In this study, there was a significant, high association between peritumoral iron deposition and the presence of a PTHR compared with the absence of a PTHR. Interestingly, there was no significant association between a fibrous capsule and a PTHR. Therefore, we found that the presence of a PTHR on T2*WIs probably showed iron deposition in the peritumoral liver parenchyma. Furthermore, the presence of a PTHR was also associated with a significantly higher density of the background liver iron deposition.
As a more objective evaluation method, we performed a quantitative analysis of iron deposition as optimal density by multispectral imaging in addition to the semiquantitative method utilized in previous reports [12, 13]. Liu et al. studied the application of multispectral imaging in the quantitative immunohistochemical analysis of breast cancer. The authors reported that, in contrast to conventional RGB images, multispectral imaging was more accurate and reliable, and it provided more information on protein expression in relation to clinicopathological characteristics . Multispectral imaging can be used for immunostaining as well as with other staining methods; therefore, this imaging method could be used without limitation with Prussian blue staining in this study of liver iron quantification. In normal color imaging (conventional RGB imaging), images are acquired as wavelength bands of light divided by the three RGB channels, whereas in multispectral imaging, images are acquired in more subdivided wavelength bands. Therefore, multispectral imaging can more accurately determine iron deposition with Prussian blue staining as the specific wavelength band. Because there was a significant correlation between the semiquantitative and quantitative methods, we analyzed the association between a PTHR and peritumoral iron deposition using the quantitative method.
Histopathologically, the capsule appearance on imaging features either a true capsule consisting of a fibrous inner layer and a prominent sinusoid outer layer, or a pseudocapsule consisting of a prominent sinusoid and/or peritumoral fibrous tissue [6, 18]. The presence of an enhancing capsule appearance is a characteristic feature of HCC, and this feature can suggest a favorable prognosis after hepatectomy and the degree of effectiveness of transcatheter arterial chemoembolization [19-21]. Therefore, assessing capsule appearance is important not only for a differential diagnosis but also for clinical management. This finding was adopted as a major feature in the LI-RADS . In contrast, a nonenhanced capsule appearance, which is an ancillary feature favoring HCC in particular, can appear as a hypointense rim on T2WIs, nonenhanced T1WIs, or during the hepatobiliary phase of gadoxetic acid-enhanced MR images [2, 6, 22]. In particular, noncontrast MR images such as T2WIs and nonenhanced T1WIs are important for patients who cannot tolerate gadolinium, as a contrast agent (such as patients with renal failure and previous hypersensitivity reactions) [4, 5]. Chen et al. reported that T2*WIs and SWIs were superior to nonenhanced T1WIs and T2WIs for describing capsule appearance. The authors observed PTHRs in T2*WIs in 27 of 41 (66%) HCCs with the histopathologic capsule features, and PTHRs on SWIs were observed in 34 of 41 (83%) HCCs with the histopathologic capsule features. The authors speculated that an outer layer in the histopathologic capsule has abundant sinusoids with high concentrations of deoxyhemoglobin, resulting in a phase difference between the vessels and the surrounding parenchyma, although this finding has not been histopathologically analyzed in detail . However, the finding may be useful for the diagnosis of HCC with a nonenhancing capsule appearance.
Although we did not histopathologically analyze the blood vessels of the capsule in this study, there was no significant difference in the histopathologic fibrous capsule between groups. However, the presence of a PTHR on T2*WIs was strongly associated with histopathologic peritumoral iron deposition. Therefore, we speculate that SWIs are more sensitive to change in magnetic susceptibility, such as with iron deposition, than T2*WIs, so a PTHR would be detected more clearly on SWIs than on T2*WIs. We attributed moderate inter-reader agreement (kappa value = 0.5363) to assessment of a PTHR on T2*WI because it was sometimes difficult to evaluate PTHR accurately owing to an insufficient signal–noise ratio. We also considered that better agreement could be obtained by SWI, which has better sensitivity to magnetic susceptibility and provides high spatial resolution.
A significantly higher degree of iron deposition in the background liver was detected in the presence of a PTHR compared with the absence of a PTHR. Hepatic iron deposition is commonly observed in CLD, and in particular, in alcoholic liver disease, nonalcoholic steatohepatitis, and hepatitis C . In CLD, hepatic iron deposits are found in hepatocytes and reticuloendothelial system cells . Furthermore, HCCs with a PTHR had a significantly higher density of the background liver iron deposition compared with HCCs without a PTHR. Therefore, we consider that the PTHR was easily observed around larger HCCs because of the compression change in the peritumoral liver parenchyma in the iron-deposited background liver.
There was a significant difference in the etiology of the HCCs, and the reason may be the significantly larger tumor size in NBNC patients compared with HCV patients. Moreover, HCCs with a PTHR were identified in 11 of 12 NBNC patients. In these 11 patients, 7 patients had alcoholic liver disease, 3 patients had nonalcoholic steatohepatitis, and one patient had an unknown etiology. Alcoholic liver disease causes hepatic iron overload from an early stage in approximately half of all patients, which is a higher degree of iron deposition than with viral liver disease . Therefore, we considered that NBNC patients, who had a high incidence of alcoholic liver disease, also had high hepatic iron deposition and significantly more HCCs with a PTHR. Although hemochromatosis was not evaluated in our study, it is highly likely that HCC caused by hemochromatosis, in which iron is deposited in hepatocytes, will also show a PTHR.
There was significant difference in the degree of fibrosis between patients with and without a PTHR. Iron overload in the liver is associated with a high degree of fibrosis (F3 or F4) . Iron deposition in the background liver was significantly increased in HCCs with a PTHR compared with HCCs without a PTHR, in this study. However, the degree of fibrosis was significantly lower in HCCs with a PTHR, which may be explained by the small sample size of our study population.
As we have mentioned, a PTHR can be considered to reflect peritumoral iron deposition, not a histopathologic fibrous capsule. Furthermore, a PTHR might correspond to the liver parenchyma with iron deposition compressed by the tumor. Iron deposition in CLD occurs mainly in Kupffer cells and hepatocytes. In this study, a PTHR was observed when there was a large amount of iron deposition in the background liver, and with large HCCs. Therefore, we considered that PTHR was not caused by microbleeding but instead, was caused by the increased iron density around the tumor owing to iron compression of the liver parenchyma by the tumor. Therefore, we presume that a PTHR can also be detected in other hepatic tumors if the tumors are larger, and if there is more iron deposition in the background liver. We consider that the future application of T2*WI depends on whether it can be useful in the diagnosis of HCC without using contrast agents. We also believe that the use of SWI can increase the detection rate of a PTHR. However, in the future, we need to study other hepatic tumors to determine whether the presence of a PTHR is a specific finding (i.e., a nonenhancing capsule appearance) for HCC.
There were no significant differences in PFS and OS between patients with and without a PTHR, and it was difficult to confirm the clinical usefulness of T2*WI in affecting the patients’ prognosis, in this study. Therefore, even if T2*WI is added as a sequence in MRI for suspected HCC, this addition does not affect the current treatment policy for HCC.
This study has several limitations. First, there was selection bias regarding the study population size because we excluded small HCCs (< 1.5 cm) and large HCCs (> 10 cm). Additionally, the start time was set for evaluation targets that were imaged using the 3T MR system. However, the evaluation time was short because of the time and costs involved in analyzing the targets with multispectral imaging. As a result, the number of HCCs was low. Second, the entire circumference of the tumor was not evaluated in the histopathologic examination.