Longitudinal Performance of Vibration Controlled Transient Elastography in Pediatric Metabolic dysfunction Associated Steatotic Liver Disease

Abstract

Objectives:

Vibration controlled transient elastography (VCTE) is increasingly used to monitor metabolic dysfunction-associated steatotic liver disease (MASLD), yet optimal interpretation of serial controlled attenuation parameter (CAP) and liver stiffness measurements (LSM) in children is undefined. We aimed to characterize longitudinal trends in CAP and LSM and assess whether changes in these measures reflect validated measures of disease regression or progression.

Methods:

Longitudinal retrospective study of youth with MASLD who underwent ≥2 VCTE studies. Disease regression and progression were defined using categorical and continuous changes in alanine aminotransferase (ALT) and gamma glutamyl transferase (GGT) previously linked to histologic improvement or worsening. Spaghetti plots illustrated trends over time, while longitudinal linear mixed-effects models and proportional odds mixed models assessed associations between CAP, LSM, and liver enzyme outcomes.

Results:

The cohort included 149 children (72% male, 43% Hispanic) with a mean age of 14±4 years and a mean follow-up of 1.4 years. CAP and LSM showed marked intraindividual variability. Independent predictors of CAP change included age, body mass index (BMI) percentile, and glycated hemoglobin (HbA1c); predictors of LSM change included ALT, BMI, and HbA1c. Changes in CAP and LSM did not predict categorical outcomes of ALT and GGT reduction, however were significantly positively associated with continuous ALT and GGT outcomes, independent of BMI percentile.

Conclusions:

Longitudinal VCTE in pediatric MASLD has the potential to identify disease progression. The intraindividual variability of CAP and LSM may limit the ability to detect short-term regression or progression, highlighting the need for more precise noninvasive biomarkers for pediatric MASLD.

Introduction

Metabolic dysfunction-Associated Steatotic Liver Disease (MASLD) is the most common cause of chronic liver disease in children, with a prevalence that has been increasing steadily over several decades(1, 2). Pediatric MASLD has strong ties to childhood obesity, poor quality (high-calorie, low-quality nutrient) diets, and sedentary lifestyle(3). MASLD increases risk for advanced liver injury, such as cirrhosis and hepatocellular carcinoma, as well as cardiometabolic comorbidities including type 2 diabetes and dyslipidemia with long-term cardiovascular morbidity(4). Concerningly, pediatric MASLD can progress rapidly (within 1–2 years) despite lifestyle counseling(4, 5) and liver disease contributes to ~50% of the early mortality seen in these patients(4). Therefore, the ability to reliably assess disease severity and its progression over time is of paramount importance.

Given the invasiveness of liver biopsy, the development of accurate non-invasive biomarkers of MASLD severity remains an area of active investigation. Clinically, several imaging modalities have been used for that purpose in pediatrics; however, they have limitations. Conventional ultrasonography has low sensitivity and specificity for MASLD, both in terms of quantifying severity of steatosis and monitoring of fibrosis development and progression(6, 7). Quantitative and semiquantitative ultrasound methodologies, such as ultrasonography with acoustic radiation force impulse imaging (ARFI) remain to be validated in large cohorts of children with MASLD(8). Cross sectional imaging with computed tomography (CT) is avoided given the ionizing radiation exposure(9), while magnetic resonance imaging proton density fat fraction and elastography are costly, not widely available, and more difficult to obtain in young children(10). As a result, clinical practice has been shifting towards the use of point-of-care vibration controlled transient elastography (VCTE), based on extensive cross-sectional and longitudinal validation data in adults (11). However, despite increasing clinical use of VCTE to determine MASLD presence and severity in children, this technique lacks well-validated cut-points in this age range and, importantly, there are no data evaluating the implications of longitudinal changes (12–15).

FibroScan^® (Echosens, Paris, France) is the most commonly used Food and Drug Administration (FDA)-approved VCTE platform that obtains controlled attenuation parameter (CAP) and transient elastography (TE) measurements, as indirect measures of steatosis and liver stiffness, respectively. CAP is expressed in decibels per meter (dB/m) and reflects the severity of steatosis. Liver stiffness measurements (LSM) are obtained in kilopascals (kPa) and predominantly reflect fibrosis. While a high (>12 kPa) LSM is concerning for progression to advanced fibrosis in adults(11), the optimal cutoff for defining an elevated LSM in pediatric MASLD has yet to be established. (16, 17). Furthermore, the degree of change in CAP and LSM that correlates with histologic change (disease progression or regression) over time is not known, nor is the optimal frequency of repeat VCTE evaluations longitudinally.

The objective of this study was to address these gaps by characterizing patterns of change in serial VCTE measures among children and young adults with MASLD and by determining the extent to which changes in CAP and LSM reflect alterations in liver disease severity as assessed by validated changes in serum aminotransferase levels.

Methods:

Ethics statement

The Institutional Review Board at Cincinnati Children’s Hospital Medical Center (CCHMC) (IRB# 2025–0284) approved this study, and all study activities were conducted following that approval. The requirement for informed consent was waived given the retrospective nature of this study.

Study cohort

This was a retrospective cohort study performed at the Steatohepatitis Center of CCHMC. Inclusion criteria were diagnosis of MASLD in children and young adults (up to the age of 25 years) followed by the Steatohepatitis Center at CCHMC between Jan 1, 2020 and April 15, 2025, with availability of at least 2 VCTE (FibroScan^®, Echosens, Paris France) examinations obtained 6 or more months apart. The vast majority of VCTE examinations had been performed by a single experienced nurse and the remaining by a small group of trained nurses. At our institution, patients are asked to arrive fasting in clinic and most often the probe used for the VCTE is the one recommended by the machine at the time of the examination. Patients were excluded if they had any additional liver disease beyond MASLD, significant alcohol consumption, or if they were taking medications with known hepatotoxic potential (e.g., chemotherapeutic/ immunosuppressive agents or antiepileptics, etc.). Lastly, patients with chronic and systemic disorders, such as cancer or chronic inflammatory conditions, were also excluded.

Data Collection

Data were collected from the electronic medical record (EMR). These included demographics (sex and ethnicity), clinical data (anthropometrics including weight, height and body mass index [BMI]), concurrent laboratory data (alanine aminotransferase [ALT], gamma glutamyl transferase [GGT], glycated hemoglobin [HbA1c]) and all available VCTE data. Available clinical and laboratory data of interest were collected within 1 month of the VCTE examination. At our institution, we follow patients with MASLD every 3–6 months, repeat anthropometrics and liver enzymes at every clinical visit and obtain LSM and CAP measurements approximately every 6 months.

Definitions

For this study, liver disease regression was defined as combined reduction in ALT of ≥40% and GGT of ≥20% from baseline, as per Newton et al.(18). Disease progression was defined as combined ALT and GGT increase by ≥20% from baseline(18). Baseline was defined as the date of the first available VCTE. Regression and progression were assessed at all available timepoints (comparing these timepoints to baseline).

Statistical Analyses

Descriptive statistics were used to summarize the study sample, including demographic, clinical, and laboratory characteristics. Categorical variables (e.g., sex, ethnicity) were reported as frequencies and percentages, while continuous variables (e.g., ALT, GGT, body mass index (BMI) z-score, CAP, and LSM) were summarized using means and standard deviations (SD) or medians and interquartile ranges (IQR), depending on distribution. To characterize longitudinal trends in CAP and LSM values across repeated transient elastography studies, we generated individual-level spaghetti plots to visually assess variability and trajectories over time. Univariate and multivariable longitudinal models were developed to examine various outcomes in relation to exposures of interest. In all these longitudinal models, subjects were included as random intercepts to account for within-subject correlation due to repeated measurements, while time between measurements was modeled as a fixed effect. Models were adjusted for age, sex, and race as covariates, with BMI percentile, HbA1c, and ethnicity considered potential confounders. Linear mixed-effects models were used to evaluate continuous outcomes (CAP, LSM, ALT, and GGT), with results reported as β-estimates and 95% confidence intervals (CI). Longitudinal proportional odds mixed models were used to examine changes in ALT/GGT, categorized as disease regression, no-change, or progression based on the aforementioned thresholds from Newton et al. (17), with CAP and LSM as primary exposures. All available follow-up timepoints were included in the models, with subjects specified as random intercepts to account for repeated measures within individuals. Results were presented as cumulative odds ratios (OR) with 95% CI. All analyses were two-sided, with p≤0.05 considered statistically significant, and were performed using SAS version 9.4.

Results

This study included 149 patients who underwent at least two VCTE examinations during the study period. Overall, 72% were male and 43% were Hispanic. The mean age at the first VCTE was 14±4 years, and the mean BMI at that time was 33±5 kg/m². The mean time between the first and last VCTE examinations was 1.4±0.8 years. Additional demographic and clinical characteristics of the study cohort are shown in Table 1.

Table 1:

Demographic and clinical characteristics of the study cohort (n = 149)

Age at first VCTE, years; mean±SD	14±4
Children with multiple VCTE, n (%)
Two VCTE	149 (100%)
Three VCTE	76 (51%)
Four VCTE	33 (22%)
Five or more VCTE	15 (10%)
Male sex, n (%)	108 (72%)
Hispanic ethnicity, n (%)	64 (43%)
Race, n (%)
White	97 (65%)
Black or African American	7 (5%)
Asian	3 (2%)
Not available	42 (27%)
Weight at first VCTE, kg; median (IQR)	84 (71, 106)
Height at first VCTE, cm; median (IQR)	161 (152, 172)
Body Mass Index at first VCTE, kg/m2; median (IQR)	33 (29, 37)
Body Mass Index percentile at first VCTE	98.7 (97.8, 99.5)
ALT at first VCTE, U/L; median (IQR)	61 (41, 109)
GGT at first VCTE, U/L; median (IQR)	33 (21, 55)
HbA1c at first VCTE, %; median (IQR)	5.4 (5.2, 5.6)
Baseline CAP, dB/m; median (IQR)	299 (268, 328)
Baseline LSM, kPa; median (IQR)	6.2 (5.3, 7.9)

Abbreviations: ALT: alanine aminotransferase, AST: aspartate aminotransferase, CAP: controlled attenuation parameter, GGT: gamma glutamyltransferase, HbA1c: glycated hemoglobin, IQR: interquartile range, LSM: liver stiffness measurement, SD: standard deviation

The longitudinal change in CAP and LSM at the individual level is shown in Supplemental Figure 1. The mean CAP and LSM values by age are shown in Supplemental Figure 2.

Among the demographic and clinical characteristics studied, changes in weight and BMI percentiles, as well as change in ALT, GGT and HbA1c were positively associated with change in CAP over time in univariate analyses (Table 2). In multivariable models, Black race was independently associated with a decrease in CAP compared to White race, while age, BMI percentile, and HbA1c remained positively associated with CAP change.

Table 2:

Univariate and multivariable analyses determining associations between change in demographic and clinical variables and change in outcomes of interest (CAP and LSM).

		Univariate results		Multivariable results
Variable	Effect	β-estimate (95% CI)	p-value	β-estimate (95% CI)	p-value
Outcome: Controlled Attenuation Parameter (CAP)
Sex	Female Male (ref category)	−7.2 (−21.5, 7.2)	0.33	−9.2 (−23.6, 5.3)	0.21
Race *	Black Other Unknown White(ref category)	−23.2 (−53.9, 7.6) 25.6 (−13.6, 64.9) 3.4 (−11.1,17.9)	0.14 0.2 0.65	−31.8 (−60.9, −2.7) 13.6 (−22.4, 49.7) 0.3(−17.6, 18.1)	0.03 0.46 0.98
Ethnicity	Hispanic Non-Hispanic (ref category)	−3.3 (−16.3, 9.8)	0.62	−6.7 (−22.9, 9.5)	0.42
Age	-	1.6 (−0.1, 3.4)	0.07	2.2 (0.4, 4.0)	0.02
Weight percentile ‡	-	3.3 (2.3, 4.3)	<0.01	~	~
Height percentile ‡	-	0.05 (−0.15, 0.24)	0.63	~	~
BMI percentile	-	5.5 (3.9, 7.0)	<0.01	4.8 (3.0, 6.6)	<0.01
ALT	-	0.12 (0.06, 0.19)	<0.01	0.06 (−0.02, 0.13)	0.12
GGT	-	0.20 (0.06, 0.34)	<0.01	0.12 (−0.06, 0.3)	0.19
HbA1c	-	17.5 (5.2, 29.7)	<0.01	12.5 (0.9, 24.0)	0.03
Outcome: Liver Stiffness Measure (LSM)
Sex	Female Male (ref category)	−0.7 (−1.4, −0.1)	0.03	−0.2 (−1.0, 0.5)	0.53
Race **	Black Other Unknown White(ref category)	0.02 (−1.45, 1.49) −0.51 (−2.35, 1.34) 0.34 (−0.35, 1.03)	0.98 0.59 0.34	0.21 (−1.35, 1.77) −0.7 (−2.62, 1.22) 0.19 (−0.76, 1.14)	0.79 0.47 0.7
Ethnicity	Hispanic Non-Hispanic (ref category)	0.56 (−0.04, 1.17)	0.07	0.43 (−0.44, 1.29)	0.33
Age	-	0.03 (−0.05, 0.11)	0.48	0.02 (−0.1, 0.12)	0.62
Weight percentile ‡	-	0.09 (0.03, 0.14)	<0.01	~	~
Height percentile ‡	-	0.005 (−0.004, 0.014)	0.29	~	~
BMI percentile	-	0.17 (0.08, 0.26)	<0.01	0.11 (0.0, 0.2)	0.03
ALT	-	0.008 (0.005,0.011)	<0.01	0.01 (0.0, 0.01)	<0.01
GGT	-	0.012 (0.005,0.020)	<0.01	0.01 (0.00, 0.02)	0.27
HbA1c	-	1.22 (0.60, 1.83)	<0.01	0.86 (0.24, 1.48)	0.01

^*Overall p value (univariate 0.23; multivariable 0.14);

^**Overall p value (univariate 0.72, multivariable 0.86).

^‡Not adjusted for in the multivariable models because of collinearity with BMI percentile

Abbreviations: BMI: body mass index, HbA1c: glycated hemoglobin; CI: confidence interval

Factors associated with liver stiffness over time included female sex (associated with lower LSM compared with males) and Hispanic ethnicity, which showed a borderline positive association (Table 2). Changes in weight and BMI percentile, as well as changes in ALT, GGT and HbA1c were also positively associated with LSM over time in univariate analysis. In multivariable models, BMI percentile, ALT, and HbA1c remained independently and positively associated with LSM, while GGT was no longer significant.

Over time, the proportion of patients who improved (disease regression, per the study definition) was 16% at one year (labs available for n=117), 6% at two years (labs available for n=64), and 29% in three years (labs available for n=7). Similarly, the proportion of patients who worsened (showed disease progression) was 13% at one year, 20% at two years and 29% at three years. In longitudinal univariate analyses including all available follow-up timepoints (1–3 years), changes in CAP (OR: 0.99, p=0.54) and LSM (OR: 0.87, p=0.46) were not associated with the cumulative odds of disease regression versus no change and progression (results not shown). In multivariable models, only race was significantly associated with disease regression (Table 3).

Table 3:

Multivariable proportional odds models examining the odds of disease regression versus not change and progression in relation to variables of interest.

Outcome: Regression vs no change or Progression
Exposure: Controlled Attenuation Parameter
	Odds Ratio (95% CI)	p-value
unit change of CAP	0.99 (0.97–1.02)	0.55
unit change of age	1.03 (0.71–1.47)	0.89
unit change of BMI percentile	0.98 (0.72–1.33)	0.91
unit change of HbA1c	2.48 (0.32–19.17)	0.38
Female vs Male	3.72 (0.19–74.50)	0.38
Race		<0.01
Black, Asian, Am Indian, Other vs White	0.001 (0.001–0.058)
Unknown vs White	7.39 (0.20–279.29)
Ethnicity Hispanic vs Non-Hispanic	0.24 (0.01–5.65)	0.37
Outcome: Regression or Progression per study definition
Exposure: Liver Stiffness Measure (LSM)
	Odds Ratio	p-value
unit change of LSM	0.93 (0.56–1.55)	0.78
unit change of age	1.02 (0.71–1.47)	0.89
unit change of BMI percentile	0.98 (0.74–1.31)	0.91
unit change of HbA1c	2.32 (0.32–16.61)	0.40
Sex Female vs Male	4.84 (0.26–90.99)	0.29
Race		0.01
Black, Asian, Am Indian, Other vs White	<0.01 (0.00–0.17)
Unknown vs White	8.04 (0.20–321.28)
Ethnicity Hispanic vs Non-Hispanic	0.25 (0.01–5.93)	0.39

Abbreviations: BMI: body mass index; CI: confidence interval; HbA1c: glycated hemoglobin

Examining ALT and GGT as continuous outcomes (rather than binary cutoffs used in the definition of disease regression/progression), univariate analyses showed that ALT was positively associated with change in BMI percentile (β-estimate: 3.8, p<0.01), HbA1c (β-estimate: 34.0, p<0.01), CAP (β-estimate: 0.25, p<0.01), and LSM (β-estimate: 5.6, p<0.01). In multivariable linear mixed models, the associations of ALT with both CAP and LSM remained significantly positive after adjustment (Table 4). Similarly, GGT over time was positively associated in univariate analyses with age (β-estimate: 2.1, p<0.01), BMI percentile (β-estimate: 1.1, p=0.05), HbA1c (β-estimate: 16.7, p<0.01), CAP (β-estimate: 0.1, p<0.03) and LSM (β-estimate: 1.7, p<0.01). In multivariable models, CAP and LSM remained significantly associated with GGT, while BMI percentile was no longer significant. (Table 4).

Table 4:

Multivariable modeling examining ALT and GGT as outcomes (continuous variables longitudinally) in relation to variables of interest using mixed linear models.

Exposure: Controlled Attenuation Parameter (CAP)
Outcome: Alanine aminotransferase (ALT)			Outcome: Gamma glutamyltransferase (GGT)
Effect	β-estimate (95%CI)	p-value	Effect	β-estimate (95%CI)	p-value
CAP	0.2 (0.02, 0.41)	0.03	CAP	0.1 (0.0, 0.2)	0.02
Age	−0.3 (−3.8, 3.1)	0.84	Age	2.0 (0.5, 3.5)	0.01
BMI percentile	1.8 (−1.2, 4.9)	0.24	BMI percentile	0.5 (−0.6, 1.5)	0.37
HbA1c	27.6 (6.9, 48.3)	0.01	HbA1c	15.1 (6.9, 23.3)	<0.01
Sex (Female vs. Male)	−21.1 (−50.1, 7.8)	0.15	Sex (Female vs. Male)	−10.6 (−24.4, 3.1)	0.13
Race		0.95	Race		0.65
Black, Asian, AmIndian, Other	4 (−43, 51)		Black, Asian, Am Indian, Other	−1.5 (−23.8, 20.9)
Unknown	−5 (−41, 32)		Unknown	−8.2 (−25.6, 9.2)
Ethnicity (Hispanic vs. Non-Hispanic)	7.8 (−25.4, 41.0)	0.64	Ethnicity (Hispanic vs. Non-Hispanic)	6.2 (−9.7, 22)	0.44
Exposure: Liver Stiffness Measure (LSM)
Outcome: Alanine aminotransferase (ALT)			Outcome: Gamma glutamyltransferase (GGT)
Effect	β-estimate (95%CI)	p-value	Effect	β-estimate (95%CI)	p-value
LSM	6.5 (3.1, 10.0)	<0.01	LSM	2.1 (0.8, 3.3)	<0.01
Age	0.3 (−3.1, 3.6)	0.87	Age	2.4 (0.9, 3.8)	<0.01
BMI percentile	2.2 (−0.6, 5.1)	0.13	BMI percentile	0.7 (−0.3, 1.7)	0.17
HbA1c	22.4 (1.7, 43.1)	0.03	HbA1c	12.9 (4.7, 21.2)	<0.01
Sex (Female vs. Male)	−19.7 (−48.4, 8.9)	0.18	Sex (Female vs. Male)	−10.4 (−24.2, 3.3)	0.14
Race		0.95	Race		0.61
Black, Asian, Am Indian, Other	0.9 (−45.7, 47.4)		Black, Asian, Am Indian, Other	−2.6 (−24.9, 19.7)
Unknown	−5.9 (−42.1, 30.3)		Unknown	−8.6 (−26.0, 8.8)
Ethnicity (Hispanic vs. Non-Hispanic)	3.1 (−29.9, 36.0)	0.85	Ethnicity (Hispanic vs. Non-Hispanic)	4.5 (−11.3, 20.4)	0.57

Abbreviations: BMI: Body Mass Index, HbA1c: glycated hemoglobin, CI: confidence interval

Discussion:

In this study, we aimed to describe longitudinal variations in VCTE-derived CAP and LSM in a real-world pediatric MASLD cohort and to assess their ability to reflect changes in liver disease severity over time. Among 149 predominantly non-Hispanic male adolescents, CAP tended to increase with age, while LSM remained relatively unchanged (Supplemental Figure 2); however, both measures showed marked intraindividual variability across timepoints (Supplemental Figure 1). Categorical ALT and GGT changes, previously linked to histologic disease regression and progression(5), were not associated with CAP and LSM changes to a clinically significant degree, with β-estimates close to zero (Table 2). However, CAP and LSM change were significantly and independently associated with continuous changes in ALT and GGT over time, which also have been shown to correlate with histological improvement versus worsening in cohorts of children with biopsy-confirmed MASLD(19).

VCTE is increasingly used clinically to monitor whether adults with MASLD have progressed or improved over time and to risk-stratify their care(20). However, to our knowledge, this is the first pediatric study assessing the performance of serial VCTE measures for the purpose of assessing change in liver disease longitudinally. Our study cohort was representative of the spectrum of disease severity seen in our clinics and the baseline clinical, laboratory and transient elastography data (median BMI percentile in class I obesity range, ALT 61 U/L, GGT 33 U/L, HbA1c in the normal range, CAP 299 dB/m and LSM 6.2 kPa) were indicative of mild to moderate liver disease severity. Overall, these patients remained relatively stable over the study period, as reflected by the proportions meeting criteria for disease regression and progression (by categorical ALT and GGT change) at the 1-, 2- and 3-year marks. The lack of an association between serial CAP and LSM measures and categorical ALT and GGT outcomes precluded the determination of the relative change in CAP and LSM that predicts categorical measures of regression versus progression. We also observed substantial intraindividual fluctuations in CAP and LSM. Our findings suggest that serial VCTE measurements in individual patients should be interpreted with caution and not relied upon in isolation for clinical decision-making.

When considering ALT and GGT as continuous outcomes, we found that change in CAP and LSM were associated with change in ALT and GGT (the β-estimate being much smaller for CAP than for LSM; Table 4). Whether changes in CAP and LSM perform better in detecting liver disease progression vs. regression over a longer period of time, or in children/young adults with more advanced liver disease at baseline remains to be determined. The majority of our study cohort had repeat VCTE and liver measures at ~1 year of follow up, with <50% having scans or liver outcomes repeated at 2 and 3 years. Additionally, in contrast to adults, the role of serial VCTE measurements in predicting long-term clinical outcomes in children with MASLD has not been examined. In adults, longitudinal changes in LSM have been linked to the risk of liver-related events over a median follow-up of 28 months(21). Specifically, an increase in LSM of >20% has been found to be associated with elevated risk, whereas a decrease of >20% is associated with reduced risk of liver-related events among high-risk patients (baseline LSM >20 kPa)(21). Longitudinal pediatric studies of adequate duration are needed to address these questions.

In this study we found that increase in CAP over time was independently associated with an increase in age, BMI percentile and ALT, while African American race was protective against longitudinal increases in CAP (Table 2). These findings are not surprising, given the known association between hepatic steatosis severity and the aforementioned variables(22). Similarly, increase in LSM was independently associated with an increase in BMI percentile, ALT and HbA1c. The latter is consistent with the findings of other pediatric studies that have highlighted an association between increasing BMI z score, ALT and HbA1c and worsening histology in the context of MASLD(19). While these results do not prove causation, they support the notion that obesity severity and diabetes prevention should remain key treatment objectives for children and young adults with this condition.

Strengths of this study include its longitudinal design and its generalizability given the inclusion of an unselected pediatric cohort with MASLD typical of those referred to pediatric gastroenterology and hepatology practices. The routine implementation of serial VCTE at our institution minimized selection bias and enhances the clinical relevance of our results. Limitations of this study include its retrospective design, which contributed to missing data. Additionally, changes in ALT and GGT over time were used to define disease progression and regression, rather than change in histology. Although validated in pediatric MASLD, these measures are surrogate rather than direct assessments of liver histology(5, 19). Their use may have introduced measurement error, limiting our ability to define VCTE change thresholds that precisely correspond to true changes in liver disease severity. In clinical practice, repeat biopsies are performed only in a small subset of high-risk patients, a group that is not representative of the broader pediatric MASLD population. Finally, although the sample size of patients with at least two VCTE examinations was adequate, relatively few underwent three or more serial studies, and the median follow-up of 1.4 years may have limited our ability to detect longer-term changes in CAP and LSM. Similarly, the sample size for certain subgroups, such as African American children, was small warranting cautious interpretation of race-specific findings.

Conclusions:

In this first longitudinal pediatric study using serial VCTE examinations, we observed considerable intraindividual variability in CAP and LSM. The relative change in CAP and LSM that signify clinically meaningful change still remain to be determined in children. Our findings support further study of serial VCTE over longer periods of time in children to better define the relationship with hepatic and other clinical endpoints. Given the substantial intra-individual variation and the current lack of validated definitions for clinically significant change in VCTE measures (CAP and LSM) in children with MASLD, serial VCTE measurements in individual pediatric patients should be interpreted with caution and not relied upon in isolation for clinical decision-making. The latter should be based on a global patient assessment that takes into consideration factors such as changes in anthropometrics, lifestyle changes, medications, laboratory and imaging findings.

Supplementary Material

Figure 1

Supplementary Figure 1: Individual variation in mean controlled attenuation parameter (CAP; panel A) and liver stiffness measures (LSM; panel B) over time. Solid black lines represent Loess lines.

Figure 2

Supplementary Figure 2: Scatterplots showing the relationship between age and (A) controlled attenuation parameter (CAP, dB/m) and (B) liver stiffness measures (LSM, kPa). Dashed lines represent fitted regression lines.

What Is Known

• Vibration-controlled transient elastography (VCTE) is increasingly used to assess disease severity in patients with metabolic dysfunction associated steatotic liver disease (MASLD).

• Pediatric thresholds for interpreting VCTE results (controlled attenuation parameter [CAP] and liver stiffness measurement (LSM) are still being defined. There are no data evaluating the implications of longitudinal VCTE changes in pediatric MASLD.

What Is New

• Serial VCTE examinations show substantial intraindividual variability in both CAP and LSM.

• Changes in CAP and LSM do not correspond to previously defined biomarkers of MASLD regression or progression.

• Changes in CAP and LSM correlate with liver enzymes.

Data availability statement:

The data that support the findings of this paper are available upon reasonable request from the corresponding author.