This multicenter prospective study revealed that iATT showed excellent diagnostic performance in detecting steatotic liver disease, as well as high reproducibility when compared with MRI-PDFF, suggesting its potential as an alternative to MRI-PDFF.

The controlled attenuation parameter, which could be obtained together with the LSM using FibroScan, measures the attenuation of the ultrasound beam. The controlled attenuation parameter is an easy-to-measure tool that has been available for more than a decade, and several studies have assessed its value in liver fat content quantification, with some meta-analyses confirming its usefulness [34,35,36]. However, the lack of a B-mode display and the measurement discrepancies between the M probe and XL probe continue to be challenges [37, 38]. Furthermore, as reported by the recent position paper on liver fat quantification by the World Federation for Ultrasound in Medicine and Biology and the AIUM-RSNA QIBA (The American Institute of Ultrasound in Medicine-RSNA Quantitative Imaging Biomarkers Alliance) initiative article, the controlled attenuation parameter is not an adequate reference standard for evaluating the accuracy of emerging ultrasound techniques for fat quantification with attenuation coefficients [38, 39].

In the last few years, several ultrasound manufacturers have developed software for liver fat content quantification using attenuation coefficients [18,19,20, 23, 24]. Among them, iATT, an upgraded version of the ATT algorithm released by Fujifilm Healthcare, Tokyo, Japan (previously Hitachi Ltd., Tokyo, Japan), was introduced a few years ago [24]. Attenuation coefficient methods based on B-mode imaging, such as iATT, are excellent for hepatic steatosis assessment because they enable the confirmation of measurement sites by B-mode imaging. There is no need to purchase additional equipment for FibroScan, and most existing echo machines in the facility can be utilized for attenuation coefficient measurement through software updates. While there have been numerous high-quality reports on the controlled attenuation parameter, to the best of the authors’ knowledge, evidence from multicenter prospective studies on B-mode-based attenuation measurement methods has only been established for UGAP. Prospective multicenter studies, such as the present study, which compared iATT to MRI-PDFF, should provide a high level of evidence for other B-mode methods.

Currently, MRI-PDFF is becoming an established alternative to biopsy. Similar to UGAP, it is important to investigate the compatibility of iATT with MRI-PDFF. In addition, liver fat content quantification has been reported to have inter-operator variability in reproducibility [40]. We believed that by adjusting our research methodology to that of Imajo et al., we could confirm whether similar results to UGAP could be achieved. In many cases in this study, liver biopsy was not available to assess steatosis. As noted above, the correlation between MRI-PDFF and histological steatosis grade decreases with increasing steatosis grade; however, overall, the correlation is at an acceptable level [41]. In patients with any degree of steatosis, the relationship between PDFF and histology was predominantly linear (r = 0.85 [95% CI 0.80–0.89]). The American Association for the Study of Liver Diseases guidelines also indicate that liver biopsy should be considered in patients with MASLD who are at increased risk of steatohepatitis and advanced fibrosis [42]. Recently, many multicenter clinical trials have used MRI-PDFF as the reference standard; furthermore, the number of trials using biopsy tissue as the reference standard is expected to decrease owing to ethical reasons [16, 18, 43]. Therefore, in the present study, MRI-PDFF instead of histology was used as the reference standard based on the previous study.

In this study, the agreement between iATT and MRI-PDFF was evaluated by performing a Bland–Altman analysis. There was almost no fixed error. In comparison, in a large-scale study, proportional errors are more likely to occur. Therefore, as in previous studies, we established two evaluation criteria and analyzed the agreement between the two modalities [18, 44]. As both criteria were met, we concluded through the Bland–Altman analysis that iATT was compatible with MRI-PDFF. High concordance was also confirmed in the intraclass correlation analysis. Therefore, iATT could serve as an alternative to MRI-PDFF.

Subsequently, diagnostic performance was analyzed. MRI-PDFF showed a strong correlation with histological fat deposition from the early stages of S grade. By determining S1, a clear distinction can be made between healthy individuals and patients with early-stage steatotic liver disease. The use of iATT calculated by the Youden index in this study yielded favorable diagnostic performance parameters, including sensitivity, specificity, positive predictive value, and negative predictive value. A previous study in Sweden showed that even simple steatotic liver disease was associated with a worse prognosis, as compared to that in healthy individuals [13]. If the cutoff value that generates a sensitivity of 90% or higher is < 0.64, the individuals are very likely to be healthy; therefore, no therapeutic intervention is considered necessary. Conversely, if the cutoff value that yields a specificity of > 90% is > 0.74, the individuals are conceivably patients with steatotic liver disease requiring therapeutic interventions to improve prognosis. The 0.74 value, which is close to the cutoff values for S2 and S3 determined by the Youden index, supports the notion that therapeutic interventions are warranted. In a meta-analysis conducted by the authors, the ATT, before it was improved to iATT, had the disadvantage of lower sensitivity than other ultrasound attenuation methods [45]. The reason for the lower diagnostic performance of the ATT is the use of the dual-frequency method wherein two frequencies are transmitted and received simultaneously as the algorithm for evaluating attenuation coefficients. The attenuation coefficient is calculated from the slope of the signal amplitude ratio curve between intervals 40 mm and 100 mm. Meanwhile, the iATT uses the reference method, as do the UGAP and the ATT. Moreover, the attenuation coefficient is calculated from the slope of the signal ratio curve between intervals 35 mm and 75 mm to acquire reproducible data [46]. The improved diagnostic performance of iATT in comparison to that of ATT was recently reported by Ogawa et al. [46]. The correlation coefficients between iATT or ATT values and MRI‐PDFF values were 0.803 and 0.533 (p < 0.001). For the detection of hepatic steatosis of ≥ S1, ≥ S2, and ≥ S3, iATT had significantly higher AUROCs than ATT (p < 0.001, p < 0.001, p = 0.001), as demonstrated in their free downloadable graphical abstract. In this study, iATT also showed good sensitivity and demonstrated high diagnostic performance in quantifying intrahepatic steatosis, similar to other B-mode diagnostic methods.

When the causes of discrepancy between iATT and MRI-PDFF in this study were examined, SCD was the only factor identified. The measurement site for iATT is fixed at a depth of 3.5–7 cm from the body surface, unlike other testing methods. Therefore, a discrepancy between iATT and MRI might have occurred in cases with thicker subcutaneous adipose tissue. Ultrasound attenuation might be affected by the tissue between the liver and skin.

Whether ultrasound attenuation measurements are influenced by the amount of liver fibrosis remains controversial. ATT has been reported to be unaffected by fibrosis or inflammation; nonetheless, analytical methods are considered inadequate [22]. In this study, a large number of cases were examined for interactions using multiple comparisons and multiple regression analysis, and iATT was not found to be influenced by the degree of fibrosis, similar to UGAP.

The present study has some limitations. First, this study was conducted in Japan only. While racial groups are believed to exhibit no differences, an international collaborative study is required. Second, the cause of divergent cases might not have been adequately searched. Hence, we should continue to search for factors other than SCD, as almost all divergent cases do not have an SCD of 25 or higher. Third, as grade S increased, the cutoff value for iATT exhibited a stepwise increase, whereas the cutoff values for grades 2 and 3 displayed minimal differences. There are two possible reasons for this phenomenon: (i) as the degree of fat deposition increases, MRI-PDFF is not well correlated to histological steatosis grade, as compared to cases with lower fat deposition, making accurate validation more difficult in severe steatosis; and (ii) it is possible that reliance on ultrasound attenuation coefficients alone may have limitations in differentiating between grades 2 and 3. While the clinical significance of differentiating between grades 2 and 3 may be limited, if differentiation is deemed necessary, future efforts should focus on developing measurement techniques that incorporate additional ultrasound signals, such as the backscatter coefficient. Fourth, the median interval between the iATT and MRI-PDFF was slightly longer. However, we confirmed that between the two tests the participants’ weight change was not more than 5%; moreover, there was no change in their physical condition, no heavy alcohol consumption, and no history of drug treatment that might have improved their fatty liver. Ogawa et al.’s series showed no significant difference in the correlation of these parameters between iATT and MRI‐derived PDFF measurements at different intervals within 3 months [46]. In our study, the interval was no longer than 3 months.

In conclusion, iATT is highly compatible with PDFF and can be enforced as an alternative modality to PDFF. Any steatosis grade shows good diagnostic performance.