The baseline characteristics of patients are summarized in Table 1. The average age of IPH patients was 39 ± 20 years, and 8 male patients and 8 female patients were included. Three patients underwent splenectomy. The underlying causes of IPH included autoimmune diseases such as psoriasis vulgaris (1 patient), the use of the Chinese herbal medicines Sedum aizoon (1 patient) and Stellera chamaejasme (1 patient), the use of oxaliplatin (2 patients), and eating snacks with food additives such as spicy strips (1 patient). No possible IPH-related causes were investigated in the other patients. Fifteen patients had splenomegaly. The average spleen pachydiameter was 5.4 cm ± 1.2 cm, and the length was 15.1 cm ± 3.2 cm. Two patients had high aspartate aminotransferase levels (41 U/L and 41.8 U/L). Three patients had low platelet counts (46 × 109/L, 58 × 109/L and 92 × 109/L); MR examinations were performed in twelve patients, CT was performed in thirteen patients, and ultrasound was performed in all sixteen patients.
The clinical diagnoses of the 16 liver cirrhosis patients recruited as the cirrhosis group included alcoholic cirrhosis (4 patients), viral hepatitis-related cirrhosis (6 patients), primary biliary cirrhosis (4 patients) and congenital hepatic fibrosis (2 patients). Three patients underwent splenectomy. The Child-Pugh classification of the patients was class A for 14 patients and class B for 2 patients. The average spleen pachydiameter was 5.2 cm ± 1.3 cm, and the length was 15.2 cm ± 3.3 cm. There were no significant differences in spleen pachydiameter or length between the IPH group and the cirrhosis group. The detailed imaging evaluation of the liver and spleen is shown in Table 2.
The imaging changes in the liver parenchyma
Diffuse nodular changes on T1WI were shown in 63% (10/16) of the patients in the cirrhosis group but were not displayed in patients in either the healthy control group or the IPH group (Fig. 1). However, a non-homogeneous change was shown in the T1WI image in the IPH group, which was different from the homogeneous change in the healthy control group. No significant differences were found in T2WI or diffusion weighted imaging (DWI) images among the three groups.
Imaging changes in the portal area
The parameters of the main, sagittal and Segment 3 (S3) branches of the portal vein were compared among IPH patients, cirrhosis patients and healthy people. The lumen diameters of the main, sagittal and S3 branches of the portal vein in the IPH group and the healthy group were smaller than those in the cirrhosis group (P < 0.05). The vessel walls of the main and sagittal portal veins in the IPH group were thicker than those in the cirrhosis group (Figs. 2 and 3) and the healthy group (P < 0.05). The ratio of wall thickness to lumen diameter of the main and sagittal portal veins in the IPH group was greater than that in the cirrhosis group and the healthy group (P < 0.05). The detailed results are shown in Table 3.
A low enhancement area along the portal vein in the delay phase is shown in contrast-enhanced MRI (ceMRI) (Figs. 2e and 3e) and contrast-enhanced CT images (Fig. 2c) of IPH patients. This sign was not observed in the ceMRI (Figs. 2f and 3f) and CT (Fig. 2d) images of cirrhosis patients.
Pathological changes in IPH and cirrhosis
Obvious fibrosis in the portal area was found in all pathological results of IPH patients (Fig. 4a), but this sign was absent in cirrhosis patients (Fig. 4b). An apparent increase in the amount of collagen fibres, particularly around blood vessels in the portal area, was shown in the livers of IPH patients (Fig. 5a), and the increase in collagen fibres was mostly distributed in the portal area (Fig. 5c). However, the collagen fibres were mostly distributed in the septa around the liver pseudolobules (Fig. 5b,d) beyond the portal area in cirrhosis patients. Interlobular vein occlusion was reported in 9 out of 16 IPH patients (Fig. 4c), dilation was reported in 10 patients, and wall thickening was reported in 6 patients. In contrast, most of the interlobular vein in cirrhosis patients showed no wall thickening or occlusion (Fig. 4d). In IPH cases, the liver cells showed atrophy or apoptosis (Fig. 4e) around the Glisson’s sheath, but necrosis was not found. In contrast, necrosis (75%), oedema and fatty degeneration (93.8%) of liver cells were observed in most of the cirrhosis patients (Fig. 4f). The results of pathological characteristics are shown in Table 4.
Differences between IPH and liver cirrhosis based on imaging and pathology
In IPH patients, the main imaging changes were portal vein wall thickening, stenosis or occlusion, a low enhancement area along the portal vein in the delay phase in contrast-enhanced imaging, and a non-homogeneous change in T1WI imaging. The pathological changes included portal vein wall thickening, stenosis, occlusion, portal area fibrosis, and atrophy or apoptosis of hepatocytes.
The main imaging characteristic of liver cirrhosis was diffuse nodular change in T1WI MR imaging, and the related pathological change was pseudolobule formation. No imaging or pathological changes were shown in the portal vein in liver cirrhosis patients except for lumen dilation in some cases.
The specific findings related to portal hypertension in the IPH group included portal vein wall thickening and a low enhancement area along the portal vein in the delay phase ceMRI and contrast-enhanced CT. Other findings, such as portal vein lumen changes, splenomegaly and portal-systemic collaterals, were non-specific and could not differentiate IPH from liver cirrhosis.