually heterogeneous and biochemical hepatic iron concentration can vary greatly from one nodule to another. MRI can examine transverse sections of the entire liver, and region of interest signal intensity measurements average larger volumes of tissue compared with liver biopsy samples. A recent study performed at a different center has implemented the algorithm used by Gandon et al. (available at http://www.radio.univ-rennes1.fr) and found good correlation of data sets from patients with iron overload [[9]Alustiza J.M. Artetxe J. Castiella A. Agirre C. Emparanza J.I. Otazua P. et al.MR quantification of hepatic iron concentration.Radiology. 2004; 230: 479-484Crossref PubMed Scopus (173) Google Scholar]. More studies are clearly needed to validate the method using different machines and calibration settings. Yet, there is no doubt that this method seems already to hold promise in offering a most valuable non-invasive tool to quantify hepatic iron overload whenever, in diagnostic settings or in the follow-up of patients treated for iron overload, hepatic iron excess is uncertain and prognostic information on underlying liver pathology is not essential. Iron excess in tissues can lead to toxicity and organ disease. The liver, as the primary site of body iron deposition, is also a principal target for its toxicity. Heavy iron overload in humans is usually associated with a genetically determined disturbance of iron homeostasis (i.e. hereditary hemochromatosis, HH) or with intensive transfusion regimens associated with hereditary anemia (i.e. β-thalassemia). Liver biopsy is the most accurate method for assessing hepatic iron burden and provides important information on underlying liver damage and diseases. However, it entails the risk for potential complications while accurate iron measurements may be biased by sampling variability in advanced chronic disease. Therefore, at least within a diagnostic workup, costs and benefits of liver biopsy should be carefully considered. In hereditary hemochromatosis, a positive genetic test for HFE or other hemochromatosis gene mutations (transferrin receptor 2, hemojuvelin, hepcidin) may allow to avoid liver biopsy in most symptomatic patients, whereas histology remains essential in homozygotes with abnormal transaminases, hepatomegaly or serum ferritin higher than 1000 ng/ml [[1]Pietrangelo A. Hereditary hemochromatosis—a new look at an old disease.N Engl J Med. 2004; 350: 2383-2397Crossref PubMed Scopus (823) Google Scholar]. In recent years, the potential role of mild iron overload as factor of comorbidity in the progression of several hepatic disorders (e.g. viral hepatitis, alcoholic liver disease, non-alcoholic fatty liver disease, porphyria cutanea tarda) has been increasingly recognized. In fact, accurate evaluation of hepatic iron content is now regarded as an important clinical manoeuvre in the management of chronic non-HH liver disease. Due to safety concerns of liver biopsy, non-invasive methods to measure hepatic iron content have been, therefore, devised. Serum ferritin may be a good surrogate marker for evaluating iron stores but suffers from low specificity. High sensitivity has been reported for the superconducting quantum interference device (SQUID) that measures magnetic susceptibilities [[2]Fischer R. Longo F. Nielsen P. Engelhardt R. Hider R.C. Piga A. Monitoring long-term efficacy of iron chelation therapy by deferiprone and desferrioxamine in patients with beta-thalassaemia major: application of SQUID biomagnetic liver susceptometry.Br J Haematol. 2003; 121: 938-948Crossref PubMed Scopus (97) Google Scholar]. SQUID, however, is not widely available and requires experienced operators. Computed tomography shows attenuation of the liver in the presence of excess iron, but seems to suffer of low sensitivity and specificity particular at mild iron overload grades and in the presence of fat [[3]Howard J.M. Ghent C.N. Carey L.S. Flanagan P.R. Valberg L.S. Diagnostic efficacy of hepatic computed tomography in the detection of body iron overload.Gastroenterology. 1983; 84: 209-215Abstract Full Text PDF PubMed Scopus (65) Google Scholar]. Magnetic resonance has been intensively exploited, until recently, as a sensitive and specific tool for assessing hepatic iron overload [4Gandon Y. Guyader D. Heautot J.F. Reda M.I. Yaouanq J. Buhe T. et al.Hemochromatosis: diagnosis and quantification of liver iron with gradient-echo MR imaging.Radiology. 1994; 193: 533-538PubMed Google Scholar, 5Ernst O. Sergent G. Bonvarlet P. Canva-Delcambre V. Paris J.C. L'Hermine C. Hepatic iron overload: diagnosis and quantification with MR imaging.AJR Am J Roentgenol. 1997; 168: 1205-1208Crossref PubMed Scopus (114) Google Scholar, 6Bonkovsky H.L. Rubin R.B. Cable E.E. Davidoff A. Rijcken T.H. Stark D.D. Hepatic iron concentration: non-invasive estimation by means of MR imaging techniques.Radiology. 1999; 212: 227-234PubMed Google Scholar, 7Gandon Y. Olivie D. Guyader D. Aube C. Oberti F. Sebille V. et al.Non-invasive assessment of hepatic iron stores by MRI.Lancet. 2004; 363: 357-362Abstract Full Text Full Text PDF PubMed Scopus (524) Google Scholar, 8St Pierre T.G. Clark P.R. Chua-Anusorn W. Fleming A.J. Jeffrey G.P. Olynyk J.K. et al.Non-invasive measurement and imaging of liver iron concentrations using proton magnetic resonance.Blood. 2004; (doi: 10.1182/blood-2004-01-017710)PubMed Google Scholar, 9Alustiza J.M. Artetxe J. Castiella A. Agirre C. Emparanza J.I. Otazua P. et al.MR quantification of hepatic iron concentration.Radiology. 2004; 230: 479-484Crossref PubMed Scopus (173) Google Scholar]. Due to the paramagnetic properties of iron, its increased hepatic content decreases both the T2 relaxation time and the liver signal's intensity (SI): this gives the typical dark liver at MRI. Gradient-recalled-echo (GRE) techniques have been shown to be more accurate in quantifying mild iron overload states than spin-echo sequences due to higher sensitivity to field inhomogeneities induced by paramagnetic substances. Most published studies agree on the fact that MRI is accurate in assessing a moderate-severe iron overload but a validated and cost effective method for taking advantage of MRI in quantifying also low grades of iron stores was still missing. In addition, an ideal method should also be simple and easy to be standardized in order to be transferred and applied to different machines. Gandon and coworkers seem to have taken a fundamental step toward this end [[7]Gandon Y. Olivie D. Guyader D. Aube C. Oberti F. Sebille V. et al.Non-invasive assessment of hepatic iron stores by MRI.Lancet. 2004; 363: 357-362Abstract Full Text Full Text PDF PubMed Scopus (524) Google Scholar]. They studied 174 patients by both percutaneous liver biopsy with biochemical assessment of hepatic iron concentration and MRI of the liver with various GRE sequences on a 1.5 T magnet. They calculated liver iron concentration by evaluating the correlation between liver to muscle (L/M) signal intensity ratio and developed an algorithm to calculate magnetic resonance hepatic iron concentration. Data obtained in a study group (n=139) that included patients with suspected iron overload or patients managed for hepatitis C, were then applied to a validation group (n=35). A highly T2-weighted GRE sequence was most sensitive, with 89% sensitivity and 80% specificity in the validation group, with an L/M ratio below 0.88. This threshold allowed detecting all clinically relevant liver iron overload from 60 μmol/g to about ten times the upper limit of normal (375 μmol/g) (normal value <36 μmol/g). The magnetic resonance signal is intrinsically dependent on multiple acquizition variables, and cannot be directly correlated with hepatic iron concentration. Thus, comparative quantitative data need to be extracted, for instance to calculate ratios of signal intensities such as liver to noise, liver to fat, or liver to muscle. Gandon et al. [[7]Gandon Y. Olivie D. Guyader D. Aube C. Oberti F. Sebille V. et al.Non-invasive assessment of hepatic iron stores by MRI.Lancet. 2004; 363: 357-362Abstract Full Text Full Text PDF PubMed Scopus (524) Google Scholar] have found more reliable, accurate and reproducible to choose muscle to compare with liver other than, for instance, using liver to noise ratio. They claim that L/M ratio would minimize inter-machine variability and increase reproducibility. In addition, since liver signal is usually more intense than muscle, a slight decline in liver signal would be easily detected by simple visual comparison. Correlation between MRI and biochemical assessment of iron overload was good. Liver fibrosis did not affect the accuracy of MRI, neither did steatosis. MRI is obviously safer and, overall, less costly than liver biopsy. A further potential advantage over liver biopsy could also relate to more accurate analysis of patients with cirrhosis: iron deposition is usually heterogeneous and biochemical hepatic iron concentration can vary greatly from one nodule to another. MRI can examine transverse sections of the entire liver, and region of interest signal intensity measurements average larger volumes of tissue compared with liver biopsy samples. A recent study performed at a different center has implemented the algorithm used by Gandon et al. (available at http://www.radio.univ-rennes1.fr) and found good correlation of data sets from patients with iron overload [[9]Alustiza J.M. Artetxe J. Castiella A. Agirre C. Emparanza J.I. Otazua P. et al.MR quantification of hepatic iron concentration.Radiology. 2004; 230: 479-484Crossref PubMed Scopus (173) Google Scholar]. More studies are clearly needed to validate the method using different machines and calibration settings. Yet, there is no doubt that this method seems already to hold promise in offering a most valuable non-invasive tool to quantify hepatic iron overload whenever, in diagnostic settings or in the follow-up of patients treated for iron overload, hepatic iron excess is uncertain and prognostic information on underlying liver pathology is not essential.