The Future Role of Abdominal US in Hepatocellular Carcinoma Surveillance

Abstract

Abdominal US is currently the best-validated surveillance strategy for hepatocellular carcinoma (HCC) in at-risk patients. It is the only modality shown to have completed all five phases of validation and can achieve high sensitivity and specificity for HCC detection, especially when conducted by expert sonographers in high-volume centers. However, US also has limitations, including operator dependency and varying sensitivity in clinical practice. Further, the sensitivity of US for early-stage HCC detection is lower in patients with obesity or nonviral liver disease, increasingly common populations undergoing surveillance. Imaging-based and blood-based surveillance strategies, including abbreviated MRI and biomarker panels, may overcome some limitations of US-based surveillance. Both strategies have promising test performance in phase II and phase III biomarker studies and are undergoing prospective validation. Considering the variation in HCC risk and test performance between patients, there will likely be a shift away from a one-size-fits-all approach and toward precision screening, in which the “best” test is selected based on individual patient characteristics. In this upcoming era of precision HCC screening among patients with cirrhosis, US will likely continue to have an important, albeit reduced, surveillance role.

© RSNA, 2024

Summary

Promising imaging- and blood-based surveillance strategies appear poised to reduce, albeit not eliminate, the role of abdominal US in hepatocellular carcinoma surveillance, but they require validation before adoption in clinical practice.

Essentials

• ■ Abdominal US can achieve high sensitivity and specificity for hepatocellular carcinoma (HCC) detection when conducted by expert sonographers in high-volume centers; however, its sensitivity for early-stage HCC detection is lower in patients with obesity or nonviral liver disease.

• ■ Emerging imaging-based and blood-based surveillance strategies, including abbreviated MRI and biomarker panels, demonstrate promising accuracy for early-stage HCC detection and may overcome some limitations of US-based surveillance.

• ■ In an era of multiple options for HCC surveillance strategies, there will likely be a shift toward precision screening, in which the “best” test is selected for each individual patient.

• ■ Precision HCC screening will likely benefit patients with cirrhosis due to variation in test performance among patients.

• ■ US will continue to play a role, albeit reduced, for HCC surveillance in patients with cirrhosis.

Introduction

Hepatocellular carcinoma (HCC) is the fourth leading cause of cancer-related mortality worldwide. It is one of the few cancers with an incidence-to-mortality ratio approaching 1 (1). However, prognosis markedly differs by tumor stage at diagnosis. Patients with early-stage tumors can achieve a median survival beyond 10 years with curative therapies, such as surgical resection and liver transplant. However, patients with more advanced tumor burden are typically treated with palliative treatments and have a median survival below 3 years (2).

Therefore, professional society guidelines recommend HCC surveillance in at-risk patients, including those with chronic hepatitis B or cirrhosis from any cause (3,4). A large randomized controlled trial among patients with chronic hepatitis B demonstrated a 37% reduction in HCC-related mortality in patients undergoing semiannual HCC surveillance, driven by increased early tumor detection (5). However, a similar randomized trial attempted among patients with cirrhosis was prematurely terminated given poor enrollment (6). Therefore, HCC surveillance in patients with cirrhosis is supported by level II data. Several cohort studies have shown a consistent association between surveillance and clinical outcomes, including early tumor detection, more frequent curative treatment, and lower mortality among patients with cirrhosis (7).

There has been increasing recognition that the overall value of HCC surveillance is not only defined by benefits but also by potential physical, financial, and psychologic harms (8–11). The benefits of surveillance are determined by test sensitivity for early-stage detection. The benefit-to-harm ratio is worse among populations with low HCC incidence or when using tests with low specificity (causing a higher false-positive rate) or suboptimal sensitivity (leading to lower early-stage tumor detection). Patients with metabolic dysfunction–associated steatotic liver disease (MASLD) can develop HCC in the absence of cirrhosis, although the annual incidence is very low, highlighting a need for risk stratification tools in this population (12–15).

Abdominal US has formed the cornerstone of HCC surveillance for more than 20 years. All hepatology societies continue to recommend abdominal US for HCC surveillance among at-risk individuals. However, there has been increasing interest in alternative imaging- and blood-based surveillance strategies, particularly for select patient populations in whom US has lower sensitivity for early-stage HCC detection (14). This review discusses pro and con perspectives regarding the potential future role of US in HCC surveillance (Table 1).

Table 1:

Summary of Pro and Con Statements for the Future Role of US in HCC Surveillance

Pro: US Should Continue to Have a Central Role in HCC Surveillance

The use of high-quality US as a component of a successful HCC surveillance program is analogous to the use of mammography in a comprehensive breast cancer program. Both imaging examinations are performed without the need for intravenous contrast material, and both identify a subset of patients for which further action may be needed. In cases where imaging is suboptimal in liver US or breast mammography, or in some patients with specific risk factors, alternative screening imaging modalities may be of benefit. Importantly, a patient-centered, data-driven, and risk factor–influenced management system supports the need for a multimodality approach to HCC surveillance, with US remaining a critical component.

US fulfills the accepted requirements for a screening test and is currently the only modality for which prospective outcomes data show a survival benefit (5,16–20). US provides acceptable sensitivity for any-stage HCC, reaching 85% (similar to that of mammography for breast cancer). However, US sensitivity may be as low as 47% for early-stage HCC, increasing to 63% with the addition of serum α-fetoprotein (AFP) measurement (21,22). Notably, many of these previous studies included examinations with suboptimal visualization and may not have included a dedicated liver imaging protocol. Therefore, they may have underestimated the true diagnostic accuracy of US.

Nonetheless, the question of whether US should remain the primary imaging modality for HCC surveillance also depends on the available alternatives. US is superior to other current imaging options in terms of affordability, availability, and safety while maintaining an acceptable level of accuracy. Despite this, surveillance utilization is low in patients at risk for HCC; only one-fourth of at-risk patients undergo surveillance (23). Thus, even if a surveillance method has greater accuracy or sensitivity, obstacles may further negatively impact surveillance adherence rates and consequently worsen outcomes. These obstacles include cost, intravenous contrast material, radiation exposure, and limited availability. Furthermore, capacity concerns and limited access may hamper the widespread and equitable application of other imaging modalities. This includes alternative modalities such as abbreviated MRI, which is discussed in more depth in the next section.

Despite universal endorsement of US, guidance for the performance or interpretation of US surveillance examinations was not provided by medical societies until 2017. That year was the release of the American College of Radiology Ultrasound Liver Imaging Reporting and Data System (US LI-RADS) algorithm for HCC screening and surveillance (24–26). The US LI-RADS recommends providing two examination-level assessments: category (US-1, negative; US-2, subthreshold; US-3, positive) and visualization score (VIS-A, no limitations; VIS-B, moderate limitations; VIS-C, severe limitations) (24). Acknowledging the important impact of US LI-RADS in standardization of technique and reporting, the updated guidelines from the American Association for the Study of Liver Diseases now incorporate US LI-RADS category and visualization score into its surveillance recommendations (3).

Admittedly, US may not provide the same benefit for all patients. US LI-RADS was the first liver imaging system to introduce the concept of stratifying examination sensitivity according to adequate visibility of the liver parenchyma (ie, examination quality) (Figure). The visualization score informs the expected level of sensitivity of the examination and is analogous to breast density in the Breast Imaging Reporting and Data System (or BI-RADS) for mammography. Although initially developed by expert opinion, the visualization score has good interreader agreement, with an intraclass correlation coefficient of 0.7 (27). Chong and Schoenberger et al (28) showed preserved sensitivity for HCC in patients with liver examinations scored as VIS-A or VIS-B (>75%) but significantly decreased sensitivity in patients with examinations scored as VIS-C (27.3%). Only 4.2% of patients in a multi-institutional study were categorized as having images with suboptimal visualization (VIS-C) (29). Thus, most US examinations performed in patients at risk for HCC are expected to offer acceptable quality and sensitivity.

(A, B) Images in a 67-year-old female patient with cirrhosis related to metabolic dysfunction–associated steatotic liver disease who presented for outpatient hepatocellular carcinoma (HCC) surveillance. Grayscale US images through (A) the left lobe in transverse orientation and (B) right lobe in longitudinal orientation show adequate visualization of liver parenchyma without obscuration by attenuation, shadowing structures, or other artifacts. Despite parenchymal coarsening, the examination was scored VIS-A according to the Liver Imaging Reporting and Data System, reflecting a high confidence for excluding HCC (high sensitivity). (C, D) Images in a 41-year-old male patient with alcohol-related decompensated cirrhosis hospitalized for transplant work-up, unable to follow breathing and positioning instructions due to hepatic encephalopathy. Grayscale US images through (C) the left lobe in transverse orientation and (D) right lobe in longitudinal orientation show markedly compromised liver parenchymal visualization due to rib shadowing and excessive parenchymal attenuation, limiting evaluation of the hepatic dome and deeper segments of the right lobe. The examination was scored VIS-C, reflecting the expected decrease in HCC sensitivity.

The US visualization score provides a means to identify individuals with suboptimal examination quality (VIS-C) who may benefit from alternative surveillance strategies. Interestingly, a single VIS-C examination does not necessarily predict future VIS-C examinations; a patient with a single VIS-C examination still has a greater than 50% chance of having subsequent examinations improve to VIS-B or VIS-A. However, if a patient has two or more VIS-C examinations, their likelihood of continued VIS-C examinations increases (30,31). In upcoming updates to US LI-RADS surveillance (version 2024), a short-term follow-up US examination is recommended for examinations scored as VIS-C, with alternative surveillance strategies recommended if VIS-C score is repeated. Exceptions to this short-term follow-up will include patients with independent risk factors for poor US quality, including MASLD and alcohol-related liver disease, Child-Turcotte-Pugh B or C cirrhosis, and body mass index of 35 or higher. These factors have been shown to predict a higher risk for poor-quality US (ie, higher odds of VIS-C) (31–33).

To optimize imaging, US should be performed with a standardized imaging protocol, by a team with a focus on high-quality liver imaging. Patients undergoing surveillance US should be adequately prepared for the examination (ie, fasting for at least 4–6 hours before the examination) and be able to follow breathing and movement directions. Providers should defer routine surveillance examinations in patients who are in the emergency department or admitted to the hospital until their acute condition has resolved. These patients are often unable to comply with instructions and have proven lower US quality compared with when imaged in an outpatient setting (33).

Although some patients may benefit from alternative surveillance strategies in the setting of suboptimal US visualization, to our knowledge, no studies to date have shown MRI- or CT-based screening to decrease HCC-related mortality, unlike that which has been shown in US (5,16–20). Moreover, it is unclear if patients with VIS-C US examinations will have a corresponding multiphase CT or MRI examination with better quality and satisfactory diagnostic accuracy. Whether a lifetime of surveillance with contrast-enhanced CT or MRI every 6 months is a cost-effective strategy with positive patient outcomes remains to be proven.

Future studies comparing the diagnostic performance of various imaging modalities for HCC surveillance should include an equivalent dedication to optimized liver imaging protocols for each included modality. This will allow more precise head-to-head comparisons (ie, following LI-RADS technical recommendations for US, CT, and MRI). In addition, studies investigating outcomes and cost-effectiveness may be strengthened by incorporating a patient-centered algorithm in which high-quality US remains the core imaging modality. The algorithm would then consider alternative strategies to US for subsets of patients with repeated or high-risk factors for poor US examination quality. In this way, a more personalized approach may help leverage the benefits of all available imaging tools, improving patient-level outcomes while minimizing system-level costs.

Con: US Will Have a Minimal Role in HCC Surveillance

US has advantages contributing to its adoption for HCC surveillance; however, it has limitations that may impact its future role, particularly considering the expanding arsenal of alternative emerging imaging- and blood-based strategies.

An important consideration when evaluating the utility of a surveillance modality is its sensitivity for early-stage HCC detection—a critical outcome given the selective availability of curative surgical therapies for early-stage disease. Although abdominal US has high sensitivity for the detection of HCC at any stage (84% [95% CI: 76, 92]), its detection of early-stage HCC in a recent meta-analysis (22) was lower, with a pooled sensitivity of only 47% (95% CI: 33, 61), which increased to 63% with the addition of AFP measurement. In this same study, the sensitivity of CT for early HCC was 62%, and the sensitivity of MRI was not reported (22). US imaging is associated with higher odds of being suboptimal in patients with morbid obesity and nonviral liver disease, which impairs sensitivity for early-stage HCC detection (28,30). Given worldwide epidemiologic shifts with an increasing proportion of HCC related to MASLD and alcohol-related liver disease, this variation in US performance may affect the future population-level sensitivity for early-stage HCC detection (34). Simmons et al (32) reported that 35% of US examinations in patients with MASLD cirrhosis are inadequate compared with 15% in those without MASLD. A meta-analysis by Hong and Kim et al (35) similarly reported that 22% of HCC surveillance US examinations are inadequate, with a higher proportion among patients with high body mass index and MASLD. Central obesity, hepatic steatosis, and regenerative nodules are believed to contribute to impaired visualization at US. This population may particularly benefit from alternative surveillance modalities, although further data characterizing their test performance and value are needed.

Published sensitivity for US varies widely (22), likely related to a combination of patient-related factors and operator dependency, a limitation that does not exist for blood-based biomarkers and exists to a lesser extent with CT- or MRI-based surveillance. A prior cohort study highlighted large variation in visualization of US by sonographer volume, reader volume, and setting where the US examination was performed (33). Although US can achieve acceptable sensitivity when conducted by expert radiologists at high-volume centers using standardized liver protocols, this does not always reflect clinical practice, where many patients undergo US imaging in local community centers. Therefore, there remains a large gap between the potential efficacy of US under ideal circumstances and its effectiveness when performed in real-world clinical settings.

US-based surveillance is associated with potential physical, financial, and psychologic harms related to false-positive or indeterminate results leading to increased CT, MRI, or biopsy (8–10). In a study of 999 patients with cirrhosis with a median follow-up of 2.2 years, 26% undergoing US-based HCC surveillance had a finding that prompted a diagnostic CT or MRI examination. Of these patients, 27% were diagnosed with HCC. The most common “harm” reported in this study was a single CT or MRI examination (36). Thus, patients with prolonged time to diagnostic resolution are exposed to greater physical, financial, and psychologic harms caused by extra imaging (8–11). In addition, over one-fourth of patients with subcentimeter liver lesions are unnecessarily imaged with CT or MRI in clinical practice, which exacerbates screening-related harms, despite low risk of HCC and guidelines recommending monitoring with repeat US (37).

Alternative approaches with multiphasic CT, contrast-enhanced MRI, and/or serum biomarkers may prove useful as complementary or alternative tools to US-based surveillance. In addition to test performance, considerations such as patient acceptance, accessibility, ease of implementation, and costs must be considered when assessing the effectiveness of each strategy. When considering these factors, a conjoint survey found over 90% of patients preferred more sensitive and convenient strategies versus a US-based approach (38).

Although a prospective cohort study demonstrated higher sensitivity and specificity of two-phase low-dose CT compared with US (39), CT-based surveillance is likely limited by cost, radiation exposure, and potential nephrotoxicity, particularly if repeated on a semiannual basis. Alternatively, there is increasing interest in MRI-based surveillance given its high test performance. A prospective cohort study of 407 patients with Child-Pugh A or B cirrhosis demonstrated high accuracy of MRI compared with US, including higher sensitivity (86% vs 28%, respectively; P < .001) and specificity (97% vs 94%; P = .004) for early-stage HCC (40). Although MRI does not have the same limitations as CT-based surveillance, concerns with MRI include high cost, challenges in imaging center capacity, need for intravenous contrast material, and technical complexities of longer scanning time, patient motion, and patient claustrophobia. These barriers have historically limited MRI-based surveillance adoption in broad populations, although the cost of MRI has notably decreased over time (41).

Abbreviated MRI using a limited number of sequences has been proposed with or without the use of contrast material. Abbreviated MRI addresses some of the limitations of MRI by shortening in-scanner time to approximately 15 minutes, thereby offering a path to reduce costs and mitigate concerns about imaging center capacity. However, radiologic capacity may remain a challenge in parts of the world with limited access to MRI scanners. A meta-analysis of 10 studies found that abbreviated contrast-enhanced MRI had significantly higher sensitivity for detection of early-stage HCC than noncontrast MRI (87% vs 79%, respectively) but significantly lower specificity (91% vs 98%, respectively) (42). A secondary analysis of the PRIUS Study from South Korea also demonstrated higher performance of surveillance abbreviated MRI compared with US (86.0% vs 27.9%; P < .001) (43). A multicenter case-control study using pathology as the reference standard similarly reported high sensitivity and specificity of 88.2% and 89.1%, respectively (44). Notably, abbreviated MRI sensitivity for early-stage HCC was consistently high among patients in whom US has poor visualization, including patients with obesity (91.1%) or nonviral liver disease (82.1% for alcohol-related liver disease and 97.2% for MASLD). These results suggest that abbreviated MRI may be a promising alternative surveillance strategy in these populations. Randomized clinical trials (eg, FAST-IRM for HCC Surveillance in Patients With High Risk of Liver Cancer, or FASTRAK, study [ClinicalTrials.gov identifier NCT05095714]; Annual MRI Versus Biannual US for Surveillance of Hepatocellular Carcinoma in Liver Cirrhosis, or MAGNUS-HCC, study [NCT02551250]; and Usefulness of Non-contrast Magnetic Resonance Imaging Versus Non-Contrast Ultrasonography for Surveillance of Hepatocellular Carcinoma, or MIRACLE-HCC, study [NCT02514434]) are ongoing to compare abbreviated MRI versus US for early-stage HCC detection in high-risk individuals.

While imaging-based surveillance strategies are the most used, adherence remains poor due to several patient and provider barriers, with surveillance completion in only one-fourth of at-risk individuals (45–47). Blood-based biomarkers are a promising tool in surveillance, as they may improve test sensitivity for early-stage HCC detection and improve adherence to HCC surveillance (48). Currently, AFP is the only sufficiently validated blood-based biomarker to have completed all five phases of biomarker validation (Table 2) (49); however, its performance is insufficient as a standalone test, particularly among patients with nonviral liver disease (50). Similarly, other single biomarkers showed suboptimal performance, likely reflecting the heterogeneity of HCC (51). More recently, biomarker panels have been developed, with enhanced performance. The best-validated biomarker panel to date is GALAD (gender, age, lens culinaris agglutinin-reactive AFP [ie, AFP-L3], total AFP, and des-γ-carboxyprothrombin). Small phase III studies showed promising performance of GALAD, and results from the large phase III Hepatocellular Early Detection Study reported a sensitivity of 63% for GALAD within 3 months of HCC diagnosis, with a specificity of 82% (52–54). The GALAD panel is now being evaluated in a multicenter randomized phase IV study in comparison with US for early-stage HCC detection among patients with cirrhosis (ClinicalTrials.gov identifier NCT06084234). There are numerous other promising blood-based biomarkers for surveillance in earlier stages of validation, including additional protein-based biomarkers, methylated DNA markers, extracellular vesicle markers, and analysis of fragmented circulating DNA (55–58); however, these require larger-scale validation before clinical implementation (59). Nevertheless, a blood-based approach overcomes many of the barriers to imaging-based screening, including convenience and reproducibility. Thus, these approaches will potentially allow for a more effective early-stage detection strategy than current paradigms. This is especially true for rural or resource-limited areas with constrained ability to conduct imaging-based surveillance.

Table 2:

Phases of Biomarker Validation

Debrief: The Future Role of US in HCC Surveillance

In summary, US is currently the best-validated HCC surveillance strategy and the only modality shown to reduce HCC-related mortality in a large randomized clinical trial. US can achieve high accuracy when conducted by expert sonographers in high-volume centers. However, it can have reduced sensitivity and specificity in other settings, particularly when performed in patients with obesity or nonviral liver disease. Alternative imaging-based and blood-based HCC surveillance strategies have shown promising test performance in large phase III biomarker studies. These strategies may overcome the recognized limitations of US-based surveillance. Emerging surveillance strategies including abbreviated MRI and biomarker panels such as GALAD are now undergoing prospective evaluation in phase IV cohort studies and clinical trials. Thus, medicine is on the cusp of entering an era with multiple options for HCC surveillance.

In parallel, there has been increasing recognition of variation in HCC risk among patients with cirrhosis. This heterogeneity is reflected by clinical risk scores and risk stratification biomarkers (12,14), as well as variation in test performance between patients. Therefore, there have been calls for a move away from a one-size-fits-all approach and toward a precision screening strategy, in which the “best” test is selected for each individual patient (60). The future of HCC surveillance will likely include a combination of surveillance modalities dependent on individual patient needs. Abbreviated MRI may be useful for high-risk patients who are proven to have suboptimal visibility at US (ie, repeat examinations scored as VIS-C) and in those with obesity or nonviral liver disease. Blood-based biomarkers may be preferred in patients with barriers to surveillance completion.

Overall, US will likely continue to play a role, albeit more limited, in HCC surveillance in patients with cirrhosis.

Conclusion

The future of HCC surveillance will likely include a combination of surveillance modalities dependent on individual patient needs. This will incorporate a precision screening strategy in which the “best” test is selected for each individual patient. US will likely continue to play a role, albeit more limited, in HCC surveillance in patients with cirrhosis.