Perfusion Characteristics of Hepatocellular Carcinoma at Contrast-enhanced Ultrasound: Influence of the Cellular differentiation, the Tumor Size and the Underlying Hepatic Condition

This study aimed to analyze the influence of the cellular differentiation, the tumor size and the underlying hepatic condition on the enhancement pattern of hepatocellular carcinoma (HCC) on contrast-enhanced ultrasound (CEUS). 276 patients with single lesion ≤ 5 cm who underwent CEUS exam and were pathologically confirmed as HCC were retrospectively enrolled. Enhancement patterns, washout patterns, wash-in time and washout time were observed and recorded. During the arterial phase, more poorly differentiated HCCs (42.5%) and lesions > 3 cm (35.2%) performed inhomogeneous enhancement (p < 0.05). More well differentiated HCCs (63.4%) performed late washout or no washout while compared with moderately (37.8%) or poorly (24.1%) differentiated HCCs (p < 0.05). Poorly differentiated HCCs showed the shortest washout time (83.0 ± 39.8 s), moderately differentiated HCCs showed the moderate washout time (100.4 ± 52.1 s), and well differentiated HCCs showed the longest washout time (132.3 ± 54.2 s) (p < 0.05). Lesions > 3 cm (97.2 ± 51.3 s) washed out more rapidly than lesions ≤ 3 cm (113.9 ± 53.5 s) (p < 0.05). The dynamic enhancement procedure of HCC was influenced by the cellular differentiation and the tumor size. While, hepatic background showed no influence on the dynamic enhancement of HCC.

frequently in HCC with poorer grades of differentiation in portal or delayed phase [10][11][12] . Nevertheless, the analysis was performed between well differentiated HCC and less differentiated HCC in the previously mentioned studies, less differentiated HCC included both moderately and poorly differentiated HCC. The moderately differentiated and poorly differentiated HCC were not divided and the comparison between them not performed [10][11][12] . Some researchers reported that wash out is observed in about half the cases of small HCC but less in very small nodules (20-30% in those 1-2 cm, 40-60% in those 2-3 cm) 5,13 . The effect tumor size on wash out needs further investigation. In addition, whether the background of liver has an effect on the dynamic enhancement progress of HCC remains unknown.
The aims of this study were: (1) to analyze the time of wash-in and wash-out among different differentiated tumors retrospectively, especially between moderately and poorly differentiated HCC; (2) to investigate the influence of the tumor size on the enhancement of HCC at CEUS; (3) to clarify whether the background of liver has influence upon the enhancement pattern of HCC at CEUS.

Materials and Methods
Patients population. According to the guidelines from Declaration of Helsinki, institutional review board approval was achieved from the ethics committee of Southwest hospital. The need for informed consent of each patient was waived in this retrospective study. Between January 2005 and June 2015, 4370 patients who had CEUS examination and confirmed pathological diagnosis after surgical resection or ultrasound guided biopsy were retrospectively identified through a review of hospital database and records of the Department of Pathology Non-HCC pathological diagnosis was established in 1444 patients. Thus, we got 2926 cases of HCC. Clinical information (gender, age, etiology of cirrhosis, serum AFP value) was retrospectively collected from our hospital clinical information system. We performed a per-patient analysis. The inclusion criteria for the present study were as follows: (1) Patients with single solid nodule ≤ 5 cm; (2) Real-time CEUS for characterization of liver lesion was performed within less than a month before resection or biopsy; (3) The final diagnosis of HCC had to be verified pathologically through evaluation of surgical specimen or biopsy.
Exclusion criteria were: (1) Pathological diagnosis of hepatic lesion was non-HCC; (2) Tumor size > 5 cm which were not considered as fulfilling the Milan criteria for HCC and large tumors often have intra-tumoural necrosis that unpredictably influences the contrast pattern on CEUS; (3) Nodule number ≥ 2 because CEUS could not scan multiple nodules simultaneously after one injection of contrast agent if the nodules are not at the same scan plane and it is difficult to correspond the pathology of each tumor to the imaging of US in patients with multiple hepatic lesions; (4) Patients with thrombosis portal vein or hepatic vein which may influence the hepatic dynamic circulation; (5) Systemic chemotherapy or targeted treatment prior to the CEUS.
According to the inclusion and exclusion criteria mentioned above, 1284 patients with HCC satisfied the inclusion criteria of the study (Fig. 1) including 96 cases of well differentiated HCC, 1108 cases of moderately differentiated HCC and 80 cases of poorly differentiated HCC. The number of moderately differentiated HCC were too large to compare with that of well differentiated and poorly differentiated HCC statistically. One hundred cases were randomly selected from 1108 cases of moderately differentiated HCC at random, by creating the sequence numbers in the SPSS 13.0 software package (SPSS Inc,Chicago, IL). Finally, a total 276 cases of HCC (96 cases of well differentiated,100 cases of moderately differentiated and 80 cases of poorly differentiated) were included in this study. Image acquisition. Each patient receiving the CEUS examination was fasting for at least 8 h. US scanning was performed by two experienced physicians (R L, and XHZ) with more than 16 years of experience of liver ultrasound examination. An Acuson Sequoia 512 ultrasound unit (Siemens Medical Solutions, Santa Clara, Calif) with a multi-frequency transducer and contrast pulse sequencing (CPS) contrast-specific software was applied in conventional US scanning and CEUS. The characteristics of the lesion at gray-scale US, including location, size, echogenicity, shape and color Doppler, were recorded. Dual imaging mode was used to show the same scanning plane of gray scale US and CEUS simultaneously. SonoVue (Bracco Imaging B.V, Geneva, Switzerland) composed of phospholipids-stabilized microbubbles filled with sulphur hexafluoride gas was used as ultrasound agent throughout the study period. Contrast agent was administered as a bolus injection (2.4 ml during 3 seconds) followed by a 5 ml saline flush during 3 seconds with relatively constant speed. Manual injection of contrast agent was used by experienced nurses and a watch used to determine time. Real-time contrast imaging setting was used with a low mechanical index of <0.2 to avoid the bubbles disruption. And hepatic lesion was scanned continuously for up to 4 min after SonoVue injection. All images of CUES were digitally stored.
Histological examination. The histological diagnosis was the gold standard of diagnosis. All cases obtained final diagnosis histologically by analyzing biopsy specimen (38 cases) or surgical section (238 cases). US-guided biopsy was done at both within the nodule and the adjacent liver tissue by18 gauge trenchant needle (Bard Peripherals Vascular Inc, Tempe, Arizona 85281, USA). The histological diagnosis of HCC was made according the International Working Party criteria 14 . The differentiation of tumor cells was graded as well differentiated, SCienTifiC REPORts | (2018) 8:4713 | DOI:10.1038/s41598-018-23007-z moderately differentiated, and poorly differentiated, primarily based on Edmondson and Steiner 15 . Liver sections or biopsy specimens were formalin-fixed, stained with haematoxylin and eosin (HE) followed by immunohistochemical staining. Final pathological diagnosis was made in consensus by two pathological experts with over 20 years experience of liver pathology (XCY and DYG) who were unaware to of the clinical information and CEUS findings. In this study, hepatic background was divided into cirrhosis and non-cirrhosis on the basis of pathology. Liver cirrhosis was histoligically proven through evaluation of hepatic parenchyma specimen after surgical or biopsy.
CUES image analysis. Images were analyzed by two physicians with over sixteen years experience in abdominal diseases diagnosis (R L and XHZ) who were blind to the pathological diagnosis and the findings of CT or MRI. Contrast enhancement analysis was made in consensus between the two physicians. The whole enhancement procedure was studied, consisting of the arterial phase (0-30 s from the SonoVue bolus injection), portal phase (31-120 s after the injection), and delayed phase (121-240 s after the injection) according to EFSUMB recommendations 6 . The echogenicity of the lesion was noted as hyperechoic, isoechoic, hypoechoic, and mixed echo with respect to the surrounding liver tissue at baseline US. Contrast wash-in is considered as hyperechoic in comparison to the surrounding liver and the enhancement area is at least more than half of the lesion cross scanning section. The initial enhancement time and time to peak enhancement were registered separately.The initial enhancement time was defined as the time when contrast agent begins to emerge in the nodule after injection of contrast agent. The enhancement patterns of HCC at arterial phase were classified as follows: (1) Homogeneous enhancement-the whole nodule shows hyperechoic homogeneously compared with the surrounding liver parenchyma. (2) Inhomogeneous enhancement-when the lesion displays mixed hyper-enhancement at the periphery and the central part of the lesion, enhancement area involves more than half of the lesion. Washout is defined as hyper-enhancement (including homogeneous or inhomogeneous enhancement) in arterial phase followed by hypoenhancement in the portal or/and the late phase. If the lesion shows inhomogeneous hyper-enhancement in the arterial phase, observation of washout is focused on the area of hyper-enhancement in the arterial phase. The lesion without hyper-enhancement in arterial phase is not defined as washout. The time of the nodule beginning to display washout was recorded. And the time of the nodule beginning to display washout was defined as the time when the nodule began to show hypoenhancement (hypo-echogenicity) compared with the surrounding hepatic parenchyma. The hepatic lesion showing similar degree or higher degree of enhancement in the delayed phase when compared with the adjacent liver parenchyma preceded by hyper-enhancement in the arterial phase is defined as no washout. According to the first appearance of hyoenhancement, washout patterns were classified as washout in the arterial phase, washout in the portal phase, washout in the delayed phase and no washout. Statistical analysis. Comparison of enhancement patterns of HCC with different histological grades, different tumor sizes and different hepatic backgrounds was done by using the chi-square test or Fisher exact method. Wash-in time, time to peak and wash-out time were expressed as mean ± standard deviation. Differences of cellular differentiation, tumor size and background of liver on wash-in time, time to peak and wash-out time were analyzed by independent-samples t test or one way of analysis of variance. The statistical evaluation was performed using the SPSS 13.0 software package (SPSS Inc, Chicago, IL). A P value less than 0.05 was considered as statistically significant for each test.

Results
Characteristics of patients. The clinical characteristics of the 276 cases of HCC are presented in Table 1.
Mean age of the 276 patients was 51.1 ± 11.4 year(median: 50 year, range: 23-84 year) and 240 patients were male (86.9%). One hundred and sixty nine patients (61.2%) had cirrhosis. The etiology of cirrhosis was viral hepatitis B infection in 120 patients (71.0%), hepatitis B infection and alcohol misuse in 41patients (24.3%), alcohol misuse in 7 patients and viral hepatitis C infection in one patient. The remaining 107 (38.8%) patients did not have cirrhosis. Among them, sixty one patients had chronic hepatitis B without cirrhosis (57.0%). Twenty six patients (24.3%) had chronic hepatitis B and alcohol misuse but without cirrhosis. Fifteen patients had chronic alcoholic liver disease (14.0%). Five patients did not have any evidence of chronic liver disease. No significant difference was observed between the tumor diameter and tumor cellular differentiation (p > 0.05). The mean size of the lesions was 3.2 ± 1.0 cm (range: 1.0-4.9 cm). Enhancement patterns in arterial phase. Two hundred and seventy out of 270 (97.8%) lesions were hyperenhanced during the arterial phase as compared to the adjacent parenchyma, including 203 (75.2%) lesions with homogeneous enhancement and 67 (24.8%) lesions with inhomogeneous enhancement. As for the remaining 6 lesions, 4 lesions (1.5%) were hyoenhancement and 2 (0.7%) lesions were peripheral enhancement in the arterial phase. The relationship between enhancement patterns in arterial phase and tumor cellular differentiation, tumor diameter and hepatic condition are presented in Table 2. There was no difference between proportion of hyperenhancement in arterial phase and cellular differentiation, tumor diameter, or hepatic condition (p > 0.05) ( Table 2). Significant difference was only observed between homogeneous enhancement lesions and inhomogeneous enhancement lesions of different cellular differentiation and tumor size (p < 0.05) ( Table 2). More well differentiated HCCs (84.4%) and moderately differentiated HCCs (77.0%) manifested homogeneous enhancement as compared with poorly differentiated HCCs (56.3%) (p < 0.05), while higher percentage of poorly differentiated HCCs (42.5%) enhanced inhomogeneously in arterial phase when compared with well differentiated HCCs (12.5%) and moderately differentiated HCCs (21.0%) (p < 0.05). The analysis between well differentiated HCCs and moderately differentiated HCCs showed no significant difference (p > 0.05). There was no significant difference between enhancement patterns in the arterial phase and hepatic condition, namely cirrhotic or non-cirrhotic liver. (Table 2).
Washout patterns. The relationship between washout patterns and cellular differentiation, tumor size and hepatic condition were presented in Table 3. Two hundred and thirty seven lesions (87.8%) with hyperenhancement (homogeneous or inhomogeneous) in arterial phase became hyoechogenicity compared with surrounding liver tissue in portal or/and delayed phase. The percentage of washout in the portal phase, washout in the delayed phase or no washout was significantly different among different cellular differentiation (p < 0.05) ( Table 3). Further comparison revealed that a significantly higher percentage of moderately differentiated HCCs (61.2%) or poorly differentiated HCCs (74.4%) showed washout in portal phase than that of well differentiated HCCs (36.6%) (p < 0.05). And more well differentiated HCCs showed washout in delayed phase or no washout (41.9%, 21.5%, respectively) when compared with moderately differentiated HCCs (27.6%, 10.2%,respectively) or poorly differentiated HCCs (20.3%, 3.8%, respectively) (p < 0.05). The proportions of washout in portal phase, washout in late phase, and no washout were not significantly different between moderately and poorly differentiated HCCs (p > 0.05). No significance was found between washout patterns and tumor diameter or underling hepatic condition (p > 0.05) ( Table 3).  Table 4 shows the relationship between dynamic enhancement time and tumor cellular differentiation, tumor size and underling hepatic condition. The initial wash-in time and time to peak showed no significant difference in terms of cellular differentiation, tumor size and underling hepatic condition (p > 0.05) (   (Fig. 3). However, underlying hepatic condition did not exert an influence on the washout time of HCC on CEUS (p > 0.05) ( Table 4) (Fig. 4).

Discussion
CEUS, with purely intravascular contrast agent and contrast specific imaging techniques, was included in the diagnostic algorithm and recommended by EFSUMB in the diagnosis of focal liver lesions 16 . CEUS had the    17 . In the malignant transition process, the supplying vessels undergo a serious changes, normal arteries and portal veins were decreasing while abnormal neoplastic arteries increasing 18 . To some extent, these changes correlated with enhancement patterns of HCC in arterial phase on CEUS. Hyper-enhancement in arterial phase was found in 97.8% of HCC lesions in the present study. The results were comparable to other groups, who found hypervascularization in 96.2% 19 , 95% 12 , 90% 20 of HCC lesions in early phase. However, Von Herbay et al. 21 reported that 70.7% of HCC nodules showed arterial hypervascularization, while 29.3% showed non-hyperenhancement. This difference may be explaned by the differentiation of HCC nodules included. Von Herbay et al.included G4 HCCs (undifferentiated HCCs) in their study while our study and other researchers mentioned above did not. The histologial basis of hypoenhancement in arterial phase of HCCs was probably the decrease of artery supply. About 1.4% of HCC lesions displayed arterial hypoenhancement in our study which was similar to Nicolau's 19 study that 3.8% of HCC lesions showed arterial hypoenhancement. In the two studies, HCCs with hypoenhancement was mainly well differentiated HCCs (75% in our study and 100% in Nicolau's study). These results differed from that of von Herbay's 21 , in which 28% HCCs manifested non-enhancement (130 cases were enrolled in their study) and non-hyperenhancement occurred in all differentiated HCCs (24.3% in G1 cases, 35.1% in G2 cases, 21.6% in G3 cases, 16.2% in G4cases,and 2.7% in HCC/CCC cases). The possible reason could be the difference of US scan unit (Sonoline Elegra sonography unit in von Herbay's study vs Acuson Sequoia 512 ultrasound unit in our and Nicolau's study). Another reason could be no interpretation about non-hyperenhancement, which maybe includ iso-enhancement, hypo-enhancement, and peripheral enhancement. 1.5% of hypo-enhancement and 0.7% peripheral rim-like enhancement are in our study. Demonstrating that hypo-enhancement and peripheral rim-like enhancement are atypical for HCC. Hypoenhancement in delayed phase was found in 87.8% of HCC lesions in present study which was comparable to other studies by Jang 10 (92.0%), Boozari 12 (77.9%) and von Herbay 21 (86.2%). However, Iavarone and colleagues found that the percentage of hypoenhancement in the late phase was 34%, and the lesions ≤ 2 cm accounted for 66.1% of all lesions with size between 0.9-5 cm (median 1.6 cm) in their study 11 . The average washout time of all lesions with hyperenhancement in arterial phase followed by wash out in portal or/and delayed phase was 104.6 ± 58.4 s, which was comparable to that reported by Iavarone et al. (121.2 ± 71.2 s), but was longer than Kong's study (59.8 ± 20.5 s) 22 . The reason of the discrepance could be that Kong et al. included 95 recurrent lesions (accounted for 57.9% of all lesions) in their study. With regard to the relationship between tumor cellular differentiation and washout patterns, well differentiated HCCs (namely G1-HCC in other literatures) showed more often a washout in delayed phase (41.9%) or no washout (21.5%) in present study. These results was similar to some published studies 10,12,23 . Jang et al. 10 reported that 29% (4/14) of well differentiated HCCs showed washout after 3 min and 50% (7/14) of well differentiated HCCs showed no washout. However, Boozari et al. 12 reported that 94.4% (17/18)of G1-HCCs showed persistent enhancement in late phase. The difference may be attributed to sample number and tumor size. In Boozari's study, they only analyzed 18 G1 differentiated HCCs with 9 lesions < 2 cm in comparison to ours with more well differentiated HCC lesions (96 lesions in all) and more larger size lesions (80 lesions > 2 cm). Besides, the average time of emergence of washout in well differentiated HCCs was 132.3 ± 54.2 s, which was similar to Loria's 24 study (138.0 ± 114.5 s) and was significantly longer than moderately or poorly differentiated HCCs. The possible reason for late washout in well differentiated HCCs could be that the well differentiated HCC consisted of a trabecular pattern of cell cords and rich sinusoids that may lead to stagnation and slow clearing of microbubbles 25 . Our results demonstrated that HCCs with poorer grades of differentiation showed more often a washout in the portal phase than in well differentiated HCCs (p < 0.05). Washout patterns between moderately and poorly differentiated HCCs were not significantly different (p > 0.05). These results were comparable to other groups, who analyzed the difference between G1 differentiated HCC and G2-G3 differentiated or G2-G4 differentiated HCC 11,12,23,24 . However, statistically different results were observed by analyzing the time of initial washout in our study. Poorly differentiated HCCs (83.0 ± 39.8 s) showed the shortest washout time, moderately differentiated HCCs (100.4 ± 52.1 s) followed, and well differentiated HCCs (132.3 ± 54.2 s) the longest washout time. In our study, we observed that percentage of washout between moderately and poorly differentiated HCCs did not reveal statistical difference, whereas time of washout emergence between moderately and poorly differentiated HCCs was statistically different. This was an important finding and first reported up to now. If a nodule displays hyperenhancement in arterial phase followed by washout in early portal phase, suspicious of poorly differentiated HCC should be included in the differential diagnosis in patients with chronic liver disease. The possible explanation for early washout in poorly differentiated HCC lesions could be that this kind of HCC contained less sinusoids and were mainly nourished by hepatic artery 25 .
Tumor diameter plays a role in the prognostic process of HCC. In our study, there existed no relationship between tumor diameter and cellular differentiation. These findings were in agreement with von Herbay et al. 21 , but in contrast to Fan et al. 23 who found that tumors with larger size were significantly less differentiated HCC. Both our present study and Nicolau et al. 19 (by using a cut-off diameter of 2 cm) demonstrated that the hyperenhancement patterns and washout patterns were not related to tumor diameter. In contrast, von Herbay 21 found that more lesions (92%) > 3 cm manifested hyoenhancement in late phase than lesions <3 cm (77%). The same result was got in G1 subgroup lesions (95% of lesions > 3 cm vs 64% lesions < 3 cm). However, their study included fewer lesions in comparison to ours (in our study: 120 lesions ≤ 3 cm vs 156 lesions > 3 cm; in Nicolau's study: 36 lesions < 2 cm vs 68 lesions > 2 cm; in von Herbay's study: 53 lesions < 3 cm vs 77 lesions > 3 cm), thus, our data may lead to a better comparison between the two groups statistically. In addition, a significant percentage of lesions > 3 cm enhanced inhomogeneously in our study. The result was comparable to Fan's study 23 , with 64.6% of less differentiated HCCs, resembling to ours (65.2% of less differentiated HCCs), indicating more necrosis in large HCCs. Further more, time of initial hyoenhancement was significantly shorter in lesions > 3 cm in our study which included more moderately to poorly differentiated HCCs when compared with lesions ≤ 3 cm. . Enhancement features of HCC with cirrhosis and non-cirrhosis. There was a well defined, hypoechoic nodule of 2.8 cm with hypo-echoic halo in baseline us in the right lobe liver in a 56 year male patient (a). The nodule shows global hyperenhancemen in the arterial phase ((b),14 s after contrast agent injection) followed by slight washout in the portal phase ((c), 85 s after contrast agent injection) and moderate washout in the late phase ((d), 135 s after contrast agent injection). The patient was diagnosed poorly differentiated HCC without cirrhosis by surgery. There was a vague margin, slightly hypoechoic nodule of 2.4 cm in the right lobe liver in a 46 year male patient (e). The nodule displays heterogeneous hyperenhancemen in the arterial phase ((f), 23 s after contrast agent injection) followed by washout in the portal phase ((g), 80 s after contrast agent injection) and marked washout in the late phase ((h), 185 s after contrast agent injection). The patient was diagnosed poorly differentiated HCC with cirrhosis by surgery. The lesion was indicated by white arrow.
SCienTifiC REPORts | (2018) 8:4713 | DOI:10.1038/s41598-018-23007-z CEUS was widely used in the diagnosis of HCC in Europe and Asia. Though most HCC occur in cirrhotic liver, it may also happen in patients with chronic hepatitis. Therefore, it is of clinical importance to know whether underlying hepatic background affect the enhancement patterns of HCC on CEUS. Jang et al. 10 found that the proportions of each washout time interval were not significantly different between the two groups (no fibrosis and serve fibrosis). In our study, underlying hepatic background did not show correlation with perfusion characteristics of HCC, including enhancement patterns in arterial phase, washout patterns, and time of first occurrence of washout. It was an important finding, because HCC could occur in the liver without cirrhosis. Therefore, the noninvasive criteria could be probably applied to the HCC cases without cirrhosis, especially in patients with chronic hepatitis.
There were several limitations in our study. First, the sensitivity and specificity were not calculated. The results of our study need to be validated in a prospective study in the future. Second, the US scan unit in our study (Acuson Sequoia 512) has a relatively higher mechanical index of CEUS mode than other unit which maybe lead to earlier disruption of the microbubbles. Third, only the lesions ≤ 5 cm were enrolled in our study. We did not know whether the conclusions we got were adaptable to the lesions > 5 cm, in which the necrosis was more easily discovered.
In conclusion, global arterial enhancement and late washout were manifested in more well differentiated HCC than moderately or poorly differentiated HCC. Poorly differentiated HCC showed shorter washout time than moderately differentiated HCC. Large HCC (>3 cm) displayed more heterogeneous enhancement in the arterial phase and faster washout when compared with smaller HCC (≤3 cm). Underlying hepatic background showed no influence on the dynamic enhancement of HCC on CEUS.
Data availability statement. Our original data cannot be made available in the manuscript, the supplemental files, or a public repository because of ethical restriction of the ethics committee of Southwest hospital.