Clinical Value of Ultrasound Fat Fraction in Grading Hepatic Steatosis: Preliminary Cut-Off Values in Obese Patients

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Dear Authors, 

 

Thank you very much for your valuable contribution to the literature. I had read you study with a great interest, and must say that your article is well-structured, clearly written, and addresses an important diagnostic challenge. 

This study explores the diagnostic utility of ultrasound-derived fat fraction (USFF) in grading hepatic steatosis in obese patients. The study provides preliminary cut-off values to classify steatosis severity and correlates them with conventional ultrasound visual grading. The topic is clinically relevant and contributes to the growing field of quantitative ultrasound in metabolic dysfunction-associated steatotic liver disease (MASLD). 

I have some suggestion to improve the content of manuscript: 

-The novelty lies in proposing preliminary USFF cut-offs specific to obese patients. However, the introduction should better clarify how this study builds upon and differs from prior work (e.g., De Robertis et al., Rónaszéki et al.). Emphasizing the gap in data for obese populations would enhance the perceived novelty.

- The authors recruited 95 obese patients, but only 84 were analyzed. The reasons for exclusion (11 patients) should be specified (e.g., technical issues, incomplete data, exclusion criteria).

- The USFF measurements were obtained using Samsung RS85 Prestige. Please specify whether quality control procedures or calibration phantoms were used, as signal consistency is essential when deriving quantitative thresholds.

- The use of ANOVA and ROC curves is appropriate. However, the authors should provide 95% confidence intervals for sensitivity/specificity of each proposed cut-off and clarify whether adjustments for multiple comparisons were applied.

- The progressive rise in USFF values across steatosis grades supports internal validity. However, external validation is lacking. The discussion should explicitly state that these cut-offs are population-specific and may not generalize across ethnicities, devices, or BMI categories.

-The limitations are acknowledged but can be elaborated:

• Absence of histological or MRI-PDFF validation.

• Small number of mild steatosis cases, affecting statistical balance.

• Potential operator and device dependence of USFF readings.

• Exclusively obese cohort may limit applicability to general or lean populations.
  Future studies should evaluate inter-system comparability and standardization across ultrasound vendors.

- The manuscript is readable but would benefit from minor English editing to improve flow (e.g., “patients underwent to conventional ultrasound” → “patients underwent conventional ultrasound”).

- Use consistent terms — “steatosis grading” vs. “severity classification.” Avoid alternating between MASLD and NAFLD unless contextually justified.

- Table 1 should include p-values for comparisons between steatosis grades.

 

Author Response

We sincerely thank the reviewer for the constructive feedback, which helped us to improve the quality and clarity of the paper. We have addressed all comments as follows:

Comment 1: The introduction should clarify how this study builds upon and differs from prior work (e.g., De Robertis et al., Rónaszéki et al.) and emphasize the data gap for obese populations.

Response: We appreciate this suggestion. We have revised the Introduction to highlight how our study differs from previous work, particularly emphasizing that most published data derive from mixed or non-obese populations. We now explicitly state that our study provides the first preliminary USFF cut-offs specific to obese patients, addressing a significant knowledge gap.

Comment 2: Specify the reasons for exclusion of 11 patients (from 95 to 84 analyzed).

Response: We thank the reviewer for noting this. We have added a sentence in the Methods section clarifying the reason for the exclusion of this patients.

Comment 3: Indicate whether calibration or quality control procedures were performed on the Samsung RS85 Prestige system.

Response: Thank you for this important point. We have specified that data were acquired and subsequently processed using proprietary algorithms developed and calibrated at the Samsung Medical Center.

Comment 4: Provide 95% confidence intervals for sensitivity/specificity and clarify if multiple comparison adjustments were made.

Response: We have now included 95% confidence intervals for all sensitivity and specificity values derived from ROC analyses. No formal adjustment for multiple comparisons was applied given the exploratory nature of the study; this is now stated explicitly in the Statistical Analysis section.

Comment 5: The discussion should clarify that these cut-offs are population-specific and may not generalize across ethnicities, devices, or BMI categories.

Response: We fully agree. The Discussion now includes a statement acknowledging that the proposed thresholds are preliminary and specific to the obese population studied, and that device-related variability may limit their generalizability.

Comment 6: Expand the limitations to include lack of histological/MRI validation, small number of mild cases, operator/device dependence, and cohort specificity.

Response: We have expanded the limitations accordingly, now explicitly addressing each of these points.

Comment 7: Minor English editing is needed (e.g., “patients underwent to conventional ultrasound” → “patients underwent conventional ultrasound”).

Response: The entire manuscript has been carefully revised for English language, improving readability and consistency.

Comment 8: Use consistent terms (“steatosis grading” vs. “severity classification”; avoid alternating MASLD/NAFLD).

Response: We have standardized terminology throughout the manuscript, now consistently using “steatosis grading” and “MASLD” in accordance with current nomenclature. The term “NAFLD” was retained only when explicitly used by the authors of cited articles to preserve the accuracy of historical references..

Comment 9: Add p-values for comparisons between steatosis grades.

Response: We have updated Table 1 to include p-values for between-group comparisons of continuous variables.

Reviewer 2 Report

Comments and Suggestions for Authors

17 Abstract

Add:

84 patients were included (exclusion criteria were morphological features of advanced liver diseases or cirrhosis, active viral hepatitis, alcohol use disorder, liver enzymes alteration and heart failure).

 

The abstract can be improved.

137-139

“Patients with clinical or morphological features of advanced liver diseases or cirrhosis, active viral hepatitis, alcohol use disorder, liver enzymes alteration and heart failure were excluded.”

 

I strongly suggest specifying how much patients are in each exclusion category (advanced liver disease or cirrhosis, active viral hepatitis, alcohol use disorders, liver enzymes alteration, hearth failure) between the 11 excluded from the study.

181-182 and 195 Table 1

“No statistically significant differences were found between the four groups regarding the degree of fibrosis measured by 2D Shearwave elastography”.

 

In my opinion, based in my knowledge and in my experience, the absence of any difference in grade of fibrosis between the different grading of steatosis, in particular between grade zero and grade 4, is at least uncommon and counterintuitive. Maybe it is justifiable with the young mean age of the population in study.

More in general, the absence of cases of hepatic fibrosis reported in all the population studied (no significant fibrosis, the same grade of fibrosis in all categories of steatosis or advanced liver disease-Cirrhosis in part – but how many?) - of patients excluded from the study) is an element of perplexity.

195 and Table 1

Summary of mean BMI….

381-396

Why in capital letters? Is it in the instructions for authors?

Author Response

Comment 1:
Add in the abstract:
“84 patients were included (exclusion criteria were morphological features of advanced liver diseases or cirrhosis, active viral hepatitis, alcohol use disorder, liver enzymes alteration and heart failure).”
The abstract can be improved.

Response:
We thank the reviewer for this useful suggestion. We have revised the Abstract to specify the inclusion of 84 patients and to list the main exclusion criteria as recommended: “Eighty-four obese patients were included (exclusion criteria were morphological features of advanced liver disease or cirrhosis, active viral hepatitis, alcohol use disorder, liver enzymes alteration, and heart failure).” This addition improves clarity and completeness of the study population description.

Comment 2:
Specify how many patients were excluded for each criterion among the 11 not included in the final analysis.

Response:
We appreciate this insightful comment. We have revised the Methods section to detail the exclusion reasons.

Comment 3:
The absence of significant differences in fibrosis degree between steatosis grades is uncommon and counterintuitive; this could be justified by the young age of the population. The absence of fibrosis or cirrhosis cases raises perplexity—please clarify.

Response:
We thank the reviewer for this thoughtful observation. We agree that the abscence of significant differences in fibrosis values among steatosis grades could reflect the relatively young mean age. However, we specified in the Result section that 3 patients (3.2%) with US features of cirrhosis were excluded. This finding is consistent with the results previously reported by Yang et al. [doi: 10.1136/gutjnl-2024-332917].

Comment 4:
Why are these lines in capital letters? Is this required by the journal’s instructions for authors?

Response:
We thank the reviewer for noticing this. The text in capital letters was unintended and not required by the journal. We have corrected this section to standard sentence case formatting in accordance with the journal’s style guidelines.

Reviewer 3 Report

Comments and Suggestions for Authors

This study is important, timely and will be of interest to a spectrum of caregivers, epidemiologists, governmental agencies and pharmaceutical companies.

Nevertheless your group should consider studying patients with a focus on alcohol ingestion even in small amounts. It has become apparent in recent times that worldwide alcohol ingestion is linked to life expectancy and to HCC. The entity of "non-alcoholic steatohepatitis" which  is a component of obesity and MASLD is a misnomer. 

Consider comparing patients who regularly drink alcohol regardless of amount  with those who drink no alcohol.

Author Response

We sincerely thank the reviewer for this insightful and forward-looking comment. We agree that alcohol consumption, even in small quantities, represents an important and evolving determinant of hepatic and metabolic health. Our present study intentionally excluded individuals with alcohol use disorder or any significant alcohol intake, in line with the diagnostic framework of MASLD applied at the time of study conception. This allowed us to isolate the effects of obesity-related steatosis without the confounding impact of alcohol consumption.

We fully acknowledge, however, that emerging evidence supports a more continuous relationship between alcohol intake and hepatic/metabolic outcomes, and that the recent redefinition of “non-alcoholic steatohepatitis” toward metabolic dysfunction-associated steatohepatitis (MASH) reflects this conceptual shift. We have added a statement in the Discussion acknowledging this evolving understanding and suggesting possible directions that the future studies could investigate.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

Dear Authors

Thank you for your effort to contribute to the literature. 

This manuscript investigates the diagnostic performance of ultrasound-derived fat fraction (USFF) in grading hepatic steatosis in obese patients, providing preliminary cut-offs for different steatosis grades. The topic is highly relevant given the global rise of MASLD, the limitations of liver biopsy, and the increasing demand for reliable, non-invasive diagnostic tools.

The study is well-structured, clearly written, and methodologically sound. The introduction provides adequate background, the methods are appropriately described, and the statistical analyses (ANOVA, ROC, Youden index) are suitable for the study aims. Overall, this is a strong and clinically meaningful contribution that fits the journal’s scope.

-

The manuscript refers to both retrospective and cross-sectional elements. Please clearly define the design.
The data collection appears cross-sectional at a single time point, but exclusion and selection criteria imply retrospective review. State explicitly: “This was a cross-sectional observational study based on retrospectively collected ultrasound and clinical data.”

 

-The largest limitation—acknowledged by the authors—is that USFF is validated only against visual ultrasound assessment, not against gold standards.

Although the authors present this as a limitation, the discussion should more clearly emphasize how this affects: diagnostic accuracy, the strength of the proposed cut-off values, ability to compare with other QUS literature.

 

-The S1 (mild) group is very small, which may affect the stability of the cut-off estimate (USFF < 11.66). The authors should discuss: whether these thresholds may change with larger samples, if bootstrap or cross-validation could support robustness (optional). 

 

USFF was measured by a single experienced operator, while visual grading was performed by different blinded operators. This creates a potential measurement imbalance.

Please clarify: years of experience or training of the USFF operator, whether inter-observer reproducibility was assessed or planned for future studies.

- Method . Include probe frequency and model used. Clarify whether fasting was standardized (≥6 hours).

- Conclusion, Consider adding a sentence stating that USFF has potential but should not replace MRI-PDFF in clinical trials until further validation.

Good lucks!

Author Response

We thank the reviewer for the valuable comments, which reflect a careful evaluation of the manuscript and substantially enhance the quality of the work.

Comment 1: The manuscript refers to both retrospective and cross-sectional elements. Please clearly define the design.
The data collection appears cross-sectional at a single time point, but exclusion and selection criteria imply retrospective review. State explicitly: “This was a cross-sectional observational study based on retrospectively collected ultrasound and clinical data.”

Response 1:
We thank the reviewer. We include the explicit sentence in the text. 

Comment 2: 

-The largest limitation—acknowledged by the authors—is that USFF is validated only against visual ultrasound assessment, not against gold standards.

Although the authors present this as a limitation, the discussion should more clearly emphasize how this affects: diagnostic accuracy, the strength of the proposed cut-off values, ability to compare with other QUS literature.

Response 2: We thank the reviewer for highlighting this important point. We agree that the absence of validation against histology or MRI-PDFF represents the major limitation of our study. We have expanded the Discussion to more explicitly state how this impacts diagnostic accuracy, the robustness of the proposed cut-offs, and comparability with existing QUS literature. Specifically, we clarify that our cut-off values should be interpreted as preliminary and internally valid within this cohort, but not directly transferable across devices, populations, or studies employing MRI-PDFF or biopsy as reference standards. A statement has been added to underscore the need for future validation studies using established gold-standard modalities.

Comment 3: -The S1 (mild) group is very small, which may affect the stability of the cut-off estimate (USFF < 11.66). The authors should discuss: whether these thresholds may change with larger samples, if bootstrap or cross-validation could support robustness (optional). 

Response 3: We thank the reviewer for this important observation. We agree that the limited size of the S1 group may reduce the stability and precision of the corresponding cut-off estimate. We have expanded the limitations section to acknowledge that the proposed thresholds—particularly for mild steatosis—may vary when assessed in larger or more balanced cohorts.

 

Comment 4: 

USFF was measured by a single experienced operator, while visual grading was performed by different blinded operators. This creates a potential measurement imbalance.

Please clarify: years of experience or training of the USFF operator, whether inter-observer reproducibility was assessed or planned for future studies.

Response 4: We thank the reviewer for the comment. We clarified the experience of the operators, and we explicitly stated that inter-observer variability was not assessed, reporting it as a limitation. 

 

Comment 5: - Method . Include probe frequency and model used. Clarify whether fasting was standardized (≥6 hours).

Response 5: We include probe frequency and model. We clarified the fasting period before the measurement.

 

Comment 6: - Conclusion, Consider adding a sentence stating that USFF has potential but should not replace MRI-PDFF in clinical trials until further validation.

Response 6: Thank you for highlighting this point. We integrated the conclusion with the proposed sentence.