Proper definitions of MRI contrast enhancement in liver tumors

Abstract

Dear Editor, As researchers working in the field of contrast agents (CAs) for use in magnetic resonance imaging (MRI), we would like to congratulate Narita et al. on the results reported in their recently published article ‘‘Expression of OATP1B3 determines uptake of Gd-EOB-DTPA in hepatocellular carcinoma’’ [1] showing that overexpression of OATP1B3 is related to Gd-EOB-DTPA accumulation or positive contrast enhancement (CE) in certain hepatocellular carcinomas (HCCs). These results were obtained using biomics technologies on clinical patients. However, a number of apparent points in the article have drawn our attention and motivated us to initiate the discussion presented here. Originally, hepatobiliary MRI CAs, including gadolinium ethoxybenzyl diethylenetriaminepentaacetic acid (Gd-EOB-DTPA), were developed to improve lesion detectability by specifically enhancing liver signal intensity (SI) and leaving the hepatic tumor unenhanced, i.e., negative tumor contrast enhancement (TCE) on T1-weighted MRI [2]. However, 15 years ago Ni et al. [3] first reported the very rare but remarkable positive TCE, which was best distinguishable during the hepatobiliary phase around 30 min following the injection of Gd-EOB-DTPA in rats with chemically induced HCCs. This phenomenon was found to be associated with certain highly differentiated HCCs and with benign hepatocytic nodules, as evident by the histopathology [3, 4]. For many years, in the early clinical trials, this Gd-EOB-DTPA-induced positive TCE was either absent or ignored [5, 6], and it has only been recently reported after this agent has been widely applied in clinical patients [1, 7, 8]. The HCC lesion with positive TCE that is shown in Fig. 1D of the article by Narita et al. [1] is to date the closest clinical finding comparable to what has been reported in rats [3, 4, 9]. In contrast, other experimental studies have failed to convincingly demonstrate the true positive TCE in rodent hepatic tumor models [10, 11]. Hence, a proper definition of CE in liver tumors becomes crucial, particularly with regard to the use of liver-specific MRI CAs. Similar to the procedure adopted by Narita et al. in their study [1], a number of other researchers have used the enhancement ratio (ER), where ER = (SIpostcontrast SIprecontrast)/SIprecontrast, or the relative enhancement (RE), where RE(%) = (SIpost SIpre)/SIpre 9 100, in their experimental studies. This ratio only takes the tumor SI into account—irrespective of the changes in the SI of the surrounding liver, which is the main target of hepatobiliary CAs [11]. Other researchers have used criteria that do not reflect the real tumor-to-liver contrast on Gd-EOB-DTPAenhanced MR images, such as a tumor said to be positively enhanced did not really display a positive TCE pattern [10, 12]. Logically, a negative or positive TCE can only be defined when the SI of the tumor is measured as being either lower or higher than that of the liver, based on visualization on the postcontrast MRI scan, as shown in Fig. 1D of Narita et al. [1] and in animal results [3, 4, 9]. The most straightforward approach is to calculate the postcontrast tumor-to-liver contrast ratio (CR), i.e., CR = SItumor/SIliver. When CR [ 1 or CR \ 1, a positive An answer to this letter to the editor is available at doi: 10.1007/s00535-009-0176-x.