Toward noninvasive prediction of treatment outcomes in patients with variceal bleeding

Abstract

We read with great interest the article by Ainora et al. Esophageal variceal bleeding is among the most serious and feared complications of cirrhosis and is responsible for considerable morbidity, high rates of rebleeding and significant mortality. Its onset reflects severe portal hypertension (pHTN) and often signals a turning point in disease progression. Its treatment requires urgent stabilization and effective strategies to prevent recurrence. The standard treatment is endoscopic variceal band ligation (EVBL) and medical treatment, which remains the cornerstone of primary and secondary prophylaxis. However, follow-up continues to rely on repeated endoscopies, which are invasive, resource intensive, and uncomfortable for patients. Although alternative tools are limited, the hepatic venous pressure gradient is the gold standard for assessing pHTN; however, methods for its measurement are invasive and not widely available. Moreover, noninvasive approaches such as elastography or the Baveno criteria, although useful for risk stratification, have not been validated for predicting short-term outcomes after EVBL. The lack of reliable predictors underscores a need for novel approaches. Multiparametric ultrasound (MP-US), which combines measurements of liver and spleen stiffness with perfusion imaging, offers a comprehensive, noninvasive assessment of pHTN. The study by Ainora et al reveals that the use of MP-US can help identify patients most likely to achieve variceal eradication, reducing the number of unnecessary procedures and guiding individualized follow-up. The aim of this editorial is to review the clinical impact of variceal bleeding; to examine the limitations of current follow-up strategies after EVBL; and to analyze the rationale, potential, and challenges related to MP-US as an emerging tool in this context.

Key Words: Esophageal and gastric varices; Liver cirrhosis; Portal hypertension; Endoscopic band ligation; Ultrasonography; Elastography

Core Tip: Esophageal variceal bleeding represents one of the most lethal complications of cirrhosis, with high rebleeding and mortality rates despite standard therapy involving endoscopic variceal band ligation. Current follow-up strategies rely on repeated endoscopy and hepatic venous pressure gradient measurements, which are both invasive and resource intensive, whereas noninvasive tools such as elastography and the Baveno criteria lack validation for predicting post-endoscopic variceal band ligation outcomes. Multiparametric ultrasound, by integrating measurement of liver and spleen stiffness with perfusion imaging, offers a comprehensive, noninvasive assessment of portal hypertension and may enable individualized follow-up, reduce procedural burden, and improve patient-centered care.

INTRODUCTION

Cirrhosis affects more than 112 million people worldwide and is a major cause of chronic liver-related morbidity and mortality[1]. One of its main hemodynamic consequences is the development of portal hypertension (pHTN), which leads to the formation of gastroesophageal varices during the course of the disease in almost 50% of affected patients. pHTN can be defined simply as abnormal venous pressure elevation in the portal system[2]. Portal vein pressure normally ranges between 7 and 12 mmHg at rest. The pathophysiology of pHTN is a complex process that involves multiple mechanisms within the liver, specifically the dysfunction of liver sinusoidal endothelial cells and hepatic stellate cells[3]. Imaging studies have revealed scan features that indicate pHTN, whereas liver stiffness and portal pressure values are usually normal or slightly elevated[4]. New noninvasive and imaging techniques can assess pHTN and predict the presence or severity of varices without a need for routine endoscopic screening. These noninvasive tests aim to estimate portal pressure, evaluate liver stiffness, and detect collateral circulation, helping clinicians identify patients at the highest risk of variceal development or bleeding[5].

Esophageal variceal bleeding (EVB) is a major clinical complication of cirrhosis, marking the transition from compensated to decompensated pHTN and reflecting an advanced disease severity. Variceal bleeding occurs at an annual rate of 5% to 15%, depending on the size of the varices and the clinical severity[6]. When bleeding occurs, short-term mortality remains high, ranging from 15% to 20% within six weeks, and these episodes accelerate hepatic decompensation[7,8], increase the likelihood of recurrent bleeding, and indicate progression to more advanced stages of liver disease that are characterized by higher morbidity and lower survival rates[9].

The pathophysiology of variceal rupture is closely related to the magnitude of pHTN, which is defined by a hepatic venous pressure gradient (HVPG) greater than 10-12 mmHg and is modulated by local changes in the vascular walls, systemic hyperdynamic circulation, and altered hemostasis[10]. This explains why variceal bleeding is not just a complication but rather a sentinel event reflecting underlying pHTN severity and systemic consequences of cirrhosis[11]. Importantly, recurrent bleeding episodes remain common, with rebleeding rates in the first two years reaching 60% in the absence of adequate secondary prophylaxis, underscoring the chronic burden of this condition[12].

Endoscopic variceal band ligation (EVBL) has become the standard procedure for both acute hemostasis and secondary prevention, as it offers superior efficacy to sclerotherapy and forms the basis of current guideline-based treatment[6]. Even when technical success is achieved, EVBL rarely achieves complete eradication in a single session and requires repeated procedures every 2-4 weeks to ensure destruction of the bleeding varices and prevent recurrence[13]. This paradigm imposes a significant clinical and economic burden: Surveillance endoscopy is invasive, resource intensive, and distressing for patients; however, incomplete elimination leaves patients vulnerable to early rebleeding and death[1,8]. Despite advances in risk stratification using elastography, platelet-based criteria, and evolving predictive models, there remains no widely validated noninvasive method for predicting the short-term outcomes of EVBL[14]. As a result, physicians continue to rely on endoscopic reassessment, which, though effective, does not comply with the principles of patient-centered and resource-efficient care. In this context, reliable and noninvasive predictors of EVBL efficacy are clearly needed.

Emerging evidence has suggested that novel imaging strategies, such as multiparametric ultrasound (MP-US), may bridge this gap by capturing both structural and hemodynamic determinants of pHTN, offering a holistic assessment of treatment response[9]. These tools not only minimize unnecessary repeat endoscopies but also enable individualized follow-up strategies, optimize the allocation of healthcare resources, and, most importantly, improve patient safety and quality of life[6]. This editorial aims to critically appraise the emerging landscape of noninvasive diagnostic innovation related to pHTN, focusing on the potential of MP-US to reshape post-EVBL surveillance and improve individualized management in patients with cirrhosis.

THE LIMITS OF THE CURRENT GOLD STANDARD

The current gold standard for the stratification and treatment of variceal bleeding is based on a combination of serial endoscopic evaluations and HVPG measurements. This measurement is used to assess pHTN severity in patients with cirrhosis. The procedure involves inserting a catheter into a hepatic vein, usually through the right internal jugular vein, and is performed under local anesthesia and fluoroscopic guidance. A balloon catheter is inserted into a large hepatic vein. The catheter is then used to occlude the vein, either by advancing it into a small branch or by inflating the balloon, which blocks outflow and allows measurement of wedged hepatic venous pressure[15]. However, each approach has notable limitations. HVPG measurement, while valuable from a prognostic standpoint, is invasive, resource intensive, and generally limited to specialized centers, which restricts its applicability in public or resource-limited healthcare settings. Furthermore, for the HVPG to be clinically meaningful, two measurements must be taken to assess the therapeutic response, further compounding its logistical and practical complexity. Its use is also limited by the need for specialized technical expertise and careful patient selection. However, in recent studies, HVPG assessment has been restricted to high-risk patients in an attempt to optimize its application, although these strategies do not address the fundamental drawback of the procedure’s invasiveness[16].

Serial endoscopy, including esophagogastroduodenoscopy, is a fundamental component in the treatment of EVB, both for acute intervention and secondary prophylaxis. An urgent endoscopic evaluation (within 12 hours of presentation) is recommended to confirm the source of bleeding and initiate endoscopic treatment (primarily EVBL)[15]. However, serial endoscopy a major burden in terms of resources; repeated procedures require specialized equipment, qualified personnel, and recovery facilities, which translates to high costs and resource consumption, especially in resource-constrained environments[17]. It also has procedural risks, such as complications related to sedation, while gastrointestinal risks include bleeding and, in rarer cases, perforation[18]. Beyond clinical complications, the procedure itself is associated with considerable discomfort. Endoscopic evaluation is also subject to variability among technical personnel, mainly regarding the estimation of the variceal diameter, which plays an important role as a predictor of bleeding risk and treatment outcome. Recent studies have demonstrated discrepancies between visual and objective measurements of variceal size, highlighting the limitations of subjective endoscopic evaluation and its ability to predict outcomes after ligation[19].

Considering the limitations of these invasive modalities, increasing attention has been given to the development of innovative, noninvasive diagnostic strategies aimed at improving risk stratification and reducing the need for repeated endoscopic procedures[20,21]. Several diagnostic methods have been developed for patients with EVB, including invasive, noninvasive, and other innovative methods, to identify patients at low risk for high-risk varices and to reduce the need for repeated endoscopies (Figure 1). These include transient elastography, spleen stiffness measurement (SSM), and the Baveno VI criteria (liver stiffness < 20 kPa and platelet count > 150000)[21]. These criteria have been shown to have high negative predictive value for ruling out high-risk varices and may prevent the need for endoscopy in a small proportion of patients with advanced compensated chronic liver disease (CLD). Despite the increasing use of noninvasive tools to predict the risk of variceal bleeding, their usefulness in ensuring the reliable prediction of outcomes after EVBL remains limited[22]. A study conducted by Protopapas et al[23] revealed that the Baveno VI and VII criteria can be used to rule out varices requiring treatment in patients with CLD; however, these criteria do not allow for the prediction of postoperative bleeding or patient outcomes after EVBL. Furthermore, analyses have revealed that parameters such as spleen stiffness, which is measured by shear wave elastography, correlate with baseline risk but do not replace risk stratification after EVBL. Although tools such as transient elastography, acoustic radiation force imaging, and two-dimensional shear wave elastography have a high area under the curve for the detection of varices (r > 0.9 in some meta-analyses), their sensitivity is drastically reduced when attempting to predict new bleeding or other complications after ligation[24,25]. Furthermore, these noninvasive tools may not be feasible in resource-limited settings and require additional prospective validation for risk prediction beyond initial screening.

Figure 1 Diagnostic approaches for esophageal varices in patients with cirrhosis: From invasive standards to emerging noninvasive tools. The figure summarizes the spectrum of available diagnostic tools for the assessment of esophageal varices and portal hypertension in patients with cirrhosis, ranging from traditional invasive standards to novel noninvasive strategies that are currently under development. Invasive methods, such as upper gastrointestinal endoscopy, hepatic venous pressure gradient measurement, and endoscopic ultrasound, remain the cornerstone for diagnosis and hemodynamic characterization but are resource intensive and associated with patient discomfort. Noninvasive methods, including elastography-based techniques, the Baveno VI-VII consensus criteria, Doppler ultrasound, and combined clinical/Laboratory indices, provide valuable surrogate markers of tissue stiffness, hemodynamics, and risk stratification, enabling a reduction in the number of unnecessary endoscopic examinations for selected patients. Nonetheless, they remain affected by operator and equipment variability and have a limited ability to predict posttreatment outcomes. Emerging approaches such as multiparametric ultrasound, artificial intelligence, and machine learning-based predictive models, as well as advanced cross-sectional imaging with computed tomography/magnetic resonance imaging computational modeling, represent promising strategies to achieve comprehensive, noninvasive evaluation of portal hypertension and the risk of variceal bleeding. These approaches offer multiparametric integration and potential improvements in precision but currently face challenges related to cost, technical expertise, and lack of standardized validation. Together, these modalities reflect an ongoing transition toward individualized, less invasive, and more precise management of cirrhotic patients. CT: Computed tomography; MRI: Magnetic resonance imaging; TE: Transient elastography; 2D-SWE: Two-dimensional shear wave elastography; ARFI: Acoustic radiation force impulse; SSM: Spleen stiffness measurement; PSR: Platelet-to-spleen ratio.

MP-US: BEYOND STIFFNESS

Approximately 2 million people worldwide die each year from liver disease. In patients with CLD, changes in liver morphology and tissue properties can be detected using imaging techniques. Hence, access to effective and accurate imaging is important for patients. MP-US is a new technique for comprehensive noninvasive diagnosis, as it allows a diagnosis to be made in a single scan, providing physicians worldwide with a broader perspective and constituting a key element among their diagnostic tools[26]. MP-US incorporates three main modalities for assessing pHTN. Liver stiffness measurement (LSM) quantifies tissue stiffness and serves as an indicator of liver fibrosis and portal pressure. SSM is strongly correlated with the HVPG and the presence of clinically significant pHTN. Dynamic contrast-enhanced US facilitates the assessment of perfusion within the liver parenchyma and portal vein and captures hemodynamic alterations after interventions such as EVBL. Vidili et al[27] demonstrated the predictive value of US for esophageal varices. In a prospective study conducted by Ainora et al[28], 41 patients underwent EVBL, and 68.3% achieved variceal eradication (VE) within one month. While innovative, this study must be interpreted within the context of its limitations, including its small sample size, single-center design, and potential selection bias. Patients with VE had lower LSM and SSM values and a significant reduction in SSM after ligation. In addition, these patients showed increases in dynamic perfusion parameters, including peak enhancement, area under the curve, and washout rate in both the hepatic parenchyma and portal vein. Collectively, these results indicate that noninvasive US markers accurately reflect the local hemodynamic response to treatment and may facilitate early identification of patients requiring retreatment. Using US to monitor treatment outcomes, rather than repeated endoscopic procedures, may reduce patient discomfort, procedural risks, and healthcare costs. An additional potential advantage of MP-US is its application in personalized risk stratification. Baseline LSM and SSM values are consistently associated with the presence and severity of pHTN, as reported by Jain et al[29].

Variceal bleeding is a life-threatening complication seen in patients with CLD and pHTN and is among the main factors contributing to the high morbidity and mortality observed in these patients[30]. It often requires repeated invasive procedures to minimize the risk of bleeding and may not be cost effective[31]. These complications and risks could be reduced if patients did not need to undergo such invasive procedures after EVBL treatment to check whether the condition had been resolved. The authors set out to identify noninvasive predictors of VE after EVBL treatment using MP-US. The secondary objective was to develop a model for predicting the success of VE on the basis of noninvasive parameters. Sixty-one patients were enrolled, and only 41 met the inclusion criteria. Enrolled patients underwent MP-US with Doppler, LSM, SSM, and dynamic contrast-enhanced-US immediately before endoscopy and 4 weeks after treatment to assess treatment efficacy or need for additional EVBL.

The clinical significance of the findings by Ainora et al[28] remains uncertain despite the reporting of significant associations between SSM and postligation outcomes, including correlations with hemodynamic parameters. HVPG measurement is widely recognized as the gold standard for direct measurement of portal pressure under controlled hemodynamic settings, whereas SSM represents a composite of multiple elements, such as fibrosis, congestion, splenic blood flow, and tissue elasticity[15,32]. Moreover, HVPG continues to be the reference standard for assessing clinically significant pHTN and evaluating therapeutic response, particularly because its prognostic thresholds are well validated across diverse populations[15]. Consequently, although the correlations found in this small cohort confirm the physiological validity of MP-US as a monitoring tool, they are not sufficient evidence to establish SSM as a stand-alone surrogate or to replace HVPG. To determine whether SSM may achieve the repeatability, precision, and outcome-predictive accuracy required to serve as a noninvasive substitute in everyday practice, larger trials are necessary.

When medical and endoscopic treatments fail, portosystemic shunts become necessary to decompress the portal system[33]. Compared with nonselective shunts, selective portocaval shunts are valuable because they decompress the portal system while maintaining some portal flow, helping to reduce the risk of hepatic encephalopathy. These shunts may be monitored using noninvasive Doppler US, which can assess shunt patency and blood flow characteristics. In these scenarios, MP-US serves as a valuable tool for achieving a thorough and standardized patient assessment, substantially limiting potential gaps or ambiguities in clinical interpretation. Other interventional options, such as TIPS, effectively control recurrent variceal bleeding and ascites but have risks, including hepatic encephalopathy[34]. Endoscopic injection sclerotherapy and EVL are established treatments for esophageal varices. Red dichromatic imaging, a newer image-enhanced modality, improves the visualization of deeper mucosal and submucosal vessels compared with that of white light imaging[35]. Within this landscape, MP-US offers a complementary advantage by enabling comprehensive, real-time assessment of vascular structures and procedural anatomy, thereby reducing diagnostic uncertainty and potentially enhancing the safety and precision of these interventions.

Traditionally, the diagnosis of liver diseases is based on the evaluation of liver histology, and liver biopsy is considered the gold standard method for identifying and classifying hepatocellular lesions. However, this procedure has notable limitations, including sampling errors, observer variability, low patient acceptability (especially during follow-up), and the risk of complications such as bleeding or even death[36]. The use of MP-US for the noninvasive assessment of EVBL outcomes in patients with liver cirrhosis is an excellent way to evaluate other diagnostic methods to prevent the need for follow-up endoscopy as a reference method[37]. At present, MP-US should be regarded as a diagnostic adjunct rather than a final diagnostic tool. The accuracy of stiffness and perfusion measurements can be compromised by several physiological and technical parameters, such as respiratory motion, obesity, ascites, and acoustic shadowing. These restrictions have been extensively documented in studies on elastography-based methods[38]. Opting for MP-US is also a way to reduce costs and minimize the need for frequent invasive endoscopic procedures while continuing to indirectly monitor pHTN and variceal recurrence. However, despite its promising potential, MP-US is still limited by technical barriers; factors such as interequipment variability, operator dependence, patient agitation, and hemodynamic instability or inability; and the absence of standardized acquisition and interpretation protocols, which may compromise reproducibility and diagnostic performance. Accessibility also remains an obstacle, as advanced MP-US platforms and contrast-enhanced US are not universally available, particularly in resource-limited hospitals[39]. These variables highlight the need for cautious clinical interpretation and the importance of integrating MP-US results with thorough clinical judgment as opposed to relying solely on them.

BRIDGING INNOVATION AND CLINICAL PRACTICE

Although research on EVBL has advanced considerably, several challenges continue to limit the clinical applicability and generalizability of the current findings. Many studies are limited by small sample sizes, which reduce statistical power and increase susceptibility to bias. In addition, most research consists of single-center studies that may not reflect the heterogeneity of clinical practices, patient populations, and procedural expertise across institutions. A notable exception is a multicenter retrospective study by Jung et al[40], in which the authors analyzed data from patients who underwent EVBL for acute variceal bleeding. This design allowed for a larger and more diverse cohort, improving the representativeness and robustness of the conclusions.

In addition, a recent multicenter Japanese study proposed a new scoring system, the Hospital Outcome Prediction after Endoscopic Variceal Ligation, which is a simple five-parameter tool (systolic blood pressure, Glasgow Coma Scale score, bilirubin level, creatinine level, and albumin level) that accurately predicts in-hospital mortality after EVBL. This model illustrates the significance of larger multicenter cohorts and consistent techniques for the design of prognostic tools for variceal hemorrhage, as demonstrated by its superior discriminatory performance compared to the model for end-stage liver disease and Child-Pugh scores[41]. While Hospital Outcome Prediction after Endoscopic Variceal Ligation provides a biochemical and clinical risk profile, MP-US offers hemodynamic and structural information; therefore, these approaches are not competitive but rather are complementary, and their combined use could yield a more integrated and precise assessment of patients undergoing EVBL.

The need for multidisciplinary recognition and technical standardization has been emphasized by multicenter retrospective studies, which have demonstrated the need for a diverse patient population and various clinical settings to verify the value and reliability of innovative tools such as virtual ruler-based diameter measurement[19]. Similarly, a retrospective study by Cao et al[42] employed an artificial intelligence-based endoscopic virtual ruler and revealed that, compared with conventional visual grading systems, the variceal diameter measurement was a noteworthy predictor of early rebleeding. By using endoscopic virtual ruler, researchers reported a threshold of 1.4 cm, indicating that the risk of rebleeding within six weeks increased considerably (area under the curve 0.85). The findings of this study reinforce the need for accessible and reproducible tools that can enhance risk stratification after EVBL and inform decisions about surveillance intensity or adjunctive interventions.

CONCLUSION

EVB remains a sentinel complication of cirrhosis, reflecting advanced pHTN and carrying substantial risks of morbidity, recurrent hemorrhage, and mortality despite recent therapeutic advances. EVBL continues to be the standard of care, yet its reliance on repeated, invasive procedures highlights an urgent need for alternative strategies to guide follow-up and predict outcomes. Traditional approaches, including serial endoscopy and HVPG measurement, are limited by their invasiveness, lack of availability, and patient burden, and existing noninvasive methods such as elastography or the Baveno criteria have not demonstrated sufficient accuracy in predicting post-EVBL prognosis. In this context, MP-US has emerged as a promising innovative technique that integrates stiffness measurements and perfusion imaging to capture both the structural and the hemodynamic dimensions of pHTN. Early evidence has indicated that MP-US may identify patients who are more likely to achieve VE, reduce unnecessary procedures, and support individualized surveillance strategies. Moving forward, validation through large, multicenter cohorts and standardized protocols will be essential to confirm its clinical applicability, but its potential to transform current paradigms of care underscores a critical step toward more patient-centered, noninvasive, and cost-effective management of variceal bleeding in patients with cirrhosis.