How Is Fatty Liver Diagnosed? Tests Doctors Use

Fatty liver disease is usually diagnosed through a combination of blood work, imaging, and sometimes specialized fibrosis tests rather than any single definitive exam. Most people first learn they have excess liver fat after a routine ultrasound or abnormal liver enzyme results, but confirming how much fat is present and whether any scarring has developed requires more targeted tools. The diagnostic process has shifted heavily toward noninvasive methods in recent years, and the specific path your doctor follows depends on your risk factors and what they find at each step.

Blood Tests and Liver Enzymes

The first clue often comes from standard blood work. Liver enzymes, particularly ALT, AST, and GGT, are routinely checked in metabolic panels, and elevated levels can signal liver inflammation or fat accumulation. In people with insulin resistance, GGT appears to be a stronger predictor of fatty changes than ALT. One study of young adults with insulin resistance found that higher GGT values carried greater odds of developing hepatic steatosis compared to ALT, with GGT correctly predicting fatty changes in over 90% of insulin-resistant participants.¹

That said, liver enzymes alone are unreliable for diagnosis. Plenty of people with fatty liver have completely normal enzyme levels, and elevated enzymes can stem from many other causes. Enzymes tell you something is irritating the liver; they do not tell you why. This is why doctors treat abnormal liver enzymes as a reason to investigate further, not as a diagnosis in themselves.

Fibrosis Scoring Systems

Once fatty liver is suspected, the most critical question becomes whether the liver has started developing fibrosis, the scarring that can progress to cirrhosis. Doctors often begin with simple scoring formulas calculated from routine blood work and basic clinical data. The most widely used are FIB-4 (which factors in age, ALT, AST, and platelet count), the NAFLD Fibrosis Score or NFS (which adds BMI, diabetes status, and albumin), and APRI (which uses AST and platelets). These cost nothing beyond the blood draw you’ve already had.

For predicting liver-related events over time, FIB-4 and NFS consistently outperform APRI. A systematic review found that FIB-4 and NFS both achieved strong prognostic accuracy for liver-related outcomes, while APRI lagged behind, especially for predicting mortality.² These scores work best as a first filter: a low FIB-4 score reliably rules out advanced fibrosis, sparing you further testing, while a high score flags you for more detailed evaluation.

The limitation is in the middle. A large proportion of people land in an “indeterminate” zone where the score cannot confidently say yes or no. Population-level screening studies have also found that FIB-4 and NFS carry a meaningful risk of overdiagnosis and a non-trivial rate of false negatives, particularly in people who already have risk factors for liver disease.³ So these scores are useful gatekeepers, but they are not final answers.

The Two-Tier Approach

Because of those indeterminate results, many liver specialists now recommend a two-step process. You start with a simple, inexpensive test like FIB-4. If it comes back low, you’re monitored but not referred for further workup. If it’s indeterminate or high, you move to a second-tier test, typically either transient elastography (FibroScan) or the Enhanced Liver Fibrosis (ELF) blood test. This approach catches people who need specialist care while keeping everyone else out of unnecessary referrals.⁴

The ELF test measures three proteins in the blood that are directly involved in the scarring process. It performs well in fatty liver disease, with studies showing diagnostic accuracy ranging from 0.78 to 0.97 for detecting advanced fibrosis.⁵ In studies of patients with alcohol-related liver disease, ELF performed comparably to another specialized blood test called FibroTest, both achieving high accuracy.⁶

Economic evaluations in the UK have found that using FIB-4 followed by either ELF or transient elastography cuts unnecessary specialist referrals by roughly a third to 40% compared to standard care, while also reducing false diagnoses and saving money per patient.⁷ A separate economic model estimated that the stratification pathway cost only about £2,138 per quality-adjusted life year gained, well within what health systems consider worthwhile.⁸

Standard Ultrasound

Conventional ultrasound is often the first imaging test ordered when fatty liver is suspected, partly because it is cheap, widely available, and involves no radiation. A fatty liver appears brighter than normal on ultrasound because fat reflects sound waves differently than healthy tissue. Radiologists look for increased echogenicity compared to the kidney, along with other visual signs of fat infiltration.

The catch is sensitivity. Standard B-mode ultrasound struggles to detect mild steatosis, defined as fat in just over 5% of liver cells. Studies report sensitivity of only about 61% to 65% for this early stage.⁹ It gets better as fat content rises, but a normal-looking ultrasound does not guarantee a fat-free liver, especially in people who are overweight or obese. This is why guidelines for children with obesity emphasize that a normal ultrasound combined with normal liver enzymes does not rule out fatty liver and that continued monitoring is warranted.¹⁰

FibroScan and Controlled Attenuation Parameter

FibroScan is the brand name for a device that performs transient elastography, and it has become one of the most commonly used tools in fatty liver workups. It does two things at once. Its liver stiffness measurement (LSM) gauges fibrosis by sending a vibration wave through the liver and measuring how fast it travels: stiffer liver, faster wave, more scarring. Its controlled attenuation parameter (CAP) estimates fat content by measuring how quickly the ultrasound signal fades as it passes through tissue.

For fibrosis, FibroScan’s accuracy improves as scarring becomes more severe. In patients with fatty liver and abnormal blood sugar, the accuracy for detecting the earliest stage of fibrosis was moderate, but it climbed steadily, reaching an area under the curve above 0.90 for cirrhosis.¹¹ A multicenter study confirmed that reliable readings were obtainable in roughly 85% to 90% of patients using either the standard M probe or the larger XL probe designed for people with higher BMI. The XL probe produced slightly different cut-off values but comparable overall accuracy.¹²

For fat quantification, CAP correlates with the degree of steatosis. One large study found that CAP distinguished significant steatosis from minimal fat with an area under the curve of 0.81, with about 76% sensitivity and 79% specificity at a threshold of 283 dB/m.¹³ Another study found even better performance for more advanced fat grades, with the area under the curve climbing to 0.93 for severe steatosis.¹⁴

One weakness worth knowing: CAP tends to overestimate fat. In one study, the CAP score flagged over 80% of patients with only mild-to-moderate steatosis as having advanced steatosis, meaning it is better at confirming fat is present than at pinpointing exactly how much.¹⁵

Shear Wave Elastography

Transient elastography is not the only way to measure liver stiffness. Two other ultrasound-based approaches, point shear wave elastography (pSWE) and two-dimensional shear wave elastography (2D-SWE), are built into conventional ultrasound machines rather than requiring a standalone device. These techniques generate shear waves inside the liver and track their speed using real-time imaging, providing a stiffness map of a broader area of liver tissue.

A meta-analysis comparing pSWE and transient elastography in fatty liver patients found both performed similarly. For detecting significant fibrosis, advanced fibrosis, and cirrhosis, pSWE achieved areas under the curve of 0.86, 0.94, and 0.95, while transient elastography reached 0.85, 0.92, and 0.94 for the same stages.¹⁶ A head-to-head study that also included supersonic shear imaging (SSI, a form of 2D-SWE) found that SSI had the highest accuracy for detecting significant fibrosis, outperforming both FibroScan and another technique called ARFI.¹⁷

The practical advantage of shear wave elastography is convenience. If you’re already getting a liver ultrasound, these measurements can be added during the same appointment, on the same machine, without booking a separate FibroScan visit.

MRI-Based Methods

When precise fat measurement matters, MRI is the top option. MRI-derived proton density fat fraction (MRI-PDFF) is now considered the most accurate and precise noninvasive method for quantifying liver fat.¹⁸ It works by using the different magnetic properties of water and fat to calculate the exact percentage of fat in the liver. A study of patients with obesity found a strong correlation between MRI-PDFF and biopsy results, with MRI-PDFF achieving about 91% sensitivity and 86% specificity for identifying moderate-or-worse steatosis at a threshold of roughly 17%.¹⁹ Because of this precision, MRI-PDFF has become the preferred tool in clinical trials for fatty liver treatments, where researchers need to detect small changes in fat content over time.²⁰

For fibrosis, MRI has its own stiffness tool: magnetic resonance elastography (MRE). MRE sends vibrations into the liver during a standard MRI scan and creates a color-coded map of tissue stiffness. A meta-analysis using individual patient data found areas under the curve of 0.88 for significant fibrosis, 0.93 for advanced fibrosis, and 0.92 for cirrhosis, and these numbers held up regardless of sex, obesity, or the underlying cause of liver disease.²¹ In a direct head-to-head comparison with transient elastography, MRE showed higher accuracy, particularly for detecting advanced fibrosis and cirrhosis, where transient elastography’s performance dropped off.²²

A comprehensive meta-analysis that compared all available methods concluded that MRE and shear wave elastography had the highest diagnostic accuracy for staging fibrosis in fatty liver disease, outperforming blood-based scores and FibroScan alike.²³ The main barriers to wider use are cost and availability. MRI scanners are expensive to operate, MRE requires special hardware attachments, and many community clinics simply do not have access to either.

CT Scanning

CT scans are not typically ordered specifically for fatty liver, but they detect it incidentally when performed for other reasons. On an unenhanced CT, fat makes the liver appear darker than it should be, and radiologists compare its density to the spleen as a reference. Unenhanced CT is highly specific for moderate-to-severe fat accumulation, meaning if it says you have fatty liver, it is almost certainly right. But its sensitivity is poor, so it misses many cases, especially mild ones.²⁴ Because CT also exposes you to radiation, it is not used as a screening tool. If fatty liver shows up on a CT you had for something else, your doctor will usually follow up with ultrasound or FibroScan rather than ordering more CTs.

Liver Biopsy

Biopsy remains the reference standard. A needle is inserted through your skin into the liver, guided by ultrasound, and a tiny cylinder of tissue is extracted for examination under a microscope. A pathologist can then grade both the amount of fat and the degree of inflammation, scarring, and cell damage, distinguishing simple steatosis from the more aggressive form of the disease that involves active liver cell injury.

That level of detail comes with real drawbacks. Biopsy carries risks including pain, bleeding, and very rarely more serious complications. There’s also sampling error: the needle captures only a tiny fraction of the whole organ, so the sample might not represent what’s happening throughout.²⁵ However, because fatty liver changes tend to be spread fairly evenly through specific zones of the liver, sampling error is less of a problem than it is with some other liver diseases.²⁶

In practice, most people with fatty liver never need a biopsy. It’s reserved for situations where noninvasive tests give conflicting results, where the doctor needs to confirm active inflammation before starting a specific treatment, or where there’s concern about overlapping liver diseases.

Sorting Out the Cause

Diagnosing fatty liver is not just about confirming fat and measuring fibrosis. Doctors also need to determine what’s driving it. The most common cause is metabolic dysfunction, linked to excess weight, insulin resistance, type 2 diabetes, and high triglycerides. But alcohol-related fat accumulation looks identical on imaging and biopsy, and many patients have a mix of both contributors.

Separating metabolic from alcohol-related fatty liver has traditionally relied on patient-reported drinking history and indirect blood markers like the AST-to-ALT ratio and mean corpuscular volume. These are crude tools. A prospective study found that phosphatidylethanol (PEth), a direct alcohol biomarker, was statistically and clinically superior to all of these indirect markers for distinguishing patients whose fatty liver involves significant alcohol consumption from those whose disease is purely metabolic.²⁷ PEth remains positive for several weeks after heavy drinking, making it harder to game than a conversation about drinking habits. This distinction matters for treatment: a patient whose fatty liver is worsened by alcohol needs different guidance than someone whose disease is entirely metabolic.

Who Should Be Screened

Not everyone needs fatty liver testing. Current guidelines focus screening on people at highest risk of having the disease in a form that could progress. People with type 2 diabetes are a priority group, given how common fatty liver is in that population and how much faster fibrosis can advance when diabetes is in the picture.²⁸ People with obesity, metabolic syndrome, or persistently elevated liver enzymes are also commonly screened. If you fall into one of these groups, your doctor may start with a FIB-4 score and a standard ultrasound, then escalate to elastography or specialist referral based on the results.

Genetic Risk and PNPLA3

Genetics are beginning to play a role in the diagnostic workup, though routine genetic testing is not yet standard practice. The most studied gene variant is PNPLA3, specifically the rs738409 variant. Carrying two copies of this variant substantially increases the risk of developing fatty liver and progressing to cirrhosis.²⁹

Where genetic testing shows the most practical promise is in patients with indeterminate fibrosis scores. Researchers found that among people whose FIB-4 landed in the gray zone, those who carried the high-risk PNPLA3 variant and also had diabetes developed cirrhosis at a rate comparable to patients whose FIB-4 was clearly high risk. Routine PNPLA3 genotyping could identify this subgroup as warranting closer monitoring and earlier treatment, rather than sitting in a follow-up limbo.³⁰ Whether this will become part of everyday clinical practice depends on how quickly guidelines adopt it, but it represents a shift toward more personalized risk assessment in fatty liver care.

Diagnosing Fatty Liver in Children

Fatty liver is no longer an adult disease. Rising obesity rates in children have made pediatric fatty liver disease increasingly common, and the diagnostic approach is not simply a miniature version of the adult workup. The European Society for Paediatric Gastroenterology, Hepatology and Nutrition notes that the same surrogates used in adults, including ultrasound, liver enzymes, and serum fibrosis markers, are used in children, but with important caveats. Enzyme levels and fibrosis thresholds validated in adults may not apply directly to growing bodies, and a single normal set of results does not rule out the disease.¹⁰ Biopsy remains the reference standard for confirmation, but given its invasiveness, clinicians are even more reluctant to use it in young patients than in adults. Most pediatric fatty liver diagnoses are made clinically, based on the combination of obesity, blood work, and imaging, with biopsy reserved for unclear or concerning cases.