Grading of Fatty Liver
Hepatic steatosis should be graded using histology as the reference standard, with MRI-PDFF as the most accurate non-invasive quantitative method, while ultrasound serves as the recommended first-line screening tool despite its limitations in mild disease.
Histology Grading (Reference Standard)
Histological grading remains the gold standard for steatosis assessment 1:
- Grade 0 (S0): <5% of hepatocytes containing fat vacuoles
- Grade 1 (S1, mild): 5-33% of hepatocytes with fat
- Grade 2 (S2, moderate): >33-66% of hepatocytes with fat
- Grade 3 (S3, severe): >66% of hepatocytes with fat
This system provides the definitive classification against which all non-invasive methods are validated.
Ultrasound Grading
Conventional ultrasound is recommended as first-line for steatosis detection 2, though it has significant limitations:
Qualitative Grading 3:
- Mild: Mild diffuse increase in liver echogenicity with clear definition of diaphragm and intrahepatic vessel walls
- Moderate: Mild diffuse increase in liver echogenicity with obscuration of diaphragm and intrahepatic vessel walls
- Severe: Marked increase in liver echogenicity with non-visualization of diaphragm and intrahepatic vessel walls
Performance Limitations:
Ultrasound performs well for moderate-severe steatosis (>30% fat content: 84.8% sensitivity, 93.6% specificity) but poorly for mild steatosis (sensitivity drops to 53.3-65%) 3. The 2021 EASL guidelines explicitly state that non-invasive scores are not recommended for steatosis diagnosis in clinical practice 2.
Quantitative Alternative - Hepatorenal Index:
The hepatorenal index offers superior performance for mild steatosis detection, with excellent correlation to MR spectroscopy and liver biopsy (AUC 99.2-99.6%), and is independent of confounding factors like BMI, inflammation, or fibrosis 3.
FibroScan CAP (Controlled Attenuation Parameter)
CAP values above 275 dB/m can diagnose steatosis with >90% sensitivity 2, though it cannot yet be recommended as first-line given limited availability:
Steatosis Grade Thresholds 4:
- S1 (mild): 248 dB/m
- S2 (moderate): 268 dB/m
- S3 (severe): 280 dB/m
Performance data shows sensitivities/specificities of 78%/79% for S1, 85%/79% for S2, and 83%/79% for S3 3. A 2025 study in NAFLD patients found mean CAP of 271.53 dB/m (moderate steatosis), with patients having BMI ≥25 kg/m² being 3.7 times more likely to have severe steatosis (CAP ≥291 dB/m) 5.
MRI-PDFF (Proton Density Fat Fraction)
MRI-PDFF is the most accurate non-invasive method for detecting and quantifying steatosis 2, recommended primarily for clinical trials due to cost and availability:
Steatosis Grade Thresholds 4:
- S1 (mild): 5% fat fraction
- S2 (moderate): 11-18% fat fraction
- S3 (severe): 16-23% fat fraction
Diagnostic Performance:
A 2017 multi-center NASH trial demonstrated excellent accuracy 6:
- Grades 2-3 vs 0-1: AUROC 0.95 (cutoff 16.3% at 90% specificity, 83% sensitivity)
- Grade 3 vs 0-2: AUROC 0.96 (cutoff 21.7% at 90% specificity, 84% sensitivity)
A 2012 study showed strong correlation (r² = 0.56) between MRI-PDFF and histology, with mean PDFF values of 8.9% (grade 1), 16.3% (grade 2), and 25.0% (grade 3) 7.
Critical caveat: Patients with advanced fibrosis (stage 4) had significantly lower steatosis by both MRI-PDFF (7.6% vs 17.8%) and histology, meaning low hepatic fat does not exclude advanced disease 7.
MRI-PDFF for Treatment Response:
PDFF reduction of ≥5.15% identifies histological steatosis improvement with 90% specificity and 58% sensitivity 6. The 2021 EASL guidelines support using MRI-PDFF to assess steatosis evolution under treatment 2.
Fibrosis Staging
Fibrosis stage, not steatosis grade, is the key predictor of liver-related outcomes and mortality 8:
Non-Invasive Fibrosis Assessment Strategy 2, 9, 8:
First-line screening with FIB-4:
- Low risk (rule-out advanced fibrosis): FIB-4 <1.3 (or <2.0 if age >65)
- Indeterminate risk: FIB-4 1.3-2.67
- High risk: FIB-4 >2.67
Second-line testing for indeterminate/high-risk patients:
- Transient elastography (TE): LSM <8 kPa rules out advanced fibrosis 2
- Enhanced Liver Fibrosis (ELF): <9.8 rules out advanced fibrosis 2
MR Elastography (MRE):
MRE is the most accurate non-invasive fibrosis staging method 2, 4, though only marginally better than other NITs for F3-F4:
- F2: 3.14 kPa
- F3: 3.53 kPa
- F4: 4.45 kPa
The MEFIB score (MRE + FIB-4) provides excellent diagnostic performance: MRE <3.3 kPa AND FIB-4 <1.6 rules out F2 fibrosis 4.
Clinical Pitfalls
Normal liver enzymes do not exclude NAFLD or significant fibrosis - ALT >2× ULN has only 50% sensitivity for NASH and 40% for advanced fibrosis 8
Ultrasound cannot reliably grade mild steatosis - sensitivity drops dramatically for <30% fat content 3, 1
Low steatosis on imaging does not mean mild disease - advanced fibrosis patients often have paradoxically lower fat content 7
FIB-4 has age-dependent cutoffs - use <2.0 (not <1.3) for patients >65 years to avoid false positives 9, 8
Repeat assessment intervals: Every 1-3 years for low-risk patients, annually for advanced-stage NAFLD 2