Cell-Free DNA as Biomarker in Oral Squamous Cell Carcinoma: Dynamics, Mutational Landscape and Clinical Implications

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This manuscript addresses an important and timely area of research, and the data generated have the potential to make a meaningful contribution to the field. However, several elements of the study would benefit from refinement to improve clarity, analytical robustness, and overall interpretability. In particular, issues relating to pre‑analytical variability and the reliance on mean rather than median cfDNA values may significantly influence the conclusions drawn. Addressing these points would strengthen the scientific rigor and enhance the readability of the manuscript. The specific comments below outline areas that require attention.

 

1. It is well established that variations in urine specific gravity and ionic strength can drive cfDNA aggregation or promote its association with crystalline or proteinaceous material, which may subsequently be pelleted during the 2,000 × g centrifugation step. This raises the possibility that cfDNA recovery from the supernatant may be inconsistent across samples. Could the authors comment on whether they assessed this source of pre‑analytical variability and indicate what measures, if any, were implemented to mitigate it (e.g., use of preservatives or stabilizers, specific gravity assessments, standardized collection protocols, or parallel analysis of pellet and supernatant fractions)? If not, perhaps this could be written as a study limitation.
2. In the results section, the authors should reconsider their use of mean values to summarise cfDNA concentrations. cfDNA measurements are well known to exhibit non‑Gaussian, skewed distributions, often influenced by biological variability and the presence of extreme high‑value outliers. Given the small sample size of the control group (n = 5) and the established skewness of cfDNA data, reporting means risks misrepresenting group central tendencies. Median values, accompanied by interquartile ranges, provide a more robust and statistically appropriate summary of cfDNA levels under these conditions. Furthermore, non‑parametric statistical tests (e.g., Mann–Whitney U) should be used unless distributional assumptions are explicitly verified Given this issue, reporting medians rather than means may substantially change the results and potentially lead to different interpretations. If the conclusions remain unchanged, please state that analyses using both mean and median values yielded consistent findings.
3. Figure 1B is unclear regarding whether the data represent patients, controls, or a combination of both. Please clarify this in the figure legend. In addition, given the small sample size, it would be more informative to present individual data points overlaid with a box‑and‑whisker plot showing the minimum, 25th percentile, median, 75th percentile, and maximum. This approach provides greater transparency of the underlying data compared with a bar chart displaying only the mean and standard deviation.
4. Figures could be improved as some fonts and datapoints are too small and not legible.
5. The paper could be improved by presenting the content more succinctly.

Comments on the Quality of English Language

The paper could be improved by presenting the content more succinctly.

Author Response

COMMENTS FOR THE REVIEWERS:

Reviewer 1

This manuscript addresses an important and timely area of research, and the data generated have the potential to make a meaningful contribution to the field. However, several elements of the study would benefit from refinement to improve clarity, analytical robustness, and overall interpretability. In particular, issues relating to pre‑analytical variability and the reliance on mean rather than median cfDNA values may significantly influence the conclusions drawn. Addressing these points would strengthen the scientific rigor and enhance the readability of the manuscript. The specific comments below outline areas that require attention.

1. It is well established that variations in urine specific gravity and ionic strength can drive cfDNA aggregation or promote its association with crystalline or proteinaceous material, which may subsequently be pelleted during the 2,000 × g centrifugation step. This raises the possibility that cfDNA recovery from the supernatant may be inconsistent across samples. Could the authors comment on whether they assessed this source of pre‑analytical variability and indicate what measures, if any, were implemented to mitigate it (e.g., use of preservatives or stabilizers, specific gravity assessments, standardized collection protocols, or parallel analysis of pellet and supernatant fractions)? If not, perhaps this could be written as a study limitation.

Author’s response: We thank the reviewer for this insightful remark. We agree that differences in urine physicochemical properties, such as specific gravity and ionic strength, may contribute to pre-analytical variability in urinary cfDNA recovery.

In the present study, we did not specifically evaluate the influence of urine specific gravity, ionic strength, or cfDNA partitioning between pellet and supernatant fractions. Our protocol focused on the analysis of the supernatant obtained after centrifugation at 2,000 × g, which is a commonly used approach in urinary cfDNA studies. Although preservatives or stabilizing agents were not used, all urine samples were processed within 2 hours of collection, centrifuged promptly, and the resulting supernatant was stored at −80 °C until cfDNA extraction in order to minimize degradation and handling-related variability.

Nevertheless, we acknowledge that differences in urine composition could possibly influence cfDNA recovery. Accordingly, we have added the following statement as a limitation in the revised manuscript: “Also, regarding the urine results, a potential limitation of this study is that urine physicochemical properties can influence cfDNA recovery. As only the supernatant fraction was analyzed, possible variability in cfDNA recovery cannot be excluded.”

 

2. In the results section, the authors should reconsider their use of mean values to summarise cfDNA concentrations. cfDNA measurements are well known to exhibit non‑Gaussian, skewed distributions, often influenced by biological variability and the presence of extreme high‑value outliers. Given the small sample size of the control group (n = 5) and the established skewness of cfDNA data, reporting means risks misrepresenting group central tendencies. Median values, accompanied by interquartile ranges, provide a more robust and statistically appropriate summary of cfDNA levels under these conditions. Furthermore, non‑parametric statistical tests (e.g., Mann–Whitney U) should be used unless distributional assumptions are explicitly verified Given this issue, reporting medians rather than means may substantially change the results and potentially lead to different interpretations. If the conclusions remain unchanged, please state that analyses using both mean and median values yielded consistent findings.

 

Author’s response: We thank the reviewer for this important comment regarding the distribution of cfDNA measurements. In our study, samples with extremely high cfDNA concentrations were excluded during data curation, as these values were considered likely to represent technical artefacts or biologically implausible measurements that could bias the analysis. After this exclusion step, the remaining dataset did not contain extreme outliers that would substantially distort the calculated means. Also during data processing we conclude that both mean and median do not change the interpretation. In addition, TapeStation analysis was performed to assess cfDNA fragment size profiles and verify that the samples were not significantly contaminated with high-molecular-weight genomic DNA. However we have clarified this point in the revised manuscript:

“Samples presenting extremely high cfDNA concentrations, considered outliers and likely to bias summary statistics, were excluded from the analysis.”

3. Figure 1B is unclear regarding whether the data represent patients, controls, or a combination of both. Please clarify this in the figure legend. In addition, given the small sample size, it would be more informative to present individual data points overlaid with a box‑and‑whisker plot showing the minimum, 25th percentile, median, 75th percentile, and maximum. This approach provides greater transparency of the underlying data compared with a bar chart displaying only the mean and standard deviation.

Author’s response: We thank the reviewer for their insightful remark. In this figure our intention was to illustrate the overall temporal trend of cfDNA concentrations across the patients. For this reason, mean values with standard deviation were retained to facilitate visualization of the average cfDNA dynamics across timepoints. To improve clarity, we have revised the figure legend to explicitly indicate that panel B represents mean cfDNA concentrations across the OSCC patient cohort at each timepoint.

4. Figures could be improved as some fonts and datapoints are too small and not legible.

Author’s response: We appreciate the recommendation and have improved the size of some figures.

5. The paper could be improved by presenting the content more succinctly.

Author’s response: We thank the reviewer for their insightful remark and the manuscript has been carefully reviewed.

Reviewer 2 Report

Comments and Suggestions for Authors

The article “Cell-free DNA as Biomarker in Oral Squamous Cell Carcinoma: Dynamics, Mutational Landscape and Clinical Implications” is interesting and shows promise for the diagnosis of oral cancer.

The paper is very well structured, and the content is easy to understand and interpret.

The limitations are clearly stated. It is a valuable contribution to literature that should be considered in future research, given the questions it raises regarding tumor stage, recurrence, and progression-free survival.

The self-citation is justified.

Thank you very much.

Author Response

COMMENTS FOR THE REVIEWERS:

Reviewer 2

The article “Cell-free DNA as Biomarker in Oral Squamous Cell Carcinoma: Dynamics, Mutational Landscape and Clinical Implications” is interesting and shows promise for the diagnosis of oral cancer.

The paper is very well structured, and the content is easy to understand and interpret.

The limitations are clearly stated. It is a valuable contribution to literature that should be considered in future research, given the questions it raises regarding tumor stage, recurrence, and progression-free survival.

The self-citation is justified.

Thank you very much.

Author’s response: We sincerely thank the reviewer for the positive evaluation of our manuscript and for recognizing the relevance of our work on cell-free DNA as a biomarker in oral squamous cell carcinoma.

Reviewer 3 Report

Comments and Suggestions for Authors

This manuscript investigates the potential role of cell-free DNA (cfDNA) as a biomarker in oral squamous cell carcinoma (OSCC) by analyzing cfDNA dynamics in plasma and urine samples from 32 patients and performing targeted next-generation sequencing (NGS) in a subset of five cases. The study explores longitudinal cfDNA kinetics during treatment and examines the mutational landscape detected through liquid biopsy.

The topic is timely and clinically relevant, as liquid biopsy technologies are increasingly explored for non-invasive cancer monitoring. The integration of plasma and urine cfDNA analysis together with mutational profiling represents a potentially valuable approach to understanding tumor heterogeneity and disease evolution in OSCC.

However, several methodological and interpretative limitations reduce the impact of the study. The small cohort size limited genomic characterization, and use of a non-OSCC-specific sequencing panel constrain the conclusions. Additionally, the statistical analysis and clinical correlation require further clarification. Addressing these issues would significantly strengthen the manuscript. My comments are described as follows:

 

Comments:

1. The authors should justify the sample size or provide a power analysis, clarify whether the five sequenced patients were randomly selected or clinically representative, and moderate claims about the generalizability of the observed mutational dynamics.
2. The manuscript refers to multiple sampling timepoints (TP1–TP5), but the precise timing of these collections relative to treatment stages is not consistently described, which may reduce clarity for readers. Providing a timeline figure summarizing the sampling schedule in relation to key clinical events and treatment phases would significantly improve the transparency and interpretability of the study design.
3. Some figures require improvement. Figure 3: The clinical events (recurrence/metastasis) indicated by arrows should be explained more clearly in the caption. On the other hand, Figure 4 heatmap: The color scale and mutation counts are difficult to interpret without a clearer legend.
4. For the detected mutations, the authors should provide additional annotation, including COSMIC identifiers, oncogenic classification based on established databases such as OncoKB or ClinVar, and the level of evidence supporting their clinical relevance. Including this information would strengthen the interpretation of the mutational data and improve the overall scientific rigor of the analysis.
5. Are any of the detected hotspot mutations validated using an orthogonal method such as digital PCR or Sanger sequencing?
6. What biological mechanisms support the presence of OSCC-derived cfDNA in urine, and how did the authors ensure that these signals were not due to contamination or systemic DNA release?
7. The study reports partial concordance between tissue and liquid biopsy mutations. How do the authors interpret mutations detected exclusively in liquid biopsy samples?
8. The study reports slightly higher cfDNA levels in early-stage patients compared to stage IV disease. Could this unexpected observation be influenced by confounding factors such as comorbidities or inflammatory conditions?
9. Considering that clonal evolution analyses are based on five patients, how confident can we be that the observed mutation dynamics reflect general OSCC tumor evolution?
10. How were potential technical variations in cfDNA extraction, library preparation, and sequencing depth controlled to ensure reproducibility?

Author Response

COMMENTS FOR THE REVIEWERS:

Reviewer 3

This manuscript investigates the potential role of cell-free DNA (cfDNA) as a biomarker in oral squamous cell carcinoma (OSCC) by analyzing cfDNA dynamics in plasma and urine samples from 32 patients and performing targeted next-generation sequencing (NGS) in a subset of five cases. The study explores longitudinal cfDNA kinetics during treatment and examines the mutational landscape detected through liquid biopsy.

The topic is timely and clinically relevant, as liquid biopsy technologies are increasingly explored for non-invasive cancer monitoring. The integration of plasma and urine cfDNA analysis together with mutational profiling represents a potentially valuable approach to understanding tumor heterogeneity and disease evolution in OSCC.

However, several methodological and interpretative limitations reduce the impact of the study. The small cohort size limited genomic characterization, and use of a non-OSCC-specific sequencing panel constrain the conclusions. Additionally, the statistical analysis and clinical correlation require further clarification. Addressing these issues would significantly strengthen the manuscript. My comments are described as follows:

Comments:

1. The authors should justify the sample size or provide a power analysis, clarify whether the five sequenced patients were randomly selected or clinically representative, and moderate claims about the generalizability of the observed mutational dynamics.

 

Author’s response: We thank the reviewer for this important methodological point. The sample size reflects the prospective recruitment capacity of a single centre over the study period. The cohort of 32 patients represents a clinically diverse sample, including variation in tumor stage (I-I-IV), treatment modality, and clinical outcome, which we believe strengthens the representativeness of the cfDNA dynamics described. The five patients selected for NGS were not randomly chosen but were identified based on the availability of sufficient high-quality cfDNA across multiple timepoints and the diversity of clinical presentations within the cohort, ensuring clinically informative data. We acknowledge that the small NGS subsample limits the generalizability of the mutational findings, and we have revised the manuscript to explicitly frame these observations as preliminary requiring validation in larger prospective cohorts.

“…and 5 were selected for further, more detailed genomic characterization, based on the availability of sufficient high-quality cfDNA across multiple timepoints and the diversity of clinical presentations represented within the cohort.” We also add: “In this cohort of five patients, recurrently detected mutations in TP53, PIK3CA, SMAD4, and FBXW7 are consistent with the established roles of these genes in oral tumorigenesis and may serve as a starting point for future investigations into their prognostic and therapeutic significance in larger OSCC cohorts.”

 

2. The manuscript refers to multiple sampling timepoints (TP1–TP5), but the precise timing of these collections relative to treatment stages is not consistently described, which may reduce clarity for readers. Providing a timeline figure summarizing the sampling schedule in relation to key clinical events and treatment phases would significantly improve the transparency and interpretability of the study design.

Author’s response: We thank the reviewer for their insightful remark. The table summarizing the different timepoints (TP1–TP5) in relation to clinical events and treatment phases is included in the supplementary materials. We opted to place it there to avoid exceeding the maximum number of figures allowed in the main manuscript. Given the variability in timing between individuals, we decided to include a summary table in the supplementary materials to convey the information clearly without overcomplicating the visualization.

 

3. Some figures require improvement. Figure 3: The clinical events (recurrence/metastasis) indicated by arrows should be explained more clearly in the caption. On the other hand, Figure 4 heatmap: The color scale and mutation counts are difficult to interpret without a clearer legend.

Author’s response: We appreciate the recommendation and have incorporated the following information in the caption of the figures: Figure 3: “Arrows indicate timepoints corresponding to clinically documented disease events: renal tumor detection in patient P3 (between TP6 and TP7), local recurrence with cervical metastasis in patient P4 (between TP4 and TP5), and cervical metastasis in patient P5 (between TP5 and TP6), illustrating the temporal relationship between cfDNA dynamics and clinically documented disease progression.”

Figure 4: “The colour scale indicates the number of mutations detected per gene, with white representing 0 mutations, light blue representing 1–3 mutations, medium blue representing 4–8 mutations, and dark blue representing more than 8 mutations.”

 

4. For the detected mutations, the authors should provide additional annotation, including COSMIC identifiers, oncogenic classification based on established databases such as OncoKB or ClinVar, and the level of evidence supporting their clinical relevance. Including this information would strengthen the interpretation of the mutational data and improve the overall scientific rigor of the analysis.

Author’s response: We thank the reviewer for this helpful suggestion. We have now included additional variant annotation in the Supplementary Table “List of hotspot variants detected across the different sources and timepoints.” Specifically, we added the OncoKB oncogenic classification for each detected hotspot mutation. We also have incorporated the following information in the manuscript: “Consistent with their known roles as cancer drivers, OncoKB classification confirmed that the majority of these hotspot variants were annotated as oncogenic or likely oncogenic. The level of clinical evidence for therapeutic actionability was assigned according to the OncoKB evidence framework, with several mutations, including TP53 hotspots, PIK3CA E542K/E545K, KRAS G12C, and NRAS Q61R, carrying level 1-3 evidence for targeted therapeutic relevance in diverse cancer contexts."

 

5. Are any of the detected hotspot mutations validated using an orthogonal method such as digital PCR or Sanger sequencing?

Author’s response: We thank the reviewer for this question. Orthogonal validation of hotspot mutations was not performed in this study. However, the detected variants showed strong technical support, including high sequencing depth and quality metrics. In addition, several hotspot mutations were consistently detected across multiple timepoints and, in some cases, in more than one patient, supporting their reproducibility and reducing the likelihood of sequencing artefacts. Based on these factors, we considered the variant calls to be reliable.

6. What biological mechanisms support the presence of OSCC-derived cfDNA in urine, and how did the authors ensure that these signals were not due to contamination or systemic DNA release?

Author’s response: We thank the reviewer for their insightful remark. In our study, several observations support that the detected urinary variants originate from tumor-derived cfDNA rather than contamination or nonspecific systemic DNA release. First, many of the mutations correspond to well-known cancer hotspot variants in established driver genes. Second, several variants were detected consistently across multiple timepoints and, in some cases, across different sample types (plasma and urine) from the same patient. Third, strict sequencing quality filters and standard laboratory procedures were applied to minimize potential contamination.

7. The study reports partial concordance between tissue and liquid biopsy mutations. How do the authors interpret mutations detected exclusively in liquid biopsy samples?

Author’s response: We thank the reviewer for their insightful remark.  We have addressed that in the discussion section of the manuscript and we included more information for clarification: “…including subclonal populations or metastatic lesions not captured by the tissue biopsy or sampling limitations”.

8. The study reports slightly higher cfDNA levels in early-stage patients compared to stage IV disease. Could this unexpected observation be influenced by confounding factors such as comorbidities or inflammatory conditions?

Author’s response: We thank the reviewer for this question.  We note that cfDNA levels were slightly higher in early-stage patients compared to stage IV disease. While our study cannot definitively determine the cause of this observation, several potential explanations are discussed: “Plasma cfDNA concentration can be influenced by non-cancerous conditions, including inflammation, cardiovascular disease, or recent physical activity [15,16].” These factors may act as confounders and could partially contribute to the observed differences.

9. Considering that clonal evolution analyses are based on five patients, how confident can we be that the observed mutation dynamics reflect general OSCC tumor evolution?

Author’s response: We thank the reviewer for their insightful remark. We acknowledge that clonal evolution analyses in this study are based on only five patients, which limits the generalizability of our findings. However, the detection of several recurrent oncogenic mutations across different patients provides additional support for the relevance of these observations. These preliminary data also demonstrate the potential of longitudinal sampling to capture tumor heterogeneity and clonal evolution

10. How were potential technical variations in cfDNA extraction, library preparation, and sequencing depth controlled to ensure reproducibility?

Author’s response: We thank the reviewer for this question. We are fully aware of the importance of reproducibility in cfDNA analyses. To minimize technical variations and ensure reproducibility, all cfDNA samples were processed by the same trained technician using standardized protocols for extraction and library preparation. Quality controls were performed at each step including during cfDNA extraction, library preparation and sequencing. These measures collectively help ensure reproducibility and comparability of the results.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The authors have not corrected the font-size issues in several figures, which still require magnification of approximately 200% to read. Despite this persistent problem, I am proceeding with acceptance of the revised manuscript.

 

 

Comments on the Quality of English Language

The paper could be improved by presenting the content more succinctly.

Reviewer 3 Report

Comments and Suggestions for Authors

Thank you for your revision. I have no more questions.