Skip Content
You are currently on the new version of our website. Access the old version .
MDPIMDPI
Search all articles
by
  • Yuan Wang1,†,
  • Yu Song2,† and
  • Zhiyong Liang1,*
  • et al.

Reviewer 1: Anonymous Reviewer 2: Gorkem Durak Reviewer 3: Anonymous

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This manuscript examines the prognostic value of PAM50 intrinsic subtypes and immune gene expression signatures in triple-negative breast cancer (TNBC) treated with adjuvant chemotherapy. The topic is highly relevant, given the marked biological heterogeneity of TNBC and the growing recognition that IHC-based surrogates insufficiently capture molecular diversity. The authors present data from 111 well-characterized patients with long-term follow-up and integrate molecular, pathological, and immune-related parameters, offering a broad and informative overview of TNBC biology.

Strengths:

The study addresses an important and relatively underexplored area. The cohort is clearly defined, treated according to consistent standards, and followed for a substantial period, which enhances the reliability of survival outcomes.

Major Comments:

  1. Small sample size and subtype imbalance: This is the major limitation of the study… Certain subgroups (such as PAM50 luminal B and HER2-enriched) contain very few patients, which limits statistical power and the strength of the conclusions. Reporting confidence intervals for survival curves and Cox models would provide a clearer understanding of uncertainty. Moreover, The patients should include in the Methods and Results section a post-hoc power analysis. This is mandatory.
  2. Unexpected ROR-outcome relationship: The observation that high ROR scores correspond to better DFS contradicts established PAM50 biology. This finding needs deeper discussion, including potential explanations such as chemotherapy sensitivity, enrichment for basal-like tumors, or differences in disease stage.
  3. RNA-based ER/PR/HER2 positivity: Reclassifying some TNBC cases as RNA-positive raises important questions about assay variability and its implications for patient inclusion. This should be more explicitly addressed.
  4. Clinical applicability: This is another major point of criticism. How the authors’ research apply in clinical practice. Are surgery or oncologic results affected by this? Please cite this study PMID: 36248756 to improve the quality of your manuscript and Discussion section.
  5. Immune score thresholds: The basis for defining immune-weak and immune-strong groups (<40 and ≥55) should be clarified, and sensitivity analyses could help demonstrate the robustness of these cutoffs.

Author Response

 Please see the file.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

This study presents findings on the combined prognostic value of PAM50 molecular subtypes and immune profiling in triple-negative breast cancer (TNBC). The concept is valuable, especially in the current era of biomarker-driven therapy. However, several key points need clarification or further development.

The main limitations include the retrospective design, a relatively small, single-center cohort, and potential changes in treatment standards since 2002–2014. Additionally, several key findings, such as the inverse relationship between ROR and DFS/OS, the lack of independent prognostic value of the immune score in multivariate analysis, and the discrepancy between IHC and PAM50, require clearer explanation and more robust data presentation.

Major Points:

  1. Clarify Clinical Relevance
    The combined use of PAM50 and immune score is promising. Please explain how this could guide treatment. Mention if similar models are being tested in clinical trials.
  2. Address ROR Paradox
    Surprisingly, high-risk ROR patients had better DFS and OS. Please acknowledge this clearly and explain—could it be due to basal-like tumors responding well to chemotherapy? Since ROR was developed for ER-positive disease, is it appropriate in TNBC?
  3. Multivariate Results
    The immune score wasn’t an independent prognostic factor in multivariate analysis. Please discuss this. Could combining PAM50 and immune score improve prediction?
  4. Feasibility of Testing
    The PAM50 assay and immune gene panel may not be widely available. Please briefly discuss cost, accessibility, and whether these tests are used in routine practice at your center.
  5. IHC vs PAM50 Discordance
    The low agreement between IHC and PAM50 (κ = 0.218) is important. Please describe what’s different about the discordant group and how their outcomes compare. Also, comment on whether IHC is still a valid surrogate where PAM50 isn’t available.
  6. Show Subgroup Results
    You say the immune score adds value within each subtype, but don’t show data. Please consider adding Kaplan–Meier curves or a table comparing survival in basal-like/i-strong vs basal-like/i-weak, etc.

Minor Points:

  1. Terminology Consistency
    Use terms like “immune-strong” or “i-strong” consistently. Also, define ROR when it is first mentioned and clarify whether your ROR cutoffs were taken from the original PAM50 study or adapted for this cohort.
  2. Statistical Presentation
    In Figure 1D, add a note explaining the unexpected better DFS in the high-ROR patient.
    Avoid using words like “trend” when differences are not statistically significant.
  3. Limitations
    Please acknowledge that the treatment era (2002–2014) predates modern TNBC management, including immunotherapy and more routine use of neoadjuvant therapy.
  4. There are additional numbers at the end of sentences in the first paragraph of the introduction (2 and 3). Were these supposed to be citations?

Author Response

Please see the file of point-by-point response.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors
  1. The study's conclusion that the basal-like subtype had the "best prognosis" after chemotherapy, while luminal A and HER2-enriched had worse prognoses, seems counterintuitive given the general understanding of TNBC aggressiveness. Could the authors elaborate on this finding, perhaps by comparing it more directly to outcomes in non-TNBC basal-like cancers or by discussing the specific chemotherapy regimens used and their known efficacy across subtypes? This point needs further clarification and contextualization to avoid misinterpretation.
  2. The sample size of 111 patients, while not extremely small, is still a limitation, especially when stratifying into multiple PAM50 subtypes, some of which (like Luminal B with only 1 patient) are severely underrepresented. This makes it difficult to draw robust conclusions for these less frequent subtypes. The authors acknowledge this as a limitation, but perhaps a more detailed discussion on how this impacts the statistical power and generalizability for each subtype would be beneficial.
  3. The discordance between IHC and PAM50 subtyping (kappa=0.218) is a significant finding. While the authors state that gene expression-based basal-like subtypes benefited from chemotherapy, it would be helpful to explore the clinical implications of this discordance further. For instance, what are the potential consequences of misclassifying a patient based on IHC versus PAM50, and how might this impact treatment decisions in a real-world setting?
  4. The study mentions that high-risk patients, as determined by PAM50 ROR, had the longest DFS, which is a "contrary prediction" to what is typically expected for ER-positive breast cancer. This is an interesting observation for TNBC. A deeper dive into the biological reasons behind this inverse relationship in TNBC, perhaps linking it to the specific characteristics of high-risk TNBCs that respond well to chemotherapy, would strengthen the discussion.
  5. The use of a 17-gene panel for immune activity assessment is a good approach. However, the definition of "immune-weak (<40)" and "immune-strong (>=55)" based on immunity scores seems somewhat arbitrary without further justification or reference to established cut-offs. Could the authors provide more context on how these thresholds were determined and validated, or discuss the sensitivity analysis around these cut-offs?
  6. While the study found that immune score was not an independent prognostic factor in multivariate analyses (Tables 4 and 5), the univariate analysis showed significant associations with DFS and OS, especially in later stages. This discrepancy warrants more discussion. What other factors might be confounding the immune score's prognostic value in the multivariate model, and how might future studies address this to better understand its independent contribution?
  7. The correlation between high immune gene expression and positivity in tumor cells' PD-L1 expression, but not TILs' PD-L1 expression, is intriguing. The authors suggest this might indicate immune evasion by tumor cells. This hypothesis could be explored further with additional mechanistic insights or by discussing how this specific PD-L1 expression pattern might influence response to immunotherapy.
  8. The follow-up period, with a median of 92 months (ranging from 7-182 months), is quite good for a retrospective study. However, the wide range suggests some patients had much shorter follow-up. Could the authors discuss if the varying follow-up duration might have introduced any bias, particularly for survival outcomes, and how they accounted for this in their statistical models?
  9. The study states that "Normal-like data were excluded from further analyses because of potential contamination from the normal breast tissue." While this is a standard practice, it would be beneficial to briefly mention the proportion of samples initially classified as normal-like and the potential implications of their exclusion on the overall representation of TNBC subtypes in the cohort.
  10. The discussion section could benefit from a more direct comparison of the study's findings with other major TNBC classification systems (e.g., Lehmann et al., Burstein et al.) beyond just mentioning the prevalence of basal-like subtypes. A table or a more structured comparison of prognostic implications and treatment responses across these systems, in light of the current study's results, would enhance the manuscript's contribution to the field.

Author Response

Please see the file of point-by-point response.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

I thank the authors for responding appropriately to comments and making the necessary improvements. The quality of the research and writing has greatly improved.

Reviewer 3 Report

Comments and Suggestions for Authors

All the queries raised have been satisfactorily clarified and resolved.