Original Research Superparamagnetic Iron Oxide (SPIO)-Enhanced Liver MRI With Ferucarbotran: Efficacy for Characterization of Focal Liver Lesions Sook Namkung, MD, 1,2 Christoph J. Zech, MD, 1 * Thomas Helmberger, MD, 1,3 Maximilian F. Reiser, MD, 1 and Stefan O. Schoenberg, MD 1 Purpose: To evaluate the efficacy of ferucarbotran in T2- weighted (T2W) fast spin-echo (FSE) and T2*W gradient- echo (GRE) sequences for characterizing focal liver lesions. Materials and Methods: In 68 patients, 46 malignant and 22 benign focal liver lesions were evaluated. Precontrast (NCE) T2W FSE images and contrast-enhanced (CE) T2W FSE and T2*W GRE images were obtained on a 1.5T MR system. Based on signal intensity (SI) measurements in focal lesions and liver parenchyma, the signal-to-noise ra- tio (SNR) and contrast-to-noise ratio (CNR) were calculated for all sequences. The percentage of SI loss (PSIL) in focal lesions after contrast agent (CA) application was calculated for the T2W FSE sequence. Qualitative analyses were per- formed to assess image quality and lesion conspicuity ob- tained with the CE-T2W FSE and CE-T2*W GRE se- quences. Results: The mean PSIL was higher in solid benign lesions than in malignant lesions (39.6% vs. 3.2%, P  0.05). With a threshold PSIL of 25%, the sensitivity and specificity for characterizing malignant lesions were 97.8% and 92.9%, respectively. The mean CNR of the malignant lesions was higher in the CE-T2*W sequence than in the CE- and NCE- T2W FSE sequences (29.9 vs. 22.7 (P  0.01) vs. 12.8 (P  0.01)). CE-T2*W images showed a superior image quality and lesion conspicuity (P  0.05) compared to the CE-T2W FSE sequence. Conclusion: The PSIL can be an accurate tool for charac- terizing benign and malignant lesions. The addition of a CE-T2*W GRE sequence is helpful for the detection and characterization of malignant lesions. Key Words: liver neoplasms; MR imaging; SPIO (ferumox- ides); liver lesion characterization; PSIL J. Magn. Reson. Imaging 2007;25:755–765. © 2007 Wiley-Liss, Inc. LIVER-SPECIFIC CONTRAST AGENTS (CAs), such as superparamagnetic iron oxide (SPIO) particles, have been developed to increase the potential of MRI for detecting and characterizing focal liver lesions (1,2). SPIO particles are taken up by reticulo-endothelial system (RES) cells (the so-called Kupffer cells) of normal liver parenchyma, as well as by macrophages of the spleen and lymph nodes. They shorten T2 and T2* re- laxation times by disturbing the local magnetic field in the liver parenchyma, resulting in a signal intensity (SI) loss of normal liver parenchyma (3– 6). Malignant liver tumors do not have a substantial number of RES cells and appear as hyperintense lesions in contrast to the hypointense liver parenchyma after application of SPIO on T2-weighted (T2W) MRI (7). In selected patients with hepatocellular carcinoma (HCC) or liver metastases, improved survival can be achieved with surgical resection, and preoperative eval- uation of the number, size, and segmental location of lesions is very important (8,9). Biphasic spiral CT is widely used for liver imaging. For the detection of focal liver lesions, however, it reportedly has lower sensitivity (74.3%) than MRI (95.4%) (10). Computed tomography during arterial portography (CTAP) is considered the most sensitive preoperative imaging modality for the detection of liver lesions, with reported sensitivities of 81% to 93% (11,12). SPIO-enhanced MRI as a noninva- sive technique is a useful method with a high preoper- ative diagnostic efficacy (13,14), and is equivalent to CTAP in detection of liver lesions (15–17). Another important aspect of liver imaging is the abil- ity to not only detect and localize a lesion but also to characterize it, and in particular to determine whether it is benign or malignant, since such information will affect the patient’s treatment. With the use of widely available extracellular CAs for both CT and MRI, char- acterization of a focal hepatic lesion is based on its vascularity and interstitial composition. MRI with tis- 1 Institute of Clinical Radiology, Munich University Hospitals– Grosshadern, Munich, Germany. 2 Department of Radiology, Hallym University Hospital, Gangwon-do, South Korea. 3 Department of Radiology, University of Luebeck, Luebeck, Germany. S. Namkung and C.J. Zech contributed equally to this manuscript. *Address reprint requests to: C.J.Z., Institute of Clinical Radiology, Munich University Hospitals–Grosshadern, Marchioninistr. 15, 81377 Munich, Germany. E-mail: Christoph.Zech@med.uni-muenchen.de Received December 11, 2005; Accepted October 10, 2006. DOI 10.1002/jmri.20873 Published online 2 March 2007 in Wiley InterScience (www.interscience. wiley.com). JOURNAL OF MAGNETIC RESONANCE IMAGING 25:755–765 (2007) © 2007 Wiley-Liss, Inc. 755