What is the clinical evidence from this study?

Study design: Observational. Population: Metabolic dysfunction-Associated Steatotic Liver Disease (MASLD) (n=40). Intervention: AI-powered ultrasound radiofrequency analysis vs. MRI-PDFF. Primary outcome: Correlation with MRI-PDFF (R²) (R² 0.81).

What question did this study set out to answer?

The goal is to develop AI methods for accurately quantifying liver fat in children with MASLD using ultrasound data.

March 19, 2026Scientific Reports0 citationsOpen Access

AI-powered ultrasound radiofrequency analysis for non-invasive pediatric liver fat quantification

GCGayoung Choi SHSungwon Ham BJBo-Kyung Je

Key Result

AI-powered ultrasound radiofrequency analysis combined with attenuation parameters and ALT accurately quantified pediatric liver fat, achieving an R² of 0.81 compared to MRI-PDFF.

Key Points

The goal is to develop AI methods for accurately quantifying liver fat in children with MASLD using ultrasound data.
Enrolled 40 pediatric patients referred for MASLD.
Conducted simultaneous ultrasound radiofrequency data collection and MRI-PDFF examinations.
Created radiomics and deep learning models that used various RF data inputs and clinical parameters.
Best radiomics model (XGBoost) and deep learning model (Mod-MHDNet) achieved R2 of 0.81 and 0.76, respectively.
Both models showed excellent agreement with MRI-PDFF within a mean bias of < 0.4% points.

Study Design

Type

Observational (n=40)

Multicenter

Structured PICO

Does AI analysis of ultrasound radiofrequency data accurately quantify liver fat compared to MRI-PDFF in pediatric patients with suspected MASLD?

Population

Pediatric patients referred for clinical evaluation of suspected liver steatosis (MASLD), mean age 12.16, 65% male

Intervention

Artificial intelligence (AI) analysis of ultrasound radiofrequency (RF) data combined with ultrasound-guided attenuation parameters (UGAP) and clinical parameters (ALT)

Comparator

Magnetic resonance imaging-proton density fat fraction (MRI-PDFF)

Outcome

Accuracy and agreement of AI-predicted liver fat fraction compared to MRI-PDFFsurrogate

AI-powered analysis of ultrasound radiofrequency data provides an accurate, radiation-free, and accessible alternative to MRI-PDFF for quantifying liver fat in pediatric patients with MASLD.

Main Result

Effect estimate: R² 0.81

Limitations

Modest sample size
Data from a single center and a single ultrasound platform
Histological confirmation was not available
Agreement analyses were conducted at both the patient and slice levels

Abstract

Metabolic dysfunction-Associated Steatotic Liver Disease (MASLD) affects 30–50% of obese children, yet accurate non-invasive quantification remains challenging. While magnetic resonance imaging-proton density fat fraction (MRI-PDFF) represents the reference standard, its limited accessibility necessitates alternative approaches. Forty pediatric patients (age 12.16 ± 2.56 years) referred for MASLD were prospectively enrolled for same-day ultrasound radiofrequency (RF) data acquisition and MRI-PDFF examination. Two artificial intelligence (AI) approaches using multiple input combinations of RF data, ultrasound-guided attenuation parameters (UGAP), and clinical parameters were developed for non-invasive pediatric liver fat quantification: radiomics-based models and deep learning models. The best radiomics model (XGBoost) and the best deep learning model (Mod-MHDNet) achieved optimal performance with multimodal inputs (R2 = 0.81 and 0.76, respectively). Bland–Altman analysis demonstrated excellent agreement with MRI-PDFF, with a mean bias of < 0.4% points for both approaches. AI analysis of ultrasound RF data enables accurate and accessible quantification of pediatric liver fat, offering a practical alternative for MASLD evaluation.

Demander à l'IA

Bookmark

View Full Paper