Cirrhosis is a pathological process characterized by a continuous diffuse process of fibrosis and parenchymal distortion and an occurrence of different types of nodules either benign regenerative nodules, premalignant dysplastic nodules, or malignant hepatocellular nodules (HCC). The incidence of malignant nodules is expected to increase mainly due to the widespread hepatitis B and C infection. HCC nodules are the fifth most common malignancy in the world [8].
MRI nowadays is considered the most accurate imaging modality for the detection of cirrhosis and its complications. Short-time sequences, better soft tissue resolution, and Triphasic contrast examinations have become a basic component of abdominal imaging [9].
DWI, in addition, helps in a better differentiation of benign and malignant nodules [10].
In our work, the study population included 60 patients with a male predominance (48/60) of 80%. The sixth decade group was the most affected group (28/60) 46.7% followed by the seventh decade group (24/60) 40%.
By MRI, out of 60 patients, most types of nodules were HCCs 40/60 (66.7%). These findings matched with Glenn et al. whose study population included 71 patients (42 males, 29 females) with 65 HCC cases and Rieko et al. whose study population included 58 patients (39 males,19 females) with 40 HCC cases [11, 12]. In this study, group A HCC nodules (72/124) (58.1%) were hypointense on T1WIs and hyperintense on T2WIs and SPAIR; this signal pattern is typically the characteristic for well-differentiated HCC nodules. These findings were also noted by Gaurav et al. and also Glenn et al. who reported that 94% of their 47 HCC nodule were hyperintense on T2WI [11, 13].
On the contrary, Van et al. published that the signal intensity of malignant nodules may be various on T2WIs and HCC nodules may be of similar signal or even low SI relative to the surrounding liver on Fat Sat T2WIs [14].
In our triphasic study, 64 out of 72 nodules (88.8%) of the group A nodules displayed the typical pattern of enhancement of malignant nodule as it showed rapid arterial enhancement and rapid contrast washout in the portal and delayed phases. These findings were similar to Robert et al., Jonathon et al., and Gaurav et al. who stated that this pattern of enhancement due to hypervascularity and they considered this pattern was essential characteristic features for HCC as the tumor recruits unpaired arteries and sinusoidal capillaries with resultant avid arterial enhancement [7, 13, 15].
Our study proved that all 28/124 were high-grade dysplastic nodules which showed hyperintense in T1WIs, T2WIs, and SPAIR. However, the 8/124 nodules of low-grade dysplastic nodules were also hyperintense in T1WI, but isointense in T2WI and SPAIR sequences. This data was also stated by Gaurav et al. and Tatsuyuki et al., who reported that T1WI dysplastic nodules characteristically demonstrate high SI which may be related to the deposition of copper, glycogen, protein, or lipid. In T2WI, most of the dysplastic nodules are usually isointense or low intensity. On the other hand, Jonathon et al. stated that dysplastic nodules may have different MR appearance, but in rare cases, they appeared hyperintense in T2WIs and STIR [7, 13, 16].
In our study, there was no significant difference between the two types of dysplastic nodules in their signal intensity on T1 and T2WIs, but Robert et al. reported that the high-grade dysplastic nodules had a slightly high signal intensity on T2-weighted images. In this case, the differentiation between HCC and the high-grade dysplastic nodule may be difficult even by pathology [15].
In our triphasic MRI study, the 28 nodules of high-grade dysplasia displayed subtle arterial enhancement. However, they became more intense in the subsequent portal and delayed phases more than the liver parenchyma. On the other hand, the 8 low-grade dysplastic nodules were similar to the liver parenchyma in all phases of the dynamic study. These findings were also noticed by Jonathon et al. Tatsuyuki et al. and Gaurav et al. who reported that as regards to the blood supply, the high-grade dysplastic nodules appear of low vascularity in the arterial phase with dominantly portal and venous blood supply, so they become more enhanced at portal and delayed phases with no contrast washout. The increased arterial vascularity was seen in a small number of high-grade dysplastic nodules that receive blood supply from the hepatic artery, and this may be confusing with HCC nodules during hepatocarcinogenesis. On the other hand, low-grade dysplastic nodules are normally supplied by the portal vein and therefore are similar to the liver parenchyma in all phases of triphasic contrast study [7, 13, 16].
Robert et al. reported that the regenerative nodules showed changeable signals on T1-weighted images. On the T2-weighted MRI, they were isointense to hypointense but were almost never hyperintense [15]. In our study, group C, 16 regenerative nodules (13%) from 124 studied nodules showed high SI on T1WIs and isointense on T2WIs and SPAIR sequences. Since MRI signals of the regenerative nodules were variable, we cannot only rely on T1 and T2WIs to diagnose them [15].
In our triphasic study, group C regenerative nodules (16 nodules) were similar to the liver parenchyma with no evidence of arterial enhancement or washout owing to a large blood supply from the portal vein and minimal contribution from the hepatic artery. These findings were also found in the publication issued by Seale et al. stating that most of the regenerative nodules were enhanced as the liver parenchyma or show very faint enhancement as uptake, and excretion of gadolinium (DTPA) by these nodules is similar to normal liver tissue. Consequently, after contrast injection, all regenerative nodules have an equivalent intensity to the liver parenchyma which gave the liver homogenous appearance [17].
In our study, in group D, 8 cases of hemangiomas (8 nodules/124) (6.5%) were characteristically hypointense on T1WIs and hyperintense on the T2WIs and SPAIR sequences. In the triphasic study, the nodules showed peripheral nodular enhancement in the arterial phase and no washout in the portal or delayed phase. This finding similar to Debees et al. who studied 5 cases of hemangioma out of 30 cases of hepatic masses on the cirrhotic liver and stated that 4 studied hemangiomas were relatively typical in the appearance and 1 of them appeared with atypical appearance. And they explained this atypical enhancement as hemangioma rarely occurs in end-stage cirrhosis, probably because of the cirrhosis obliterates existing hemangioma.
Our study was done using b value of 1000 s/mm2 for DWIs to overcome the perfusion of the capillary and diffusion of water to the extracellular extravascular space; this high b value was needed for reduction of the signal from protons movement in the nearby structure. This will lead to an increase in contrast between the nodule and the liver parenchyma. Moreover, the differentiation between malignant and benign nodules was increased with using high b value. This b value was the same that used in studies done by Demir et al. and Hosny [18, 19].
All of group A HCC nodules (72 nodules) and (64/72) nodules (88.8%) had restricted diffusion and 8/72 nodules were partially restricted. This finding was similar to that published by Gaurav et al., who reported that a nodule in the cirrhotic liver with restricted diffusion would be confirmatory to be a malignant nodule, especially when combined with other MRI features of HCC lesions [13]. Mean ADC value of HCCs nodule was 1.40 × 10−3 ± 0.20 × 10−3 mm2/s.
In our study, dysplastic nodules (28 high grade, 16 low grade) were not restricted with mean ADC value of 1.09 × 10−3 ± 0.01 × 10−3 mm2/s. Regenerating nodules (8 nodules) showed no restriction with a mean ADC value of 1.2 × 10−3 ± 0.17 × 10−3 mm2/s. Benign nodules (8 nodules) showed no restriction with a mean ADC of 1.77 × 10−3 ± 0.35 × 10−3 mm2/s.
Our results are in match with Debees et al. who found an ADC value of HCCs nodule in cirrhotic liver 0.9 × 10−3 ± 1.3 × 10−3 with a mean value of 1.095 × 10−3 ± 0.108 × 10−3 and that of regenerating nodules of 1 × 10−3–1.3 × 10−3 with a mean of 1.98 × 10−3 ± 0.19 × 10−3 and that of hemangioma of 1.8 × 10−3 ± 2.3 × 10−3 with a mean value of 1.98 × 10−3 ± 0.192 × 10−3 and also found the result of ADC statistically significant. Many other studies found significant differences between the ADC value in between benign and malignant nodules but not between various types of malignant lesions or between various types of benign lesions [20,21,22,23,24,25]. This was matched with our results which succeed in discrimination between benign and malignant nodules with a cutoff value of 1.22 × 10−3 and with a sensitivity of 80% and specificity of 85%.
Other studies by Elbadway et al. found an insignificant difference between the benign and malignant lesions and suggest no definite cutoff value [25].
In our study, we focused on small nodules ≤ 2 cm and found different types of nodules by MRI and confirmed by pathology while Robert et al. stated that lesions with a small diameter were more likely to be benign than malignant [15].
The major limitation of our study was relatively a small number of patients especially those with benign nodules, so future study with much more number of patients is recommended. Also, another limitation was using one kind of contrast medium (Gd-DTPA) to identify different types of nodules. Another promising contrast material is reticuloendothelial agents; these agents are taken by Kupffer cells. Most liver tumors which are deficient in Kupffer cells do not accumulate this agent [25]. And so, liver tumors appear relatively hyperintense as the background liver darkens. This agent are used most routinely to help in the detection of HCC in high risk patients as in the detection of HCC in cirrhotic patients which may be difficult with gadolinium alone, because of several cirrhosis parenchymal changes (fibrosis and regeneration) and alteration of liver perfusion (collaterals, increased hepatic arterial flow relative to portal venous flow) [26]. The use of this agent may help improve HCC detection in such patients when combined with gadolinium to create a double-contrast effect. With this technique, the reticuloendothelial agent was infused first and then followed by gadolinium. The two agents act complementary to improve the contrast of lesion to the liver background on dynamic T1-weighted images because the background liver is darkened by the reticuloendothelial agent while the lesion of interest became more lightened by gadolinium [26, 27]. Further research with newly developed contrast material is recommended.