Noninvasive multimodal management of portal hypertension: from early screening to precision clinical decision-making

Introduction

Portal hypertension (PH) represents the defining hemodynamic complication of progressive chronic liver disease, primarily arising from a combination of increased intrahepatic vascular resistance and a subsequent hyperdynamic circulatory state (1). It is a primary driver of life-threatening complications, including gastroesophageal variceal bleeding, ascites, and hepatic encephalopathy, and remains a leading cause of morbidity and mortality in cirrhosis worldwide. Recent epidemiological data underscore the growing public health crisis, with global cirrhosis-related deaths to the liver rising from 1.02 million in 1990 to approximately 1.43 million in 2021 (2). This escalating burden necessitates a refined approach to early detection and long-term management to mitigate the transition from compensated to decompensated disease.

For decades, the measurement of the hepatic venous pressure gradient (HVPG) has served as the reference standard for diagnosing clinically significant portal hypertension (CSPH), defined by a threshold of 10 mmHg or more (3). However, the clinical utility of HVPG is significantly constrained by its invasive nature, requirement for specialized radiological expertise, high procedural costs, and potential risks such as bleeding and infection (4). These limitations preclude its routine use in primary care and limit its utility for the longitudinal, repeated measurements required for chronic disease management (4). In response, a major paradigm shift toward “noninvasive risk-modulated care” has been championed by recent international consensus statements, including the Baveno VII consensus and the American Association for the Study of Liver Diseases (AASLD) recommendations (3,5). These frameworks prioritize noninvasive tests (NITs) to identify patients at high risk of decompensation while sparing low-risk individuals from unnecessary invasive procedures.

Despite the rapid proliferation of NITs—ranging from simple serum biomarkers to advanced elastography—several critical research gaps and translational barriers remain. Current methodologies often struggle with “diagnostic grey zones,” where results are neither sufficiently high to confirm CSPH nor low enough to exclude it (6). Furthermore, the diagnostic accuracy of established thresholds is frequently challenged by the shifting etiology of liver disease, particularly the global rise in metabolic dysfunction-associated steatotic liver disease (MASLD), in which obesity and steatosis can confound stiffness measurements (7). There is also a lack of standardized protocols for integrating these diverse tools into a cohesive, stepwise clinical pathway that spans the entire disease continuum, from early screening to post-treatment surveillance.

To address these challenges, this review provides a comprehensive synthesis of the current noninvasive measures for PH evaluation. We critically evaluate the diagnostic performance and limitations of serum biomarkers, elastography, sophisticated hemodynamic imaging, and emerging artificial intelligence (AI)-based models. Beyond a simple survey of tools, we propose a “noninvasive-first, multimodal integration” framework designed to facilitate closed-loop clinical decision-making. By identifying practical application scenarios and outlining a tiered management pathway, this review aims to bridge the gap between theoretical research and evidence-based clinical practice, ultimately fostering a more precise and accessible strategy for PH management.

Serum-based biomarkers

Serum-based biomarkers provide a cost-effective, accessible, noninvasive evaluation of PH. These surrogates capture the essential pathophysiological drivers of PH, including hepatic fibrosis, hypersplenism-induced thrombocytopenia, and systemic endothelial activation. Due to their simplicity and high reproducibility, they are particularly suited for early screening and longitudinal monitoring. Among these, the von Willebrand factor-to-platelet ratio (VITRO) score—a composite index of endothelial dysfunction and splenic sequestration—has emerged as a robust tool for risk stratification and clinical decision-making.

Conventional serum-based scores: aspartate aminotransferase to platelet count ratio index (APRI) and fibrosis-4 (FIB-4)

The APRI and the FIB-4 index, which consist of age, platelet count, aspartate aminotransferase, and alanine aminotransferase measurement, are the most widely implemented first-line tests for initial risk assessment in non-specialist and primary care settings (8). A single-center retrospective cohort study showed a moderate correlation between APRI and HVPG (r=0.45). An APRI cutoff of 0.876 identifies severe PH (HVPG >12 mmHg) with a diagnostic accuracy of 73%, sensitivity of 71%, and a high positive predictive value (PPV) of 94% (9). A meta-analysis of over 100,000 patients demonstrated that, while these scores have limited “rule-in” specificity for definitive diagnosis, they possess excellent negative predictive value (NPV) (10). Specifically, an FIB-4 score <1.45 effectively excludes clinically significant PH (10). Consequently, AASLD guidelines recommend these indices for initial fibrosis screening and risk stratification (11). The specificity of APRI and FIB-4 is frequently compromised by splenic sequestration, hepatic necroinflammation, and systemic inflammatory states. These markers are less sensitive to early or mild PH. They are best utilized for initial exclusion rather than definitive diagnosis or treatment choice, especially in patients with active inflammation, severe hypersplenism, or decompensated disease. Notably, in routine practice, FIB-4 and APRI are validated for initial screening of liver fibrosis rather than as direct predictors of PH. They are rarely used alone to assess HVPG or CSPH. Instead, they serve as accessible, frontline gatekeepers to identify at-risk advanced chronic liver disease populations. Suspected patients must then be triaged to specific PH surrogates for accurate hemodynamic characterization.

Von Willebrand factor (vWF) and the VITRO score: from endothelial dysfunction to precision assessment

Diagnostic and prognostic performance

vWF is a well-established indicator of endothelial activation and the hyperdynamic circulatory state central to PH pathogenesis. The VITRO score (vWF-to-platelet ratio) integrates endothelial injury with PH-driven hypersplenism, significantly enhancing the diagnostic accuracy for CSPH (12).

vWF levels correlate strongly with CSPH, though its specificity can be reduced by systemic infection, acute inflammation, or advanced decompensation (12,13). Therefore, vWF-based indices are more appropriately interpreted as markers of risk rather than as standalone diagnostic tools. In patients with compensated advanced chronic liver disease (cACLD), vWF-based models predict hepatic decompensation and liver-related mortality with accuracy comparable to invasive HVPG (14,15). The role of vWF in acute-on-chronic liver failure (ACLF) remains complex. While vWF levels are markedly elevated in ACLF and assist in estimating short-term mortality, they have not consistently emerged as an independent predictor of ACLF onset or liver-related death in decompensated patients (14,16). Consequently, vWF may be more effective when integrated into composite models that include biomarkers such as ADAMTS13 and inflammatory indices, thereby enhancing predictive accuracy for ACLF prognosis.

The prognostic utility of the VITRO score often exceeds its diagnostic performance for identifying CSPH. Evidence from several head-to-head cohort studies indicates that VITRO is reliable and agrees with invasive HVPG measurements in predicting clinical outcomes. Specifically, among patients with cACLD, elevated VITRO values correlate positively with the risk of future hepatic decompensation and liver-related mortality. Furthermore, VITRO values increase significantly with worsening Child-Pugh class, indicating its ability to reflect overall disease severity (17). Given its strong correlation with long-term prognosis, the VITRO score is uniquely suited to longitudinal follow-up and dynamic monitoring of the disease trajectory.

Clinical application strategies: the sequential approach

To optimize the diagnostic utility of noninvasive tools, a sequential strategy combining Baveno VII criteria with VITRO-based stratification has been proposed to address the inherent limitation of using liver stiffness and platelet count alone. While Baveno VII is highly effective for screening for high-risk varices, it frequently leaves up to 45.7% of patients in a “diagnostic grey zone” in which CSPH cannot be definitively excluded. By applying the VITRO score as a secondary stratifier for those failing to meet the Baveno VII “low-risk” criteria, diagnostic uncertainty was reduced to less than 15% (18). This two-step approach achieves a PPV exceeding 90% for confirming CSPH, thereby significantly minimizing unnecessary endoscopic procedures.

In resource-constrained environments where elastography is unavailable, a similar tiered strategy using FIB-4 for initial screening followed by VITRO offers a practical alternative. Furthermore, the repeatability of the VITRO score makes it an ideal tool for longitudinal follow-up, enabling clinicians to monitor risk fluctuations and assess therapeutic responses to etiological treatment or nonselective beta-blockers. However, practitioners should exercise caution in the context of severe infection or ACLF; in such cases, non-hepatic endothelial activation may cause false-positive elevations, necessitating correlation with spleen stiffness measurements (SSMs) or imaging-based assessments.

Other emerging serum biomarkers

Beyond established indices, several novel biomarkers offer complementary insights into the pathological drivers of PH, including inflammatory activation, hepatocellular injury, extracellular matrix remodeling, and gut barrier dysfunction. Golgi protein 73 (GP73), a marker of hepatocellular injury and fibrosis, demonstrates moderate diagnostic accuracy for CSPH, with an area under the curve (AUC) of approximately 0.75. While its performance is enhanced when integrated into the IP73 score (combining GP73, platelet, and international normalized ratio; AUC =0.85) (19), its clinical utility is hampered by poor disease specificity, with frequent false positives in cases of active hepatitis or hepatocellular carcinoma (20). Similarly, soluble CD163 (sCD163), a marker of Kupffer cell activation, correlates positively with HVPG and reflects systemic inflammation linked to disease progression (21-24). However, the absence of standardized assays and validated cutoffs currently restricts its routine clinical application.

Other markers target specific niches of PH pathophysiology but remain in the exploratory phase. For instance, type V procollagen (Pro-C5) reflects extracellular matrix remodeling and has been associated with elevated portal pressure in alcoholic cirrhosis, though these findings lack generalizability across broader etiologies (25). Indicators of intestinal barrier dysfunction, such as intestinal fatty acid-binding protein (I-FABP), have been linked to an increased risk of gastroesophageal varices in PH, yet they await validation in large multicenter cohorts (26). Furthermore, while indocyanine green retention at 15 minutes (ICG-r15) assists in identifying advanced PH by reflecting hepatic functional reserve, its lack of specificity to portal pressure necessitates its use as a functional adjunct rather than as a standalone marker (27).

In summary, while these emerging biomarkers provide valuable mechanistic data, their clinical accessibility and evidence base remain limited. A comparative summary of their underlying mechanisms and diagnostic or prognostic value is provided in Table 1. Current risk stratification protocols should prioritize the VITRO score, using markers such as sCD163 as adjunctive indicators, while reserving more specialized indices for research settings or for composite predictive models.

Imaging-based noninvasive assessment

While transjugular HVPG remains the gold standard, endoscopic ultrasound (EUS)-guided HVPG measurement has emerged as a novel, minimally invasive alternative (28). Using a fine-gauge needle and a digital transducer under real-time Doppler guidance, EUS enables direct vascular access. This approach is highly advantageous in specific scenarios, such as patients with severe anatomic distortions, failed transjugular access, or those undergoing concurrent endoscopic variceal screening, providing direct pressure readings where traditional methods carry high clinical risks (29). Imaging-based techniques evaluate PH by quantifying tissue stiffness and hemodynamic changes, providing a direct biophysical measure of intrahepatic resistance and portal pressure. These methods have moved to the forefront of clinical practice, serving as indispensable tools for risk stratification, complex diagnosis, and longitudinal therapeutic monitoring.

Elastography techniques: the standard of care

Unlike serum biomarkers, elastography provides precise anatomical localization, which is essential for noninvasive assessment from initial screening to precision risk stratification. The choice between transient elastography (TE), shear wave elastography (SWE), and magnetic resonance elastography (MRE) is typically dictated by diagnostic performance, applicability, and resource availability (Table 2).

TE

TE is the most widely adopted modality due to its procedural simplicity and cost-effectiveness. It demonstrates a strong correlation with HVPG and reliable diagnostic performance for CSPH (32). According to the Baveno VII consensus, a liver stiffness measurement (LSM) ≤15 kPa combined with a platelet count ≥150×10⁹/L effectively rules out CSPH, while an LSM ≥25 kPa is sufficient for diagnosis (3). However, the precision of TE is limited in patients with obesity, ascites, hepatic inflammation, or significant steatosis. In MASLD, technical failure rates remain high, and the PPV of the 25 kPa cutoff is diminished, suggesting that models integrating body mass index (BMI) and platelet count may offer superior diagnostic accuracy in these populations (30,33). Beyond static diagnosis, longitudinal changes in LSM provide critical prognostic significance. In patients with cACLD, including those with alcohol-related etiologies, a ≥20% increase in LSM is associated with a higher risk of hepatic decompensation and mortality, whereas a reduction to <20 kPa is associated with improved outcomes (34,35). These findings support the use of serial TE measurements in routine clinical practice for dynamic risk monitoring.

SWE

SWE offers a distinct advantage in patients with decompensated cirrhosis, as it is not affected by ascites. Meta-analyses indicate that two-dimensional (2D) SWE achieves diagnostic performance comparable to TE, with a liver stiffness cutoff of 15.2 kPa demonstrating high specificity and PPV for confirming CSPH (31,36,37). Furthermore, integrating SWE with clinical scores, such as the MELD score, has been shown to predict survival in advanced chronic liver disease (38). Despite these strengths, SWE faces technical constraints in obese patients due to reduced tissue penetration, though the use of deep abdominal transducers has significantly improved technical success rates in this population (33). Currently, the widespread clinical adoption of SWE is limited by higher equipment costs and greater operator dependence than TE. Consequently, while SWE is a powerful tool for specialized assessment, large multicenter studies are warranted to validate it further.

MRE

MRE provides superior diagnostic accuracy for advanced fibrosis and cirrhosis, consistently outperforming ultrasound-based methods (39,40).

A key advantage of MRE is its resilience to confounding factors; measurements remain unaffected by obesity or hepatic steatosis, making it the preferred tool for risk stratification in MASLD populations, particularly for identifying high-risk gastroesophageal varices (40,41). MRE is highly effective in predicting hepatic decompensation and mortality across various etiologies, even in patients with early or minimal ascites. Furthermore, longitudinal increases in MRE-derived liver stiffness serve as a critical prognostic indicator, correlating strongly with decompensation risk, hepatocellular carcinoma, and other liver-related events (42). While MRE can be cost-effective when targeted toward high-risk groups (e.g., those with elevated FIB-4), its widespread implementation in primary care is currently hindered by high costs, limited availability, and prolonged examination times (43). Additionally, its relatively low sensitivity for detecting early-stage fibrosis suggests it is best reserved for the management of advanced chronic liver disease.

Pathophysiologically, liver stiffness primarily reflects structural intrahepatic resistance and correlates strongly with portal pressure within the lower HVPG range. As collateral circulation develops and hyperdynamic circulation becomes dominant, this relationship becomes non-linear, which explains the reduced sensitivity of LSM at higher pressure levels. Consequently, a scenario-based approach to elastography selection is recommended. TE is the preferred first-line tool for routine screening in primary care. SWE is particularly indicated for patients with ascites, where TE accuracy is compromised. MRE should be reserved for obese patients with MASLD or for cases in which ultrasound-based techniques yield inconclusive results.

SSM: a hemodynamic proxy

SSM provides a more direct assessment of portal hemodynamics and splenic venous congestion than liver stiffness does, as it is largely unaffected by hepatic inflammation or steatosis. Notably, SSM shows a relatively strong linear correlation with HVPG, making it an important complement to LSM in optimizing the noninvasive diagnosis of PH (44). The Baveno VII consensus established clinically relevant cutoff values for SSM, defining thresholds of <21 kPa for excluding CSPH and >50 kPa for diagnosing CSPH (specifically using a 100 Hz probe, which yields a PPV of 98% (43,45,46). In addition, SSM ≤40 kPa safely rules out high-risk varices, thereby reducing unnecessary endoscopic examinations (3). In clinical practice, SSM serves as a powerful adjunct to the Baveno VII criteria, particularly for resolving diagnostic uncertainty in patients with indeterminate LSM findings or suspected high-grade PH. However, because SSM requires specialized equipment and trained operators, and cutoff values may differ across technical platforms, results must be considered alongside LSM, platelet count, and the broader clinical context.

Emerging imaging techniques for precision assessment

Emerging imaging technologies are shifting the diagnostic paradigm from static stiffness measurements toward dynamic hemodynamic modeling, offering high-precision alternatives for complex clinical scenarios.

Subharmonic-aided pressure estimation (SHAPE)

SHAPE is a novel contrast-enhanced ultrasound technique that estimates portal pressure from the nonlinear oscillations of microbubbles in an acoustic field. SHAPE correlates more closely with portal hemodynamics than conventional elastography. Using a cutoff value of −3.5 dB, it has demonstrated an AUC of 0.92, a sensitivity of 80.0%, and a specificity of 89.0% for identifying high-risk varices. It outperforms both LSM and SSM, with an NPV of 88% (47,48). Despite promising results (AUC 0.89–0.93 for CSPH), current evidence is limited to small cohorts and lacks standardized protocols (49). Consequently, SHAPE is presently reserved for research or specialized assessment in tertiary centers.

Multiparametric magnetic resonance (mpMRI)

Advanced mpMRI parameters, integrating 2D and three-dimensional (3D) MRE with dynamic contrast-enhanced (DCE) MRI, provide a comprehensive evaluation of PH (50). mpMRI parameters, particularly the combination of MRE-based spleen stiffness and hepatic uptake rates, have been shown to outperform bedside SWE for diagnosing CSPH and predicting hepatic decompensation (50). This multimodal approach offers a superior tool for precise assessment of PH and prognostic stratification in patients with cirrhosis.

Computed tomography (CT)-based vascular radiomics

CT-based vascular radiomics combined with computational fluid dynamics (CFD) modeling enables virtual estimation of HVPG, with reported accuracy reaching 94%. While highly accurate for detailed anatomical and hemodynamic mapping, the associated radiation exposure makes this approach less suitable for the longitudinal or repeated monitoring required in chronic care (51).

In summary, emerging imaging techniques such as SHAPE, mpMRI, and CT-based vascular radiomics are driving PH management toward a precision-oriented paradigm. By enabling high-resolution hemodynamic modeling, these technologies lay the foundation for future integration with AI and digital twin technologies, moving beyond simple risk stratification toward personalized clinical decision-making.

International guideline recommendations

Current international guidelines provide complementary frameworks for the noninvasive assessment of PH, reflecting differences in clinical priorities, evidentiary standards, and target patient populations (Table 3) The Baveno VII Consensus prioritizes clinical viability and the “noninvasive-first” paradigm. It establishes the combination of LSM and platelet count as the foundation for risk stratification. Notably, Baveno VII is the first to officially advocate for SSM to resolve diagnostic uncertainty and refine high-risk varices. On the other hand, the AASLD guidelines adopt a more conservative evidentiary stance, emphasizing the use of noninvasive methods primarily for risk stratification and exclusion rather than for definitive diagnosis. They caution against the broad extrapolation of diagnostic thresholds across diverse etiologies and disease stages, favoring an individualized, component-based interpretation. The European Association for the Study of the Liver (EASL) guidelines are more imaging-based, particularly for the MASLD and obese populations. EASL places greater emphasis on advanced imaging procedures, including MRE and multiparametric MRI, for prognostic evaluation in selected populations, while remaining more cautious about SSM and emerging serum-based models. In sum, these nuances underscore the need for a multimodal, situational assessment of PH. While Baveno VII provides the most practical pathway for routine screening, the EASL and AASLD framework justify the use of specialized imaging and etiology-specific interpretation in complex cases.

AI and data-driven paradigm

The increasing availability of multimodal noninvasive data has catalyzed the development of AI-based models that integrate imaging features with clinical variables. These tools enhance diagnostic precision by extracting complex imaging patterns that correlate with portal pressure and disease severity. Current AI strategies can be classified into four primary domains.

Deep learning-based imaging models

Deep learning models use convolutional neural networks (CNNs) on CT or MRI data to extract imaging features for CSPH identification automatically. They demonstrate exceptional diagnostic accuracy, with AUCs ranging from 0.90 to 0.94, comparable to invasive HVPG (52-54). Nevertheless, these “black-box” models typically lack physiological explainability and their decision-making processes are not transparent, which remains a significant barrier to physician confidence and clinical adoption.

Radiomics-based HVPG prediction models (aHVPG)

Radiomics-based models utilize quantitative features extracted from CT images and apply ML algorithms to predict HVPG values. These models provide continuous HVPG estimations and maintain high discrimination power (AUC >0.80) across critical thresholds (10, 12, and 16 mmHg). However, its performance is sensitive to image quality and inter-center variability, and the biological interpretation of radiomic features remains challenging (55).

Vascular geometry-based interpretable models

Vascular geometry-based models are interpretable and rely on measurable parameters, such as portal vein diameter and branching angles. They have demonstrated good diagnostic performance in internal and external validation cohorts (AUC: 0.85–0.90) (56). Unlike deep learning, the model logic is grounded in vascular anatomy, making it highly interpretable and more likely to be translated into real-world clinical AI applications.

Machine learning (ML) for variceal screening

Gastroesophageal varices screening tools based on ML incorporate routine parameters, including liver stiffness, platelet count, and bilirubin levels, to rule out high-risk varices. These models effectively minimize unnecessary endoscopies while maintaining low miss rates for high-risk varices in multicenter validation studies. They are best utilized as decision-making aids to refine traditional diagnostic standards.

A systematic comparison of AI models with respect to data sources, interpretability, validation status, and translational readiness is provided in Table 4. To provide a clear overview of these digital advancements, the specific workflows of various AI incorporation models—alongside a balanced synthesis of their technical limitations and future clinical possibilities—are pictorially summarized in Figure 1.

Figure 1 AI incorporation models in portal hypertension: workflows, limitations, and clinical possibilities. AI, artificial intelligence.

From research to clinical translation

Overall, AI-based approaches demonstrate substantial diagnostic potential for noninvasive PH assessment, although their clinical maturity varies considerably. Currently, models that offer external validation and structural interpretability are the most transition-ready, whereas black-box deep learning approaches largely remain in the research domain. Future studies should focus on improving interpretability, ensuring robustness across centers, and evaluating real-world implementation costs. Taken together, noninvasive assessment strategies address different clinical needs and disease stages in PH. A concise comparison of their diagnostic performance, applicability, resource requirements, and clinical maturity is summarized in Table 5, providing a practical framework for evidence-based selection.

Multimodal assessment and longitudinal management pathway

The multifactorial pathophysiology of PH necessitates a transition from single-modality testing toward a standardized, stepwise workflow. By combining screening, risk assessment, treatment guidance, and ongoing surveillance, we propose a scalable management pathway adaptable to varying healthcare capabilities (Figure 2). Primary care employs cost-effective, serum-based screening (e.g., FIB-4) to identify patients requiring specialist referral. Specialized centers utilize elastography-based risk stratification (LSM and SSM) to guide therapeutic decision-making and variceal screening. Tertiary referral centers reserve high-tech imaging (MRE, mpMRI) and AI-driven applications for complex cases or indeterminate results.

Figure 2 Stepwise noninvasive management pathway for portal hypertension. APRI, aspartate aminotransferase to platelet count ratio index; CSPH, clinically significant portal hypertension; EUS, endoscopic ultrasound; FIB-4, fibrosis-4 index; LSM, liver stiffness measurement; MRE, magnetic resonance elastography; PLT, platelet count; SSM, spleen stiffness measurement; VITRO, von Willebrand factor-to-platelet ratio.

Stepwise noninvasive management pathway

Step 1: primary care screening

Initial risk assessment is performed using inexpensive, serum-based indices (APRI or FIB-4). Individuals with low-risk results continue routine primary surveillance, while those with elevated scores are referred for specialist assessment.

Step 2: specialist risk stratification

Referred patients undergo risk stratification using the Baveno VII criteria based on LSM and platelet count. To minimize diagnostic uncertainty, clinicians may integrate SSM or the VITRO score, allowing for clear classification into low-risk, high-risk, or indeterminate groups.

Step 3: precision diagnosis in tertiary centers

For indeterminate cases or patients with confounding factors (e.g., obesity, MASLD, or ascites), precision diagnostic assessments are employed. These include MRE, EUS-guided portal pressure gradient measurement, or AI-based imaging models to achieve a definitive diagnosis (57).

Step 4: longitudinal monitoring

Serial assessments using LSM, SSM, or the VITRO score create a closed-loop management system to track disease progression, evaluate response to nonselective beta-blocker therapy, and refine prognostic estimation.

Recommended multimodal strategies by clinical scenarios

In compensated disease, it is recommended to combine the Baveno VII criteria with SSM or VITRO-based stratification to maximize NPV/PPV and minimize unnecessary endoscopies (3,17). In patients with MASLD or obesity, MRE is preferred as it remains unaffected by steatosis or elevated BMI, offering superior diagnostic accuracy (30). In decompensated disease or during treatment monitoring, it is recommended to utilize 3D MRE or multiparametric MRI for dynamic functional assessment and monitoring therapeutic response (50,58). In resource-limited settings, a two-step strategy using FIB-4 for initial screening, followed by VITRO-based stratification, offers a practical alternative to elastography-based approaches, achieving comparable risk classification without reliance on specialized equipment (59).

Challenges and future research agenda

While noninvasive PH assessment has evolved into a robust multimodal framework, significant barriers to standardized clinical implementation remain. Future research should focus on the following priority areas to facilitate a transition from diagnostic assessment toward comprehensive, precision-based disease management.

Standardization and analytical validation

A major challenge is the lack of standardization across diagnostic platforms. For serum-based biomarkers, standardized assays and validated cutoffs are required for markers such as sCD163 and vWF to ensure cross-center reproducibility and affordability. For elastography, unified operational protocols and cross-platform calibration models are essential for reconciling technical inconsistencies across manufacturers and device-specific thresholds.

Etiology-specific validation: the MASLD gap

Most existing evidence supporting noninvasive assessment strategies is derived from viral or alcohol-related liver disease, which may not translate to MASLD, which has emerged as the leading cause of chronic liver disease worldwide. In patients with MASLD, hepatic steatosis and obesity reduce the PPV of the Baveno VII criteria 25 kPa cutoff to 67% (30). In addition, optimal SSM thresholds for MASLD have not yet been established. This diagnostic gap is further compounded in patients with metabolic dysfunction-associated alcohol-associated liver disease (Met-ALD). In this subgroup, active alcohol-induced hepatic inflammation triggers acute cellular swelling, which artificially inflates LSM and drives false-positive overestimations of CSPH. Overcoming these confounding inflammatory flares requires a dual-triage approach: implementing serial longitudinal testing following a documented period of alcohol abstinence (typically 4–12 weeks), and prioritizing SSM or macrophage-specific serum panels (such as the VITRO score), which reflect congestive portal hemodynamics rather than intrinsic liver inflammation. Ultimately, large-scale, multicenter prospective studies are required to redefine and validate etiology-specific cutoff values. Furthermore, noninvasive assessment must be expanded to evaluate noncirrhotic PH, including portal vein thrombosis and schistosomiasis.

Cost-effectiveness and resource allocation

Beyond diagnostic accuracy, future studies should more rigorously evaluate the cost-effectiveness and real-world feasibility of noninvasive strategies. While MRE offers superior diagnostic accuracy, its higher cost and equipment dependency limit its use in primary or resource-constrained settings. Stratified cost-effectiveness analyses comparing MRE with TE across disease stages and patient subgroups are required to define appropriate clinical boundaries. Similarly, integrating AI requires significant investment in data infrastructure, software maintenance, and cross-center calibration. Future studies must evaluate the economic impact and workflow integration of AI to ensure sustainable adoption at scale.

From diagnosis to therapeutic monitoring

Research must shift from cross-sectional diagnostic performance toward longitudinal, a closed-loop management involving treatment decisions and therapeutic monitoring. The use of dynamic changes in SSM, the VITRO score, and associated markers as surrogate endpoints in response to nonselective beta-blocker therapy has yet to be fully established. It is essential to establish direct connections between noninvasive measurements and clinical events such as variceal bleeding and liver decompensation to enable real-time therapeutic adjustments.

Data-driven precision management

The integration of explainable AI systems and digital twin technology represents a transformative shift toward personalized PH management. By synthesizing multimodal imaging and clinical data, explainable AI models can extract hemodynamic features of high physiological significance, ensuring that AI-driven insights are both accurate and transparent to clinicians. These data serve as the foundation for patient-specific digital twins, which are virtual replicas of the hepatic vascular system. Digital twin models provide a simulation platform for validating and optimizing AI predictions by modeling hemodynamic responses to various clinical interventions. These systems enable noninvasive HVPG prediction by providing real-time virtual estimation of pressure gradients without invasive catheterization. Furthermore, it enables virtual therapeutic testing by simulating dose-response interactions for nonselective beta-blockers and post-TIPS hemodynamics (56), which can be useful in making accurate decisions in complicated scenarios.

Summary of core barriers to clinical translation

To bridge the gap between research advances and routine clinical practice, the structural limitations and translational hurdles across current noninvasive strategies are consolidated in Table 6.

Conclusions

The management of PH has transitioned toward a noninvasive, multimodal paradigm that prioritizes longitudinal, risk-stratified care. By integrating serum biomarkers, elastography, advanced imaging, and AI, clinicians can now implement a tiered framework encompassing early screening, risk stratification, treatment monitoring, and prognostic assessment. Pragmatic tools like the FIB-4 index and Baveno VII criteria, augmented by SSM or the VITRO score, effectively identify high-risk patients while minimizing unnecessary endoscopic procedures. In complex cases, such as obese patients with MASLD or those with indeterminate results, advanced technologies, including MRE and AI-based vascular modeling, provide the diagnostic depth needed to optimize clinical outcomes.

Despite this progress, standardized implementation is hindered by cross-platform variability, a lack of etiology-specific validation (particularly in MASLD), and limited evidence linking noninvasive markers to therapeutic response. Future research must address these gaps while also considering cost-effectiveness and real-world feasibility. Importantly, noninvasive assessment is not intended to replace the HVPG measurement as the gold standard. Instead, its value lies in establishing a tiered, clinically oriented management framework that balances accessibility, precision, and resource allocation to support effective long-term care for PH.

Acknowledgments

None.

Footnote

Peer Review File: Available at https://tgh.amegroups.com/article/view/10.21037/tgh-2026-0055/prf

Funding: This work was supported by Guizhou Provincial Key Laboratory for Digestive System Diseases [No. ZSYS(2025)021].

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tgh.amegroups.com/article/view/10.21037/tgh-2026-0055/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Guixé-Muntet S, Quesada-Vázquez S, Gracia-Sancho J. Pathophysiology and therapeutic options for cirrhotic portal hypertension. Lancet Gastroenterol Hepatol 2024;9:646-63. [Crossref] [PubMed]
2. Duo H, You J, Du S, et al. Liver cirrhosis in 2021: Global Burden of Disease study. PLoS One 2025;20:e0328493. [Crossref] [PubMed]
3. de Franchis R, Bosch J, Garcia-Tsao G, et al. Baveno VII - Renewing consensus in portal hypertension. J Hepatol 2022;76:959-74. [Crossref] [PubMed]
4. Møller S, Danielsen KV, Hobolth L, et al. Diagnosis of Portal Hypertension. Diagnostics (Basel) 2025;15:2774. [Crossref] [PubMed]
5. Rockey DC, Alsawas M, Duarte-Rojo A, et al. Noninvasive liver disease assessment to identify portal hypertension: Systematic and narrative reviews supporting the AASLD Practice Guideline. Hepatology 2025;81:1086-104. [Crossref] [PubMed]
6. Feng G, Wong VW, Targher G, et al. Non-invasive tests of fibrosis in the management of MASLD: revolutionising diagnosis, progression and regression monitoring. Gut 2025;74:1741-50. [Crossref] [PubMed]
7. Sukocheva O, Ow TW, Harding D, et al. Liver stiffness measurements in patients with metabolic dysfunction-associated steatotic liver disease: Updates on the method effectiveness and perspectives. World J Hepatol 2025;17:106675. [Crossref] [PubMed]
8. Kirnake V, Arora A, Sharma P, et al. Non-invasive aspartate aminotransferase to platelet ratio index correlates well with invasive hepatic venous pressure gradient in cirrhosis. Indian J Gastroenterol 2018;37:335-41. [Crossref] [PubMed]
9. Patel K, Asrani SK, Fiel MI, et al. Accuracy of blood-based biomarkers for staging liver fibrosis in chronic liver disease: A systematic review supporting the AASLD Practice Guideline. Hepatology 2025;81:358-79. [Crossref] [PubMed]
10. Duarte-Rojo A, Patel K, Rockey DC. Noninvasive assessment of liver fibrosis and portal hypertension. Curr Opin Gastroenterol 2024;40:148-55. [Crossref] [PubMed]
11. Simbrunner B, Villesen IF, Scheiner B, et al. Von Willebrand factor processing in patients with advanced chronic liver disease and its relation to portal hypertension and clinical outcome. Hepatol Int 2023;17:1532-44. [Crossref] [PubMed]
12. Pomej K, Scheiner B, Balcar L, et al. Clinical significance of substantially elevated von Willebrand factor antigen levels in patients with advanced chronic liver disease. Dig Liver Dis 2022;54:1376-84. [Crossref] [PubMed]
13. van den Boom BP, Stamouli M, Timon J, et al. Von Willebrand factor is an independent predictor of short-term mortality in acutely ill patients with cirrhosis. Liver Int 2023;43:2752-61. [Crossref] [PubMed]
14. Hofer BS, Semmler G, Simbrunner B, et al. Course of Portal Hypertension and Applicability of Noninvasive Tests After Recompensation in Alcohol-related Cirrhosis. Clin Gastroenterol Hepatol 2026;24:1175-8. [Crossref] [PubMed]
15. Dominik N, Scheiner B, Zanetto A, et al. Von Willebrand factor for outcome prediction within different clinical stages of advanced chronic liver disease. Aliment Pharmacol Ther 2024;59:1376-86. [Crossref] [PubMed]
16. Schwarzer R, Reiberger T, Mandorfer M, et al. The von Willebrand Factor antigen to platelet ratio (VITRO) score predicts hepatic decompensation and mortality in cirrhosis. J Gastroenterol 2020;55:533-42. [Crossref] [PubMed]
17. Jachs M, Hartl L, Simbrunner B, et al. The Sequential Application of Baveno VII Criteria and VITRO Score Improves Diagnosis of Clinically Significant Portal Hypertension. Clin Gastroenterol Hepatol 2023;21:1854-1863.e10. [Crossref] [PubMed]
18. Gatselis NK, Tornai T, Shums Z, et al. Golgi protein-73: A biomarker for assessing cirrhosis and prognosis of liver disease patients. World J Gastroenterol 2020;26:5130-45. [Crossref] [PubMed]
19. Liu S, Ma J, Chen P, et al. Novel serum biomarker of Golgi protein 73 for the diagnosis of clinically significant portal hypertension in patients with compensated cirrhosis. J Med Virol 2024;96:e29380. [Crossref] [PubMed]
20. Costa D, Trebicka J, Ripoll C, et al. Interaction of inflammation and portal hypertension in cirrhosis progression. Nat Rev Gastroenterol Hepatol 2025;22:846-65. [Crossref] [PubMed]
21. Tiede A, Stockhoff L, Liu Z, et al. Insertion of a transjugular intrahepatic portosystemic shunt leads to sustained reversal of systemic inflammation in patients with decompensated liver cirrhosis. Clin Mol Hepatol 2025;31:240-55. [Crossref] [PubMed]
22. Lamatsch S, Hassan M, Kappert K, et al. Soluble urokinase plasminogen activator receptor is a prognostic biomarker in decompensated cirrhosis. JHEP Rep 2026;8:101677. [Crossref] [PubMed]
23. Takami M, Namisaki T, Shibamoto A, et al. The 7S Domain of Type IV Collagen as a Noninvasive Surrogate of Portal Hypertension. Hepatol Res 2026;56:795-805. [Crossref] [PubMed]
24. Cao Z, Yao Y, Cai M, et al. Blood markers for type-1, -2, and -3 inflammation are associated with severity of acutely decompensated cirrhosis. J Hepatol 2025;82:836-50. [Crossref] [PubMed]
25. Miuma S, Miyaaki H, Taura N, et al. Elevated intestinal fatty acid-binding protein levels as a marker of portal hypertension and gastroesophageal varices in cirrhosis. Sci Rep 2024;14:25003. [Crossref] [PubMed]
26. Kobayashi K, Uldry E, Kokudo T, et al. Correlation Between Portal Pressure and Indocyanine Green Retention Rate is Unaffected by the Cause of Cirrhosis: A Prospective Study. World J Surg 2021;45:2546-55. [Crossref] [PubMed]
27. You MW, Kim KW, Pyo J, et al. A Meta-analysis for the Diagnostic Performance of Transient Elastography for Clinically Significant Portal Hypertension. Ultrasound Med Biol 2017;43:59-68. [Crossref] [PubMed]
28. Kim S, Lewey S, Meuller L, et al. Endohepatology in clinical practice: EUS-guided portal pressure measurement combined with EUS-guided liver biopsy and variceal screening and treatment in outpatients. Endosc Ultrasound 2024;13:89-93. [Crossref] [PubMed]
29. Pavic T, Mikolasevic I, Kralj D, et al. Role of Endoscopic Ultrasound in Liver Disease: Where Do We Stand? Diagnostics (Basel) 2021;11:2021. [Crossref] [PubMed]
30. Bauer DJM, Nixdorf L, Dominik N, et al. The deep abdominal ultrasound transducer (DAX) increases the success rate and diagnostic accuracy of shear wave elastography for liver fibrosis assessment in patients with obesity-A prospective biopsy-controlled study. Aliment Pharmacol Ther 2024;60:70-82. [Crossref] [PubMed]
31. He R, Liu C, Grgurevic I, et al. Validation of Baveno VII criteria for clinically significant portal hypertension by two-dimensional shear wave elastography. Hepatol Int 2024;18:1020-8. [Crossref] [PubMed]
32. Bañares J, Pons M, Reiberger T, et al. Refining the Baveno VII criteria for clinically significant portal hypertension: An individual patient data meta-analysis. J Hepatol 2026;84:543-53. [Crossref] [PubMed]
33. Semmler G, Yang Z, Fritz L, et al. Dynamics in Liver Stiffness Measurements Predict Outcomes in Advanced Chronic Liver Disease. Gastroenterology 2023;165:1041-52. [Crossref] [PubMed]
34. Abdulrasak M, Ahmed M, Hootak S. Utility of splenic transient elastography in assessing the presence of portal hypertension: A review. World J Methodol 2025;15:107411. [Crossref] [PubMed]
35. Thorhauge KH, Semmler G, Johansen S, et al. Using liver stiffness to predict and monitor the risk of decompensation and mortality in patients with alcohol-related liver disease. J Hepatol 2024;81:23-32. [Crossref] [PubMed]
36. Dong B, Chen Y, Chen Y, et al. Diagnostic accuracy of liver stiffness on two-dimensional shear wave elastography for detecting clinically significant portal hypertension: a meta-analysis. Expert Rev Med Devices 2023;20:141-9. [Crossref] [PubMed]
37. Trebicka J, Gu W, de Ledinghen V, et al. Two-dimensional shear wave elastography predicts survival in advanced chronic liver disease. Gut 2022;71:402-14. [Crossref] [PubMed]
38. Duarte-Rojo A, Taouli B, Leung DH, et al. Imaging-based noninvasive liver disease assessment for staging liver fibrosis in chronic liver disease: A systematic review supporting the AASLD Practice Guideline. Hepatology 2025;81:725-48. [Crossref] [PubMed]
39. Liang JX, Ampuero J, Niu H, et al. An individual patient data meta-analysis to determine cut-offs for and confounders of NAFLD-fibrosis staging with magnetic resonance elastography. J Hepatol 2023;79:592-604. [Crossref] [PubMed]
40. Kim BK, Bergstrom J, Loomba R, et al. Magnetic resonance elastography-based prediction model for hepatic decompensation in NAFLD: A multicenter cohort study. Hepatology 2023;78:1858-66. [Crossref] [PubMed]
41. Kobayashi T, Iwaki M, Nogami A, et al. Prediction of outcomes in patients with metabolic dysfunction-associated steatotic liver disease based on initial measurements and subsequent changes in magnetic resonance elastography. J Gastroenterol 2024;59:56-65. [Crossref] [PubMed]
42. Sangha K, Chang ST, Cheung R, et al. Cost-effectiveness of MRE versus VCTE in staging fibrosis for nonalcoholic fatty liver disease (NAFLD) patients with advanced fibrosis. Hepatology 2023;77:1702-11. [Crossref] [PubMed]
43. Zhang X, Zhou L, Liang W, et al. Identification of Clinically Significant Portal Hypertension in cACLD Individuals With Spleen Stiffness Measurement. Liver Int 2025;45:e16241. [Crossref] [PubMed]
44. Gaspar R, Mota J, Almeida MJ, et al. Spleen Stiffness Predicts the Risk of Liver-related Complications in Patients With Compensated Advanced Chronic Liver Disease. Clin Gastroenterol Hepatol 2025;23:2519-2528.e1. [Crossref] [PubMed]
45. European Association for the Study of the Liver. Electronic address: easloffice@easloffice; . EASL Clinical Practice Guidelines on vascular diseases of the liver. J Hepatol 2026;84:399-456.
46. Prakash JH, Anirvan P, Gupta S, et al. Diagnostic Accuracy of Spleen-Dedicated 100 Hz Transient Elastography to Predict High-Risk Esophageal Varices. Am J Gastroenterol 2026;121:1659-66. [Crossref] [PubMed]
47. Duan Y, Dong B, Li S, et al. Diagnostic accuracy of contrast-enhanced ultrasound for detecting clinically significant portal hypertension and severe portal hypertension in chronic liver disease: a meta-analysis. Expert Rev Gastroenterol Hepatol 2023;17:489-97. [Crossref] [PubMed]
48. Gupta I, Eisenbrey JR, Machado P, et al. Diagnosing Portal Hypertension with Noninvasive Subharmonic Pressure Estimates from a US Contrast Agent. Radiology 2021;298:104-11. [Crossref] [PubMed]
49. Ozkaya E, Bane O, Bhuiyan E, et al. Multiparametric MRI for Diagnosing Clinically Significant Portal Hypertension and Predicting Liver Decompensation. Liver Int 2025;45:e70368. [Crossref] [PubMed]
50. Xiong Z, Wang X, Yan Y, et al. A new computational fluid dynamics based noninvasive assessment of portacaval pressure gradient. J Biomech 2024;167:112086. [Crossref] [PubMed]
51. Liu Y, Ning Z, Örmeci N, et al. Deep Convolutional Neural Network-Aided Detection of Portal Hypertension in Patients With Cirrhosis. Clin Gastroenterol Hepatol 2020;18:2998-3007.e5. [Crossref] [PubMed]
52. Erratum regarding previously published articles. J Affect Disord Rep 2022; Epub ahead of print. [Crossref]
53. Yu Q, Huang Y, Li X, et al. An imaging-based artificial intelligence model for non-invasive grading of hepatic venous pressure gradient in cirrhotic portal hypertension. Cell Rep Med 2022;3:100563. [Crossref] [PubMed]
54. Wang C, Huang Y, Liu C, et al. Diagnosis of Clinically Significant Portal Hypertension Using CT- and MRI-based Vascular Model. Radiology 2023;307:e221648. [Crossref] [PubMed]
55. Huang Y, Li J, Zheng T, et al. Development and validation of a machine learning-based model for varices screening in compensated cirrhosis (CHESS2001): an international multicenter study. Gastrointest Endosc 2023;97:435-444.e2. [Crossref] [PubMed]
56. Kolb JM, Monachese M, Rubin RA, et al. Endoscopic Ultrasound-Guided Portosystemic Pressure Gradient Correlates With Clinical Parameters and Liver Histology. Clin Gastroenterol Hepatol 2025;23:2167-2175.e4. [Crossref] [PubMed]
57. Semmler G, Hartl L, Mendoza YP, et al. Simple blood tests to diagnose compensated advanced chronic liver disease and stratify the risk of clinically significant portal hypertension. Hepatology 2024;80:887-900. [Crossref] [PubMed]
58. Roy A, Verma N, Jajodia S, et al. Magnetic resonance elastography (MRE) outperforms acoustic force radiation impulse (ARFI) in predicting oesophageal varices in obese NAFLD cirrhosis. Abdom Radiol (NY) 2024;49:3088-95. [Crossref] [PubMed]
59. Dajti E, Ravaioli F, Marasco G, et al. A Combined Baveno VII and Spleen Stiffness Algorithm to Improve the Noninvasive Diagnosis of Clinically Significant Portal Hypertension in Patients With Compensated Advanced Chronic Liver Disease. Am J Gastroenterol 2022;117:1825-33. [Crossref] [PubMed]