Endohepatology – Current Status

Jerome C. Edelson; Natalie E. Mitchell; Don C. Rockey


Curr Opin Gastroenterol. 2022;38(3):216-220. 

In This Article

Abstract and Introduction


Purpose of Review: This article discusses the most recent studies regarding the emerging field of endohepatology – the use of diagnostic and therapeutic endoscopic tools for the management of patients with liver disease and portal hypertension.

Recent Findings: New research has shown that liver biopsy specimens obtained by each Endoscopic ultrasound (EUS)-guidance, the percutaneous approach, and the transjugular approach contained sufficient portal triads to adequately analyzed by experienced pathologists – suggesting that any of these routes of liver biopsy is clinically acceptable; further, all had similar rates of adverse events. An initial prospective study showed that EUS guided portal pressure measurement was safe, effective, and accurate. A recent metanalysis showed that EUS-guided cyanoacrylate injection and coil embolization was statistically more efficacious and with less complications than EUS guided cyanoacrylate injection and EUS guided coil injection alone, suggesting that combination therapy appears to be the preferred approach for gastric varices (GV) bleeding. A prospective study evaluating focal liver lesions showed that the use of artificial intelligence had up to 100% sensitivity and 81% specificity for identifying malignant focal liver lesions.

Summary: EUS guided liver biopsy is safe and enables accurate diagnosis of underlying liver disease. EUS guided portal pressure measurement is also safe and is accurate. Combination therapy of EUS guided cyanoacrylate injection and coil embolization is more efficacious and has less complications than injection or coil therapy alone when used for GV bleeding. Artificial intelligence is highly sensitive and specific when used in conjunction with EUS in the diagnosis of malignant focal liver lesions. Endohepatology is a rapidly expanding field with great potential.


The anatomical relationship of the liver to the stomach makes it an easy organ to visualize with echoendoscopes, making evaluation of hepatic parenchyma both feasible and effective. Further, endoscopic ultrasound utilizes real-time imaging to provide both diagnostic and interventional access to the liver.[1] A wide range of both diagnostic and therapeutic options have been developed for use in patients with liver disease and portal hypertension (PH); this approach has led to an emerging field of endohepatology (Table 1).

Endosonographic Evaluation of the Liver

Evaluation of the Hepatic Parenchyma. There are many advantages to the utilization of EUS to evaluate the hepatic parenchyma. When compared to transabdominal ultrasound, the EUS probe is able to be placed extremely close to the liver (closer to the hepatic parenchyma that with transabdominal US), and as such, avoids overlying structures such as bowel loops, the gallbladder, or ribs, enabling superior image quality. The EUS probe is also able to better evaluate deep structures such as the hepatic hilum, caudate lobe, and left lobe of the liver. Several studies, including a prospective study enrolling 132 patients showed that EUS had a higher diagnostic accuracy for detecting metastatic disease at 98% compared to computerized tomography (CT) scan at 92% (P = 0.0578).[2] A retrospective study demonstrated that EUS can detect lesions that may be missed by CT and magnetic resonance imaging (MRI), including small lesions less than 3 mm in size.[3] In addition to evaluation with direct visualization, EUS also has the advantage of being able to acquire tissue through fine needle aspiration. A study evaluating 574 consecutive patients undergoing EUS with a history or suspicion of primary gastrointestinal or metastatic pulmonary tumor to the liver found that EUS could not only identify small focal lesions not detected by CT but also change the treatment course by making the cytologic diagnosis of malignancy in patients who previously did not have evidence of metastatic disease – thus avoiding unnecessary surgery.[4] However, it is important to note that EUS is limited in its ability to visualize the entire liver, and should therefore not be considered appropriate as a screening tool for metastatic disease.[5]

EUS guided transient elastography (EUS-TE) has been proposed as an approach to assess the hepatic parenchyma, particularly to assess fibrosis and steatosis. A potential advantage to EUS-TE is its ability to accurately assess fibrosis regardless of the presence of increased abdominal adiposity or ascites – both of which are important limitations of transabdominal TE techniques. An early prospective study evaluated echogenicity measured via EUS in 11 patients without known liver disease and 8 patients with cirrhosis found that echogenicity measured using pixel density could be used to segregate cirrhotic from non-cirrhotic patients with 100% sensitivity and 100% specificity.[6] A later prospective study enrolled 50 patients who underwent EUS-TE and found that realtime elastography obtained via EUS significantly correlated with abdominal imaging and can distinguish between normal, fatty, and cirrhotic appearing livers.[7] Although this approach is promising, larger prospective studies are required.

Artificial intelligence has also been proposed to evaluate for focal liver lesions. A prospective database including cases of focal liver lesions sampled from EUS (including a total of 210,685 unique images) in 256 patients was reviewed and used to train, validate, and test a convolutional neural network model.[8] The model found that when evaluating a random still image, there was a 90% sensitivity and 71% specificity for identifying malignant focal liver lesions (area under the receiver operating characteristic (AUROC) 0.861) and a 100% sensitivity and 80% specificity for identifying malignant liver lesions on full-length video clips (AUROC 0.904), suggesting a promising role for artificial intelligence (AI) in this area.[8]

Liver Sampling

Biopsy. Although there has been an emergence of noninvasive testing for various liver diseases due to advances in imaging resolution and elastography, histologic evaluation of hepatic tissue via liver biopsy still remains the gold standard and is critically important for many diagnoses and determination of subsequent treatment plans.[9] Percutaneous liver biopsy (PCLB) or trans-jugular liver biopsy (TJLB) have typically been considered to be the standard approaches; however, EUS-guided liver biopsy (EUSLB) has emerged as a promising alternative technique. For most patients, PCLB has traditionally been the preferred approach whereas TJLB is often utilized for individuals with large volume ascites, coagulopathy or those who require hepatic venous pressure gradient (HVPG) measurement to assess for PH.[10] EUSLB presents a number of advantages, including real-time imaging and localization, acquisition of liver biopsy specimens during endo-sonography, the ability to perform several needles passes with single puncture, and multilobar sampling regardless of body habitus. It is also faster than TJLB and results in significantly less pain compared to either PCLB or TJLB.[11,12]

A recent meta-analysis assessed the adequacy of liverspecimens obtained by PCLB, TJLB, andEUSLB by assessing the number of complete portal triads (CPT), total specimen length (TSL), and length of the longest piece (LLP) of liver tissue.[13] Although the optimal definition of specimen adequacy for EUSLB has not been established, the American Association for the Study of Liver Diseases (AASLD) currently recommends >11 CPTs as adequate – regardless of the route of sampling. Additionally, the AASLD guidelines also recommend a minimum TSL >15 mm, although it considers the ideal size to be 30 mm9. Cumulative biopsy adequacy rates were 98%, 98%, and 94% for PCLB, TJLB, and EUSLB, respectively.[13] In a similar meta-analysis, the successful histologic diagnosis was made in 94% of EUSLB patients.[14]

A subgroup analysis limited to EUS needles currently utilized in clinical practice demonstrated that EUSLB was, again, similar to PCLB and TJLB in biopsy adequacy.[13] There was no difference in both CPT and LLP between PCLB and EUSLB and no difference in LLP but fewer CPT compared to TJLB. Interestingly, TSL >30 mm occurred more often with EUSLB compared to PCLB and TJLB.[13]

Another study also suggested that specimen adequacy with EUSLB was similar to PCLB and TJLB, and TSL with was greater with EUSLB.[15]

The pooled rate of adverse events, specifically bleeding and postprocedural abdominal pain, was comparable for EUSLB and PCLB (16.5% vs. 12.2%, P = 0.516), respectively and to TJLB (16.5% vs. 4.9%, P = 0.071).[13] An adverse rate of 0.9% has been reported for EUSLB performed with a 19 gauge fine needle aspiration (FNA) needle – suggesting type of needle may improve safety outcomes compared to other core biopsy needles.[14] Additionally, it was noted that a 19 gauge needle was shown to provide significantly better biopsy specimens.[14]

Overall, currently available data suggest that EUSLB is a safe and effective technique, achieving comparable adequacy of diagnostic specimen sampling to PCLB and TJLB. Future prospective studies are necessary to determine cost-effectiveness of each modality as well as which subset of patients may most benefit from EUSLB.

Portal Pressure Measurement. Due to the proximity of the gastrointestinal tract to the major vessels of interest and relative ease of portal vein (PV) access, EUS-guided PV catheterization is attractive. Here, we have focused on measurement of portal venous pressures, albeit it should be noted that there are broad potential clinical applications of EUS-guided vascular access, including angiography as described further below, EUS-guided transhepatic intrahepatic portosystemic shunt creation and PV sampling for circulating tumor cells for pancreaticobiliary malignancies.[16–19]

PH is a complication of cirrhosis and portal pressure measurements are utilized clinically to aid in determining the stage, progression and prognosis of disease. There are a number of significant sequelae of PH, including esophageal varices, portal hypertensive gastropathy, and ascites. The HVPG reflects the degree of PH and appears to be the best prognostic indicator in advanced liver disease, making it a useful diagnostic tool with which to guide clinical decision-making. For example, an HVPG ≥10 mm Hg is associated with the development of varices and ≥ 12 mm with variceal hemorrhage.[20] Further, a reduction of HVPG by 20% or <12 mm Hg with pharmacotherapy can reduce the risk of future bleeding or rebleeding episodes.[20–22]

Currently, measurement of the HVPG is done via the jugular or femoral vein and is considered the gold standard. The free venous hepatic pressure is recorded and subtracted from a wedged hepatic venous pressure to determine the HVPG. However, this is an indirect measurement that may not truly reflect true PV pressures. Additionally, HVPG is typically only routinely performed at experienced tertiary medical centers, limiting its widespread applicability and usefulness[16,23] (Table 2).

EUS portal pressure measurements were first described in porcine models[23,24] and subsequently reported in a human clinic report in 2014.[25] An initial human study involving 28 patients demonstrated no adverse effects and 100% successful portal pressure gradient (PPG) measurements using a simple set up of a linear endoscope, 25G FNA needle, and a compact manometer.[26] In a small prospective study, EUS-PPG measurements were been shown to be comparable to HVPG measurements.[27] An advantage of an EUS-PPG approach compared to the traditional HVPG technique potentially offers patients with liver disease the ability to complete variceal screening, liver biopsy, and PPG in a single diagnostic study. Additionally, compared to HVPG, it is less invasive, minimizes radiographic exposure and obtains a direct PPG, which may be a more accurate reflection of true PH, particularly in patients with presinusoi-dal PH.[16]


EUS-guided Angiotherapy. Conventional endoscopic treatment of gastric varices (GV) typically involves the injection of cyano-acrylate glue via a sclerotherapy needle. EUS-guided angiotherapy improves the treatment of GV by enabling direct visualization of the network of varices, controlled and targeted delivery of therapeutic agents (liquid and nonliquid), along with real-time Doppler monitoring of response.[28,29] Cyanoac-rylate (CYA) with and without coil embolization has been demonstrated to be successful in achieving hemostasis during active bleeding and for secondary prophylaxis[30–32] and in a single-center study was shown to be effective for primary prophylaxis of high-risk GV.[32] A meta-analysis of EUS-guided angiotherapy for GV obliteration (EUS-guided CYA, EUS-guided coil injection and EUS-CYA þ coil embolization) demonstrated excellent technical success rates, clinical success rates, and low adverse events at 100%, 97% and 14%, respectively. Additionally, in a subgroup analysis, EUS-CYA þ coil embolization was more effective and had fewer complications than EUS-guided CYA and EUS-guided coil injection alone.[33] Although EUS-guided combination therapy appears to be the preferred approach for bleeding GVs, its current use is limited by cost and access to an experienced sonog-rapher. Future studies are needed to determine the optimal type and number of coils, liquid injectables and cost-effectiveness. Additionally, the use of EUS-guided angiotherapy for primary prophylaxis of GV warrants further study.

Future Directions. The future of endohepatology is bright, with many new applications evolving. EUS sampling of portal blood in peri-ampullary or pancreatic cancers has been shown to predict liver metastasis and has the potential to help further understand hematogenous spread of gastrointestinal cancers through the portal system.[28,34] EUS guided PV injection of chemo-therapeutic agents can result in twice the hepatic concentration and half the systemic concentration as well as shown that EUS guided tumor ablation could be a feasible approach in patients refractory to percutaneous approaches.[28] Animal models have shown technical success in achieving portal system decompression, suggesting a role for EUS guided intrahepatic portosystemic shunting in the future, as more purpose-made vascular devices are developed.[28]