Advances in endoscopic ultrasound-guided shear wave elastography: A comprehensive review of its clinical applications

Abstract

Endoscopic ultrasound-guided shear wave elastography (EUS-SWE) represents a significant advancement in non-invasive tissue characterization, enabling objective assessment of quantitative tissue stiffness in real-time with potential clinical relevance across a variety of gastrointestinal disorders. Recent developments in EUS-SWE have expanded its application beyond hepatic fibrosis to include pancreatic diseases and the evaluation of solid tumors. EUS-SWE has demonstrated diagnostic accuracy comparable to vibration-controlled transient elastography in assessing fibrosis stages, positioning it as a potential alternative to liver biopsy. Moreover, EUS-SWE has shown promise in evaluating pancreatic tissue stiffness, aiding in the diagnosis and monitoring of chronic pancreatitis and pancreatic cancer. This technique offers a distinct advantage by allowing tissue stiffness measurements during the same procedure, thereby reducing the need for additional imaging studies and biopsies. Despite its clinical potential, challenges remain, including the need for standardized protocols, optimal cutoff values, and validation across diverse patient populations. This minireview provides a comprehensive analysis of the current literature on EUS-SWE, examining its diagnostic performance, reproducibility, and limitations. Furthermore, we discuss the future directions of EUS-SWE, including its integration into routine clinical practice and its evolving role in precision medicine, emphasizing the necessity of large-scale studies to solidify its clinical utility and establish standardized guidelines for its use.

Key Words: Endoscopic ultrasound; Shear wave elastography; Chronic pancreatitis; Autoimmune pancreatitis; Liver fibrosis

Core Tip: Endoscopic ultrasound (EUS)-guided shear wave elastography provides real-time, quantitative assessment of tissue stiffness during standard EUS examinations. Current evidence supports its value in staging chronic pancreatitis, monitoring autoimmune pancreatitis, and assessing liver fibrosis with diagnostic accuracy comparable to vibration-controlled transient elastography. While feasibility and reproducibility are consistently high, the lack of standardized protocols and validated cut-off values limits routine application. Large-scale prospective studies are essential to define its role in clinical practice and establish EUS-shear wave elastography as a reliable tool in precision gastroenterology.

INTRODUCTION

Endoscopic ultrasound (EUS) has markedly advanced the management of numerous conditions, with well-defined clinical indications, particularly in the field of biliopancreatic diseases[1-3]. Despite its high resolution, conventional B-mode imaging alone may not always ensure accurate characterization, and differentiating between disease entities often remains challenging[4,5]. To overcome these limitations, advanced EUS-based imaging modalities have been developed, improving diagnostic performance in several clinical settings[6]. Among these, EUS-guided elastography has emerged as one of the most promising tools for the evaluation of pancreatic and hepatic disorders.

Elastography provides information on tissue stiffness, with two main approaches applicable to EUS: Strain elastography (SE) and shear wave elastography (SWE). Initial applications of EUS elastography were limited to SE, a non-invasive real-time method that evaluates tissue deformation induced by slight probe compression or cardiovascular pulsation[7]. Its results can be interpreted qualitatively using color maps or quantitatively through strain ratio and histogram analysis[8]. Although SE demonstrated a good diagnostic yield in biliopancreatic disorders, its major limitation is the lack of objectivity, as it provides only relative elasticity within a region of interest (ROI) and is influenced by operator-dependent pressure, resulting in reduced reproducibility and limited quantification[9]. Until 2019, SE represented the sole elastographic modality available in the EUS setting.

More recently, SWE has been introduced as a complementary technique, offering objective and reproducible quantification of tissue stiffness. SWE is widely applied in abdominal imaging and is particularly established for staging chronic liver disease, where it quantifies stiffness as a surrogate marker of fibrosis[10,11]. Vibration-controlled transient elastography (VCTE) was the first SWE-based technique and remains one of the most important reference in hepatology[12,13]. Subsequently, multiple transabdominal platforms integrated SWE technology, with both point-SWE and two-dimensional SWE now endorsed by international societies[14,15]. These modalities provide the additional advantage of visualizing the exact region of measurement, overcoming the blind sampling inherent to VCTE, and extending applicability to organs such as the pancreas and intestinal wall[16].

Nevertheless, the role of SWE in the pancreas remains less well defined, likely due to technical challenges related to organ depth, complex morphology, and interference from obesity or bowel gas, which may also affect hepatic assessments[17,18]. To address these limitations and broaden elastographic capabilities, SWE has recently been incorporated into EUS platforms, representing an important step forward beyond SE and currently the focus of active clinical investigation.

SEARCH DETAILS

We conducted a structured narrative review of EUS-guided elastography with a focus on EUS-SWE. We searched MEDLINE/PubMed, Google Scholar and the Cochrane Library using combination of terms including “endoscopic ultrasound”, “EUS”, “shear wave elastography”, “SWE”, and “EUS-SWE”. We also screened reference lists of key reviews and primary studies. We included original human studies and high-quality reviews reporting technical, diagnostic, or clinical data on EUS-SWE of pancreas, liver and spleen, and excluded non-EUS-SWE studies, non-English papers, abstracts without full data, and case reports. Given the heterogeneity in study designs, equipment, ROI settings, and reference standards, we synthesized the evidence narratively rather than performing a meta-analysis.

TECHNICAL FOUNDATIONS OF EUS-SWE

SWE relies on the generation and propagation of shear waves within biological tissues to estimate their mechanical properties[19]. In practice, shear waves are created by applying a localized acoustic radiation force impulse generated by a focused ultrasound pulse. This brief “push pulse” induces a minute displacement of tissue particles, which in turn produces laterally propagating shear waves oriented perpendicular to the initial ultrasound beam[20]. The propagation velocity of these shear waves (Vs, expressed in m/second) is directly related to tissue stiffness: Stiffer tissues transmit waves faster, whereas softer tissues slow down propagation[21]. Ultrasound systems detect these waves using high-frame-rate “tracking pulses,” which monitor the minute displacements induced by shear wave propagation at different points within a predefined ROI[22]. Based on the time-distance relationship, the Vs can be calculated. Through physical models of tissue mechanics, this velocity can also be converted into the Young’s modulus (E, kPa) using the equation E = 3ρVs², where ρ represents tissue density[23]. Note that the conversion of Vs to Young’s modulus assumes a near-incompressible, isotropic, and linearly elastic medium with constant tissue density. In vivo, tissues are viscoelastic and heterogenous, so these assumptions may not hold, potentially leading to overestimation of stiffness. Therefore, Vs should be considered the primary quantitative parameter, while E should be interpreted cautiously and reported together with acquisition setting and reliability indices. To ensure internal validity, each Vs measurement is accompanied by a reliability index (net effective shear wave velocity, VsN, %), which reflects the proportion of effective shear waves contributing to the calculation; a VsN above 50% is generally considered indicative of a reliable acquisition. Furthermore, reproducibility across multiple measurements can be quantified by the interquartile range-to-median ratio (IQR/M), a quality index whereby lower values indicate greater consistency. While universal cut-offs have not yet been validated for EUS-SWE, criteria commonly extrapolated from liver elastography suggest IQR/M < 15% for Vs and < 30% for E as indicative of high reliability[18].

Importantly, shear wave generation and detection are highly sensitive to factors such as depth, probe pressure, and tissue heterogeneity[24-26]. In EUS-SWE, additional variables, including respiratory motion, proximity of vessels or ducts, and the distance between probe and target organ, may influence wave propagation and measurement reliability[5,18,27]. The integration of indices such as VsN and IQR/M into EUS-SWE software provides a standardized framework for evaluating measurement quality and reproducibility, thereby facilitating the clinical adoption of this technique[28]. In particular, respiratory motion can have an impact in measurement during SWE and to limit this issue some authors suggest to perform measurement during minimal respiratory fluctuations.

SWE is considered very safe; contraindications are limited to those regarding performing EUS more than SWE itself. SWE has been proven to be safe also during pregnancy. There are however factors considered to be associated with technical limitations of SWE such as respiratory motion artifacts, as already mentioned, and intervening structures such as vessels, fluid or very dense structures such as cyst or bones. Finally, it is not possible to assess deep structures as the penetration of the wave is limited. On the other hand, EUS-SWE is less affected by obesity and ascites compared to standard transabdominal SWE.

EQUIPMENT AND PLATFORMS CURRENTLY AVAILABLE

One important element to consider regarding accessibility is that at the moment EUS processors featuring SWE are limited in number and usually available only on the most expensive top-of-the-line products that have obtained Food and Drug Administration clearance (for tissue stiffness quantification). Most commonly the ARIETTA 750/850 platform (Fujifilm, previously Hitachi, Japan), in combination with different convex-array echoendoscopes. Recently, the new Aplio i800 (Canon, Japan) coupled with linear EUS probes by Olympus allows also to perform 2D-SWE. Several manufacturers provide compatible linear or curvilinear scopes, which allow the generation and tracking of shear waves during EUS examinations. The most widely used are Olympus models such as the GF-UCT260, extensively adopted in Japanese studies for pancreatic applications[29-35] and its more recent evolution, the GF-UCT180, employed in Western cohorts for hepatic and pancreatic assessment[36,37]. A further Olympus option is the GR-UCT180, used in large multicenter feasibility studies, which demonstrated high reliability and reproducibility across different pancreatic regions[38]. Fujifilm convex scopes (EG-740UT) have also been applied for pancreatic elastography, sometimes in combination with radial probes (EG-580UR) for diagnostic evaluation[39]. Pentax devices, such as the EUS-J10 linear echoendoscope, represent an alternative platform, although less frequently reported in the literature[40]. While radial echoendoscopes are primarily diagnostic and not suited for SWE quantification, linear or curvilinear scopes remain the standard for elastographic measurements as well as for EUS-guided interventions.

CLINICAL APPLICATIONS OF EUS-SWE, PANCREATIC APPLICATIONS

EUS-SWE is an emerging technology for the quantitative characterization of pancreatic stiffness. When integrated with EUS-guided fine-needle aspiration/biopsy and contrast-enhanced EUS, it holds the potential to provide a comprehensive, multiparametric assessment of pancreatic parenchymal architecture and focal lesions. Early studies suggest its potential role in the diagnosis and staging of chronic pancreatitis (CP), in monitoring autoimmune pancreatitis (AIP), and in the characterization of pancreatic lesions.

CP

Several studies have investigated the relationship between EUS-SWE and the clinicopathological features of CP, with a focus on disease stage and pancreatic function. The progressive fibrotic remodeling that characterizes CP leads to increased parenchymal stiffness, suggesting that quantitative elastographic assessment may serve as an objective biomarker of disease progression.

Consistently, multiple studies have demonstrated that Vs increase across diagnostic categories, regardless of the classification system applied. Yamashita et al[34] reported a strong positive correlation between EUS-SWE values and the number of Rosemont criteria (RC) features (rs = 0.72, P < 0.001), a finding subsequently confirmed in larger cohorts and by Shintani et al[31], who also showed a significant association between Vs and the total number of EUS features (rs = 0.626, P < 0.001)[31,32]. In multivariate analyses, a history of acute pancreatitis, the presence of ≥ 2 EUS features, and lobularity with honeycombing in the pancreatic body[30] emerged as independent predictors of increased stiffness.

From a diagnostic standpoint, EUS-SWE has consistently demonstrated good performance in distinguishing CP from normal pancreas. Yamashita et al[33] showed that Vs values increased progressively across Japanese Pancreas Society stages (1.36 m/second for normal pancreas, 2.07 m/second for early CP, 2.78 m/second for probable CP, and 3.08 m/second for definite CP; P < 0.001), with an area under the receiver operating characteristic (AUROC) of 0.92 and an optimal cut-off of 1.96 m/second. Similar results were observed by Shintani et al[31], with significantly higher Vs in CP compared to controls (2.31 ± 0.67 m/second vs 1.59 ± 0.40 m/second; P < 0.001). Stratification by RC further confirmed that SWE values in “consistent with CP” and “suggestive of CP” categories were markedly higher than in normal pancreas[32]. Reported diagnostic thresholds vary across studies, ranging from 1.96 to 2.33 m/second, with AUROC values between 0.82 and 0.97 and sensitivities/specificities consistently above 80%[30-34]. The best performance was reported by Yamashita et al[32], with an AUROC of 0.97 using a cut-off of 2.19 m/second (sensitivity 100%, specificity 94%). This variability likely reflects differences in study population, ROI selection, and operator technique, as most measurements were performed in the pancreatic body but under non-uniform depth and angle settings. Inter-operator variability and the absence of standardized reliability thresholds (e.g., VsN ≥ 50%-60%) further influence stiffness values and diagnostic accuracy. Moreover, etiology specific differences may alter parenchymal fibrosis patterns, contributing to divergent Vs ranges. Despite the overall high accuracy, variability among studies underscores the need for standardized acquisition protocols and defined population selection.

The relationship between pancreatic stiffness and functional impairment has also been explored. EUS-SWE values were significantly higher in patients with exocrine insufficiency and correlated with quantitative pancreatic function tests[33,34]. However, its accuracy in detecting endocrine dysfunction associated with diabetes mellitus was modest (AUROC: 0.63-0.75)[32,33], and Shintani et al[31] found no significant differences in Vs according to pancreatic insufficiency status.

Finally, when directly compared with SE, EUS-SWE showed superior diagnostic accuracy across multiple reference standards. In a prospective study of 49 patients undergoing both techniques, Yamashita et al[34] demonstrated that SWE significantly outperformed SE for the diagnosis of CP defined by computed tomography, RC, Japanese Pancreas Society 2019, and exocrine dysfunction (AUROCs: 0.77-0.85 for SWE vs 0.53-0.61 for SE; P < 0.001 in all comparisons). These findings highlight SWE as a more objective and reproducible method for quantifying pancreatic stiffness and for supporting the diagnosis and staging of CP.

AIP

EUS-SWE may play a role not only in the diagnosis of AIP but also in monitoring treatment response to corticosteroids. In a prospective exploratory study including 160 patients, 14 of whom were diagnosed with AIP, Ohno et al[29] reported significantly higher stiffness in the pancreatic body of AIP patients compared with controls (median Vs 2.57 m/second vs 1.89 m/second; P = 0.0185). In six patients re-evaluated two weeks after corticosteroid therapy (prednisone 0.6 mg/kg/day), mean Vs decreased from 3.32 m/second (IQR: 2.93-3.59) to 2.46 m/second (IQR: 1.84-2.96; P = 0.0234). No correlation was found between Vs and serum IgG4 levels[29]. In contrast, Kojima et al[39] retrospectively analyzed 17 patients with AIP and 14 controls and found no significant difference in Vs between lesions, background pancreas, and controls. However, the reliability index (VsN) was significantly lower in AIP lesions compared with background parenchyma and controls (17% vs 71.1% vs 86.7%; P < 0.001). A VsN threshold of 50% yielded an AUROC of 0.929 for distinguishing lesions from other groups. In three cases followed longitudinally, both Vs and VsN improved stepwise at 4, 8, and 16 weeks after corticosteroid initiation. Notably, VsN did not differ between AIP lesions and pancreatic neoplasm, highlighting the limited discriminative value of this parameter for lesion characterization[39] (Table 1). Overall, although based on small cohorts, these findings suggest a potential role of EUS-SWE in both the diagnosis and monitoring of AIP, warranting further studies before stiffness-based remission criteria can be established.

Table 1 Endoscopic ultrasound-guided shear wave elastography in pancreatic diseases.

Ref.	Design	Population	Techniques	Reference standard	Results
Kojima et al[39]	Single-center, retrospective observational study (pilot + SWE sub-study)	Diagnostic EUS in 16 cases (normal pancreas, IPMN, CP, PC, etc.); SWE study: 32 PC, 17 AIP, 14 controls	EUS-SWE (10 measurements, ROI 5 mm × 10 mm; Vs and VsN measured)	Clinical diagnosis (PC, AIP; normal controls)	PC: Vs not significantly different vs controls (2.22 m/second vs 1.96 m/second, P = 0.063), but VsN significantly lower in lesions (21% vs 87%, P < 0.001); AUROC 0.918 for VsN cut-off 50%. AIP: Vs not significantly different vs controls (2.35 m/second vs 1.96 m/second, P = 0.065), but VsN significantly lower (17% vs 87%, P < 0.001); AUROC 0.929. Treatment monitoring: In AIP, Vs and VsN improved after 16 weeks of steroid therapy
Shintani et al[31]	Retrospective, single-center	657 screened; 22 patients with ECP vs 22 matched controls (JPSC)	EUS-SWE (≥ 5 measurements in pancreatic body, ROI 10 mm × 5 mm; VsN > 50% considered reliable)	JPSC 2019 for ECP	Mean Vs significantly higher in ECP vs normal (2.31 ± 0.67 m/second vs 1.59 ± 0.40 m/second, P < 0.001). Optimal cut-off 2.24 m/second (AUROC 0.82, 95%CI: 0.69-0.94). Vs strongly correlated with number of EUS findings (rs = 0.626, P < 0.001). Independent predictors of high Vs: History of acute pancreatitis and ≥ 2 positive EUS findings
Shintani et al[30]	Retrospective, single-center	50 patients with CP or suspected CP	EUS-SWE in three pancreatic regions (head, body, tail; ≥ 5 measurements per site, ROI 10 mm × 5 mm, VsN > 50%)	RC	Vs values significantly higher in CP vs non-CP across all regions. Diagnostic accuracy highest in pancreatic body: AUROC 0.87 (cut-off 2.33 m/second; sensitivity 824%, specificity 87.5%). Head and tail AUROCs 0.79 and 0.81, respectively. Vs correlated with number of RC features in all regions (strongest in body, rs = 0.55). Lobularity with honeycombing was independent predictor of stiffness
Yamashita et al[34]	Prospective, single-center	49 patients with suspected/confirmed CP	EUS-SWE vs EUS-SE	CT, RC, JPSC 2019, pancreatic exocrine dysfunction	EUS-SWE values positively correlated with RC and JPSC severity grades and number of EUS features; EUS-SE showed no correlation. Diagnostic accuracy (AUROC SWE vs SE): CT 077 vs 0.61; RC 0.85 vs 0.56; JPSC 0.83 vs 0.53; exocrine dysfunction 0.78 vs 0.61
Abboud et al[38]	Prospective, single-center	117 consecutive patients undergoing EUS for various indications (cysts, CP, recurrent acute pancreatitis, mass, screening)	EUS-SWE (10 measurements in head, body, tail; ROI 5-15 mm; VsN > 50% considered reliable)	Not disease-specific (technical feasibility/reproducibility study)	Safety/feasibility: 3320 measurements, 100% success, no peri-procedural complications. Reliability: Higher in head (85.1% reliable), vs body (75.5%) and tail (64.2%). Reproducibility: ICC good across all sites (0.80-0.89)
Yamashita et al[33]	Prospective, single-center	40 patients (normal, early, probable, and definite CP)	EUS-SWE (10 measurements in pancreatic body; ROI 5 mm × 10 mm)	JPSC 2010; BT-PABA for exocrine function; DM for endocrine dysfunction	EUS-SWE positively correlated with JPSC stages and with pancreatic function diagnostic test. Diagnostic accuracy: CP AUROC 0.92 (cut-off 1.96 m/second; sensitivity 83%, specificity 100%); exocrine dysfunction AUROC 0.78 (cut-off 1.96 m/second; sensitivity 90%, specificity 65%); DM AUROC 0.63 (cut-off 2.34 m/second; sensitivity 75%, specificity 64%)
Ohno et al[35]	Retrospective, single-center	64 patients with solid pancreatic lesions (43 PC, 9 MFP, 9 PanNEN, 3 metastases)	EUS-SWE vs EUS-SE	Histology or ≥ 12 months clinical follow-up	Vs: No significant difference between PC, MFP, PanNEN, metastases (PC 2.19 m/second vs MFP 2.56 m/second, P = 0.57); AUROC for PC vs non-PC = 0.56. VsN: Low across SPLs (PC 19%, MFP 13%, PanNEN 18%). Strain histogram: Mean strain significantly lower in PC vs MFP (45.4 vs 74.5, P = 0.0007); AUROC = 0.68
Yamashita et al[32]	Retrospective, single-center	52 patients (normal, indeterminate, suggestive, and consistent CP)	EUS-SWE (10 measurements in pancreatic body; ROI 5 mm × 10 mm)	RC; DM as marker of endocrine dysfunction	EUS-SWE values significantly correlated with RC stages (rs = 0.81) and number of EUS features (rs = 0.72). Diagnostic performance: CP AUROC 0.97 (cut-off 2.19 m/second; sensitivity 100%, specificity 94%); DM AUROC 0.75 (cut-off 2.78 m/second; sensitivity 70%, specificity 56%)
Ohno et al[29]	Prospective, single-center exploratory study	160 patients undergoing EUS; subgroup: 14 with AIP, 16 normal controls	EUS-SWE (≥ 5-10 measurements in pancreatic head, body, tail; ROI 5 mm × 10 mm; VsN reliability index, ≥ 50% accepted)	International Consensus Diagnostic Criteria for AIP	Overall feasibility: 97.6% success rate (3743/3837 measurements). Median Vs in AIP group significantly higher than controls (2.57 m/second vs 1.89 m/second, P = 0.0185). In 6 patients, Vs decreased after steroid therapy (3.32 m/second to 2.46 m/second, P = 0.0234), paralleling pancreatic size reduction

SWE: Shear wave elastography; EUS: Endoscopic ultrasound; IPMN: Intraductal papillary mucinous neoplasm; CP: Chronic pancreatitis; PC: Pancreatic cancer; AIP: Autoimmune pancreatitis; ECP: Early chronic pancreatitis; MFP: Mass-forming pancreatitis; PanNEN: Pancreatic neuroendocrine neoplasms; ROI: Region of interest; Vs: Shear wave velocity; VsN: Net effective shear wave velocity; SE: Strain elastography; JPSC: Japanese diagnostic criteria 2019; RC: Rosemont criteria; CT: Computed tomography; BT-PABA: N-benzoyl-L-tyrosyl-p-aminobenzoic acid; DM: Diabetes mellitus; AUROC: Area under the receiver operating characteristic; CI: Confidence interval; ICC: Intraclass correlation coefficient; SPLs: Solid pancreatic lesions.

Pancreatic cancer

Current evidence does not show significant differences in Vs between malignant lesions, surrounding pancreatic parenchyma, and normal pancreas, thereby limiting the immediate diagnostic utility of EUS-SWE for characterizing solid pancreatic masses. In particular, the study conducted by Kojima et al[39], including 32 patients with pancreatic cancer (PC) and 14 with normal pancreas, found that Vs values did not differ significantly between cancer lesions, background parenchyma, and controls (2.22 m/second vs 2.26 m/second vs 1.96 m/second). In contrast, VsN was markedly lower in cancer lesions compared with background and normal parenchyma (21.1% vs 66.4% vs 86.7%; P < 0.001). A VsN threshold of 50% yielded an AUROC of 0.918 for distinguishing lesions from other groups.

Similarly, Ohno et al[35] reported in a cohort of 64 patients that Vs values were comparable across PC, mass-forming pancreatitis, and pancreatic neuroendocrine neoplasms (2.19, 2.56, and 1.31 m/second, respectively), with no significant differences among groups. As a result, the diagnostic performance of Vs alone was limited [area under the curve (AUC) 0.56]. By contrast, the use of SE with histogram analysis in the same study provided superior discrimination, showing significantly lower strain values in PC compared with inflammatory lesions and a higher diagnostic yield (AUC 0.68).

The limited diagnostic value of EUS-SWE in PC could be explained by the heterogeneity of tumor tissue, which includes variable proportions of necrosis, fibrosis and viable cancer cells. This structural complexity leads to unstable shear-wave propagation and wide intra-lesional variability in Vs and VsN values, reducing reproducibility and contrast with surrounding parenchyma. Future perspectives may rely on multimodal approaches, combining SWE with contrast-enhanced EUS or artificial intelligence-based image analysis to integrate stiffness maps with vascular and texture data, thereby improving diagnostic performance and lesion differentiation. Representative EUS-SWE images of the normal pancreas, CP and PC are shown in Figure 1.

Figure 1 Endoscopic ultrasound-guided shear wave elastography images. A: Normal pancreas; B: Chronic pancreatitis; C: Pancreatic cancer. In endoscopic ultrasound-guided shear wave elastography, shear wave velocity values were similar in normal parenchyma and chronic pancreatitis, but markedly higher in pancreatic cancer. The net effective shear wave velocity shows a decreasing trend across these conditions, reflecting increasing tissue heterogeneity and stiffness.

Feasibility, reliability and reproducibility

EUS-SWE showed high feasibility in pancreatic assessments, with reported success rates consistently high, ranging from 96%-99%[29] to 100%[38], across the pancreatic head, body, and tail. Reliability is commonly defined as a VsN (valid shear wave measurements) ≥ 50%, and both Abboud et al[38] and Ohno et al[29] reported high reliability in the pancreatic head (median VsN 85% and 83%, respectively). However, other evidence suggests that the pancreatic body provides the most reproducible measurements: Shintani et al[30] reported VsN rates of 91.6%, 93.5%, and 82.8% in the head, body, and tail, respectively. Reproducibility has also been demonstrated with intraclass correlation coefficients up to 0.94[38]. Moreover, significant correlations between Vs and the number of EUS criteria, most pronounced in the pancreatic body (Spearman’s rs = 0.34, 0.55, and 0.34 in head, body, and tail, respectively), further support preferential acquisition in this region[30].

Limitations and future directions

The clinical use of EUS-SWE remains limited by methodological heterogeneity and the absence of universally validated cut-off values. Most pancreatic EUS-SWE studies are small, single-center series with heterogenous inclusion criteria and variable reference standards, which may overestimate diagnostic accuracy. Across published studies, acquisition protocols vary considerably in terms of measurement site, number of acquisitions, and diagnostic criteria used for CP, while patient cohorts remain relatively small, particularly for AIP and pancreatic neoplasms.

Technical factors also influence measurement accuracy. EUS-SWE is sensitive to the distance between the pancreas and the gastrointestinal wall, probe stability, and respiratory motion, all of which may lead to variability in Vs across different pancreatic regions. In particular, pancreatic stiffness has been shown to vary significantly with the respiratory phase, with higher values recorded during deep inspiration compared with free breathing (5.7 kPa vs 4.1 kPa; P < 0.001)[41]. Likewise, excessive probe compression can artificially increase elasticity values, with experimental studies demonstrating up to several-fold overestimations due to the nonlinear mechanical behavior of soft tissues[42]. Despite these challenges, most studies have focused on the pancreatic body, with a limited number of measurements (typically 5-10), leaving the optimal number and regional distribution of acquisitions uncertain. Lastly, the broader clinical implementation of EUS-SWE remains constrained by the scarcity of SWE-capable equipment and the requirement for specialized operator training.

Further research is needed to establish standardized acquisition protocols and reliable cut-off values that can be applied across centers. Key priorities include multicenter validation studies using harmonized acquisition parameters (e.g., ROI depth < 2 cm, ROI height 10-15 mm) and predefined quality metrics such as VsN ≥ 50%-60% and IQR/M thresholds. In the future, artificial intelligence-assisted frame selection, ROI optimization, and system calibration may help minimize operator dependence and enhance reproducibility.

HEPATIC APPLICATIONS

Endo-hepatology is a rapidly evolving discipline that integrates both diagnostic and therapeutic EUS to assist clinicians in the evaluation and management of liver diseases[43]. Nonetheless, current evidence for EUS-SWE in this setting remains limited, being restricted to diffuse liver disorders, the quantification of fibrosis and evaluation of portal hypertension (Table 2). Representative EUS-SWE images of the normal liver, cirrhotic liver, and spleen are shown in Figure 2.

Figure 2 Endoscopic ultrasound-guided shear wave elastography images. A: Normal liver; B: Cirrhotic liver; C: Spleen. In endoscopic ultrasound-guided shear wave elastography, shear wave velocity is higher in cirrhotic liver compared to normal liver, reflecting increased parenchymal stiffness. Spleen stiffness measurements may provide additional information on the degree of portal hypertension.

Table 2 Endoscopic ultrasound-guided shear wave elastography in liver diseases.

Ref.	Design	Population	Techniques	Reference standard	Results
AbiMansour et al[27]	Prospective, single-center	Patients undergoing EUS for various indications; subset analyzed for liver disease staging	EUS-SWE performed in the left lobe of the liver during EUS	VCTE and clinical diagnosis of advanced liver disease	Patients with advanced liver disease showed higher values of EUS-SWE compared to healthy control: LHL 17.6 kPa vs 12.7 kPa (P < 0.001); RHL 24.8 kPa vs 11.0 kPa (P < 0.001)
Yousaf et al[45]	Prospective + retrospective, single-center	25 patients with suspected chronic liver disease	EUS-SWE (right lobe), EUS-PPG, and concomitant EUS-liver biopsy	Histology (METAVIR), clinical/radiologic criteria	Mean EUS-SWE 24.1 kPa. Significant fibrosis (F2-F4) associated with higher PPG (5.9 mmHg vs 2.5 mmHg, P = 0.003) and higher stiffness (30.0 kPa vs 15.6 kPa, P = 0.02). Advanced fibrosis (F3-F4) vs non-advanced: PPG 6.0 mmHg vs 3.4 mmHg (P = 0.01), EUS-SWE 32.0 kPa vs 18.8 kPa (P = 0.04)
Del Valle et al[40]	Single-center, cross-sectional diagnostic trial (case-control)	59 patients: 29 cirrhosis, 30 controls	EUS-SWE (ROI 10 mm, ≥ 10 measurements in both lobes) vs VCTE	Gold standard: 2-lobe EUS-guided liver biopsy (Kleiner/METAVIR)	Cirrhosis 27 kPa (RHL)/25 kPa (LHL) vs controls 5.6/6.5 kPa (P < 0.001). Accuracy: AUROC 0.97 (RHL), 0.98 (LHL), 0.95 (VCTE). Performance: Se 966% (both lobes), Sp 90% (RHL), 96.7% (LHL). Cut-offs: 10.7 kPa (RHL), 14 kPa (LHL). Agreement: Higher with LHL-EUS-SWE (96.6%) vs RHL (93.2%) and VCTE (91.5%)
Diehl et al[36]	Prospective pilot cohort	52 patients undergoing EUS-guided liver biopsy for abnormal liver function tests or fibrosis staging	EUS-SWE (10 measurements per lobe, VsN > 60% reliable); compared with VCTE	Bilobar EUS-guided liver biopsy (histology, Kleiner/METAVIR)	Correlation with histology: RHL SWE r = 0.57 (P < 0.0001), LHL r = 0.37 (P = 0.0079). Accuracy: AUROC advanced fibrosis F3-4: 0.98 (RHL), 0.95 (LHL), 0.99 (VCTE), no significant differences. Cut-offs: RHL 25.5 kPa (Se 60%, Sp 98%), LHL 22.5 kPa (Se 60%, Sp 83%). Reproducibility: Variance 35 × higher in LHL vs RHL; RHL more reliable
Wang et al[37]	Multicenter, cross-sectional (prospective data)	62 obese patients (BMI ≥ 30) with suspected MASLD/MASH undergoing liver biopsy	EUS-SWE (ROI < 2 cm from transducer, 10 measurements, ROI ≤ 1.5 cm; minimal probe pressure)	Liver biopsy (Brunt criteria); subset comparison with VCTE (n = 19) and FIB-4	Stiffness increased with fibrosis stage (Kruskal-Wallis P < 0.0001). Accuracy (AUROC): F ≥ 20.87, F ≥ 30.93, F4 0.95. Cutoffs (90% Se/90% Sp): F ≥ 27.50/9.82 kPa (PPV/NPV 0.74/0.89 at Se-prioritized; 0.91/0.78 at Spprioritized); F ≥ 38.48/10.20 kPa (PPV/NPV 0.73/0.94; 0.80/0.88); F4 11.30/14.60 kPa (PPV/NPV 0.56/1.00; 0.67/0.98). Comparators: EUSSWE superior to FIB4 for F ≥ 2 and F ≥ 3 (AUROC 0.87 vs 0.61, P = 0.0048; 0.93 vs 0.63, P < 0.0001); vs VCTE in reliable subgroup (n = 19): EUSSWE AUROC 0.93/0.89/0.93 (F ≥ 2/F ≥ 3/F4) vs VCTE 0.84/0.61/0.60 with superiority for F ≥ 3 (P = 0.0067) and F4 (P = 0.0022). Feasibility/safety: No adverse events; IQR/M > 30% in 29%, VsN < 50% in 12.9% (accuracy unchanged in Se analyses)
AbiMansour et al[51]	Prospective, single-center	142 patients undergoing clinically indicated EUS; 13 (9.2%) with CSPH	EUS-SWE of spleen (ROI 1-1.5 cm from probe, ≥ 10 reliable readings, VsN > 60%)	Clinical and radiologic criteria for CSPH	Mean spleen stiffness significantly higher in CSPH vs non-CSPH (37.6 ± 8.5 kPa vs 29.1 ± 9.9 kPa; P = 0.003). AUROC 0.74 for CSPH detection; optimal cut-off 37.6 kPa (Se 692%, Sp 79.8%). Spleen stiffness correlated with liver stiffness measured by EUS-SWE (P = 0.001) and MRE (P = 0.003)
Kohli et al[44]	Single-center, prospective tandem study	42 patients undergoing liver biopsy	EUS-SWE (left and right lobes separately) vs VCTE	Liver biopsy (histology, fibrosis staging)	AUROC for advanced fibrosis: VCTE 0.87, EUS-SWE left lobe 080, right lobe 078. AUROC for cirrhosis: VCTE 0.90, EUS-SWE left lobe 096, right lobe 090. VCTE unreliable in 8 patients who all had successful EUS-SWE

EUS: Endoscopic ultrasound; BMI: Body mass index; MASLD: Metabolic dysfunction-associated steatotic liver disease; MASH: Metabolic dysfunction-associated steatohepatitis; CSPH: Clinically significant portal hypertension; SWE: Shear wave elastography; PPG: Portal pressure gradient; ROI: Region of interest; VCTE: Vibration-controlled transient elastography; VsN: Net effective shear wave velocity; FIB-4: Fibrosis-4 index; RHL: Right hepatic lobe; LHL: Left hepatic lobe; AUROC: Area under the receiver operating characteristic; Se: Sensitivity; Sp: Specificity; PPV: Positive predictive value; NPV: Negative predictive value; IQR/M: Interquartile range-to-median ratio; MRE: Magnetic resonance elastography.

Fibrosis evaluation

In a single-center prospective study, Kohli et al[44] compared the diagnostic performance of EUS-SWE and VCTE in distinguishing significant and advanced fibrosis in 42 patients who underwent liver biopsy. EUS-SWE demonstrated comparable accuracy to VCTE in detecting advanced fibrosis, with AUROCs of 0.87 for VCTE, 0.80 for EUS-SWE of the left liver lobe, and 0.78 for the right lobe. Similarly, for cirrhosis, the AUROCs were 0.90 for VCTE, 0.96 for EUS-SWE of the left lobe, and 0.90 for the right lobe.

In line with these findings, Diehl et al[36] reported a strong correlation between EUS-SWE measurements in the right liver lobe and histological fibrosis (ρ = 0.571, P < 0.0001), and a moderate correlation in the left lobe (ρ = 0.368, P = 0.0079). No significant differences were observed in AUROCs between EUS-SWE and VCTE for identifying significant fibrosis. In this context, the authors proposed EUS-SWE cut-off values to distinguish F0-2 from F3-4 fibrosis stages: For the right liver lobe, a threshold of 25.5 kPa yielded a sensitivity of 60%, specificity of 98%, positive predictive value (PPV) of 75%, and negative predictive value (NPV) of 96%; for the left liver lobe, a cut-off of 22.5 kPa resulted in a sensitivity of 60%, specificity of 83%, PPV of 27%, and NPV of 95%[36].

In another single-center cross-sectional study, Del Valle et al[40] compared EUS-SWE values between 29 cirrhotic and 30 non-cirrhotic patients. In the right liver lobe, EUS-SWE values were significantly higher in cirrhotic patients compared to the non-cirrhotic group (27.0 kPa vs 5.6 kPa, P < 0.001). Similarly, in the left liver lobe, EUS-SWE values were also significantly elevated in cirrhotic patients (25.0 kPa vs 6.45 kPa, P < 0.001). When compared to established VCTE cut-offs, EUS-SWE demonstrated superior performance in distinguishing cirrhosis, with AUROC values of 0.971 for right lobe and 0.986 for left lobe, compared to 0.952 for VCTE. The authors proposed the following EUS-SWE cut-offs to differentiate F0-2 from F3-4 stages: For the right lobe, 10.7 kPa (sensitivity 96.6%, specificity 90%, PPV 90.3%, NPV 96.4%); for the left lobe, 14.0 kPa (sensitivity 96.6%, specificity 96.7%, PPV 96.6%, NPV 96.0%)[40].

Notably, EUS-SWE measurements were successfully obtained in all patients in the studies of Del Valle et al[40] and Kohli et al[44], and in over 96% of patients in the study by Diehl et al[36]. In contrast, VCTE produced unreliable results in 8 patients in the Kohli et al’s study[44] and in 3 patients in Diehl et al’s study[36]. These findings suggest that EUS-SWE may play a valuable role in assessing liver stiffness, particularly in patients with obesity and metabolic dysfunction-associated steatotic liver disease (MASLD) or metabolic dysfunction-associated steatohepatitis, in whom percutaneous non-invasive techniques such as transient elastography and transabdominal SWE often face technical limitations.

In a multicenter cross-sectional study with prospectively collected data in 62 obese patients with body mass index ≥ 30 kg/m² with suspected MASLD/metabolic dysfunction-associated steatohepatitis, Wang et al[37] compared EUS-SWE measurements among histological determined liver fibrosis categories. EUS-SWE exhibited increased stiffness values at higher fibrosis stages, with an AUC of 0.87, 0.93 and 0.95 to distinguish significant fibrosis (F0-1 vs F2-4), advanced fibrosis (F0-2 vs F3-4) and cirrhosis (F0-3 vs F4), respectively. The 90% sensitivity cutoffs for EUS-SWE were 7.50 for F2, 8.48 for F3, and 11.30 for F4. In contrast, the 90% specificity cutoffs for EUS-SWE were 9.82 for F2, 10.20 for F3, and 14.60 for F4. Interestingly, EUS-SWE outperformed both fibrosis-4 index and VCTE in predicting significant and advanced fibrosis, although the latter comparison was based on a subgroup analysis of only 19 patients.

Recently, Yousaf et al[45] demonstrated in a prospective single-center study that concomitant EUS-liver elastography, EUS-guided portal pressure gradient (PPG) measurement, and EUS-guided liver biopsy can be performed safely in a single session, with a 100% technical success rate. Both EUS-liver elastography and EUS-PPG correlated with histological fibrosis stage, with significantly higher values in patients with advanced fibrosis (mean PPG 6.0 mmHg vs 3.4 mmHg; liver stiffness 32.0 kPa vs 18.8 kPa). This study highlights the potential of EUS as a comprehensive “one-stop” endohepatology approach.

Portal hypertension

Recent data also suggest that EUS-SWE can be extended beyond liver parenchyma to assess portal hypertension. The rationale for splenic stiffness measurement lies in the pathophysiology of portal hypertension: Increased portal pressure leads to splenic congestion, sinusoidal dilatation, and progressive fibrogenic changes within the spleen, all of which contribute to increased tissue stiffness[46]. Consistently, non-invasive assessment of spleen stiffness has been shown to correlate more closely with portal pressure than liver stiffness, and meta-analyses confirm its utility in predicting clinically significant portal hypertension (CSPH) and varices[47,48]. Reflecting this evidence, the Baveno VII consensus recommends spleen stiffness thresholds to stratify risk: Values ≥ 50 kPa indicate a high probability of CSPH, whereas values ≤ 40 kPa reliably rule out high-risk varices, supporting its integration into non-invasive diagnostic algorithms[49].

In this context, Robles-Medranda et al[50] investigated EUS-elastography of the liver and spleen in 61 patients with cirrhosis and portal hypertension. Using strain-based techniques, they reported AUROC values above 80% for both organs, with spleen strain ratio achieving high sensitivity (87.5%) and NPV (83.3%) and the strain histogram demonstrating high specificity (89.7%) and PPV (85.7%). Moreover, the integration of azygos vein hemodynamic parameters (diameter, velocity, blood-flow index) further improved diagnostic accuracy, particularly in alcohol-related cirrhosis, where sensitivity and NPV reached 100%. In line with these observations, AbiMansour et al[51] prospectively demonstrated in 142 patients that spleen stiffness measured with EUS-SWE was significantly higher in those with CSPH (37.6 kPa vs 29.1 kPa, P = 0.003), with an AUROC of 0.74, supporting the feasibility of incorporating splenic assessment into endohepatology practice.

Feasibility, reliability and reproducibility

EUS-SWE is a highly feasible and well-tolerated technique that adds only a few minutes to endoscopic procedures. In Wang et al’s pilot study[37] on 62 patients with obesity and MASLD, most acquisitions were successful, with unreliable results limited to a minority of cases (approximately 13%). Similar findings were reported by Diehl et al[36], who identified only three unreliable measurements in a cohort of 52 patients. In cirrhotic and control patients, Del Valle et al[40] confirmed excellent applicability, with diagnostic agreement exceeding 93% and AUROC values above 0.95. Reliability, defined by a VsN ≥ 50%-60%, was consistently high across studies. Reproducibility appeared greater in the right hepatic lobe, as Diehl et al[36] observed over threefold higher variability in the left, although diagnostic accuracy remained robust in both sites. Concordance with VCTE was strong, particularly in the left lobe[40], while Wang et al[37] underlined that technical factors such as probe pressure may influence cut-off values. Overall, EUS-SWE emerges as a feasible, reliable, and reproducible tool for liver stiffness assessment, with the right lobe generally preferred for consistency. Standardized acquisition protocols are still needed to reduce variability and define universally applicable cut-offs.

LIMITATION AND FUTURE DIRECTION

Although these studies have demonstrated the potential role of EUS-SWE for liver fibrosis staging, particularly in patients with obesity and MASLD, current evidence remains preliminary. Available hepatic EUS-SWE data are derived mainly from small single-center cohorts and feasibility studies. These investigations report high AUROC values for advanced fibrosis (0.87-0.98) when compared with VCTE or biopsy, but direct comparison with magnetic resonance elastography, the current noninvasive gold standard, are lacking. Variability in ROI positioning, measurement depth, and VsN criteria may influence stiffness values and hinder cross-study comparison. Reproducibility also differs between hepatic lobes, with higher variance reported in left-lobe measurements. These methodological inconsistencies and the lack of large, prospective validation studies highlight the need for protocol harmonization and standardized reporting.

Moreover, formal cost-effective analyses or comprehensive evaluations of accessibility for EUS-SWE are still lacking. Current economic evidence is limited to SWE in other contexts, such as liver or prostate imaging, and to cost studies of EUS in different procedural applications. Consequently, while EUS-SWE demonstrates promising clinical utility, robust health-economic models and real-world accessibility data are still absent, underscoring the need for future research in this area.

Standardizing acquisition protocols represents a crucial step toward improving measurement consistency and facilitating broader clinical implementation. In this regard, an initial effort has been made by establishing a protocol based on benchtop and animal studies using porcine models[52]. Future studies in larger cohorts, ideally with comparisons against established reference standards such as magnetic resonance elastography, with which EUS-SWE has already shown good correlation, will be essential to consolidate its role in clinical practice[27].

CONCLUSION

EUS-SWE has rapidly evolved into a promising adjunct to EUS, offering quantitative and reproducible assessment of tissue stiffness in both pancreatic and hepatic diseases. Current evidence indicates high feasibility and diagnostic performance in CP, potential applications in AIP and solid pancreatic lesions, and strong accuracy in staging liver fibrosis, particularly in obese patients or in those with technically limited transabdominal approaches. Nevertheless, methodological heterogeneity, lack of universally validated cut-offs, and variability in acquisition protocols remain major obstacles to routine clinical use. Standardization of technique and large multicenter prospective studies are urgently needed to define reliable thresholds and consolidate its clinical role. Looking ahead, integration of EUS-SWE into comprehensive endoscopic examinations may not only refine the diagnosis and staging of gastrointestinal diseases but also contribute to precision medicine by enabling real-time, minimally invasive tissue characterization during a single procedure.