Alteration of contrast enhanced ultrasound (CEUS) of hepatocellular carcinoma in patients with cirrhosis and transjugular intrahepatic portosystemic shunt (TIPS)

Transjugular intrahepatic portosystemic shunt (TIPS) can treat portal hypertensive complications and modifies hepatic hemodynamics. Modification of liver perfusion can alter contrast enhancement dynamics of liver nodules. This study investigated the diagnostic performance of contrast-enhanced ultrasound (CEUS) to diagnose hepatocellular carcinoma (HCC) in cirrhosis with TIPS. In this prospective monocentric observational study, CEUS was used to characterize focal liver lesions in patients at risk for HCC with and without TIPS. Times of arterial phase hyperenhancement (APHE) und washout were quantified. Perfusion-index (PI) and resistance-index (RI) of hepatic artery and portal venous flow parameters were measured via doppler ultrasonography. Diagnostic gold standard was MRI/CT or histology. This study included 49 liver lesions [23 TIPS (11 HCC), 26 no TIPS (15 HCC)]. 26 were diagnosed as HCC by gold standard. Sensitivity and specificity of CEUS to diagnose HCC with and without TIPS were 93.3% and 100% vs. 90.9% and 93.3%, respectively. APHE appeared significantly earlier in patients with TIPS compared to patients without TIPS. TIPS significantly accentuates APHE of HCC in CEUS. CEUS has good diagnostic performance for diagnosis of HCC in patients with TIPS.


Methods
Patient cohort. This is an analysis including focal liver lesions of cirrhotic patients with and without TIPS from the prospective NICETIES cohort from our tertiary center (clinicaltrials.gov identifier: NCT03746210). CEUS was used to characterize focal liver lesions that were identified during HCC screening via B-mode ultrasound. Liver lesions were then confirmed by at least one more contrast-enhanced imaging (CT and/or MRI). Diagnostic gold standard was MRI/CT or histology. The choice of CT vs. MRI was made according to the physician's discretion and mostly depended on the patient's renal function and history of radiation exposure. For MRI, a commercially available clinical 3.0 T MR imaging system (Ingenia 3.0 T; Philips Healthcare, Best, Netherlands) or a 1.5 T MR imaging system (Ingenia 1.5 T; Philips Healthcare, Best, Netherlands) and for CT commercially available clinical CT imaging systems (Philips Brilliance 64 or Philips Brilliance 256 iCT, both Philips Healthcare, Best, The Netherlands) were used. For diagnosis of focal liver lesions standardized imaging protocols for MRI and CT were used according to our institution's standards. All imaging data were reviewed and interpreted by the same experienced radiologist. Histology slides were assessed by the same pathologist. This study was approved by the ethics committee of the University of Bonn, Faculty of Medicine (number 133/18) and performed in accordance to the declaration of Helsinki. Informed written consent has been obtained from all participants.

Contrast enhanced ultrasound (CEUS).
Contrast enhanced ultrasound (CEUS) was used to examine liver lesions in patients at risk of HCC (liver cirrhosis). Baseline sonography and CEUS were performed using the same ultrasound scanner (Supersonic Aixplorer, Supersonic Imaging). All patients were investigated after a 6-h fasting period. First, the liver was examined with baseline ultrasound to locate focal liver lesions. A sulphur hexafluoride-filled microbubble contrast agent, SonoVue (Fa. Bracco, Italy) was injected into a peripheral vein in a common manner. The dynamics of the contrast enhancement were observed for at least 5 min. The examination was performed by an experienced DEGUM (German Society of Ultrasound in Medicine) certified specialist in Internal Medicine.
Arterial phase hyperenhancement (APHE) and washout of contrast medium was quantified. A recent multicenter study showed that 90-99% nodules classified as CEUS LI-RADS 4 and 5 were HCC 9 . Moreover, APHE is key feature for the correct diagnosis of HCC in CEUS 23 . Therefore, in this study focal liver lesions classified as CEUS LI-RADS-4 and -5 were considered HCC in this study.
Hepatic perfusion parameters. Perfusion-index (PI) and resistance-index (RI) of hepatic artery and portal venous flow parameters were measured via doppler ultrasonography. Statistical analysis. Data  www.nature.com/scientificreports/ groups. Continuous variables are expressed as median (range). Categorical variables are presented as absolute cases or percentage. The sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) were analyzed. A two-tailed p-value < 0.05 was considered statistically significant.

Results
Patient characteristics. In total, 49 focal liver lesions from 33 patients with liver cirrhosis were included in this study (median age 61 years, 16 male). 18 patients had TIPS. Median MELD and Child-Pugh score were 13 (6-24) and 7 (5-11) respectively. Etiology of liver cirrhosis was mostly referred to alcohol in 25 (76%) patients. The indication for TIPS was refractory ascites in 12 (67%) and variceal bleeding in 6 (33%) patients. AFP was not significantly different between both groups (Table 1). There was no significant difference in the distribution of patients with solitary/multiple lesions, the number of patients with HCC, the number of lesions and number of HCC lesions between the TIPS and no-TIPS groups.
The diagnostic performance of CEUS was analyzed with the gold standard as reference. Sensitivity, specificity, positive predictive value, negative predictive value and diagnostic value to diagnose HCC in TIPS patients was 93.3%, 100%, 100%, 88.9%, 96% and in non-TIPS patients 90.9%, 93.3%, 90.9%, 93.3%, 92% ( Table 4). The rate of correctly classified HCC was similar in both groups, 14 of 15 (93%) lesions and 10 out of 11 (91%) in the TIPS and the no-TIPS group, respectively.
Quantitative assessment of contrast enhancement. Quantitative analysis of APHE timing showed eight seconds earlier APHE in patients with TIPS compared to non-TIPS patients, which was statistically significant ( Table 5). The start of venous washout time and time to be washed out completely was later in the TIPS group. Moreover, there was a trend towards higher rates of lesions without washout in non-TIPS patients ( Table 5). CEUS of all benign lesions, except for 3 referred to biopsy, showed characteristic patterns for the respective lesions (regenerative nodule, hemangioma, abscess).

Discussion
This study is the first to highlight the significantly altered APHE in CEUS of HCC in patients with TIPS (Supplementary Figure 1). Moreover, it confirms the diagnostic performance of CEUS for HCC and expands it to patients with TIPS. It is of special importance, because TIPS is indicated in selected patients with decompensated liver cirrhosis 24 , which are at high risk of developing HCC 4 . HCC is one of the few malignancies with an increasing mortality rate. Early diagnosis is crucial for prognosis. Therefore, good diagnostic and easily accessible tools in patients with TIPS that distinguish between malignant and benign lesions are an unmet need. Current international guidelines recommend regular HCC screening twice a year via B-mode ultrasound 25 . However, due to conflicting reports of low specificity, CEUS is debated across national and international guidelines 4,7,8,26 . Nonetheless, in clinical practice some patients have contraindications for contrast media and CT/MRI scans are not widely available in all health care systems. In these situations, CEUS plays an important role as diagnostic tool. In this study, we confirm the good diagnostic performance of CEUS for HCC in cirrhosis, underlining the robustness of our data 9,10 . Moreover, we expand the use of CEUS to patients with TIPS without significant change in diagnostic performance, suggesting routine use in these patients. This is not necessarily expected because diagnostic algorithms such as CEUS LI-RADS rely on APHE and washout in later phases. Concerning liver perfusion, TIPS severely reduces portal venous perfusion of the liver, resulting in a compensatory increase of hepatic arterial perfusion 17,19 . This could lead to the assumption that perfusion of liver focal liver lesions in patients with TIPS might be altered from "regular" perfusion. Indeed, HCC perfusion mainly relies on arterial neovascularization, which is reflected in CEUS LI-RADS criteria and even more emphasized in other algorithms 23 . According to our data, APHE is not only maintained but rather significantly accentuated after TIPS, making it possible for HCC detection with CEUS in patients with TIPS. However, CEUS performers need to be aware of the earlier timing of APHE in these patients.
An important issue for the treatment and prognosis is the distinction of HCC from other malignancies, cholangiocellular carcinoma in particular. Conflicting data led to degrading of CEUS in guidelines in the past 2 . Marked washout of the contrast agent in the portal venous phase or late phase is a common feature of most metastatic lesions and cholangiocellular carcinoma, while APHE patterns can differ between different kinds of malignant lesions. The washout pattern of HCC is typically mild, gradual and occurs in later phases. In intrahepatic cholangiocellular carcinoma and other malignancies, the washout appears earlier and more pronounced 27 . Venous washout of HCC begins later and requires a longer time to be washed out completely in our patients with TIPS compared to those without. Nevertheless, median start of washout was in the late phase after 120 s. The longer period of washout is in line with previous observations with rather mild (not fast and marked) washout in HCC lesions. The trend of longer washout in the TIPS group vs. no-TIPS group might also be due to the changed hepatic hemodynamics and -perfusion after TIPS placement. However, more studies with more detailed quantification of venous washout, i.e. via computerized time intensity curve analysis are needed to confirm our data.
Of note, in our study, two biopsied HCC that did not show washout were G1 HCC. In these cases in/correct classification by CT/MRI and CEUS were identical, suggesting the diagnostic performance to be associated to tumor grading. More studies including more patients are needed to investigate whether tumor grading and subclassification of HCC show different enhancement patterns.
In both of our study cohorts with and without TIPS, the non-HCC lesions showed characteristic contrastenhancement patterns, suggesting no significant impact of TIPS in those lesions. However, larger studies to  www.nature.com/scientificreports/ characterize non-HCC liver lesions in patients with TIPS and their differentiation are needed as well to confirm our results. To date, routine HCC screening is done by B-mode ultrasound. Though CEUS has been successfully described in the context of surveillance after HCC treatment [28][29][30][31] , data regarding the use of CEUS for routine HCC screening is at best scarce 32,33 . Currently, CEUS is recommended as a good diagnostic tool for the work up of focal liver lesions found in routine HCC screening via B-mode. Our data support the good diagnostic accuracy for HCC and expands it to patients with TIPS, who often are at risk for HCC development.
There are several limitations to this study, it is a monocentric study and the sample size is relatively small. However, despite the small sample size, this study is the first, aiming to establish CEUS as an accurate diagnostic tool for characterizing HCC in patients with TIPS. This should stimulate future prospective multicenter studies with larger cohorts for validation. Moreover, the gold standard would ideally be histology in all focal liver lesions. However, biopsy is not the standard of diagnostic for HCC according to guidelines. Interestingly, in this cohort MRI was performed more frequently in the TIPS group, which can be attributed to the higher rate of renal dysfunction being a contraindication for CT contrast medium in those patients.
In conclusion, this study shows first data of CEUS as an accurate diagnostic tool for characterizing HCC lesions in patients with TIPS. It shows an alteration of the APHE, and possibly venous washout pattern compared to patients without TIPS.