Fatty liver shows up on ultrasound as a brighter-than-normal liver, a pattern caused by fat droplets inside liver cells scattering the ultrasound beam more than healthy tissue does. This increased echogenicity is the hallmark finding, and when it is pronounced enough, it also obscures deeper structures like the diaphragm and portal vein walls. Though ultrasound cannot measure fat percentage with the precision of MRI, its accessibility and low cost have made it the frontline imaging tool for detecting and roughly grading hepatic steatosis. The imaging patterns are more varied than many people expect, and the technology behind them has evolved considerably in the past decade.
What a Fatty Liver Looks Like on Standard Ultrasound
In a normal liver, the tissue has a similar brightness (echogenicity) to the kidney cortex sitting next door. When fat accumulates in liver cells, the tissue becomes noticeably brighter than the adjacent kidney. This brightness difference is the single most recognizable sign of steatosis on a conventional ultrasound scan. The fat-laden liver also tends to attenuate the ultrasound beam as it travels deeper, meaning the far edge of the liver and the diaphragm behind it can look washed out or disappear entirely in more advanced cases.
A sonographer evaluating for fatty liver looks at several things simultaneously: how bright the liver parenchyma is compared with the right kidney, whether the walls of the intrahepatic portal vein branches are still visible, and whether the diaphragm can still be clearly seen through the liver. These visual landmarks form the basis of the grading system used in clinical practice.
The Grading Scale for Steatosis
Most radiology departments use a four-tier system, scored from 0 to 3, to categorize what they see. A score of 0 means the liver looks normal. Grade 1 (mild) means there is a slight, diffuse increase in liver brightness, but the diaphragm and portal vein walls are still clearly visible. Grade 2 (moderate) means the liver is moderately brighter, the portal vein walls start to blur, and the diaphragm becomes harder to make out. Grade 3 (severe) means the liver is markedly bright, with poor or no visualization of the portal vein wall, the diaphragm, or the deeper portions of the right lobe.1PubMed Central. Ultrasound-based techniques for the diagnosis of liver steatosis
This grading system is intuitive and quick, which is why it persists in everyday practice. But it relies heavily on the sonographer’s subjective judgment. Two experienced operators scanning the same patient can occasionally disagree on whether the liver is grade 1 or grade 2, particularly in borderline cases. The system also struggles to detect very early fat accumulation, when the liver is only slightly brighter than normal and the difference from the kidney is subtle.
The Hepatorenal Index and More Objective Measurement
To reduce the subjectivity problem, researchers developed the hepatorenal index (HRI), a ratio of the liver’s brightness to the kidney’s brightness measured from the same ultrasound image. Rather than relying on a sonographer’s eye, the HRI is calculated by placing measurement regions on the liver and the adjacent kidney cortex and comparing their pixel intensities. A higher ratio means a brighter liver relative to the kidney, which in turn suggests more fat.
One study comparing the HRI against biopsy-confirmed fat percentages found a strong correlation between the two. An HRI of about 1.28 or above had perfect sensitivity for detecting more than 5% liver fat, meaning it caught every case, though specificity was lower at 54%. The researchers estimated that using the HRI as a screening step could have spared roughly a third of patients from unnecessary biopsies.2PubMed. Hepatorenal index as an accurate, simple, and effective tool in screening for steatosis A separate pathology-correlated study found that an HRI at or below 1.17 predicted less than 5% fat with 97% specificity, while an HRI of 1.4 or above predicted 10% or more fat with 99% specificity, with an indeterminate zone in between where clinical judgment or biopsy may still be needed.3PubMed Central. Ultrasound Stratification of Hepatic Steatosis Using Hepatorenal Index
The HRI is a useful add-on to standard scanning, but it has a practical limitation: it requires a visible right kidney for comparison. In some patients, the kidney may be difficult to image clearly due to body habitus or anatomical variation, which makes the ratio unreliable.
Atypical Patterns That Can Cause Confusion
Not all fatty liver looks like a uniformly bright organ. Fat can deposit unevenly, producing patterns that trip up even experienced clinicians. Two of the most common atypical appearances are focal fatty infiltration (a patch of fat in an otherwise normal-looking liver) and focal fatty sparing (a patch of normal-looking tissue surrounded by a diffusely fatty liver). Both can mimic tumors on ultrasound because they appear as areas that look different from the surrounding parenchyma.
Focal fatty sparing tends to occur in predictable locations, particularly around the gallbladder fossa and near the portal hilum in segment 4, though it can appear in virtually any segment.4PubMed Central. Focal fatty sparing areas of the pediatric steatotic liver: pseudolesions on hepatobiliary phase magnetic resonance images On contrast-enhanced ultrasound, these areas usually enhance the same as the surrounding liver in the later venous and delayed phases, which helps distinguish them from actual masses. But the picture is not always tidy. One case report documented nodular focal fatty sparing that mimicked hepatocellular carcinoma even on contrast-enhanced imaging, requiring additional workup before the correct diagnosis was reached.5PubMed Central. Nodular focal fat sparing of liver mimicking hepatocellular carcinoma in contrast-enhanced ultrasound: A case report
When researchers looked more closely at how these focal fatty variations behave on contrast-enhanced ultrasound, they found that a quarter or more of focal fat-sparing areas appeared darker than the surrounding liver in the arterial phase, not uniformly matching the surrounding tissue the way guidelines traditionally describe.6PubMed. Contrast-Enhanced Ultrasound Quantification Assessment of Focal Fatty Variations in Liver Parenchyma This means the old teaching that “if it enhances the same as the surrounding liver, it’s just focal fat” is an oversimplification. In ambiguous cases, MRI with chemical-shift sequences or proton density fat fraction mapping is the problem-solving tool.
What Doppler Adds to the Picture
Beyond the brightness of the liver tissue itself, ultrasound can evaluate blood flow using Doppler techniques, and fatty liver changes the flow patterns in detectable ways. The hepatic veins, which drain blood from the liver back toward the heart, normally show a characteristic triphasic waveform on Doppler: the flow undulates in sync with the cardiac cycle, reflecting the way pressure waves from the heart travel backward through the venous system. In a fatty liver, this waveform often flattens.
Studies have consistently shown a high rate of abnormal hepatic vein waveforms in patients with steatosis. In one study, about 43% of patients with diffuse fatty infiltration had an abnormal (biphasic or monophasic) waveform, compared with just 2% of healthy controls.7PubMed. Hepatic vein Doppler waveform in patients with diffuse fatty infiltration of the liver Another found abnormal waveforms in over 55% of patients with nonalcoholic fatty liver disease versus about 3% of controls, along with reduced portal vein flow velocity and a lower vascular pulsatility index, suggesting decreased vascular compliance throughout the liver.8PubMed Central. Comparison of portal vein doppler indices and hepatic vein doppler waveform in patients with nonalcoholic fatty liver disease with healthy control A third study found the proportion of abnormal waveforms climbing from 10% in mild steatosis to 80% in severe steatosis.9PubMed. Correlation of hepatic vein Doppler waveform and hepatic artery resistance index with the severity of nonalcoholic fatty liver disease
An important caveat: the degree of fat on standard brightness imaging and the severity of the Doppler waveform abnormality do not always move in lockstep. Some patients with heavily fatty livers maintain normal triphasic flow, and some with only moderate fat show flattened waveforms. So Doppler is a supportive clue, not a standalone grading tool.
Newer Quantitative Ultrasound Techniques
The biggest limitation of conventional ultrasound for fatty liver has always been its qualitative nature. Calling a liver “grade 2” does not tell you whether 15% or 30% of the liver cells contain fat. In response, ultrasound manufacturers have developed quantitative tools that measure how much the ultrasound beam weakens as it passes through fatty tissue.
Attenuation imaging (ATI) is one of the most studied of these tools. It measures the rate at which ultrasound signal strength drops with depth, expressed in decibels per centimeter per megahertz. Fat absorbs and scatters more ultrasound energy, so a fattier liver produces a steeper attenuation curve. One study comparing ATI against liver biopsy found near-perfect discrimination between normal and fatty livers, with an area under the curve of 0.97 for detecting any steatosis and 0.99 for moderate-or-worse steatosis.10PubMed Central. Attenuation Imaging with Ultrasound as a Novel Evaluation Method for Liver Steatosis A large prospective multicenter study using MRI-based fat fraction as the reference found that ATI outperformed the controlled attenuation parameter (CAP) from FibroScan at every steatosis grade, with areas under the curve of roughly 0.90 to 0.94 across mild, moderate, and severe categories.11PubMed Central. Diagnostic performance of attenuation imaging versus controlled attenuation parameter for hepatic steatosis with MRI-based proton density fat fraction as the reference standard
Another emerging approach is ultrasound-derived fat fraction (UDFF), which attempts to estimate the actual percentage of fat in the liver using speed-of-sound or backscatter techniques. When compared head-to-head with MRI proton density fat fraction, UDFF showed good agreement, with areas under the curve above 0.85 for mild steatosis and above 0.95 for moderate and severe steatosis at specific cutoff values.12PubMed. Comparing ultrasound-derived fat fraction and MRI-PDFF for quantifying hepatic steatosis: a real-world prospective study These quantitative approaches are still finding their footing in routine clinical workflows, but they represent a genuine step toward making ultrasound a measurement tool rather than just a pattern-recognition exercise.
Sensitivity Limits and Where Ultrasound Falls Short
Ultrasound’s overall sensitivity for detecting fatty liver is good but not perfect. When the liver–kidney brightness difference was used as the primary criterion and MRI fat fraction served as the reference standard, one study found a sensitivity of about 97% and a negative predictive value of about 98% for identifying any steatosis. In practical terms, a normal-looking liver on ultrasound is very reassuring that meaningful fat is not present.13Scientific Reports. Comparison of conventional sonographic signs and magnetic resonance imaging proton density fat fraction for assessment of hepatic steatosis Specificity, however, was lower at around 75%, meaning the test sometimes flags livers as fatty when they are not, particularly when other conditions increase liver echogenicity.
The presence of fibrosis complicates things further. In patients with fibrosis, one study found that ultrasound’s sensitivity for detecting coexisting steatosis dropped to about 75% and specificity fell to 60%.14PubMed Central. Diagnostic Performance of Ultrasonography in Detecting Fatty Liver Disease in Comparison with Fibroscan in People Suspected of Fatty Liver The same study found no significant correlation between ultrasound-determined fat grade and fibrosis stage, reinforcing that these are two separate processes that do not necessarily march in parallel. Elastography-based stiffness measurements, meanwhile, can be confounded by fat and inflammation: fat tends to lower measured stiffness while inflammation raises it, and when both coexist the net effect is unpredictable.15PLOS ONE. Quantitative ultrasound, elastography, and machine learning for assessment of steatosis, inflammation, and fibrosis in chronic liver disease
Body size is another real-world obstacle. Ultrasound image quality deteriorates as body mass index rises, because the sound waves must travel through more subcutaneous fat before even reaching the liver. In a study of imaging quality across different BMI ranges, ultrasound quality scores worsened progressively above a BMI of about 35, though the use of newer high-performance probes partially offset the degradation.16PubMed Central. The application of high-performance ultrasound probes increases anatomic depiction in obese patients This matters because the patients most likely to have fatty liver are often the same patients in whom ultrasound performs worst.
Fatty Liver Imaging in Children
Pediatric fatty liver disease has become increasingly common, and ultrasound is usually the first test ordered for children because it avoids radiation and sedation. The imaging patterns are broadly the same as in adults: diffuse increased echogenicity is the norm, but focal, heterogeneous, and even multinodular forms can appear, sometimes creating diagnostic confusion with pediatric liver tumors.17PubMed Central. Imaging patterns of fatty liver in pediatric patients
Quantitative tools show promise in this population. A prospective study of children using both attenuation coefficient and hepatorenal index measurements found that an HRI of about 2.1 achieved perfect sensitivity and specificity for identifying moderate steatosis. The attenuation coefficient similarly discriminated well, with a threshold of 0.60 dB/cm/MHz reaching 80% sensitivity and 98% specificity for moderate fat.18PubMed. Liver Fat Quantification by Ultrasound in Children: A Prospective Study Machine learning models trained on ultrasound texture features have also shown accuracy in differentiating pediatric fatty liver from normal in external validation cohorts, outperforming simpler intensity-based indices like the HRI alone.19PubMed. Digital image analysis of ultrasound images using machine learning to diagnose pediatric nonalcoholic fatty liver disease
Artificial Intelligence and Automated Grading
Human grading of liver fat on ultrasound is inherently subjective, so there has been strong interest in training AI systems to do it. Deep learning algorithms can now take standard B-mode ultrasound images and produce steatosis scores without requiring any special hardware. One large-scale deep learning study reported areas under the curve of 0.85, 0.91, and 0.93 for classifying mild, moderate, and severe steatosis, respectively. The algorithm matched or outperformed FibroScan’s controlled attenuation parameter on histology-proven cohorts.20PubMed Central. Accurate and generalizable quantitative scoring of liver steatosis from ultrasound images via scalable deep learning A separate deep learning system designed to classify severity into three grades using standard ultrasound images similarly demonstrated strong discrimination across grades.21PubMed Central. Deep learning for abdominal ultrasound: A computer-aided diagnostic system for the severity of fatty liver
The broader promise of AI in liver ultrasound goes beyond fat grading. Researchers are exploring its use for fibrosis staging, distinguishing benign from malignant focal lesions, and even predicting treatment response in liver cancer.22PubMed Central. Artificial intelligence in liver ultrasound For fatty liver specifically, the appeal is clear: if a standard ultrasound machine can produce a reliable, reproducible fat score without a specialized probe or expensive add-on software, screening becomes far more practical in primary care settings.
Screening Recommendations and Who Should Get Scanned
Major diabetes societies recommend that all patients with type 2 diabetes be screened for fatty liver disease, given the strong overlap between the two conditions.23PubMed Central. Nonalcoholic Fatty Liver Disease Screening in Type 2 Diabetes Mellitus Patients in the Primary Care Setting A proposed screening algorithm for primary care uses liver enzymes and abdominal ultrasound as the initial tier, followed by clinical prediction rules for fibrosis risk if steatosis is detected, and referral to a specialist when advanced fibrosis is suspected.24PubMed Central. Screening for Nonalcoholic Fatty Liver Disease in the Primary Care Clinic The logic is straightforward: ultrasound is cheap, available, and has a high negative predictive value. If it shows no fat, you can move on. If it shows fat, you know to dig deeper for fibrosis.
Point-of-care ultrasound (POCUS), performed at the bedside by the treating clinician rather than in a dedicated imaging suite, is gaining traction as a screening tool for fatty liver. A systematic review found that POCUS examinations were roughly four minutes faster on average than formal ultrasound studies, with potential cost savings as well, though the authors noted that more research is needed to fully quantify the economic impact.25Archives of Gastroenterology Research. Exploring the Use of Point of Care Ultrasound in Screening for Non-Alcoholic Fatty Liver Disease: A Systematic Literature Review and Meta-Analysis For a condition as common as fatty liver, even modest savings in per-scan cost and scheduling logistics add up when multiplied across millions of at-risk patients.
When Stiffness Measurements Accompany Fat Assessment
Because fatty liver can progress to inflammation and fibrosis, many ultrasound exams now bundle a steatosis assessment with an elastography measurement of liver stiffness in the same sitting. Shear wave elastography and transient elastography (the technology behind FibroScan) both estimate stiffness by measuring how fast mechanical waves travel through the liver, with stiffer tissue transmitting waves faster. In a head-to-head comparison in patients with fatty liver disease and a median BMI of about 31, transient elastography was somewhat better than two-dimensional shear wave elastography for detecting fibrosis overall. Both techniques performed well for more advanced fibrosis stages, with areas under the curve above 0.84 for stage 2 or higher.26SAGE Journals (Therapeutic Advances in Gastroenterology). Diagnostic accuracy of two-dimensional shear wave elastography and transient elastography in nonalcoholic fatty liver disease
The practical takeaway is that a single ultrasound visit can now assess both fat content and tissue stiffness, giving clinicians a more complete snapshot of liver health. This matters because fat alone is relatively benign, while fat plus fibrosis carries a substantially higher risk of progression to cirrhosis and liver failure. The combination of a quantitative fat score and a stiffness measurement is increasingly seen as the minimum information a clinician needs to decide whether a patient requires monitoring, lifestyle intervention, or referral.