HCC in MASLD: A New 2-Step Non-Invasive Approach

Impact and Implications HCC emerges as the most frequent liver-related complication in MASLD —surpassing hepatic decompensation and challenging current paradigms of disease progression. We identified a clinically actionable liver stiffness threshold (<7.4 kPa by 2D-SWE) below which the risk of HCC is negligible, potentially reducing unnecessary surveillance in low-risk patients. Conversely, patients with LSM ≥7.4 kPa, particularly those with type 2 diabetes and elevated GGT, are at significantly higher risk of HCC, as confirmed by multivariate analysis—allowing for targeted monitoring. To enhance clinical applicability, we developed two-step risk stratification algorithms (eg, LSM-SSI combined with the validated PLEASE score), achieving excellent diagnostic accuracy and a 100% negative predictive value—providing a practical, evidence-based tool to guide more efficient and sustainable HCC surveillance in MASLD. Introduction Metabolic dysfunction-associated steatotic liver disease (MASLD) progresses to metabolic dysfunction-associated steatohepatitis (MASH) in up to 20% of cases.1 Among those, around 10% eventually develop cirrhosis.2 Over time, cirrhosis leads to severe complications, which may represent the most frequent long-term clinical outcomes in this population. Nevertheless, 23% to 50% of hepatocellular carcinoma (HCC) cases linked to MASLD arise in non-cirrhotic livers.3–8 This paradox positions MASLD as a significant cause of HCC even in the absence of cirrhosis, challenging traditional paradigms of liver cancer development. In these non-cirrhotic cases, HCC is often diagnosed at a more advanced stage, with larger tumors and poorer clinical outcomes, due to the lack of clinical features identifying patients requiring surveillance.8 This pattern underscores the challenges in early detection and emphasizes the urgent need for improved surveillance strategies.9 In cirrhotic MASLD, an annual HCC incidence of 1.5% is considered the minimum threshold for cost-effective surveillance.10,11 The global burden of MASLD, affecting approximately 38% of the adult population above 48 years,12 renders widespread HCC surveillance impractical and cost-ineffective, differently from the decompensating events which only occur in patients with cirrhosis.13,14 Although a threshold of 0.5% per year has been proposed for non-cirrhotic liver patients in general,15 a specific threshold for non-cirrhotic MASLD is still to be established.16 Accurately identifying the subset of non-cirrhotic MASLD patients who would benefit from HCC surveillance remains a complex and unresolved challenge.17 Fibrosis progression is the main driver of HCC onset in chronic liver diseases.18 Nevertheless, emerging evidence supports the existence of an alternative carcinogenic pathway in MASLD, distinct from the classical fibrosis–cirrhosis–HCC sequence.13 Populations at increased risk for MASLD-HCC include males, Hispanics, older adults, and individuals with features of metabolic syndrome (MS). Additional contributing factors may involve moderate to high alcohol intake, tobacco use, elevated liver enzymes, and genetic predisposition.13,18,19 Several models are currently emerging to stratify patients according to their risk of developing HCC. Among them, the PLEASE (Platelet, Elastography, Age, Sex, Etiology) score has been recently proposed to assess HCC risk in individuals with advanced chronic liver disease (ACLD).20 Scoring systems and prediction models are promising, particularly when validated across diverse clinical settings and when their risk strata is linked to tailored surveillance strategies.21 In non-cirrhotic MASLD, comprehensive and individualized HCC risk assessment remains essential; however, the target population for surveillance must still be more clearly defined.17 Accordingly, the primary aims of this study were1 to investigate the incidence of liver-related complications occurring during follow-up in a real-world cohort of MASLD patients, with a special focus on those without cirrhosis, and2 to search for a combination of variables able to identify patients with non-cirrhotic MASLD at higher risk of developing HCC deserving surveillance. Patients and Methods Study Design This study presents a post-hoc analysis of a consecutive prospective cohort of MASLD patients evaluated at a third-level hepatology center in Bologna, Italy. The cohort included patients referred from primary care (general practitioners, diabetologists, or obesity clinics) for further assessment, based on consistent evidence of hepatic steatosis (confirmed by imaging or biopsy), associated metabolic risk factors, or abnormal liver function tests, according to the EASL Guidelines.22 Patients were included if seen between October 2014, and February 2023 [time of availability of SuperSonic equipment to measure liver stiffness], with the last follow-up recorded on January 30, 2024, in order to guarantee a minimal possible follow-up of one year even for the most recent patients. Ethical Approval The study was conducted in compliance with the Declaration of Helsinki, Istanbul, and Good Clinical Practice guidelines. It received approval from the Institutional Ethical Committee of IRCCS Sant’Orsola-Malpighi University Hospital in Bologna, Italy (BOMASH study, 0044335/2024). Informed consent was obtained and signed by each alive participant. Patients, Inclusion, and Exclusion Criteria Patients were eligible for inclusion if they met the following criteria: Diagnosis of MASLD according to the updated definition.23 No other known causes of liver disease, including viral hepatitis, autoimmune liver diseases, cholestatic disorders, alcohol-related liver disease, inherited conditions (alpha-1 antitrypsin deficiency, hemochromatosis, Wilson disease), drug-induced liver injury, and vascular liver disorders.23 Age ≥18 years. Availability of semiquantitative grading of liver steatosis via ultrasound (US),4 Availability of a reliable baseline LSM-SSI measurement. Comprehensive clinical and laboratory evaluation at baseline. Exclusion criteria included: Presence of HCC at or prior to enrollment. Presence of hepatic decompensation at or prior to enrollment. Absence of a reliable baseline LSM-SSI measurement. Follow-up duration of less than 6 months. Patients with missing data were excluded from the analysis. The patient flow is illustrated in Figure S1. Baseline Demographic, Clinical, and Laboratory Data Collection The inclusion period began with the first LSM-SSI performed as part of a comprehensive multiparametric US evaluation at baseline. Demographic and clinical data were systematically collected for each patient, including laboratory parameters such as blood count, coagulation profile, electrolytes, renal and liver function tests, inflammatory markers, and lipid and glycemic profiles. Lifestyle factors were documented, including alcohol consumption, smoking status, and body mass index (BMI). Clinical Outcomes and Follow-Up The endpoints were the development of HCC diagnosed according to the EASL clinical practice guidelines,24,25 the development of the first decompensation event (ascites, variceal bleeding, overt hepatic encephalopathy) according to the BAVENO VII guidelines,26 or the development of portal vein thrombosis (PVT). Follow-up assessments were conducted every 6 months in patients with cirrhosis and every 12 months in non-cirrhotic patients, in line with institutional protocols. Liver Stiffness Measurement Using Two-Dimensional Shear Wave Elastography with the SuperSonic Imagine Ultrasound System (LSM-SSI) Ultrasound Examination LSM-SSI was performed by experienced ultrasound operators using the Aixplorer ultrasound system (SuperSonic Imagine SA, Aix-en-Provence, France). Patients were examined in a fasting state, placed supine with the abducted right upper limb. The operator selected the region of interest using B-mode ultrasound, avoiding large vascular structures and targeting an area at least 15 mm below the liver capsule through a right intercostal space. An adequate SWE color map with complete and homogeneous filling was acquired, followed by image capture for stiffness measurement using the Q-box tool. The Stability Index (SI) was calculated for each acquisition, aiming for values ≥0.90. For each patient, at least five reliable measurements were obtained during neutral respiratory apnea, and their median and interquartile range (IQR) was reported in kilopascals (kPa). Measurement reliability was defined according to the latest EFSUMB Guidelines and Recommendations on the Clinical Use of Ultrasound Elastography.27 Optimized LSM-SSI Cut-off Values for Advanced Fibrosis (≥F3) and Cirrhosis (F4) Optimized LSM-SSI cut-off values, derived from our recently published meta-analysis28 and adapted for the SuperSonic Imagine ultrasound system, were applied as follows: <7.4 kPa to exclude advanced fibrosis (≥F3), and ≥15.6 kPa to confirm cirrhosis (F4). These thresholds were used to stratify the study population, highlighting fibrosis stage as the most significant predictor of clinical events in MASLD.29 Other HCC Risk Stratification Scores PLEASE Score To evaluate the risk of HCC occurrence in patients with non-cirrhotic MASLD, we calculated the recently developed PLEASE score20 using the online tool available at https://www.medizin.uni-muenster.de/en/med-b/please-calculator.html. Additional non-invasive risk stratification scores, using predefined thresholds according to guidelines,22 were also calculated, with detailed methodologies provided in the Supplementary Materials. Assessment of Additional Clinical Events During Follow-Up The occurrence of dysplastic nodules (DN), clinically significant portal hypertension (CSPH), portal vein thrombosis (PVT), and major adverse cardiovascular events (MACE) were also assessed during follow-up, with definitions of these conditions provided in the Supplementary Materials. Statistical Analysis Continuous variables were assessed for normality using the Kolmogorov–Smirnov test. Based on their distribution, the data were summarized as either the median with the first and third quartiles (Q1-Q3) or the mean with the standard deviation (SD). Categorical variables were presented as counts and percentages. To compare means between two groups, we applied the independent samples t-test, with Levene’s test used to assess homogeneity of variances. When the assumptions for parametric tests were not met, the Mann–Whitney U-test was employed as a non-parametric alternative. For categorical variables, the Chi-square test (χ2) was used, or Fisher’s Exact Test for smaller sample sizes. The Kruskal–Wallis test was applied to compare more than two independent groups with ordinal or non-normally distributed data. Competing risks regressions based on Fine and Gray’s proportional sub-hazards models were performed to identify risk factors for HCC occurrence, with results expressed as sub-hazard ratios (sHRs) and 95% confidence intervals (CIs) for each predictor. The competing risk events for HCC occurrence included the first decompensation event—defined by the Baveno VII consensus as ascites, variceal bleeding, or hepatic encephalopathy26 - and the occurrence of portal vein thrombosis (PVT), if initially detected by ultrasound. Subsequently, we conducted univariable and multivariable Cox proportional hazards models without accounting for competing risks. Variables significantly associated with HCC occurrence in the univariable analyses were then entered into the multivariable model, and the optimal model was ultimately identified using a backward elimination procedure. Follow-up ended with the diagnosis of HCC, the occurrence of a competing risk event (first decompensation event or PVT), or patient censoring at the last available follow-up. PVT presence was confirmed by a CT scan. A receiver operating characteristic (AUROC) curve analysis was conducted to assess the predictive performance of continuous parameters. The Youden Index was applied to identify optimal cut-off values, while additional cut-off points were established based on pre-defined sensitivity (Se) and specificity (Sp) criteria. The Kaplan-Meier method was used to illustrate and compare the cumulative incidence functions of HCC in MASLD by LSM-SSI category. The cumulative incidence of HCC was calculated, accounting for decompensation and the occurrence of portal vein thrombosis (PVT) as competing risk events. The Log rank test and Gray’s test were used to assess group differences for survival and HCC cumulative incidence. To evaluate the performance of various predictors and two-step algorithms for predicting HCC onset in MASLD at different pre-set occurrence,30 key metrics were assessed, including sensitivity (Se), specificity (Sp), negative predictive value (NPV), positive predictive value (PPV), negative likelihood ratio (LR-), positive likelihood ratio (LR+), and accuracy (Acc). The algorithms were tested for the following HCC occurrence rates: 0.3% (retrieved in our non-cirrhotic population, <15.6 kPa), 0.5% (surveillance proposed for non-cirrhotic liver patients in general,15 0.8% (retrieved in our overall population), and 2% (suitable for a high risk scenario).30 The evaluations were conducted using both optimized and previously reported cut-off values20,22 across different strategies. The added value of a second examination for patients with indeterminate results or those falling within the “grey zone” was also examined. All statistical tests were two-sided, with a significance level set at p < 0.05. Analyses were performed using STATA 18 or R software (version 4.3.2). Results Patients’ Characteristics Following the application of the study criteria, 352 patients were included in the final study population (Figure S1). Baseline population characteristics are detailed in Tables 1 and S1. The median age of the cohort was 58.2 (IQR: 48.5–66) years, with 191 (54%) male participants. The median body mass index (BMI) was 29.1 (IQR: 26.4–31.7) kg/m2. Type 2 diabetes mellitus (T2DM) and arterial hypertension (HTA) were present in 94 patients (27%) and 241 patients (68%) at baseline, respectively. The cohort had a median LSM-SSI of 6.2 (IQR: 5.2–7.7) kPa. A total, of 257 patients (73%) had LSM-SSI values below 7.4 kPa, while 95 (27%) had values above this threshold. Patients with LSM-SSI > 7.4 kPa were significantly older, with a median age of 64.2 (IQR: 58.8–70.3) vs 56 (IQR: 45.2–63.2) years, had a higher prevalence of T2DM, 54.7% vs 42%, and HTA 82.1% vs 63%, as well as higher median blood glucose levels 125 (IQR: 90–149) mg/dL vs 100 (IQR: 84–108) mg/dL. The complete comparison is presented in Tables 1 and S1. Follow-Up, Liver-Related Complications, and HCC Incidence Rates The median follow-up duration for the entire cohort was 31 months (IQR: 14.1–57.8). During follow-up, 9 patients (2.6%) developed HCC, 6 (1.7%) experienced a first decompensation event, and 2 (0.6%) developed portal vein thrombosis (PVT). No patient with a baseline LSM-SSI <7.4 kPa developed liver-related clinical outcomes. Conversely, among patients with LSM-SSI ≥7.4 kPa, the following event rates were observed: 9.5% (9 patients) developed HCC, and 6.3% (6 patients) experienced a first decompensation event, including 5 cases of grade II ascites and 1 case of overt hepatic encephalopathy. One patient concurrently developed both HCC and a new decompensation episode (ascites). Additionally, 2.1% (2 patients) developed de novo PVT, with one case coinciding with HCC diagnosis. De novo dysplastic nodules were observed exclusively in patients with LSM-SSI ≥7.4 kPa, with a global occurrence of 5.3% (5 patients). Similarly, 13.7% (13 patients) developed de novo clinically significant portal hypertension (CSPH), identified by gastroesophageal varices on upper endoscopy or collateral circulation on cross-sectional imaging (Table 1). Regarding MACE, 10.5% (10 patients) with LSM-SSI ≥7.4 kPa experienced events, compared to 7% (18 patients) among those with LSM-SSI <7.4 kPa (Table 1). Baseline LSM-SSI values among patients who developed MACE ranged from 3.7 to 28.8 kPa, with events occurring regardless of fibrosis severity (Table 1). The median time to HCC occurrence was 22.6 months (IQR: 12.3–40.8). Nine patients with MASLD developed HCC over 1128 PY of follow-up, resulting in an annual incidence rate of 0.8 per 100 PY. When patients with LSM-SSI between 7.4–15.6 kPa were considered only, the HCC incidence rate increased to 1.3 per 100 PY. When patients with LSM-SSI ≥15.6 kPa (F4) were excluded, the HCC incidence rate in the subgroup <15.6 kPa decreased to 0.3 per 100 PY. Additional incidence rates are reported in Table S2. Kaplan–Meier estimates for HCC occurrence by LSM-SSI categories are presented in Figure 1. Figure S2A–C show the Kaplan–Meier curves for the first decompensation event, MACE, and onset of CSPH, respectively, stratified by LSM-SSI categories. Competing-Risk Regression Analysis for HCC Occurrence Univariate analysis identified age (HR 1.075, 95% CI 1.030–1.121, p=0.001), white blood cells (HR 0.999, 95% CI 0.998–0.999, p=0.013), platelets (HR 0.987, 95% CI 0.975–0.997, p=0.020), aspartate aminotransferase (HR 1.023, 95% CI 1.005–1.041, p=0.011), gamma glutamyl transferase (HR 1.004, 95% CI 1.001–1.006, p=0.001), alkaline phosphatase (HR 1.009, 95% CI 1.005–1.011, p=0.000), total bilirubin (HR 2.147, 95% CI 1.154–3.990, p=0.016), international normalized ratio (HR 3.873, 95% CI 1.188–12.619, p=0.025), blood glucose (HR 1.008, 95% CI 1.004–1.012, p=0.000), T2DM (HR 4.952, 95% 1.236–19.835, p=0.024), LSM-SSI (HR 1.052, 95% CI 1.027–1.076, p<0.0001) as independent risk factors for HCC occurrence in MASLD in the overall cohort. Multivariate analysis identified LSM-SSI (HR 1.052, 95% CI 1.030–1.075, p<0.001), T2DM (HR 4.555, p=0.038), and gamma-glutamyl transferase (HR 1.004, p=0.003) as independent predictors for HCC occurrence in our MASLD overall cohort (Table 2). The results of the univariable and multivariable Cox proportional hazards models, conducted without accounting for competing risk events, are presented in Table S3. Table S4 presents a comparison of the performance of various scores and clinical and laboratory variables in predicting HCC occurrence in MASLD. Validation of PLEASE Score for HCC in Non-Cirrhotic MASLD The predictive performance of the PLEASE score for HCC occurrence was indicated by an AUROC of 0.94 [0.92–0.97] for the whole population and 0.80 [0.70–0.87] for those with LSM-SSI ≥7.4 kPa. For the entire cohort, using our proposed rule-out (Se>90%) cut-off of 0.076, the PLEASE score showed Se=100%, Sp=89.21%, PPV=19.6%, NPV=100%. By applying our proposed rule-in cut-off (Sp>90%) of 0.152, the PLEASE score exhibited a Se=55.6%, Sp=95.9%, PPV=26.3%, and NPV=98.8%, respectively. Clinical Algorithms for HCC Surveillance in Patients with MASLD The diagnostic performance of FIB-4, LSM-SSI, the PLEASE score, and our multivariate model were first assessed individually using standardized cut-off values to rule out (sensitivity >90%) and rule in (specificity >90%) HCC, across predefined HCC occurrence scenarios of 0.3%, 0.8%, and 2%. Based on the diagnostic performance of each individual score, two-step algorithms were developed and tested to enhance HCC risk stratification. In the first step, standardized thresholds were applied to rule out HCC (sensitivity >90%), identify high-risk individuals (specificity >90%), and define a “grey zone” of indeterminate cases. A second stratification step was applied within this grey zone to refine risk assessment and improve diagnostic accuracy. All possible single and sequential combinations of these approaches were systematically evaluated to determine the most effective strategy. The diagnostic performance of the algorithms across different pre-test HCC occurrence scenarios (0.3%, 0.5%, 0.8%, and 2%) is summarized in Tables 3, 4, and S5, with an overview illustrated in Figure 2 and a clinical interpretation in Figure 3. At a predefined HCC occurrence of 0.3%, a FIB-4 value <1.3 effectively ruled out HCC occurrence, with a Se=88.9% and a NPV=99.9%, outperforming the age-adapted FIB-4 ≥2.0 strategy in patients aged ≥65 years (Table 3). In the same setting, LSM-SSI <7.4 kPa ruled out HCC with perfect Se (100%) and NPV (100%), and demonstrated superior overall accuracy compared to FIB-4 (75.6% vs 64.5%). Similarly, the PLEASE score, using a proposed rule-out threshold of <0.076 (Se >90%), matched the performance of LSM-SSI <7.4 kPa (Se=100%, NPV=100%), and achieved an accuracy of 89.2% (Table 3). In rule-in scenarios using standard thresholds (eg, FIB-4 ≥2.67, LSM-SSI ≥15.6 kPa), the diagnostic performance was modest. FIB-4 ≥2.67 yielded a Sp=91% and a PPV=2.5%, while LSM-SSI ≥15.6 kPa showed a Se=93.6% and a PPV=3.0%. When non-invasive tests (NITs) were integrated into two-step algorithms to enhance diagnostic accuracy and improve risk stratification, the combination of LSM-SSI (<7.4 kPa, ≥15.6 kPa) with a PLEASE score <0.076 achieved excellent rule-out performance (Se =100%, NPV=100%) and a high overall accuracy (Acc=89.5%). This algorithm outperformed other tested combinations, such as FIB-4 (<1.3 and ≥2.67) followed by LSM-SSI <7.4 kPa (Se=88.9%, NPV=100%, Acc=81.8%), or FIB-4 followed by PLEASE score <0.076 (Se=88.9%, NPV=100%, Acc=86.6%). This trend remained consistent across higher HCC occurrence scenarios, including 0.8% (Table 4) and 2% (Table S5). Discussion In our cohort of MASLD patients, hepatocellular carcinoma (HCC) emerged as the most prevalent liver-related event, with an incidence rate of 0.8 per 100 person-years in the overall cohort. This surpassed the incidence of both first hepatic decompensation events and portal vein thrombosis. Among the subgroup with liver stiffness measurement (LSM-SSI) values ranging from 7.4 to 15.6 kPa, the HCC incidence rate was higher at 1.3 per 100 person-years, approaching the 1.5% threshold, for which HCC surveillance is generally recommended across all liver disease etiologies.31,32 This findings underscore the importance of establishing effective HCC surveillance strategies in MASLD and highlights the distinct clinical course of this condition compared to other chronic liver disease etiologies.33–35 Recent guidelines emphasize that patients with MASLD and advanced fibrosis, but without cirrhosis, are at higher risk of developing HCC.21 However, routine HCC surveillance is not currently recommended for non-cirrhotic MASLD due to insufficient supporting evidence.21 Nevertheless, etiology-specific surveillance could support more personalized approaches by distinguishing between low-risk and high-risk patients. A study from the ITALICA Study Group highlighted the challenge of delayed HCC diagnosis in MASLD, as HCC is often detected at more advanced stages and may occur in the absence of cirrhosis.8 Timely detection of MASLD-HCC during its often-prolonged asymptomatic phase is critical to improve treatment decision-making and expand the range of available therapeutic options.8,17 Cost-effectiveness analyses suggest that HCC surveillance is justified when the annual incidence reaches at least 1.5%, regardless of the underlying etiology,31,32 as mentioned above. However, recent modeling studies have shown that an annual HCC incidence of approximately 0.8% may also be cost-effective for initiating surveillance, even when accounting for potential harms such as high false-positive rates and the need for additional diagnostic procedures.36 As previously reported, the benefits of HCC screening remain uncertain for MASLD patients with a low annual incidence, such as 0.4%.36 This finding aligns with previous studies that assessed HCC incidence or composite liver-related events (LREs), using vibration-controlled transient elastography (VCTE).37–43 Recent evidence suggests that individuals with significant or advanced fibrosis, combined with additional risk factors, may benefit from cost-effective HCC surveillance strategies.36 Consistently, our univariate and multivariate analyses identified LSM-SSI, among other variables (T2DM, and GGT) as independent predictors of HCC. Moreover, LSM-SSI demonstrated a progressively increasing risk trajectory for HCC occurrence on Kaplan–Meier analysis (Figure 1), in alignment with previous findings reported for vibration-controlled transient elastography (VCTE).44 Emerging concerns about the potential harms of surveillance—such as false positives leading to unnecessary testing and patient anxiety—have raised important questions regarding the overall value of HCC surveillance in MASLD.45,46 In response, our study aimed to refine the criteria for identifying patients who may not require surveillance, achieving excellent negative predictive values (NPVs). In our cohort, LSM-SSI outperformed FIB-4 in guiding HCC surveillance decisions, demonstrating superior rule-out capability and higher overall diagnostic accuracy. We found that HCC surveillance is unnecessary in MASLD patients with LSM-SSI <7.4 kPa (rule-out ≥F3), for at least around 3 years, time at which patients could be reassessed. Furthermore, we proposed a two-step algorithm based on LSM-SSI and PLEASE score, which achieved a negative predictive value (NPV) of 100% and an overall accuracy of 89.5% in identifying patients who should avoid HCC surveillance, assuming a predefined HCC occurrence of 0.3% (Table 3). We also developed and tested similar algorithms combining various available tools—including FIB-4, the PLEASE score, LSM-SSI, and our multivariate model (LSM-SSI, GGT, T2DM)—to assess their performance under different clinical conditions and HCC occurrence assumptions. These strategies aim to provide clinicians with practical, adaptable tools for real-world decision-making, tailored to varying risk profiles and surveillance settings. Such an approach may prove both feasible and acceptable for implementation in the MASLD population.17 Since elastography modules are now integrated into most modern ultrasound systems, 2D-SWE offers substantial potential for evaluating patients with MASLD in both primary and tertiary care settings;47 in this context, LSM-SSI emerges a promising alternative to VCTE-based LSM. While PPV increased with higher HCC occurrence—reaching up to 20% at a 2% pre-set occurrence rate (Tables 3, 4, S4)— the PPV remained relatively low overall. For this reason, we recommend using these algorithms primarily to identify individuals who do not require HCC surveillance. This is a critical consideration, especially given the widespread prevalence of MASLD, which affects more than one-third of the global adult population above 48 years.12 Conversely, to better define the subgroup of patients most likely to benefit from HCC surveillance, future studies are needed in larger cohorts with more reported HCC events. Such studies would support improved PPV, maintain diagnostic accuracy, and enhance the performance of algorithms for disease confirmation. Many at-risk individuals may remain unidentified until advanced stages of MASLD. Our proposed two-step non-invasive algorithms offer a practical approach to bridge this gap by combining readily available biochemical scores with two-dimensional ultrasound–based fibrosis assessment. This modality has the potential to become more widely accessible in the future as standardization progresses, offering a feasible alternative to other liver stiffness measurement techniques.48 Unlike existing risk models that often rely on single metrics or are not routinely applicable in everyday practice,49–51 our approach enables flexible and scalable risk stratification tailored to the tests available in different clinical settings. Limitations and Strengths The limitations of this study stem from its post-hoc nature and monocentric design. Importantly, the threshold of elastography we defined may not apply to all other elastographic ultrasound methods. Nevertheless, its strength lies in the use of a prospective consecutive cohort of 352 patients, retrospectively evaluated, all of whom underwent standardized LSM-SSI evaluation using the same ultrasound system (SuperSonic Aixplorer), thereby ensuring data consistency and measurement reliability. In developing the algorithm, standardized cut-off values derived from previously published studies were applied, enhancing the robustness and reproducibility of the model. While the algorithms show promise, external validation in independent populations would be necessary to ensure broader applicability. Interestingly, our cohort included a lower proportion of patients with elevated LSM, compared to prior studies,20 which may reflect a more representative group of the entire MASLD population. Nevertheless, our cohort consisted of 352 patients from a single-center experience, of whom 9 (2.6%) developed HCC during follow-up. This limited number of events may have reduced the statistical power of our analysis and should be taken into consideration when interpreting the findings. External validation would further support the generalizability and clinical applicability of our proposed algorithms. Conclusions HCC represents the most common liver-related complication in non-cirrhotic MASLD. As supporting evidence expands and the method becomes further standardized, two-dimensional shear wave elastography (2D-SWE) is increasingly becoming more available in clinical practice and has the potential to make a meaningful contribution to HCC risk stratification in MASLD. An LSM-SSI value <7.4 kPa (rule-out ≥F3) reliably identifies patients at negligible risk of developing HCC, for whom routine surveillance may be safely omitted. In contrast, patients with LSM-SSI ≥7.4 kPa, particularly those with additional risk factors such as T2DM or elevated GGT, may benefit from regular HCC surveillance. To support clinical decision-making, a two-step algorithm combining LSM-SSI with the PLEASE score demonstrated excellent rule-out performance and high overall accuracy, offering a practical tool for real-time risk stratification. This personalized approach reduces unnecessary examinations in low-risk individuals while enhancing the efficiency of surveillance in higher-risk groups. Future validation in larger and more diverse populations will be crucial to confirm its widespread implementation and to further optimize HCC surveillance strategies in MASLD.