Vibration‐controlled transient elastography in pediatric metabolic dysfunction‐associated steatotic liver disease

Abstract

Objectives

Metabolic dysfunction‐associated steatotic liver disease (MASLD) is a prevalent disease in children. Vibration‐controlled transient elastography (VCTE) offers a noninvasive alternative to liver biopsy, using controlled attenuation parameter (CAP) to estimate steatosis and liver stiffness measurement (LSM) for fibrosis. However, pediatric data with histological validation are limited. This prospective, multicenter study evaluated the accuracy of CAP and LSM in pediatric MASLD.

Methods

Children with histologically confirmed MASLD from the Nonalcoholic Steatohepatitis Clinical Research Network (NASH CRN) Database 3 underwent VCTE within 6 months of liver biopsy. CAP was evaluated for correlation with steatosis grades, and LSM for correlation with fibrosis stages. The diagnostic performance of LSM in distinguishing fibrosis stages was analyzed using histological findings as the reference standard.

Results

Among 92 children with MASLD (mean age 13 ± 3 years), CAP values were similar across steatosis grades (median 325, 310, and 323 dB/m for grades 1–3, respectively) and showed no significant correlation with histologic steatosis (p = 0.422). Median LSM values increased with fibrosis stage (6.0–8.8 kPa), but significant differences were detected only between stage 0 and stage 3 fibrosis (p = 0.037). For advanced fibrosis (stages 3–4), area under the receiver operating characteristic curve was 0.67, with sensitivity 67%, specificity 76%, positive predictive value 40%, and negative predictive value 90%.

Conclusion

In this prospective, multicenter cohort, VCTE showed modest accuracy for grading steatosis or staging fibrosis in pediatric MASLD. Improved noninvasive methods are urgently needed for evaluation and monitoring in this population.

What is Known

• Vibration‐controlled transient elastography (VCTE) is a noninvasive, point‐of‐care tool used in adults to assess hepatic steatosis and fibrosis.

• Pediatric validation of VCTE in metabolic dysfunction‐associated steatotic liver disease (MASLD) remains limited.

What is New

• Controlled attenuation parameter did not reliably assess the amount of liver fat in children with MASLD.

• Liver stiffness measurement showed a modest relationship with histologic fibrosis and limited ability to distinguish fibrosis stages.

• The use of VCTE in pediatric MASLD should be approached with caution.

1. INTRODUCTION

Metabolic dysfunction‐associated steatotic liver disease (MASLD) is the most prevalent liver disease in children in the United States, affecting up to 10% of the pediatric population. ¹ Characterized by abnormal accumulation of lipid droplets in hepatocytes, MASLD can progress to steatohepatitis, fibrosis, and cirrhosis. Children with MASLD are at increased risk for type 2 diabetes and premature mortality. ² , ³ , ⁴ Yet, awareness and diagnostic efforts among healthcare providers remain limited, leaving many cases undetected. ⁵

The diagnosis and staging of MASLD in children often require liver histology, which remains the clinical standard for grading steatosis and staging fibrosis—key elements for prognosis and management. ⁶ However, liver biopsy is limited by its invasiveness, high cost, and inaccessibility. Magnetic resonance imaging proton density fat fraction (MRI‐PDFF), while a validated imaging biomarker for hepatic steatosis, is similarly restricted by high expense and availability. ⁷ , ⁸ Risk stratification remains essential, yet noninvasive fibrosis scores developed for adults, including the Fibrosis‐4 (FIB‐4) index and AST‐to‐platelet ratio index (APRI), perform poorly in children, with reported area under the curves (AUCs) ranging from 0.51 to 0.67. ⁹ Pediatric models such as the nonalcoholic fatty liver disease (NAFLD) fibrosis index (PNFI) and NAFLD fibrosis score (PNFS) have likewise shown limited accuracy. Although the more recent Fibro‐PeN model may improve pediatric fibrosis risk stratification, it has not been externally validated. ⁹ , ¹⁰ These limitations reinforce the continued need for reliable noninvasive diagnostic alternatives.

Vibration‐controlled transient elastography (VCTE) is a noninvasive, point‐of‐care tool that uses a skin surface ultrasound probe and low‐frequency vibrations to create shear waves in the liver. The controlled attenuation parameter (CAP) estimates hepatic steatosis by measuring ultrasound signal attenuation, while liver stiffness measurement (LSM) assesses fibrosis based on the velocity of shear wave propagation. ¹¹ , ¹² According to the 2023 American Association for the Study of Liver Diseases (AASLD) practice guidance, VCTE can be used as a point‐of‐care option in adults to identify steatosis and screen for advanced fibrosis. ¹³ Although VCTE is feasible in pediatric populations, data on its diagnostic accuracy in children are limited and mixed, with few studies using histologic reference standards. ¹¹ , ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ , ²⁰

Given these gaps, we conducted this study to evaluate the diagnostic performance of CAP and LSM in children with MASLD, using liver histology as the reference standard. Our primary aim was to assess the association between CAP and histologic steatosis grade, and between LSM and fibrosis stage. For LSM, we further examined its ability to discriminate between clinically relevant fibrosis thresholds.

2. METHODS

2.1. Ethics statement

The study protocol was approved by a Single Institutional Review Board (sIRB) for Multi‐Site Research at Johns Hopkins Medicine, with reliance agreements in place at each participating center. Written informed consent was obtained from a parent or guardian, and written assent from all children aged 8 years or older.

2.2. Study design and participants

Study participants included children diagnosed with MASLD and enrolled in the Clinical Research Network in Nonalcoholic Steatohepatitis (NASH CRN) NAFLD Database 3, a prospective cohort study conducted across eight pediatric clinical centers in the United States (NCT04454463). For this cross‐sectional analysis, we evaluated the accuracy of CAP and LSM in assessing liver steatosis and fibrosis among these patients. Study visits for this analysis took place between January 2021 and January 2024. Participants were under 18 years of age and had MASLD confirmed by clinical history, laboratory tests (e.g., liver enzymes, metabolic markers), and liver histology within 90 days of enrollment. Exclusion criteria included the presence of other liver diseases or significant alcohol consumption. VCTE measurements had to be performed within 180 days of the liver biopsy.

2.3. Study visits and procedures

Demographic, anthropometric, and laboratory data were collected at baseline using standardized procedures. A structured interview gathered demographic information, including age, sex, and relevant medical history. Anthropometric measurements—weight and height—were recorded to the nearest 0.1 kg and 0.1 cm, respectively. Body mass index (BMI) was calculated as weight (kilograms [kg]) divided by height (meters [m]) squared, and BMI z‐scores were computed to allow for comparisons across different ages and sexes. Laboratory data were obtained following an overnight fast and included fasting serum glucose, insulin, lipid profiles, liver chemistries (alanine aminotransferase [ALT], aspartate aminotransferase [AST], alkaline phosphatase, gamma‐glutamyl transferase [GGT], bilirubin, albumin, total protein), coagulation markers (prothrombin time [PT], international normalized ratio [INR]), and complete blood count (CBC).

2.4. Liver histology

Liver histology was evaluated to achieve consensus scores by the Pathology Committee of the NASH CRN using hematoxylin and eosin (H&E) and trichrome stained slides, according to the NASH CRN scoring system. ²¹ , ²² The pathologists were masked to the clinical study, age, sex, or specific clinical features of the children at the time of review. The degree of steatosis was graded as follows: grade 0 (none to <5%), grade 1 (5%–33%), grade 2 (34%–66%), and grade 3 (>66%). Fibrosis was staged as follows: stage 0 (no fibrosis), stage 1a (mild zone 3 perisinusoidal requiring trichrome stain), stage 1b (moderate zone 3 perisinusoidal fibrosis without trichrome stain), stage 1c (portal/periportal fibrosis only), stage 2 (zone 3 perisinusoidal and periportal fibrosis), stage 3 (bridging fibrosis), and stage 4 (cirrhosis). ²¹ , ²² For the purposes of our study, fibrosis stages 1a, 1b, and 1c were all considered to be in one group under stage 1.

2.5. VCTE exams

VCTE was performed using the FibroScan® device (Echosens, France) by trained operators following a standardized protocol. All operators completed documented training by a manufacturer‐certified trainer and were required to demonstrate competency through at least 10 supervised patient exams before performing study procedures independently. ²³ , ²⁴ In addition, each site conducted pilot scans under supervision before enrollment to ensure fidelity to acquisition protocols. Operators were blinded to histology results at the time of scanning. To minimize variability, sites maintained consistent device calibration and quality control procedures throughout the study.

Patients fasted for a minimum of 3 h and were examined in the supine position with the right arm maximally abducted. ²⁵ The probe was placed in the intercostal space over the right hepatic lobe. For probe selection, BMI was used initially to guide workflow across sites (<30 kg/m² for M probe, ≥30 kg/m² for XL probe), with final probe choice confirmed or modified by the device′s automated probe selection tool based on skin‐to‐liver capsule distance. A minimum of 10 valid measurements was required, and the median was recorded as the final LSM (kPa). CAP values were acquired simultaneously (dB/m). Exams with fewer than 10 valid measurements or with an interquartile range (IQR) > 30% of the median were deemed unreliable and excluded from analysis. ²³ , ²⁴

2.6. Data analysis

Continuous variables were summarized as means with standard deviations (SD) or medians with IQR, depending on distribution; categorical variables were reported as frequencies and percentages. Normality of continuous variables, including CAP and LSM, was assessed using the Shapiro–Wilk test. Non‐parametric methods were used when normality assumptions were not met. Comparisons across steatosis grades and fibrosis stages were conducted using analysis of variance (ANOVA) for normally distributed data and Kruskal–Wallis tests for non‐normally distributed data. Post hoc comparisons used Tukey's honestly significant difference (HSD) following ANOVA and Dunn′s test following Kruskal–Wallis, as appropriate.

Ordinal logistic regression was used to evaluate the association between CAP and steatosis grades (grades 1–3) and between LSM and fibrosis stages (stages 0–4), with secondary models adjusting for age and BMI z‐score. The proportional odds assumption was assumed to hold for all models. Model fit was assessed using log‐likelihood and Akaike information criterion (AIC), and results were reported as regression coefficients, standard errors, z‐statistics, p‐values, and 95% confidence intervals (CIs).

Receiver operating characteristic (ROC) curve analysis was performed to assess the diagnostic performance of both CAP and LSM. For CAP, areas under the ROC curve (AUROC) were calculated for distinguishing histologic steatosis grade 1 versus grades 2–3 and grades 1–2 versus grade 3. For LSM, AUROCs were calculated for distinguishing fibrosis stages: nonfibrotic versus fibrotic (0 vs. 1–4), no/mild versus clinically significant fibrosis (0–1 vs. 2–4), and nonadvanced versus advanced fibrosis (0–2 vs. 3–4). Balanced cutoff points were identified using the Youden index to maximize sensitivity and specificity. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated at these points to describe the diagnostic performance in practical terms. Additional cut points were derived to meet predefined performance targets (e.g., 90% sensitivity or specificity), reflecting different clinical priorities. Spearman′s rank correlation coefficients were also used to assess the association between CAP and steatosis grade and between LSM and fibrosis stage.

All statistical analyses were conducted using Python 3.10.12 (Python Software Foundation) and R 4.4.3 (R Core Team). Statistical significance was defined as p < 0.05, with 95% CIs reported.

3. RESULTS

3.1. Study population

A total of 103 pediatric participants underwent VCTE within 6 months of liver biopsy. Eleven children were excluded, including seven due to invalid measurements and four whose IQR exceeded 30, leaving 92 participants (89%) with successful VCTE scans for analysis. Of these, 27 (29%) were examined with the M probe and 65 (71%) with the XL probe. The mean age of participants was 13 ± 3 years, and the median interval between liver biopsy and VCTE was 82 days (IQR: 62). Histological evaluation revealed that all participants had steatosis, with 23% having grade 1, 26% grade 2, and 51% grade 3 steatosis. Fibrosis staging indicated that 17% of participants had stage F0 fibrosis, 42% F1, 22% F2, 17% F3, and 2% F4. Detailed demographic and clinical characteristics of the study population are summarized in Table 1.

Table 1.

Baseline characteristics of study population.

Characteristic	Study population (N = 92)
Age, at enrollment, years, median (IQR)	13 (3)
Sex
Male	78 (85%)
Female	14 (15%)
Weight, kg, median (IQR)	88 (35)
Height, cm, median (IQR)	164 (16)
BMI, kg/m², median (IQR)	33 (8)
BMI z‐score, median (IQR)	3.0 (1.4)
Systolic blood pressure, mmHg, median (IQR)	121 (16)
Diastolic blood pressure, mmHg, median (IQR)	70 (13)
HDL‐C, mg/dL, median (IQR)	38 (11)
Triglycerides, mg/dL, median (IQR)	144 (101)
HbA1C, %, median (IQR)	5 (4)
Glucose, mg/dL, median (IQR)	94 (14)
Steatosis
Grade 1	21 (23%)
Grade 2	24 (26%)
Grade 3	47 (51%)
Fibrosis
Stage 0	16 (17%)
Stage 1	38 (42%)
Stage 2	20 (22%)
Stage 3	16 (17%)
Stage 4	2 (2%)

Note: Demographic, anthropometric, metabolic, and histologic characteristics of children with biopsy‐confirmed MASLD included in the study. Continuous variables are presented as median with IQR; categorical variables are presented as number (percentage).

Abbreviations: BMI, body mass index; HbA1C, hemoglobin A1C; HDL‐C, high‐density lipoprotein cholesterol; IQR, interquartile range; MASLD, metabolic dysfunction‐associated steatotic liver disease.

3.2. Correlation of CAP and steatosis grade

CAP values demonstrated substantial overlap across steatosis grades, with median CAP scores of 325 dB/m (IQR 27), 310 dB/m (IQR 53), and 323 dB/m (IQR 54) for grades 1, 2, and 3, respectively (Table S1). CAP scores were not normally distributed within any grade, as confirmed by the Shapiro–Wilk test (p‐values ranging from 0.002 to 0.015;). Ordinal logistic regression showed no significant relationship between CAP and steatosis grade (β = 0.0018, SE = 0.004, z = 0.483, p = 0.629), a finding further supported by Spearman′s rank correlation, which revealed no significant association (ρ = 0.033, 95% CI: −0.173 to 0.232, p = 0.422). Sensitivity analyses, which adjusted for age (β = −0.0032, p = 0.308) and BMI z‐score (β = 0.0013, p = 0.737), did not significantly affect these results.

3.3. Diagnostic accuracy of CAP and steatosis grade

The diagnostic performance of CAP for differentiating between histologic steatosis grades was poor (Table S2 and Figure S1). When distinguishing grade 1 from grades 2–3, the AUROC was 0.55 with sensitivity of 0.54, specificity of 0.24, PPV of 0.70, NPV of 0.14, and diagnostic accuracy of 0.47. For distinguishing grades 1–2 from grade 3, AUROC was likewise 0.55, sensitivity 0.48, specificity 0.67, PPV 0.60, NPV 0.56, and diagnostic accuracy of 0.57.

3.4. Correlation of LSM and fibrosis stage

Median LSM values increased across fibrosis stages, ranging from 6.0 kPa (IQR 1.9) in stage F0 to 8.8 kPa (IQR 4.7) in stage F3 and 8.5 kPa (IQR 0.1) in stage F4 (Table S3). The distribution of values across fibrosis stages is shown in Figure 1. Ordinal logistic regression revealed a significant association between LSM and fibrosis stage (β = 0.1235, SE = 0.048, z = 2.598, p = 0.009). Sensitivity analyses adjusting for age (β = 0.1258, p = 0.009) or BMI z‐score (β = 0.1197, p = 0.009) did not materially alter the relationship between LSM and fibrosis stage. A Dunn′s test showed that significant differences in LSM values were observed only between children with no fibrosis (F0) and those with stage F3 fibrosis (p = 0.037). Spearman′s rank correlation indicated a weak, nonsignificant association between LSM and fibrosis stage (ρ = 0.306, 95% CI: 0.113–0.470, p = 0.115).

Figure 1.

CAP values by steatosis grade and LSM values by fibrosis stage. (A) A yellow box‐and‐whisker plot illustrating the distribution of CAP values across steatosis grades, and (B) is a blue box‐and‐whisker plot depicting the distribution of LSM values across fibrosis stages. The boxes represent the IQR, with whiskers indicating the range of nonoutlier values. The median is shown as a black line within the IQR boxes, individual outliers are plotted as separate points. This visualization highlights the variability of CAP values within each steatosis grade, and the variation in LSM values across different stages of fibrosis. CAP, controlled attenuation parameter; IQR, interquartile range; LSM, liver stiffness measurement.

3.5. Diagnostic accuracy of LSM for fibrosis stage

The diagnostic performance of LSM for identifying fibrosis was moderate to limited (Table 2, Figure 2). The AUROC for distinguishing nonfibrotic (F0) from fibrotic (F1–F4) stages was 0.70. Using a balanced cutoff point of 7.30 kPa, sensitivity was 50%, specificity 88%, PPV 95%, and NPV 27%, with an overall diagnostic accuracy of 43%. For differentiating no or mild fibrosis (F0–F1) from clinically significant fibrosis (F2–F4), the AUROC was 0.63, with a balanced cutoff of 7.80 kPa, yielding a sensitivity of 53%, specificity 76%, PPV 61%, NPV 69%, and diagnostic accuracy of 36%. For identifying advanced fibrosis (F3–F4) versus none to moderate fibrosis (F0–F2), the AUROC was 0.67. At a balanced cutoff of 8.00 kPa, sensitivity was 67%, specificity 76%, PPV 40%, NPV 90%, and diagnostic accuracy 33%.

Table 2.

Diagnostic Performance of LSM for fibrosis staging in children with MASLD.

Fibrosis stage	AUROC	Balanced threshold (kPa)	Sensitivity	Specificity	PPV	NPV	Diagnostic accuracy
0 versus 1–4	0.70 (0.57, 0.84)	7.3	0.50 (0.38, 0.62)	0.88 (0.62, 0.98)	0.95 (0.88, 0.99)	0.27 (0.16, 0.41)	0.43 (0.33, 0.54)
0–1 versus 2–4	0.63 (0.51, 0.75)	7.8	0.53 (0.36, 0.69)	0.76 (0.62, 0.86)	0.61 (0.42, 0.77)	0.69 (0.56, 0.81)	0.36 (0.26, 0.46)
0–2 versus 3–4	0.67 (0.53, 0.82)	8.0	0.67 (0.41, 0.87)	0.76 (0.64, 0.85)	0.40 (0.23, 0.59)	0.90 (0.80, 0.96)	0.33 (0.23, 0.43)

Note: AUROC, balanced threshold (kPa), sensitivity, specificity, PPV, NPV, and overall diagnostic accuracy for LSM in distinguishing between fibrosis stages using liver histology as the reference standard. Thresholds were derived using the Youden index. Values in parentheses represent 95% confidence intervals.

Abbreviations: AUROC, area under the receiver operating characteristic curve; LSM, liver stiffness measurement; MASLD, metabolic dysfunction‐associated steatotic liver disease; NPV, negative predictive value; PPV, positive predictive value.

Figure 2.

ROC curves LSM by fibrosis stage. ROC curves illustrating the diagnostic accuracy of LSM in differentiating between fibrosis stages. The light blue curve represents nonfibrotic (Stage 0) versus fibrotic (Stages 1–4), the medium blue curve represents no or mild fibrosis (Stages 0–1) versus clinically significant fibrosis (Stages 2–4), and the dark blue curve represents none to moderate fibrosis (Stages 0–2) versus advanced fibrosis (Stages 3–4). AUROC values are displayed for each comparison, highlighting the moderate diagnostic performance of LSM and the overlap in fibrosis stages. AUROC, area under the ROC curve; ROC, receiver operating characteristic; LSM, liver stiffness measurement.

To further assess diagnostic thresholds for advanced fibrosis (F3–F4), cut points were identified to achieve predefined sensitivity and specificity targets. An LSM threshold of 5.7 kPa provided 90% sensitivity for detecting advanced fibrosis, with a specificity of 27%, PPV of 24%, NPV of 93%, and diagnostic accuracy of 40%. Conversely, a threshold of 12.6 kPa provided 90% specificity for advanced fibrosis, with a sensitivity of 26%, PPV of 37%, NPV of 82%, and diagnostic accuracy of 77%.

4. DISCUSSION

In this prospective, multicenter study, we evaluated the diagnostic accuracy of VCTE in assessing liver steatosis and fibrosis in pediatric patients with MASLD. We found no significant association between CAP and histologic steatosis grade and only a weak association between LSM and fibrosis stage. Additionally, the diagnostic accuracy of LSM in differentiating fibrosis stages was modest, with AUROC ranging from 0.63 to 0.70. These findings highlight the challenges in applying VCTE as a reliable, noninvasive diagnostic tool in pediatric MASLD.

In this cohort of children with biopsy‐proven MASLD, CAP was not significantly associated with histological steatosis grade, and diagnostic accuracy for distinguishing grades was poor. Prior pediatric studies have reported variable but generally moderate associations between CAP and liver fat. Runge et al. studied 60 adolescents with severe obesity and reported an AUROC of 0.80 for detecting steatosis >S1 defined by an MRS‐PDFF threshold of >4.14%. CAP values in their cohort also showed wide dispersion at low fat levels and plateaued at higher levels, consistent with signal saturation. ¹⁸ Tas et al., in 82 adolescents with obesity, found a weak correlation with MRI‐PDFF (R² = 0.11) and an AUROC of 0.65. ²³ Anand et al. similarly found a moderate correlation (r = 0.53) and 69% accuracy in 108 children with overweight or obesity, ¹⁷ while Shin et al. noted a comparable correlation (r = 0.49) that was weaker among children with obesity (r = 0.35). ¹⁹ By contrast, Chaidez et al. reported an AUROC of 0.98 for detecting histological steatosis ≥S1 in a mixed‐etiology cohort. This result was largely driven by differences between children with other liver diseases who had no steatosis and normal BMI, and children with MASLD who predominantly had S2–S3 steatosis and high BMI, rather than discrimination across grades within NAFLD. ²⁴ For context, adult biopsy‐controlled studies summarized in the AASLD Practice Guideline on noninvasive assessment report AUROCs for CAP clustering around 0.75–0.82 for detecting ≥S1 steatosis. ²⁵ These findings indicate that CAP has limited utility for assessing the severity of steatosis in pediatric MASLD.

In this study, LSM values demonstrated a general trend of increasing with fibrosis stage, but the correlation was weak and diagnostic accuracy modest. AUROC values for fibrosis staging ranged from 0.63 to 0.70, consistent with prior pediatric studies. A small study from Cincinnati Children′s Hospital limited to pediatric NAFLD found no significant correlation between LSM and fibrosis stage on histology (n = 15) or with magnetic resonance elastography (MRE; n = 27). ²⁰ In contrast, a much larger study from Children′s Hospital Colorado that included a wide spectrum of liver diseases reported only a modest correlation within the NAFLD subgroup (r = 0.33; AUROC 0.70 for advanced fibrosis), whereas performance was stronger in non‐NAFLD cases (r = 0.47; AUROC 0.77). ²⁴ Similarly, Lee et al. evaluated a mixed‐etiology pediatric cohort in which only 11% had MASLD. ²⁶ Although prespecified LSM thresholds for advanced fibrosis performed well in the derivation cohort (AUROC 0.85), diagnostic accuracy fell markedly in the independent validation cohort (sensitivity 70.8%, specificity 65.6%, accuracy 67.1%). This pattern is consistent with adult data showing that LSM performs better in mixed‐etiology or viral‐predominant populations than in MASLD alone. Taken together, our findings reinforce the broader pediatric literature: while LSM increases with fibrosis, its diagnostic accuracy in MASLD remains modest.

Although VCTE is firmly established in adult hepatology guidelines, its role in pediatric MASLD remains uncertain. ²⁷ , ²⁸ Current AASLD guidance recommends CAP and LSM for assessing steatosis and advanced fibrosis in adults, but these recommendations are based entirely on adult data, with insufficient validation in pediatrics. ¹³ In children, CAP has not shown reliable accuracy for grading the severity of steatosis, even though some studies report fair performance for detecting its presence, and LSM has demonstrated only modest diagnostic performance for advanced fibrosis. Studies also indicate that LSM values may be influenced by age, as Tokuhara et al. demonstrated substantial overlap in LSM values between healthy adolescents and those with suspected liver disease. ¹⁶ In our study, the low PPV of 0.40 shows that in children a high LSM is insufficient to confidently diagnose advanced fibrosis. For a noninvasive test to be appropriate for monitoring disease or serving as a clinical trial endpoint, it should demonstrate a strong correlation with the reference standard across the full spectrum of disease severity; in our study, CAP did not show such a relationship with histologic steatosis, and these findings do not support its use for tracking change in pediatric MASLD. Moreover, practical challenges such as body habitus, narrow intercostal spaces, and operator variability further constrain reliability in pediatric populations. ⁶ Consistent with these concerns, the 2025 AASLD Clinical Practice Statement on Pediatric MASLD recommended against the use of VCTE in children, citing insufficient evidence for its reliability or clinical utility. ²⁹ Collectively, the available evidence indicates that despite its established role in adult hepatology, VCTE cannot be recommended as a routine diagnostic tool for pediatric MASLD without further validation.

This study has several notable strengths. The prospective, multicenter design of this study enhances the generalizability of our findings on the diagnostic performance of VCTE in children with MASLD. The use of a well‐characterized cohort from the NASH CRN Database 3, along with standardized protocols for VCTE and histologic assessment and rigorous operator training, further strengthens the validity of our results. However, there are important limitations to consider. We did not evaluate intra‐ or interobserver variability; therefore, undetected operator‐related differences could have contributed to measurement variability. Since the study population was restricted to children with MASLD, we were unable to assess CAP′s ability to distinguish between the presence and absence of steatosis, a key diagnostic challenge in broader clinical contexts. Although our sample size was sufficient to detect overall trends, the relatively small number of participants with advanced fibrosis may have limited our statistical power to detect significant associations between LSM values and fibrosis stages. Furthermore, liver biopsies were performed as part of clinical care and VCTE was obtained at the time of study enrollment, resulting in a median interval of 82 days between assessments. While this reflects real‐world practice across multiple centers, some change in hepatic steatosis may have occurred during this period, as children often begin lifestyle modification after. ³⁰ However, given the short duration between assessments, the slower biological time course of fibrosis progression, and prior evidence showing relative stability of liver fat over similar intervals without targeted intervention, this interval is unlikely to have meaningfully influenced the study findings.

5. CONCLUSION

In conclusion, our findings do not support the use of VCTE as a reliable tool for grading or staging MASLD in pediatric populations. The weak correlations between VCTE measurements and histologic findings raise concerns about its ability to accurately assess disease severity. Strong and consistent correlation with histology is generally considered a prerequisite for a biomarker to be useful in monitoring change over time; therefore, our cross‐sectional results suggest that VCTE may have limited value for tracking changes in steatosis or fibrosis in this population. However, we did not perform longitudinal assessments, and prospective studies are required to directly evaluate this question. Large‐scale pediatric studies that include the full spectrum of fibrosis and use rigorous methodology are needed to accurately determine the diagnostic accuracy of VCTE for fibrosis staging. While VCTE remains a promising technology, its clinical application in pediatric MASLD should be approached with caution, and further research is urgently needed to develop and validate more accurate noninvasive methods for assessing and monitoring liver disease in this population.

Members of the Nonalcoholic Steatohepatitis Clinical Research Network Pediatric Clinical Centers

Baylor College of Medicine, Houston, TX: Paula M Hertel, MD; Alberto Ayala Aguilar, MD; Laurel Cavallo, BS; Donna Garner, CPNP; Krupa R Mysore, MD; Alison Shaw, BS; Mary Elizabeth Tessier, MD; Nicole Triggs, CPNP; Cynthia Tsai, BS.

Cincinnati Children′s Hospital Medical Center, Cincinnati, OH: Stavra Xanthakos, MD; Ana Catalina Arce‐Clachar, MD; Kristin Bramlage, MD; Kim Cecil, PhD; Nicole Chaaban, BS; Marialena Mouzaki, MD; Ann Popelar, MPH, CCRP; Andrew Trout, MD.

Emory University, Atlanta, GA: Miriam Vos, MD, MSPH; Adina Alazraki, MD; Carmen Garcia; Jorge Jara‐Garra; Saul Karpen, MD, PhD.

Indiana University School of Medicine/Riley Hospital for Children, Indianapolis, IN: Jean P. Molleston, MD; Oscar W. Cummings, MD; Kathryn Harlow Adams, MD; Ashley Hartman, CMA; Kelley S. Jackson, RN; Chaowapong Jarasvaraparn, MD; Sandie Kennedy, NP; Ann Klipsch, RN; Wendy Morlan, RN; Emily Ragozzino, CCRC; Kyla Tolliver, MD.

Northwestern University Feinberg School of Medicine/Ann & Robert H. Lurie Children′s Hospital of Chicago: Mark H. Fishbein, MD; Angela Anthony, BA, CRC; Catherine Chapin, MD.

Saint Louis University, St Louis, MO: Ajay K. Jain, MD; Danielle Carpenter, MD; Theresa Cattoor, RN; Paige Puricelli, RN.

University of California San Diego, San Diego, CA: Jeffrey B. Schwimmer, MD; Amy Alba, MPH; Cynthia Behling, MD, PhD; Nidhi Goyal, MD, MPH; Leila Keyvan; Michael S. Middleton, MD, PhD; Rebecca Morfin; Kimberly Newton, MD; Claude Sirlin, MD; Jaret Skonieczny; Patricia Ugalde‐Nicalo, MD, MAS; Karenina Valdez, MD; Andrew Wang, MD.

University of California San Francisco, San Francisco, CA: Ryan Gill, MD, PhD. University of Washington Medical Center and Seattle Children′s Hospital, Seattle, WA: Niviann Blondet, MD; Camila Khorrami, BS; Randolph Otto, MD; Matthew Yeh, MD, PhD; Melissa Young, CCRC.

RESOURCE CENTERS

National Cancer Institute, Bethesda, MD: David E. Kleiner, MD, PhD.

Data Coordinating Center, Johns Hopkins University, Bloomberg School of Public Health, Baltimore, MD: James Tonascia, PhD; Peggy Adamo, BS; Patricia Belt, BS; Jeanne M. Clark, MD, MPH; Jennifer M. DeSanto, RN, BSN, MS; Jill Meinert; Laura Miriel, BS; Emily P. Mitchell, MPH, MBA; Carrie Shade, BA; Jacqueline Smith, AA; Alice Sternberg, ScM; Annette Wagoner; Laura A. Wilson, ScM; Tinsay Woreta, MD, MPH; Katherine P. Yates, ScM.

CONFLICT OF INTEREST STATEMENT

Lauren B. Nichols, Sebastian G. J. Oakes, Cynthia Behling, Kathryn Harlow Adams, Mark H. Fishbein, Paula Hertel, Chao Jarasvaraparn, Marialena Mouzaki, Laura A. Wilson: No conflicts of interest to report. Jean Molleston: Research grants to Indiana University from Mirum, Abbvie, Albireo, Gillead. Claude Sirlin: Reports payment to institution for research grants from ACR, Bayer, Foundation of NIH, GE, Gilead, Pfizer, Philips, Siemens, V Foundation; payment to institution for lab service agreements with OrsoBio, Enanta, Gilead, ICON, Intercept, Nusirt, Shire, Synageva, Takeda; payment to institution for institutional consulting for BMS, Exact Sciences, IBM‐Watson, Pfizer; Personal consulting for Altimmune, Ascelia Pharma, Blade, Boehringer, Epigenomics, Guerbet, Livivos, and Novo Nordisk; payment to self for royalties and/or honoraria from Medscape, Wolters Kluwer, and HealthProMatch; ownership of stock options in Livivos; unpaid advisory board position in Quantix Bio; executive position for Livivos (Chief Medical Officer, unsalaried position with stock options and stock) through June 28, 2023; Principal Scientific Advisor to Livivos (unsalaried position with stock options and stock) since June 28, 2023; payment to self for serving as a speaker for HealthProMatch; support for attending meetings and/or travel from Fundacion Santa Fe, Congreso Argentino de Diagnóstico por Imágenes, Stanford, Jornada Paulista de Radiologia, Ascelia Pharma, RSNA, Sociedad Radiológica de Puerto Rico, Hospital Español Auxilio Mutuo de Puerto Rico, Paris MASH, and Liver Forum; member (no payment) of Data Safety Monitoring board for National Cancer Institute funded Early Detection Research Network; equipment loans to institution from Butterfly, GE, Siemens, and Mayo; provision of contrast material to institution from Bayer. Miriam Vos: Research grants to Emory from Target NASH, Quest, Labcorp, and Sonic Incytes Medical Corp. Consultant to Boehringer Ingelheim, Novo Nordisk, Eli Lilly, Intercept, Takeda, and Alberio. Has stock or stock options in Thiogenesis and Tern Pharmaceuticals. Stavra A. Xanthakos: Research grants to Cincinnati Children′s from Target NASH. Jeffrey B. Schwimmer: Grant support to UC San Diego from Seraphina, Intercept. Consultant for Merck.

Supporting information

Supplemental Figure S1.

Supplemental Table S1.

Supplemental Table S2.

Supplemental Table S3.

DATA AVAILABILITY STATEMENT

Deidentified data for NASH CRN studies are made available by the NIDDK Central Repository at https://repository.niddk.nih.gov/home/.