08628847

July 24, 2025, 15:51

Reader’s expertise on the topic of the manuscript

Does the reader have a conflict of interest?

The overall quality of the manuscript, based on the above-listed criteria, should be evaluated and classified according to the following five categories

Language quality (style, grammar, and spelling) should be evaluated and classified according to the following five categories.

This paper provides an insightful contribution to the field of MAFLD by developing a machine learning-based model for early detection in high-metabolic-risk populations. The integration of clinical data and TCM features adds a unique, holistic perspective, which could enhance the model's diagnostic power. The use of advanced statistical methods, such as LASSO and RFE, strengthens the reliability of the predictive model. The XGBoost algorithm demonstrated superior performance, making it a promising tool for non-invasive MAFLD screening in clinical settings. Overall, the study presents a significant step towards improving MAFLD detection and has promising clinical implications, especially in cost-effective screening and early intervention for at-risk populations.

Thank you very much for your comments.

05773185

July 24, 2025, 12:28

Reader’s expertise on the topic of the manuscript

Does the reader have a conflict of interest?

The overall quality of the manuscript, based on the above-listed criteria, should be evaluated and classified according to the following five categories

Language quality (style, grammar, and spelling) should be evaluated and classified according to the following five categories.

Tian et al. (2025) present a compelling prospective observational study employing machine learning (ML) to identify biochemical and clinical markers predictive of hepatic steatosis in individuals at high metabolic risk. Metabolic-associated fatty liver disease (MAFLD) is prevalent yet frequently underdiagnosed, especially in early stages when traditional screening modalities like ultrasonography lack sensitivity. Thus, the authors’ development of a noninvasive, cost-effective ML-based predictive model is timely and clinically relevant, aiming for early identification and improved resource utilization. A methodological strength of the study is its comprehensive use of multiple ML algorithms—XGBoost, Random Forest (RF), Support Vector Machine (SVM), and Logistic Regression (LR)—representing diverse modeling approaches. This allowed for robust comparative analysis, with each algorithm selected for specific strengths: XGBoost for nonlinear interactions, RF for overfitting resistance, SVM for high-dimensional data, and LR for baseline interpretability. This methodological breadth effectively balances predictive performance and clinical practicality. To manage high-dimensional data (156 initial candidate features), the study employed dual feature selection techniques: recursive feature elimination (RFE) and Least Absolute Shrinkage and Selection Operator (LASSO) regression. Combining these strategies yielded a robust set of ten core features, significantly reducing complexity while ensuring predictive accuracy. These final predictors, including the AST/ALT ratio, triglycerides, and waist circumference, are clinically meaningful and closely correlate with hepatic steatosis, reinforcing the validity of their selection. Among the ML models evaluated, XGBoost demonstrated superior performance with an area under the ROC curve (AUC) of approximately 0.82, surpassing RF, SVM, and LR. XGBoost’s predictive strength was further validated through cross-validation (mean AUC ~0.918), suggesting robust internal consistency. Importantly, the model achieved high accuracy (84%) and an F1-score of 0.84, reflecting a balanced sensitivity and specificity crucial for clinical implementation. Another innovative aspect was incorporating Traditional Chinese Medicine (TCM) indicators alongside conventional clinical metrics. Two TCM-derived features, greasy tongue coating and tongue edge redness, emerged prominently among the top predictors, bridging traditional holistic diagnostics and contemporary data analytics. These TCM indicators likely represent underlying metabolic dysregulation consistent with “damp-heat” conditions in TCM theory, aligning with inflammation and metabolic dysfunction commonly observed in MAFLD. Incorporating TCM features provides valuable additional diagnostic perspectives, potentially capturing subtle clinical signs not represented in standard biomedical metrics. The quantitative integration of these features through digital analysis and expert validation represents a significant methodological innovation. However, several practical challenges must be considered. TCM diagnostic standardization and reproducibility outside controlled research settings may pose limitations. Differences in tongue appearance influenced by diet, hydration, and oral hygiene could introduce variability. Thus, ensuring consistent reproducibility and gaining broad clinical acceptance remain critical tasks for future research. Clinically, the authors’ ML tool holds significant promise for noninvasive screening, particularly valuable in routine healthcare settings where expensive or invasive procedures like MRI-PDFF or liver biopsies are impractical. This model’s reliance on common biochemical and clinical data enhances its applicability, particularly in resource-limited settings. Stratification by predicted risk allows targeted use of advanced imaging or specialist follow-up, optimizing clinical workflows and resource allocation. Nevertheless, the model’s limitations must be acknowledged. First, external validation in multicenter cohorts is essential, given the study’s single-center design and potential population-specific biases. Moreover, the lack of MAFLD subtyping overlooks disease heterogeneity, possibly limiting predictive accuracy for specific patient subsets (e.g., diabetic vs. non-diabetic patients, mild vs. severe steatosis). Addressing these nuances through subgroup-specific analyses or longitudinal follow-up studies will enhance interpretability and clinical utility. Future research could expand the feature set to include advanced biomarkers or imaging modalities to enhance predictive precision. Further, refining TCM integration through sophisticated image analysis or correlating TCM features with biochemical markers would provide stronger mechanistic insights. Prospective studies evaluating clinical outcomes of model-driven interventions, such as lifestyle modification or TCM-based therapies, could demonstrate tangible patient benefits and validate real-world clinical impact. In conclusion, Tian et al. present a robustly designed study introducing an innovative ML-based approach for early hepatic steatosis detection in high-risk populations. The inclusion of TCM diagnostic features exemplifies a novel integration of traditional medicine within contemporary predictive frameworks. Although initial results are promising, external validation, refinement for disease heterogeneity, and practical clinical integration remain critical next steps. This pioneering work significantly contributes to metabolic medicine and predictive hepatology, paving the way for improved, patient-centered management of fatty liver disease.

First, thank you very much for your professional comments on the article published in World Journal of Gastroenterology. Second, we read your comments with great interest. You are welcome to format your valuable comments into a Letter to the Editor and submit it online to World Journal of Gastroenterology at https://www.f6publishing.com. There are no restrictions on the number of words, figures (color, B/W) or authors for a Letter to the Editor. In addition, the article processing charge will be exempted for this Letter to the Editor. As with all articles published by the Baishideng Publishing Group, the Letter to the Editor will be published online after completing peer review. The guidelines for a Letter to the Editor can be found at: https://www.wjgnet.com/bpg/GerInfo/219. Finally, we look forward to receiving your high-quality Letter to the Editor, which will promote academic communication and lead the development of this discipline.

03664074

July 22, 2025, 07:25

Reader’s expertise on the topic of the manuscript

Does the reader have a conflict of interest?

The overall quality of the manuscript, based on the above-listed criteria, should be evaluated and classified according to the following five categories

Language quality (style, grammar, and spelling) should be evaluated and classified according to the following five categories.

This study has applied a combined method of machine learning and regression analysis to evaluate the critical biomarkers in the diagnosis of Metabolic-associated fatty liver disease in patients. There are many correlated factors in this subsequent calculation under authors' efforts. We have considered these results to find a phenomenon that all these correlated factors coefficient values are not high. Hope that the author can explain them with scientific viewpoint.

Thank you very much for your comments.