Authors: Dong Ho Lee
Categories: Liver Imaging in Focus, Fatty Liver, Metabolic Dysfunction-Associated Steatotic Liver Disease (MASLD), Ultrasound, Quantitative Imaging
Source: Journal of the Korean Society of Radiology
Hepatic steatosis is a key characteristic of metabolic dysfunction-associated steatotic liver disease and may progress to cirrhosis and end-stage liver disease through metabolic dysfunction-associated steatohepatitis in 10%–30% of patients. Consequently, early detection and accurate assessment are essential. Grayscale US has traditionally been used for evaluating hepatic steatosis; however, its limitations include reduced sensitivity for mild cases and subjective interpretation. In addition to controlled attenuation parameter, two-dimensional attenuation-based techniques have been recently introduced to quantify US beam attenuation for assessing hepatic steatosis, demonstrating high diagnostic accuracy and reproducibility while addressing some of the limitations of grayscale US. To enhance diagnostic precision and reduce measurement variability, the standardization of examination protocols, including measurement processes, is necessary. Additionally, further research is needed to explore the role of a multiparametric approach combining attenuation-based techniques with elastography in predicting disease progression. This review examines the principles, recent research findings, and future perspectives on attenuation-based quantitative evaluation of hepatic steatosis using US.
Steatotic liver disease is characterized by an excessive and abnormal accumulation of triglycerides within hepatocytes and represents a key diagnostic feature of metabolic dysfunction-associated steatotic liver disease (MASLD) (1). MASLD is the most common cause of chronic liver disease, with an estimated global prevalence of approximately 30% (23). The incidence of MASLD has risen in parallel with the increasing prevalence of type 2 diabetes and obesity, both of which are closely associated with metabolic syndrome (4). Similar to other chronic liver diseases such as viral hepatitis, MASLD can progress to advanced stages, including liver cirrhosis and hepatocellular carcinoma (HCC). While many individuals with MASLD remain at the stage of simple steatosis, 10%–30% of patients may develop liver fibrosis, cirrhosis, and eventually HCC through an intermediate stage known as metabolic dysfunction-associated steatohepatitis (MASH). MASH is characterized by the coexistence of hepatic steatosis and hepatocellular injury, including lobular inflammation and hepatocyte ballooning.
Currently, MASH-related cirrhosis is the second leading indication for liver transplantation in North America (5), and its contribution to transplantation continues to grow. Early-stage MASLD is recognized as a potentially reversible condition, whereas advanced stages such as cirrhosis are generally irreversible (2). Therefore, early detection and appropriate management are critical to prevent disease progression to cirrhosis and end-stage liver disease (67).
Traditionally, liver biopsy has served as the reference standard for diagnosing and quantifying steatotic liver disease; it is the only method capable of differentiating MASH from simple steatosis and providing a definitive diagnosis of MASH (89). However, the invasive nature of liver biopsy and its associated risks, including potentially life-threatening complications such as bleeding, limit its utility for routine monitoring of MASLD (10). Moreover, liver biopsy samples represent only a small fraction of the total liver volume—typically about 1/50000—raising concerns regarding sampling error. Steatotic liver disease requires regular monitoring as it is a dynamic condition that can potentially progress to fibrosis or cirrhosis; therefore, the limitations of liver biopsy pose significant challenges in clinical practice. Consequently, there are ongoing efforts to develop noninvasive methods for diagnosing and monitoring MASLD.
Among noninvasive imaging modalities, MRI–proton density fat fraction (PDFF), derived using either MR spectroscopy or chemical shift–based techniques, is widely regarded as the most accurate method for quantifying hepatic steatosis. The recently published World Federation for Ultrasound in Medicine and Biology (WFUMB) guidelines also recommend MRI-PDFF as the reference standard noninvasive imaging tool for the quantitative evaluation of hepatic steatosis (1). However, the clinical utility of MRI is limited by its high cost and restricted accessibility, which preclude its routine use in general practice, except in specialized settings such as clinical trials. US, in contrast, offers several advantages including the absence of ionizing radiation, broad availability, and relatively low cost when compared to CT or MRI. In clinical practice, liver US is commonly employed as the first-line imaging modality in patients with suspected liver disease or abnormal liver function tests. Furthermore, in individuals at high risk of HCC, such as those with chronic hepatitis B, chronic hepatitis C, or cirrhosis, clinical guidelines recommend biannual US for monitoring HCC (111213). US is also frequently used to assess hepatic steatosis. In gray-scale US imaging, hepatic steatosis typically manifests as increased parenchymal echogenicity, with the liver appearing brighter. However, gray-scale US is limited by its qualitative nature, operator dependency, and reduced reproducibility. It also demonstrates suboptimal sensitivity in detecting mild steatosis and cannot provide quantitative assessments.
To address these limitations, US attenuation-based quantitative technique have recently been developed and adopted in clinical practice. These techniques allow for the quantitative assessment of hepatic steatosis by measuring the degree of US beam attenuation as it traverses the liver parenchyma. Early studies have shown promising diagnostic performance. In this review, we aimed to present the fundamental principles of US attenuation-based quantitative technique, summarize recent evidence from literature, and explore potential future directions for its clinical application.
Traditionally, gray-scale US has been widely used for evaluating diffuse liver diseases, including MASLD. In particular, US is a safe imaging modality with no associated risks, as it does not involve radiation exposure or the use of contrast agents. Compared to other imaging techniques such as CT or MRI, it is easily accessible and more affordable, making it suitable for repeated use in managing patients with MASLD. The evaluation of hepatic steatosis using gray-scale US is primarily based on alterations in hepatic parenchymal echogenicity. The abnormal accumulation of triglycerides within hepatocytes increases the reflection of US beams, resulting in increased echogenicity of the liver parenchyma (910). Consequently, hepatic steatosis is typically visualized as a “bright liver” on gray-scale US images. For diagnostic purposes, the echogenicity of the liver parenchyma is frequently compared to that of the right renal cortex. In healthy individuals, the echogenicity of hepatic parenchyma is similar to that of the renal cortex. If the liver parenchyma appears brighter than the right renal cortex because of increased echogenicity, it is interpreted as an indication of hepatic steatosis (Fig. 1).
Under normal conditions, the walls of intrahepatic portal veins appear brighter than the surrounding liver parenchyma on gray-scale US, allowing for clear delineation between the two. However, as hepatic steatosis progresses, increased parenchymal echogenicity causes the distinction between echogenic portal vein walls and the surrounding parenchyma to fade (Fig. 1). Along with increased reflection, hepatic steatosis also leads to increased attenuation of the US beam. With increasing degree of steatosis, this attenuation intensifies, resulting in reduced beam penetration to deeper structures such as the diaphragm. Consequently, in cases of advanced steatosis, the diaphragm and other deep anatomical landmarks may be poorly visualized on gray-scale US. These changes in hepatic parenchymal echogenicity enable qualitative assessment of steatosis severity using a four-point grading Grade 0 (normal), hepatic parenchymal echogenicity is similar to that of the right renal cortex; Grade 1 (mild steatosis), hepatic echogenicity is increased compared to the renal cortex, but portal vein walls remain clearly visible; Grade 2 (moderate steatosis), hepatic echogenicity is further increased, and the portal vein walls appear indistinct, though the diaphragm remains visible; and Grade 3 (severe steatosis), marked increase in hepatic echogenicity with significant beam attenuation, resulting in poor visualization of the diaphragm and other deep structures (1415). Although the diagnostic accuracy of the above-mentioned method is inconsistent across studies, it is generally recognized as having good diagnostic performance, particularly in the detection of moderate-to-severe steatosis, with reported sensitivity and specificity of approximately 90% and 95%, respectively (1416).
However, the assessment of hepatic steatosis using gray-scale US has several limitations. First, patients with renal disease may have increased echogenicity of the right renal cortex, complicating the evaluation of hepatic parenchymal echogenicity. Additionally, subjectivity owing to the qualitative nature of the assessment criteria limits its reproducibility. According to a study by Strauss et al. (17), when hepatic steatosis was assessed using the aforementioned four-point scale based on gray-scale US images, intraobserver reproducibility was 54.7%–64.9%, and interobserver agreement was 47.0%–63.7%, indicating moderate reproducibility. Another significant limitation is the reduced diagnostic efficacy of gray-scale US in detecting mild steatosis. Previous studies have reported that gray-scale US has a sensitivity of approximately 55.3%–66.6% and a specificity of 77.0%–93.1% in detecting mild degree hepatic steatosis (10151618). Given the importance of early detection and management of MASLD, the limited sensitivity of gray-scale US in identifying mild steatosis may pose a significant limitation in clinical practice. Nevertheless, gray-scale US is still recommended as the first-line screening imaging modality for patients suspected of having MASLD, particularly those with type 2 diabetes, overweight or obesity, other metabolic risk factors, or abnormal liver function tests (19). This is primarily owing to advantages such as excellent accessibility, lack of radiation or contrast agent exposure, and relatively low cost.
To address the inherent limitations of gray-scale US in the assessment of hepatic steatosis, several novel quantitative imaging techniques have been developed, among which attenuation-based quantification is considered the most significant. In liver US examinations, the US beam must traverse the hepatic parenchyma. As hepatic steatosis progresses, the accumulation of intracellular fat vacuoles leads to increased reflection and attenuation of the US beam. The degree of attenuation is influenced not only by the severity of steatosis but also by the frequency of the US beam, with higher frequencies resulting in greater attenuation. Moreover, the attenuation caused by fat components is greater than that observed in normal hepatic parenchyma. Based on these observations, the fundamental hypothesis of attenuation-based quantification methods is that the degree of hepatic steatosis can be objectively assessed by calculating the extent of attenuation of an US beam of a given frequency as it passes through the liver parenchyma (Table 1).
Controlled attenuation parameter (CAP), obtained during transient elastography (TE) examinations, was the first noninvasive method that enabled quantitative assessment of hepatic steatosis by measuring US attenuation. In TE using the M-probe, the CAP value is derived by calculating the degree of attenuation of an US beam with a central frequency of 3.5 MHz along the depth of liver tissue and is expressed in decibels per meter (dB/m). In human subjects, CAP values typically range from 100 to 400 dB/m, with higher values indicating greater attenuation and, consequently, a more advanced degree of hepatic steatosis. By measuring the area under the receiver operating characteristic curve (AUROC), previous studies reported that CAP has a diagnostic performance of approximately 0.81–0.84 for detecting mild steatosis, 0.85–0.88 for moderate steatosis, and 0.86–0.91 for severe steatosis (202122). Similar to TE, CAP is a one-dimensional (1D) measurement and does not provide gray-scale US images of the liver. Thus, a major limitation is its inability to localize the precise site of measurement within the liver. Furthermore, CAP measurement may be challenging or less accurate in patients with obesity who have an increased skin-to-capsule distance or in patients with ascites, which limits its clinical utility. To improve the success rate and accuracy of CAP measurements in individuals with obesity, the XL-probe, utilizing a lower central frequency of 2.5 MHz to enable deeper US penetration, has been introduced. Studies have reported that the XL-probe achieves an AUROC of approximately 0.82 and 0.75 in detecting mild and moderate steatosis, respectively (23).
In addition to CAP, 2D gray-scale US attenuation-based quantitative technique were recently developed and introduced into clinical practice, with most major US vendors now offering these technologies (Fig. 2). In 2D gray-scale US attenuation-based quantitative technique, the attenuation coefficient is measured and expressed in units of decibels per meter per megahertz (dB/m/MHz). Higher attenuation coefficients indicate greater US beam attenuation, which is suggestive of more advanced degree of hepatic steatosis. In contrast to CAP, 2D gray-scale US attenuation-based quantitative technique simultaneously provides a gray-scale image of the liver, allowing for easy integration with conventional liver US examinations. In addition, the measurement site for attenuation coefficient can be precisely located within the hepatic parenchyma under real-time gray-scale image guidance, offering theoretical advantages in terms of improved diagnostic performance and measurement reproducibility compared to CAP. To accurately calculate the attenuation coefficient, radio-frequency (RF) raw data acquired during the US examination must be used. Conventional US images undergo various corrections, such as depth-dependent intensity compensation and internal gain adjustments, to generate a uniform liver image. However, these corrections must be excluded when computing the attenuation coefficient. All major US vendors currently derive attenuation coefficients from RF raw data, though the center frequency of the US beam, computational methods, and algorithms used differ by manufacturer (1).
A critical initial step in performing 2D gray-scale US attenuation-based quantitative technique is the acquisition of a high quality gray-scale liver US image. Although images of the left hepatic lobe or subcostal scans of the right hepatic lobe may be used for attenuation coefficient measurements, intercostal scans of the right hepatic lobe are generally preferred, as they provide the widest and most homogeneous depiction of the hepatic parenchyma (2). Furthermore, because the accuracy of the attenuation coefficient calculated in 2D gray-scale US attenuation-based quantitative technique is highly dependent on the quality of the acquired RF raw data, obtaining an optimal quality gray-scale liver image is critical for successful evaluation. In most US systems, a gray-scale liver image is first obtained, followed by activation of the 2D gray-scale US attenuation-based quantitative technique mode. Upon activation, a measurement box and region of interest (ROI) are placed within the gray-scale image to perform attenuation coefficient measurement (Fig. 2). It is generally recommended to avoid areas immediately beneath the liver capsule, as these regions are prone to reverberation artifacts, which may compromise measurement accuracy (2). Similarly, deeper liver parenchyma adjacent to the diaphragm should be avoided, as extensive US beam attenuation at greater depths results in reduced signal amplitudes and increased noise, thereby negatively impacting the accuracy of attenuation measurements (2). Moreover, large vessels should be excluded when positioning the measurement box and ROIs, as most US systems automatically identify these regions as signal voids and exclude them during attenuation coefficient calculations. Regarding the optimal number of measurements, most vendors recommend performing three to five measurements, with either the median or mean value of the obtained measurements commonly used for subsequent analysis.
Among the 2D gray-scale US attenuation-based quantitative technique developed by various vendors, Attenuation Imaging (ATI) from Canon was the first to be commercialized. In 2019, a Korean research group first reported the diagnostic performance of ATI in evaluating hepatic steatosis, using MRI-PDFF as the reference standard (24). Since then, at least 20 studies have evaluated the diagnostic accuracy of ATI in assessing hepatic steatosis (1). The attenuation coefficient obtained using ATI has demonstrated robust diagnostic performance, with reported AUROC of 0.76–0.97 for detecting mild steatosis, 0.86–0.99 for moderate steatosis, and 0.79–0.97 for severe steatosis (12526272829). Following the introduction of ATI, US Guided Attenuation Parameter (UGAP) (30313233) from GE and Tissue Attenuation Imaging (TAI) (343536) from Samsung Medison were sequentially commercialized. Subsequent studies have reported that these techniques provide diagnostic performance comparable to that of ATI in evaluating hepatic steatosis. Since the introduction of 2D gray-scale US attenuation-based quantitative techniques by multiple vendors, comparative studies between 2D attenuation imaging techniques and CAP have been actively conducted. Evaluations of ATI (3738394041), UGAP (3033), and TAI (36) have consistently demonstrated that 2D gray-scale US attenuation-based quantitative technique offer diagnostic performance that is at least comparable to, and in some studies superior to, that of CAP, highlighting the potential advantages of these techniques over CAP.
In addition to diagnostic performance, examination reproducibility is an important factor, as the attenuation coefficient obtained through these 2D gray-scale US attenuation-based quantitative techniques is a quantitative value. In 2020, a Korean research group first reported the reproducibility of ATI (42), and since then, multiple studies have assessed the reproducibility of various 2D gray-scale US attenuation-based quantitative techniques. Most of these studies have reported intra-class correlation coefficients >0.8, indicating excellent reproducibility (1). The high reproducibility demonstrated by 2D gray-scale US attenuation-based quantitative techniques represents a substantial advancement over conventional gray-scale liver US, suggesting that these techniques may help overcome the intrinsic subjectivity of gray-scale US in the assessment of hepatic steatosis.
As previously described, 2D gray-scale US attenuation-based quantitative techniques have demonstrated excellent diagnostic performance in noninvasive assessment of hepatic steatosis. However, several limitations remain that warrant consideration. Although prior studies have consistently reported that 2D gray-scale US attenuation-based quantitative techniques offered by various vendors achieve an AUROC >0.8 for the evaluation of hepatic steatosis, the cutoff values for detecting each grade of steatosis exhibit considerable variability. For instance, the reported cutoff values for ATI from Canon for detecting mild steatosis range from 0.59 to 0.69, representing a relatively broad interval. Similarly, the cutoff ranges for moderate and severe steatosis are 0.67–0.78 and 0.68–0.86, respectively. Comparable values have been observed in other 2D gray-scale US attenuation-based quantitative techniques, including the UGAP from GE and TAI from Samsung Medison. In addition to the wide variation in cutoff values, the significant overlap among thresholds presents a critical challenge. For example, if an ATI examination yields an attenuation coefficient of 0.68 in a given patient, the clinical interpretation could differ substantially depending on the reference study used—ranging from normal liver to mild, moderate, or even severe steatosis.
One of the major reasons for the wide variation in reported cutoff values across studies is the lack of standardized and consistent measurement protocols for 2D gray-scale US attenuation-based quantitative techniques during the respective study periods. In practice, the methodologies for attenuation coefficient measurement vary considerably, including differences in the size and placement of the measurement box and ROIs, as well as the scanning techniques used to acquire liver US images. In particular, the depth at which the measurement box and ROIs are positioned can significantly influence the measured attenuation coefficient. Ferraioli et al. (43) reported that deeper placement of the measurement box and ROIs was associated with lower attenuation coefficient values, even in the same patient. This observation suggests that, even when using the same 2D gray-scale US attenuation-based quantitative technique, variations in measurement depth could lead to variation in the cutoff values applied for detecting hepatic steatosis. In addition to the measurement location, liver fibrosis may also affect the attenuation coefficient values obtained using 2D gray-scale US attenuation-based quantitative techniques. Although US beam attenuation within the liver parenchyma is primarily influenced by the degree of steatosis, advanced fibrosis is also known to contribute to increased attenuation. Recently, Kumada et al. (44) reported that the attenuation coefficient measured using UGAP may be elevated even in the absence of hepatic steatosis if advanced fibrosis is present. Given these considerations, the development and implementation of standardized measurement protocols for assessing attenuation coefficient are critical to enhance the reliability and comparability of future studies. Recent findings have indicated that the highest reproducibility for attenuation coefficient measurements using ATI was achieved when a 3-cm measurement box was placed in the right hepatic lobe image obtained via right intercostal scanning, with the upper boundary positioned 2 cm below the liver capsule (45). Based on these results, WFUMB has recently proposed a standardized protocol for attenuation coefficient measurement (Table 2). Accordingly, to enhance measurement accuracy and facilitate comparability across studies, examinations utilizing 2D gray-scale US attenuation-based quantitative techniques are recommended to adhere to this standardized protocol.
In addition to the establishment of standardized measurement protocols, another critical consideration is whether attenuation coefficients obtained using 2D gray-scale US attenuation-based quantitative techniques from a specific vendor can be directly compared or applied across systems from other vendors. Theoretically, differences in the central US frequency and algorithms used to calculate the attenuation coefficient from RF raw data among different vendors preclude direct comparability of attenuation coefficients across platforms. In 2022, a Korean research group conducted a comparative study in which commercially available ATI, UGAP, and TAI examinations from different vendors were applied to the same patient to assess inter-platform variability. The study demonstrated significant differences in measured attenuation coefficients between platforms, even in the same patient, highlighting substantial inter-platform variability and indicating that attenuation coefficients obtained from one vendor’s system cannot be directly applied to that of another (46). These findings suggest that, for longitudinal monitoring of patients with hepatic steatosis, consistent use of the same vendor’s 2D gray-scale US attenuation-based quantitative technique system is necessary. However, this requirement poses a significant challenge in clinical practice. To address this issue, efforts toward standardization, such as the development of calibration phantoms capable of providing consistent attenuation coefficient references across different platforms, are necessary.
In managing patients with hepatic steatosis, identifying individuals at risk of progression to cirrhosis or end-stage liver disease is more critical than just detecting the presence and severity of steatosis. Previous studies have demonstrated that in patients with MASLD, the presence of advanced fibrosis (stage F2 or higher) is significantly associated with increased overall and liver-related mortality (47). Accordingly, in patients with hepatic steatosis, it is essential not only to accurately assess the degree of steatosis but also to evaluate the presence of advanced fibrosis (≥F2 stage). Liver stiffness measurement using elastography techniques, including TE and shear wave elastography (SWE), has excellent diagnostic performance in assessing liver fibrosis. Currently, most vendors offering 2D gray-scale US attenuation-based quantitative techniques also provide 2D-SWE for liver stiffness evaluation. A multiparametric approach that combines 2D gray-scale US attenuation-based quantitative techniques for steatosis assessment with elastography for fibrosis evaluation may thus offer greater clinical utility for comprehensive evaluation of patients with steatotic liver disease (262729). Moreover, further long-term studies are warranted to determine whether attenuation coefficients obtained through 2D gray-scale US attenuation-based quantitative techniques can serve as prognostic indicators for overall survival and liver-related mortality in patients with hepatic steatosis.
Hepatic steatosis is a key feature defining MASLD and is currently the most common cause of chronic liver disease, with a prevalence of approximately 30%. Consequently, it represents a substantial disease burden in modern society. Given the potential for hepatic steatosis to progress beyond simple steatosis to MASH, and ultimately to cirrhosis, end-stage liver disease, and HCC, early detection and accurate evaluation are essential for effective patient management and for preventing further liver injury. Gray-scale US remains the first-line imaging modality for evaluating liver diseases, including hepatic steatosis, owing to its safety, accessibility, and lower cost compared to other imaging modalities. However, its diagnostic performance is limited, particularly in detecting mild steatosis, and its reproducibility is reduced because of its subjective nature. Recently, 2D gray-scale US attenuation-based quantitative techniques were developed and introduced into clinical practice. By calculating the attenuation coefficient from RF raw data, these techniques enable the quantitative assessment of hepatic steatosis and have excellent diagnostic performance in detecting and grading steatosis, while also improving reproducibility. To enhance measurement accuracy and minimize inter-examination variability, the standardization of measurement protocols, including detailed assessment methodologies, remains a critical goal for future research and clinical application. Moreover, the integration of 2D-SWE with 2D gray-scale US attenuation-based quantitative techniques for multiparametric assessment may further aid in the early identification of patients with hepatic steatosis who are at an increased risk of progression to cirrhosis or end-stage liver disease.