Hepatocellular carcinoma (HCC), the most common primary malignant tumor of the liver, is strongly associated with hepatitis B and liver cirrhosis [
Gadolinium ethoxybenzyl diethylenetriaminepentaacetic acid (Gd-EOB-DTPA; Primovist, Bayer Schering, Pharma AG, Berlin) is a hepatocyte-specific contrast agent commonly used to facilitate the distinction between HCC from FNH. However, about 10% of high-grade HCCs can also take up this contrast agent and have a slightly higher signal intensity in the hepatobiliary phase [
The iterative decomposition of water and fat with the echo asymmetrical and least-squares estimation quantitation sequence (IDEAL-IQ) developed by GE Healthcare (USA), could be used to measure the iron and fat content in tissues. The whole liver can be scanned within one breath-hold (about 16 s) without contrast agents. The images are then reconstructed, and a fat fraction (FF) and R2* maps can be obtained simultaneously to measure the fat and iron content within the tissue of interest. This technique eliminates the effect of the T2*, magnetic field inhomogeneity, and T1 magnetic relaxation, and therefore, the fat and iron content within the tissue can be quantified more accurately [
This study aimed to use the IDEAL-IQ MRI imaging technique to measure the iron-deficiency and fatty metamorphosis in HCC and FNH. In addition, the diagnostic efficiency and reliability of the IDEAL-IQ and IVIM-DWI in distinguishing between HCC from FNH were also compared.
This retrospective study was approved by our institutional review board. Between July 2019 and December 2021, 28 HCC patients and 11 FNH patients were examined, and their diagnoses were confirmed by operation and pathology. Another 4 FNH patients were diagnosed according to the American Association for the Study of Liver Diseases (AASLD) guidelines. Since AFP is an important marker for HCC, only those patients with negative AFP were included in this study. Furthermore, only patients with a single hepatic lesion- not exceeding 1 cm and without portal vein thrombosis and hepatic lymph node metastasis were included in the study. Patients with incomplete clinical and imaging data, multiple hepatic lesions, and those who had previous liver treatment were excluded.
All patients fasted for 6 h before the MRI examination. All MRI scans were acquired on a 3.0-Tesla(T) MRI scanner (Discovery MR750W, GE Healthcare) with a 48-element body coil. The abdomen was bound with a respiratory gating device to trigger the scanning. The scanning sequences included an axial T1-weighted imaging (T1WI), an axial T2-weighted imaging (T2WI), IDEAL-IQ, and IVIM-DWI. The IVIM-DWI images were obtained using the following parameters: repetition time/echo time (TR/TE), 6666.67/66.3 ms; slice thickness, 5.0 mm with a gap of 1.0 mm; field of view, 44 cm × 44 cm; matrix size, 96 × 128; twelve b values ranging from 0 to 1200 s/mm2 (0, 10, 20, 50, 100, 150, 200, 400, 600, 800, 1000, and 1200) were used, and the number of excitations (NEX) for each b was 2, 2, 2, 2, 2, 2, 2, 3, 3, 4, 4, 6. A three-dimensional liver acquisition with volume accelerated extended volume (LAVA-XV) multi-phase dynamic contrast-enhanced scanning was performed by injecting GD-EOB-DTPA (Primovist, Bayer Schering, Pharma AG, Berlin) through the cubital vein with a high-pressure syringe. The contrast agent was first injected at a dose of 0.1 ml/kg at a flow rate of 1–2 ml/s followed by a 10-ml saline flush. Images of the early and late arterial phase, portal phase, delayed phase, and hepatobiliary phase were collected at 15 s, 60 s, 180 s and 20 min, respectively, after the contrast agent injection.
All images were transferred to the AW4.7 workstation (GE Medical System) for image processing, analysis, and measurement of the relevant parameters. Two radiologists, one with 9 years and one with 15 years of clinical experience in reporting abdominal MRIs, evaluated the images visually and quantitatively. The radiologists were blinded to the diagnosis of the evaluated patients. The visual assessment involved evaluating the contrast enhancement and the signal intensity of the lesion on the T1W1 and T2WI for all 4 phases. For the quantitative assessment, the 2 radiologists were asked to select regions of interest (ROI) within the lesions to perform the quantitative measurements as described by Luo et al. [
The statistical package for the social sciences (SPSS) software version 19.0 was used to analyze the data. Unless otherwise specified, data are reported as mean ± standard deviation. The intraclass correlation coefficient (ICC) was used to evaluate the agreement between the two radiologists. An ICC value above 0.75 was deemed to be consistent and acceptable. The independent two-sample
The demographic and clinical characteristics of the patients are summarized in
Variables | HCC | FNH | t/χ2 | |||
---|---|---|---|---|---|---|
Gender, n(%) | 4.725 | 0.03 | ||||
M |
23 (82.1%) |
7 (46.7%) |
||||
Age (years old) | 61.8±4.4 | 28.1±5.9 | 13.309 | <0.001 | ||
Tumor diameter (cm) | 5.5±3.1 | 5.3±2.1 | 0.229 | 0.820 | ||
HbsAg, n(%) |
8 (28.6%) |
11 (73.3%) |
5.4 | 0.021 | ||
Cirrhosis | 10 (35.7%) | 0 (0.0%) | 0.08 | |||
Tumor location |
8 (28.6%) |
9 (60.0%) |
4.036 | 0.045 |
Note: HCC, hepatocellular carcinoma; FNH, focal nodular hyperplasia; Hbs Ag, Hepatitis B surface antigen.
Due to obvious fatty metamorphosis, 6 cases in the HCC group showed a mixed signal intensity on T1WI and T2WI. The signal intensity decreased on the T1WI in-phase/opposed-phase (IP/OP) images. The remaining 22 HCC cases and all FNH cases revealed slight hypo-intensity on the T1WI and slight hyper-intensity on the T2WI. In the HCC group, a central scar was observed in 9 cases and a capsule was noted in 19 cases. On the other hand, in the FNH group, a central scar was noted in 9 cases and a capsule was noted in 2 cases. The enhancement pattern was persistent in 9 HCC cases. A “washin-washout” enhancement pattern was noted in the rest of the HCC lesions. GD-EOB-DTPA enhanced MR scanning was performed for 13 of the HCC cases, and 4 showed hyper-intensity in the hepatobiliary phase. All patients in the FNH group underwent GD-EOB-DTPA enhanced MR scanning and a persistent-enhancement pattern and hyper-intensity in the hepatobiliary phase (
HCC (n) | FNH (n) | χ2 | ||
---|---|---|---|---|
Hypo-intensity on T1WI/slightly hyper-intensity on T2WI | 22 (78.6%) | 15 (100%) | – | 0.076 |
Persistent-enhancement pattern | 9 (32.1%) | 15 (100%) | – | <0.001 |
Capsule | 19 (67.9%) | 2 (13.3%) | 10.37 | 0.001 |
Central scar | 6 (21.4%) | 9 (60.0%) | 6.397 | 0.011 |
Hyper-intensity in hepatobiliary phase | 4/13 (30.8%) | 15/15 (100%) | – | <0.001 |
The ICCs of the parameters obtained by the two observers were all >0.75, indicating near perfect agreement between the two radiologists, and that the measurements were suitable for further statistical analysis (
HCC | ADC (×10−3 mm2/s) | D (×10−3 mm2/s) | FFle (%) | FFli (%) | R2*le (Hz) | R2*li (Hz) |
---|---|---|---|---|---|---|
Radio 1 | 1.27±0.34 | 0.89±0.19 | 8.51±6.34 | 3.94±2.4 | 20.89±5.59 | 57.69±10.49 |
Radio 2 | 1.35±0.27 | 0.93±0.19 | 8.04±6.02 | 3.69±2.21 | 21.33±5.95 | 58.06±10.06 |
ICC | 0.867 | 0.877 | 0.973 | 0.954 | 0.924 | 0.969 |
95% CI | 0.63–0.945 | 0.702–0.946 | 0.942–0.988 | 0.897–0.979 | 0.843–0.964 | 0.934–0.985 |
Note: HCC, hepatocellular carcinoma; FNH, focal nodular hyperplasia; ICC, intragroup correlation coefficient; ADC, apparent diffusion coefficient; D, diffusion coefficient; FFle, fat fraction of lesions; FFli, fat fraction of liver; R2*le, R2* value of lesions; R2*li, R2* value of liver; 95% CI, 95% confidence intervals.
FNH | ADC (×10−3 mm2/s) | D (×10−3 mm2/s) | FFle (%) | FFli (%) | R2*le (Hz) | R2*li (Hz) |
---|---|---|---|---|---|---|
Radio 1 | 1.66±0.31 | 1.22±0.29 | 2.47±1.16 | 4.4±3.33 | 23.56±6.45 | 42.54±5.45 |
Radio 2 | 1.62±0.35 | 1.18±0.23 | 2.7±1.24 | 4.05±3.07 | 22.6±5.86 | 41.99±5.78 |
ICC | 0.949 | 0.955 | 0.962 | 0.986 | 0.933 | 0.901 |
95% CI | 0.852–0.983 | 0.839–0.986 | 0.761–0.99 | 0.921–0.996 | 0.804–0.978 | 0.843–0.983 |
Note: HCC, hepatocellular carcinoma; FNH, focal nodular hyperplasia; ICC, intragroup correlation coefficient; ADC, apparent diffusion coefficient; D, diffusion coefficient; FFle, fat fraction of lesions; FFli, fat fraction of liver; R2*le, R2* value of lesions; R2*li, R2* value of liver; 95% CI, 95% confidence intervals.
ADC and D were significantly lower in the HCC group than in the FNH group, while the liver R2* and lesion FF were significantly higher in the HCC groups (all
Variables | HCC | FNH | t | |
---|---|---|---|---|
R2*lesion (Hz) | 21.146±5.802 | 22.799±6.069 | 0.858 | 0.396 |
R2*liver (Hz) | 57.867±10.365 | 40.758±4.070 | −5.922 | <0.001 |
FFleion (%) | 8.284±5.756 | 2.559±1.247 | −5.032 | <0.001 |
FFliver (%) | 3.58±2.55 | 4.38±3.97 | −0.664 | 0.512 |
ADC (×10−3 mm2/s) | 1.310±0.253 | 1.624±0.304 | 3.540 | 0.001 |
D (×10−3 mm2/s) | 0.909±0.192 | 1.230±0.314 | 4.115 | <0.001 |
D* (×10−3 mm2/s) | 25.722±15.543 | 33.224±5.417 | −1.916 | 0.066 |
f (%) | 0.279±0.121 | 0.336±0.105 | 1.203 | 0.22 |
Note: IVIM, intravoxel incoherent motion; IDEAL-IQ, Iterative decomposition of water and fat with echo asymmetrical and least-squares estimation quantitation sequence; HCC, hepatocellular carcinoma; FNH, focal nodular hyperplasia; AFP, alpha fetoprotein; R2*le, R2* value of lesions; R2*li, R2* value of liver; FFle, fat fraction of lesions; FFli, fat fraction of liver; ADC, apparent diffusion coefficient; D, diffusion coefficient; D*, pseudo-diffusion coefficient; f, perfusion fraction.
Variables | Cutoff | Sensitivity (%) | Specificity (%) | Youden index | AUC (95% CI) | PPV (%) | NPV (%) |
---|---|---|---|---|---|---|---|
FF (%) | 3.6 | 82.14 | 86.67 | 0.688 | 0.923 (0.799–0.982) | 92 | 72.22 |
ADC (×10−3 mm2/s) | 1.44 | 75 | 86.67 | 0.617 | 0.854 (0.712–0.943) | 91.3 | 65 |
D (×10−3 mm2/s) | 1.09 | 85.71 | 80 | 0.657 | 0.864 (0.725–0.949) | 88.89 | 75 |
Note: ROC, receiver operating characteristics; AFP, alpha fetoprotein; HCC, hepatocellular carcinoma; FNH, focal nodular hyperplasia; FF, fat fraction of lesions; ADC, apparent diffusion coefficient; D, diffusion coefficient; AUC, area under the curve; 95% CI, 95% confidence intervals; PPV, positive predictive value; NPV, negative predictive value.
HCC and FNH are focal hypervascular lesions of the hepatic system that tend to be associated with different medical histories. The AFP level is important for the differential diagnosis of HCC and FHN. Although AFP generally tends to be higher in HCC, the AFP levels of the HCC cases evaluated in our study were normal. This means that AFP is not always a reliable tumor marker for HCC, and thus there is a need to develop new non-invasive methods to characterize liver lesions. Therefore, this study aimed to evaluate the diagnostic efficacy of two different MRI techniques (IDEAL-IQ and IVIM-DWI) for the differential diagnosis of HCC and FHN lesions.
Consistent with our findings, previous studies reported that hepatitis B was more common in HCC patients than in FNH patients [
Our study showed that the ADC and D extracted from the IVIM images could detect histological differences accurately and reliably (ICC > 0.75). The mean values for both parameters significantly lower in HCC lesions when compared the FNH lesions. However, D* and f showed no significant difference between the two patient groups (
The average R2* liver value was higher in the HCC group than in the FNH group, while the average R2* lesions values were similar in the HCC and FNH groups and lower than the average R2* liver values. These results indicated that the liver in the HCC group was subject to iron overload while iron deficiency was present in both HCC and FNH lesions. Although iron is an essential trace element for human metabolism, excessive iron liver deposits may cause hepatocyte edema, degeneration, necrosis, an increase in the liver matrix, and accumulation of collagen fibers. This eventually leads to liver fibrosis, cirrhosis, and the development of HCC. Moreover, liver fibrosis will promote further iron deposition in the liver [
FNH is a nodule composed of benign-appearing hepatocytes and is not a real neoplastic lesion. It is mainly composed of normal hepatocytes, blood vessels, bile ducts, and Kupffer cells. Not all patients diagnosed with FNH have chronic liver disease. Therefore, there is typically no iron overload in the liver in FNH patients. We speculate that the mechanism behind the R2* decrease in FNH is different from that in HCC. In HCC, the iron deficiency is caused by the tumor, while in FNH, it is caused by other clinical factors such as poor nutrition and malabsorption. In FNH, the ROIs contained hepatocytes, bile ducts, Kupffer cells, as well as other non-iron-containing components. The non-iron-containing components reduced the R2* value within the lesion. Furthermore, most of the patients in the HCC group had a history of hepatitis B and older, leading to a further reduction in the liver iron content. However, the R2* value within the HCC and FNH lesions did not differ significantly.
The FF differed significantly between the HCC and FNH groups (
This study has several limitations that have to be acknowledged. The patient population in this study was small, and therefore larger studies are needed to confirm our findings. Moreover, in this study, one-third of the cases in the HCC group were well-differentiated, and only one case was poorly differentiated. The difference in HCC grade results in different diffusion and perfusion-related values. However, due to the small sample size, we could not analyze the impact of the pathological grade on the different MRI parameters. Finally, as the post-processing methods for IDEAL-IQ and IVIM-DWI were different, the location and size of the ROIs may not have been exactly the same. However, the ICCs between the two observers were all above 0.75, which indicates good agreement.
AFP-negative HCC and FNH are hypervascular lesions on contrast-enhanced CT and MRI scans, and their imaging features are sometimes similar. The parameters of FF derived from IDEAL-IQ, and ADC and D derived from IVIM-DWI, showed a good diagnostic efficiency for differentiating AFP-negative HCC and FNH. The diagnostic efficiency of FF was higher than ADC and D, which indicates that IDEAL-IQ may be a good complement to IVIM-DWI. The obvious fatty metamorphosis in the HCC lesions can be observed on routine MRI sequences, but in cases with limited fatty metamorphosis, we can obtain accurate results through FF value. The density of the tumor cells tends to be lower in high-grade HCC, while the ADC and D values tend to be higher when compared with low-grade HCC. As a result, HCC can be difficult to distinguish from FNH [
We would like to thank TopEdit (