Systemic Therapies for Desmoid Tumors: A Review of Past, Present, and Future Treatments

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This is a review article which traces the evolution of desmoid tumor (DT) management from "surgery-dominant" paradigms to individualized, multimodal treatment algorithms. This is logically organised into four principal classes of medical therapy, making it a valuable reference for clinicians. Especially, it states that the recent Phase III data, particularly for sorafenib and nirogacestat, which provides a high level of contemporary clinical utility. I think this manuscript will be suitable for publication in “Cancers” after minor revisions based on the reviewers’ suggestions.

 

1. The literature search strategy is generally well described. However, it is unclear whether this review was conducted in accordance with PRISMA guidelines. The authors should clarify this point and, if applicable, include a PRISMA flow diagram and provide details regarding study selection and risk of bias assessment to improve transparency and reproducibility.
2. The authors highlight cardiotoxicity as a limiting factor for doxorubicin. Brief advice on screening or monitoring protocols used at "sarcoma centers of excellence" would be a practical addition for readers.
3. For the section on nirogacestat, the authors correctly note the high incidence of ovarian dysfunction (75%) in women of childbearing age. Highlighting the importance of fertility counselling before initiating gamma-secretase inhibitors would be a meaningful clinical addition.
4. The RINGSIDE trial data are mentioned as "preliminary" at this moment. In the case updated results would be disclosed before final publication, adding it as citation would be highly beneficial to be most up-to-date review.

 

 

Minor comments (Technical/Formatting Notes)

1. Table Consistency: Ensure that all abbreviations in the tables (e.g., NR, PR, CR) are consistently defined in the legends throughout the manuscript.
2. “Familial Adenomatous Polyposis (FAP).“ in line 160 should be included in table 3 abbreviations.
3. The font size of “Table 4. Key reported trials on targeted therapies for desmoid Tumor/desmoid Fibromatosis (DT/DF).” seems to be larger than required.
4. There is an unnecessary space “f DTs, γ-secretase” in line 213.

Author Response

Reviewer 1

We thank Reviewer 1 for their careful reading of the manuscript and their thoughtful, specific suggestions. Their comments helped us augment and add to improve the utility of the review for a clinician audience.

Comment 1.1: “The literature search strategy is generally well described. However, it is unclear whether this review was conducted in accordance with PRISMA guidelines. The authors should clarify this point and, if applicable, include a PRISMA flow diagram and provide details regarding study selection and risk of bias assessment to improve transparency and reproducibility.”

Response: We appreciate this important observation and agree that the original wording of the Methods section could create the impression that the manuscript was intended as a formal systematic review conducted under PRISMA methodology. This manuscript was designed as a narrative review informed by a structured, comprehensive literature search, rather than a PRISMA-based systematic review or meta-analysis.

To address this concern, we have revised the Methods section to state this explicitly and to better describe how the literature was identified, screened, and synthesized. Specifically, we clarify that the search was conducted across major databases using predefined search terms and eligibility considerations to ensure broad and rigorous coverage of the contemporary and historical desmoid tumor systemic therapy literature, but that the review did not follow a formal PRISMA protocol. [Page 2]

We have also included language to detail how evidence selection and interpretation were guided by clinical relevance, landmark studies, practice-changing prospective trials, and expert multidisciplinary judgment. Our goal was to provide a balanced and clinically useful synthesis of the field rather than a pooled quantitative evidence assessment. Importantly, we have clarified that this approach was intended to minimize selection bias through a broad predefined search strategy while remaining appropriate to the aims of a narrative review focused on the evolution of systemic therapies for desmoid tumors.

Comment 1.2: “The authors highlight cardiotoxicity as a limiting factor for doxorubicin. Brief advice on screening or monitoring protocols used at "sarcoma centers of excellence" would be a practical addition for readers.”

Response: We agree that a practical note regarding cardiac monitoring adds important clinical context to the discussion of anthracycline-based therapy. To address this, we revised the Conclusions section to note that, at high-volume sarcoma centers, pegylated liposomal doxorubicin is generally used with baseline cardiac assessment and risk-adapted interval monitoring during treatment, allowing for dose adjustment or discontinuation if clinically meaningful cardiotoxicity develops. [Page 9]

Comment 1.3: “For the section on nirogacestat, the authors correctly note the high incidence of ovarian dysfunction (75%) in women of childbearing age. Highlighting the importance of fertility counselling before initiating gamma-secretase inhibitors would be a meaningful clinical addition.”

Response: We sincerely thank the reviewer for these important and clinically meaningful comments. We agree that ovarian dysfunction associated with nirogacestat is a major treatment-selection and survivorship consideration, particularly given the predominance of desmoid tumor in young women. In response, we revised the Conclusions section to more explicitly frame ovarian dysfunction as a major toxicity of γ-secretase inhibition, note that recovery after discontinuation may occur but is not predictable at the individual-patient level, and emphasize the importance of pre-treatment counseling, fertility preservation discussion for patients desiring future childbearing, early gynecology or reproductive endocrinology involvement when appropriate, and ongoing monitoring during therapy. We agree that these additions improve the clinical applicability of the review and better balance discussion of treatment efficacy with toxicity and survivorship considerations. [Page 9]

Comment 1.4: “The RINGSIDE trial data are mentioned as "preliminary" at this moment. In the case updated results would be disclosed before final publication, adding it as citation would be highly beneficial to be most up-to-date review.”

Response: We appreciate this note and are in complete agreement. If there are any further publications pertaining to the RINGSIDE trial prior to the release of this review, we would appreciate the opportunity to update the manuscript and citations. As it currently stands, we have included the data and citation (citation 35, Kasper et al.) that are most recent published material from the RINGSIDE group. We will continue to monitor in case this opportunity arises.

Minor Comments: “Table Consistency: Ensure that all abbreviations in the tables (e.g., NR, PR, CR) are consistently defined in the legends throughout the manuscript.

“Familial Adenomatous Polyposis (FAP).“ in line 160 should be included in table 3 abbreviations.

The font size of “Table 4. Key reported trials on targeted therapies for desmoid Tumor/desmoid Fibromatosis (DT/DF).” seems to be larger than required.

There is an unnecessary space “f DTs, γ-secretase” in line 213”

Response: Thank you for these granular comments- tables 1-4 reviewed and all define the included abbreviations including NR, PR, and CR in their respective legends. FAP was not defined in table three and now is (Page 6). The font size of Table 4 title was reduced to match the other tables (Page 6), and extra space removed (Page 8).

Reviewer 2 Report

Comments and Suggestions for Authors

3.1. NSAIDs 

Section 3.1 examines the historical and contemporary applications of NSAIDs. However, a comprehensive revision is necessary to address critical inconsistencies and enhance scientific rigor before publication. 

1. Discrepancy between Narrative and Evidence: The authors claim that NSAIDs have not demonstrated regression rates surpassing the natural progression of the disease. Yet, Table 2 indicates an ORR of 57% for sulindac and 24% for meloxicam. This fundamental contradiction between the narrative and the tabulated data necessitates a critical re-evaluation.
2. Integration of Molecular Stratification: The discussion on CTNNB1 mutations is insightful. It is recommended that the authors elaborate on the CTNNB1 S45F mutation and its specific association with reduced meloxicam efficacy, emphasizing the pivotal findings by Hamada.
3. Clinical Guidance and Biological Rationale: The authors should provide a practical guide detailing the recommended adjuncts for pain management. Additionally, a more in-depth exploration of the role of COX-2 inhibitors is warranted, given their overexpression as the primary biological rationale for this class of therapeutics. Although the NCCN guidelines incorporate NSAIDs primarily for pain management or as adjuncts, this section lacks a practical guide on agent selection. To serve as an effective therapeutic guide, this review must bridge the gap between the biological foundation of COX-2 and molecular stratification, specifically addressing how mutations such as S45F influence the effectiveness of this inhibition.

3.3. Tyrosine Kinase Inhibitors 

3. Hierarchical Evidence vs. Numerical Rates: In Section 3.3, the authors characterize sorafenib as more effective. However, they cite a review study reporting disease control rates of 67–96% for sorafenib, 78–92% for imatinib, and 84% for pazopanib. The lower limit of sorafenib is numerically lower than that of its competitors.
4. Correction of Narrative: It is suggested to review this data to indicate that sorafenib’s effectiveness is primarily related to the robustness of its Phase III clinical evidence and the demonstrated 2-year PFS of 81%.

3.4. Gamma-Secretase Inhibitors 

1. Damage Mitigation Strategies: The authors highlight nirogacestat as a drug that improves the quality of life for patients. However, they also mention a high incidence of ovarian dysfunction (75%) in women of reproductive age. It is suggested that the authors integrate damage mitigation strategies into their review, such as hormonal follow-up protocols and endocrinology consultation. It is crucial to clarify whether side effects, such as ovarian dysfunction, are reversible post-discontinuation or if they require a formal risk assessment prior to the initiation of therapy.
2. Fertility Preservation: In Section 3.4, the authors should explicitly emphasize fertility preservation due to the high percentage of ovarian dysfunction associated with nirogacestat.

Table 2 

The authors must review the descriptions in Section 3.1 and Table 2, as there are significant discrepancies between the text and the table regarding response rates that must be resolved or appropriately explained.

Author Response

Reviewer 2

We thank Reviewer 2 for their careful reading of the manuscript and their thoughtful, specific suggestions. Their comments helped us identify areas where greater clarity and rigor were warranted.

Comment 2.1: “Section 3.1 examines the historical and contemporary applications of NSAIDs. However, a comprehensive revision is necessary to address critical inconsistencies and enhance scientific rigor before publication.

Discrepancy between Narrative and Evidence: The authors claim that NSAIDs have not demonstrated regression rates surpassing the natural progression of the disease. Yet, Table 2 indicates an ORR of 57% for sulindac and 24% for meloxicam. This fundamental contradiction between the narrative and the tabulated data necessitates a critical re-evaluation.

Integration of Molecular Stratification: The discussion on CTNNB1 mutations is insightful. It is recommended that the authors elaborate on the CTNNB1 S45F mutation and its specific association with reduced meloxicam efficacy, emphasizing the pivotal findings by Hamada.

Clinical Guidance and Biological Rationale: The authors should provide a practical guide detailing the recommended adjuncts for pain management. Additionally, a more in-depth exploration of the role of COX-2 inhibitors is warranted, given their overexpression as the primary biological rationale for this class of therapeutics. Although the NCCN guidelines incorporate NSAIDs primarily for pain management or as adjuncts, this section lacks a practical guide on agent selection. To serve as an effective therapeutic guide, this review must bridge the gap between the biological foundation of COX-2 and molecular stratification, specifically addressing how mutations such as S45F influence the effectiveness of this inhibition.”

Response: Re NSAIDS: We thank the reviewer for this thoughtful and important critique. We agree that the original wording did not explicitly contextualize the reported NSAID response rates against the known variable natural history of desmoid tumors and therefore risked creating an apparent discrepancy between the narrative text and Table 2. In response, we revised Section 3.1 to more explicitly describe the natural history of desmoid tumors, including their well-recognized potential for prolonged stability, spontaneous regression, and regression following an initial period of progression, and we incorporated prospective active-surveillance data to provide this context. We then clarified that, when interpreted against these baseline rates of untreated stabilization and regression, the published NSAID response rates have not been clearly shown to exceed what may occur as part of the natural course of disease. We agree this revision improves the scientific clarity and interpretability of the section.

Re CTNNB1 and characterized mutations (S45F) with treatment implications: We agree that the discussion of molecular stratification is strengthened by more explicitly highlighting the clinical relevance of CTNNB1 mutation subtype, particularly the findings of Hamada et al. regarding S45F and reduced meloxicam efficacy. In response, we revised the Future Directions section to specifically note that S45F was associated with poor response to meloxicam in the prospective cohort reported by Hamada and to emphasize that CTNNB1 mutation subtype may have predictive as well as prognostic value in treatment selection. We agree that this addition improves the translational and clinical relevance of the discussion. [Page 10]

Re NSAIDS for S45F: The mechanistic work by Hamada et all demonstrated that Meloxicam may be less effective in S45F-mutant desmoid tumors because S45F appears to be associated with stronger stabilization and nuclear accumulation of β-catenin, leading to more dominant Wnt/β-catenin pathway signaling. As such, the conclusion wasn’t necessarily that S45F mutated tumors are “NSAID resistant”, but rather, may be more aggressive and thus more resistance to COX inhibition (and in the setting of Meloxicam, selectively COX-2 inhibition). To convey this, we highlight the presence of NSAIDS in the NCCN guidelines, and simultaneous importance of ongoing molecular stratification, to the reviewers point, to generate more therapeutically implicated variants that may help guide prognostication, and hopefully personalized therapy plans.

Comment 2.2: “Tyrosine Kinase Inhibitors

Hierarchical Evidence vs. Numerical Rates: In Section 3.3, the authors characterize sorafenib as more effective. However, they cite a review study reporting disease control rates of 67–96% for sorafenib, 78–92% for imatinib, and 84% for pazopanib. The lower limit of sorafenib is numerically lower than that of its competitors.

Correction of Narrative: It is suggested to review this data to indicate that sorafenib’s effectiveness is primarily related to the robustness of its Phase III clinical evidence and the demonstrated 2-year PFS of 81%.”

Response: We agree that the original wording overstated the strength of cross-study numerical comparisons among tyrosine kinase inhibitors, particularly given the overlapping disease control ranges reported across heterogeneous series. In response, we revised Section 3.3 to clarify that sorafenib is highlighted not because it can be definitively said to have uniformly superior response rates across all available studies, but because it is supported by the most robust prospective evidence within the TKI class, including the landmark randomized phase III trial demonstrating a 2-year progression-free survival of 81% versus 36% with placebo. [Page 7, Page 8] We agree that this revision improves the precision of the discussion and more appropriately distinguishes evidentiary strength from indirect numerical comparison.

Comment 2.3: “Gamma-Secretase Inhibitors

Damage Mitigation Strategies: The authors highlight nirogacestat as a drug that improves the quality of life for patients. However, they also mention a high incidence of ovarian dysfunction (75%) in women of reproductive age. It is suggested that the authors integrate damage mitigation strategies into their review, such as hormonal follow-up protocols and endocrinology consultation. It is crucial to clarify whether side effects, such as ovarian dysfunction, are reversible post-discontinuation or if they require a formal risk assessment prior to the initiation of therapy.

Fertility Preservation: In Section 3.4, the authors should explicitly emphasize fertility preservation due to the high percentage of ovarian dysfunction associated with nirogacestat.”

Response: The importance of this critique is reflected in its presence in all three reviewers notes. We have added a full section to incorporate this feedback, and detailed our response in Response 1.3. [Page 9 of manuscript]

Comment 2.4:. “Table 2

The authors must review the descriptions in Section 3.1 and Table 2, as there are significant discrepancies between the text and the table regarding response rates that must be resolved or appropriately explained.”

Response: We agree that the original wording in Section 3.1 could be read as insufficiently reconciling the numerical response rates shown in Table 2, particularly the early sulindac series, with our interpretation of the broader evidence base. In response, we revised the text to more explicitly explain that desmoid tumors have a highly variable natural history, including spontaneous stabilization and regression, and that although the reported sulindac response rate appears numerically notable, subsequent NSAID studies have shown more modest activity. We also clarified that the available NSAID literature remains limited by small series and heterogeneity, such that a reproducible treatment effect beyond the expected natural history of disease has not been clearly established. [Page 4] This revision was intended to better align the narrative interpretation with the tabulated data.

Reviewer 3 Report

Comments and Suggestions for Authors

This narrative review surveys systemic therapies for desmoid tumors across four drug classes. While the topic is clinically relevant and the organizational framework is logical, the manuscript has substantial deficiencies in scientific rigor, evidence synthesis, and scholarly completeness that require major revision.

1. The Methods section (lines 61–73) describes a systematic search strategy with inclusion/exclusion criteria, which creates an expectation of systematic review methodology — yet the manuscript provides no PRISMA flow diagram, no study quality assessment (e.g., Cochrane RoB, Newcastle-Ottawa), no accounting of studies screened versus included, and no heterogeneity analysis. This is a fundamental mismatch: either downgrade the Methods to explicitly state this is a narrative/scoping review and remove the systematic search framing, or execute actual systematic review methodology. As written, the manuscript risks misleading readers about the rigor of evidence selection.
2. The manuscript repeatedly compares response rates across non-randomized, single-arm studies with different patient populations, RECIST versions, follow-up durations, and enrollment criteria, yet presents these comparisons as if they establish relative efficacy. Line 153 states methotrexate disease control rate was "71–100%, compared to… Doxil with 90–100%," and line 172 claims Doxil "outperforms its competitors." Line 189–191 ranks sorafenib disease control at "67–96% compared to imatinib (78–92%) or pazopanib (84%)." These unadjusted cross-study comparisons are scientifically inappropriate without explicit caveats about confounding by indication, selection bias, and heterogeneous endpoints. The imatinib range (78–92%) actually overlaps with sorafenib (67–96%), yet the text frames sorafenib as clearly "superior" — this is misleading.
3. The ovarian dysfunction rate with nirogacestat (75% of childbearing-age women, line 225) is mentioned in a single subordinate clause despite being arguably the most consequential safety signal in the entire review, given DT's demographic predominance in young women. By contrast, sorafenib toxicity is described only as "grade 1 and grade 2 events" (line 207) without specifying the actual 47% grade ≥3 rate reported in Table 4. The doxorubicin cardiotoxicity discussion (lines 169–179) references only 2 of 35 patients from a single study, without citing cumulative dose-dependent cardiotoxicity data that are well-established in the anthracycline literature. A review claiming to guide treatment selection must present toxicity data with the same rigor as efficacy data, particularly for a non-lethal disease where quality-of-life tradeoffs are central.
4. The review acknowledges that DTs can spontaneously regress (lines 52–53, 83–84, 96–98) and that active surveillance is the recommended first step (line 53), yet never quantifies spontaneous regression rates or disease stabilization rates from prospective observation cohorts. Without this baseline, readers cannot contextualize whether reported ORRs from single-arm studies (e.g., sulindac 57%, meloxicam 24%) represent true treatment effects or natural history. The tamoxifen section appropriately raises this concern (lines 96–98, 107–108) but fails to apply the same critical lens to NSAID, methotrexate, or even single-arm TKI data.
5. Reference 13 (Quast et al.) is dated 2015 in the reference list but 2016 in Table 1. The Pires de Camargo study (Ref 19, 2010) is cited on line 174 as demonstrating Doxil's superiority over TKIs, but this study predates the 2018 sorafenib phase III trial and the 2023 nirogacestat phase III trial — characterizing it as a definitive comparison to TKIs is anachronistic. Reference 12 (Riedel & Agulnik, 2022) is cited for both COX-2 overexpression in DTs (line 127) and methotrexate adverse effects (line 150–151) — the reader should verify that a single general review is the most appropriate primary source for both claims rather than citing the original research that established COX-2 expression in DTs.

Author Response

Reviewer 3

We are equally appreciative of Reviewer 3’s engagement with the work. Their observations prompted us to reconsider several key points in our analysis and discussion, and we believe the manuscript is more precise as a result.

Comment 3.1: “The Methods section (lines 61–73) describes a systematic search strategy with inclusion/exclusion criteria, which creates an expectation of systematic review methodology — yet the manuscript provides no PRISMA flow diagram, no study quality assessment (e.g., Cochrane RoB, Newcastle-Ottawa), no accounting of studies screened versus included, and no heterogeneity analysis. This is a fundamental mismatch: either downgrade the Methods to explicitly state this is a narrative/scoping review and remove the systematic search framing, or execute actual systematic review methodology. As written, the manuscript risks misleading readers about the rigor of evidence selection.”

Response: We recognize the crucial importance of this feedback, highlighted by its presence in the notes from both reviewer 1 and 2. We have made changes to directly address this weakness, and detailed them in Response to 1.1. [Page 2 of manuscript]

Comment 3.2: “The manuscript repeatedly compares response rates across non-randomized, single-arm studies with different patient populations, RECIST versions, follow-up durations, and enrollment criteria, yet presents these comparisons as if they establish relative efficacy. Line 153 states methotrexate disease control rate was "71–100%, compared to… Doxil with 90–100%," and line 172 claims Doxil "outperforms its competitors." Line 189–191 ranks sorafenib disease control at "67–96% compared to imatinib (78–92%) or pazopanib (84%)." These unadjusted cross-study comparisons are scientifically inappropriate without explicit caveats about confounding by indication, selection bias, and heterogeneous endpoints. The imatinib range (78–92%) actually overlaps with sorafenib (67–96%), yet the text frames sorafenib as clearly "superior" — this is misleading.”

Response: We thank the reviewer for this careful and important comment. We agree that unadjusted cross-study comparisons among heterogeneous single-arm studies can overstate relative efficacy, particularly in a rare disease literature that spans different patient populations, treatment eras, endpoints, and follow-up durations. In response, we revised the relevant sections to remove language implying definitive superiority based on numerical comparisons alone, and instead added clearer caveats regarding the limitations of indirect cross-study comparison, including heterogeneity in study design and potential confounding by selection and indication. We also revised the discussion to emphasize that agents such as sorafenib are highlighted primarily on the basis of the strength of prospective evidence, rather than simple comparison of reported response ranges across non-randomized series. We agree that these changes improve the scientific precision and balance of the manuscript. [Page 6, Page 9]

Comment 3.3: “The ovarian dysfunction rate with nirogacestat (75% of childbearing-age women, line 225) is mentioned in a single subordinate clause despite being arguably the most consequential safety signal in the entire review, given DT's demographic predominance in young women. By contrast, sorafenib toxicity is described only as "grade 1 and grade 2 events" (line 207) without specifying the actual 47% grade ≥3 rate reported in Table 4. The doxorubicin cardiotoxicity discussion (lines 169–179) references only 2 of 35 patients from a single study, without citing cumulative dose-dependent cardiotoxicity data that are well-established in the anthracycline literature. A review claiming to guide treatment selection must present toxicity data with the same rigor as efficacy data, particularly for a non-lethal disease where quality-of-life tradeoffs are central.”

Response: Regarding the reference to ovarian dysfunction with nirogacestat, the importance of this critique is reflected in its presence in all three reviewers notes. The original wording did not give this issue sufficient emphasis, nor did it adequately detail recommended management. We have added a full section to incorporate this feedback, and detailed our response in Response 1.3. [Page 9 of manuscript]

Regarding the reference to the sorafenib toxicity data, additional details from the study were included, and written as follows: “While broadly the medication was well tolerated, the trial described the safety profile in detail. The most commonly reported adverse events were grade 1 and grade 2 events (53% sorafenib v 69% placebo), and the group reported 47% of patients developed grade 3 or 4 adverse events (25% placebo). The grade 3 and 4 adverse events were most commonly cutaneous or gastrointestinal, and led to a significantly higher rate of discontinuation of the trial regimen in the sorafenib group than in the placebo group.”  [Page 8]

Regarding the reference to doxorubicin cardiotoxicity, we agree that, in a disease such as desmoid tumor where treatment decisions are heavily influenced by long-term toxicity and quality-of-life considerations, adverse-event discussion should be presented with the same clinical seriousness as efficacy data. In response, we revised the Conclusions section to strengthen the discussion of anthracycline-related toxicity by emphasizing cumulative toxicity and patient-specific comorbidities, and by adding practical language noting that, at high-volume sarcoma centers, anthracycline-based therapy is typically accompanied by baseline cardiac assessment and risk-adapted interval monitoring during treatment, with dose adjustment or discontinuation if clinically meaningful cardiac dysfunction emerges. We agree that this revision improves the balance and clinical applicability of the manuscript.

 

Comment 3.4: “The review acknowledges that DTs can spontaneously regress (lines 52–53, 83–84, 96–98) and that active surveillance is the recommended first step (line 53), yet never quantifies spontaneous regression rates or disease stabilization rates from prospective observation cohorts. Without this baseline, readers cannot contextualize whether reported ORRs from single-arm studies (e.g., sulindac 57%, meloxicam 24%) represent true treatment effects or natural history. The tamoxifen section appropriately raises this concern (lines 96–98, 107–108) but fails to apply the same critical lens to NSAID, methotrexate, or even single-arm TKI data.”

Response: Spontaneous regression and disease stabilization under active surveillance provide essential context for interpreting response rates reported in single-arm studies, particularly in desmoid tumor. In response to this feedback, we revised the first treatment-specific section to explicitly quantify the variable natural history of desmoid tumors using prospective observational data, including rates of RECIST progression and spontaneous regression during surveillance. We chose to introduce this framework early in the review so that it would inform interpretation of response data throughout the manuscript. We included this revision to improve the rigor of the review by helping readers interpret observed response rates against an appropriate untreated baseline. [Page 4]

Comment 3.5:.“Reference 13 (Quast et al.) is dated 2015 in the reference list but 2016 in Table 1. The Pires de Camargo study (Ref 19, 2010) is cited on line 174 as demonstrating Doxil's superiority over TKIs, but this study predates the 2018 sorafenib phase III trial and the 2023 nirogacestat phase III trial — characterizing it as a definitive comparison to TKIs is anachronistic. Reference 12 (Riedel & Agulnik, 2022) is cited for both COX-2 overexpression in DTs (line 127) and methotrexate adverse effects (line 150–151) — the reader should verify that a single general review is the most appropriate primary source for both claims rather than citing the original research that established COX-2 expression in DTs.”

Response: Thank you for the detailed edits- the citation list updated to the correct date of the Quast publication: 2016. The Doxil comparison discussion was also updated to highlight that cross-era comparisons with newer targeted agents should be interpreted cautiously, and in the study by Pires de Camargo et al., Doxil demonstrated favorable progression-free survival relative to several systemic options available at that time. Lastly, added the following basic science manuscript as foundational evidence for over expression (>80%) of COX in desmoid tumors [Mignemi, N. A., D. M. Itani, J. H. Fasig, V. L. Keedy, K. R. Hande, B. W. Whited, K. C. Homlar, H. Correa, C. M. Coffin, J. O. Black, et al. "Signal transduction pathway analysis in desmoid-type fibromatosis: Transforming growth factor-beta, cox2 and sex steroid receptors." Cancer Sci 103 (2012): 2173–80. 10.1111/cas.12037. https://www.ncbi.nlm.nih.gov/pubmed/23035734.]

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

Dear authors, thank you for adequately addressing the suggestions. The manuscript does not require any changes.

Author Response

Comment 2.1: “Dear authors, thank you for adequately addressing the suggestions. The manuscript does not require any changes.”

Response:

Thank you very much to reviewer 2 for their detailed comments in round one.

Reviewer 3 Report

Comments and Suggestions for Authors

This narrative review tracing the evolution of systemic therapy for desmoid tumors (DTs) across four medication classes: NSAIDs, cytotoxic chemotherapy, tyrosine kinase inhibitors (TKIs), and γ-secretase inhibitors. The manuscript appropriately contextualizes earlier antiestrogen-based approaches, emphasizes the pivotal phase III sorafenib and nirogacestat trials, and closes with biomarker-informed future directions centered on CTNNB1 subtype stratification. The scope is clinically relevant and the narrative arc is coherent, but the current draft has substantive citation errors, structural inconsistencies, and methodological underspecification that should be corrected before publication.

1. The most serious example is in Section 3.3: the "landmark randomized phase III trial by Gounder et al." (lines 241, 243) is cited as [28], which is actually Garbay et al. on MTX/vinblastine (Ann Oncol 2011). The correct reference for the sorafenib phase III trial is [32]. Similarly, line 227 attributes a 2023 systematic review to "Tsukamoto et al." but cites [19] (Colombo et al., active surveillance); the intended reference is [23]. Line 236's citation cluster "[23, 31, 35, 37]" supporting sorafenib as first-line TKI includes two imatinib studies (Penel [35], Kasper GISG [37]) that do not support the specific claim. Given the topic, these errors materially affect the reader's ability to verify the evidence base.
2. Both the Conclusions (line 274) and Future Directions (line 320) are labeled as Section 4.
3. The Methods describe a "narrative review with structured literature search", yet no study-selection numbers (identified/screened/included), no exclusion tally, and no risk-of-bias consideration are provided. If the authors intend a narrative framing, align with SANRA reporting standards.
4. Table 1 (hormonal therapy) and Table 2 (NSAIDs) list only two studies each, omitting influential series (e.g., Nishida meloxicam 2020). Table 4 conflates TKIs and γ-secretase inhibitors despite being labeled "targeted therapies"; consider splitting, or retitling and adding a separating row.
5. The statement that NSAID use followed observation of COX-2 overexpression  is anachronistic, Tsukada 1992 sulindac data (ref [20]) predates that rationale. The doxorubicin mechanism description (line 199, "topoisomerase inhibitor … believed to function via DNA intercalation") conflates two mechanisms imprecisely; topoisomerase II poisoning is the primary accepted action. The review also omits emerging agents (tegavivint/β-catenin-directed therapies) and real-world comparative sorafenib/pazopanib data (Noujaim 2024, already in the reference list as [30] but not discussed).

Author Response

Reviewer 3

We sincerely thank reviewer three for a second round of careful, thoughtful, and constructive evaluation of our manuscript. We appreciate the reviewer’s recognition of the clinical relevance and overall narrative structure of the review, as well as the specific recommendations to improve citation accuracy, structural consistency, and methodological clarity. We found these comments highly valuable and have revised the manuscript accordingly. Below, we provide a point-by-point response detailing the changes made in response to each of the reviewer’s observations.

Comment 3.1: “The most serious example is in Section 3.3: the "landmark randomized phase III trial by Gounder et al." (lines 241, 243) is cited as [28], which is actually Garbay et al. on MTX/vinblastine (Ann Oncol 2011). The correct reference for the sorafenib phase III trial is [32]. Similarly, line 227 attributes a 2023 systematic review to "Tsukamoto et al." but cites [19] (Colombo et al., active surveillance); the intended reference is [23]. Line 236's citation cluster "[23, 31, 35, 37]" supporting sorafenib as first-line TKI includes two imatinib studies (Penel [35], Kasper GISG [37]) that do not support the specific claim.”

Response: Thank you for these detailed observations. We have corrected the two, highlighted, incorrect citations. We also removed the Penel and Kasper citations after the sentence discussing sorafenib: “Sorafenib, an oral multitargeted receptor tyrosine kinase inhibitor, has taken a leading role as the primary TKI utilized as a first line agent for desmoid tumor.” The references were initially included as context and a comparator, but we agree it is less clear and could introduce confusion, so they have been removed. [Page 7-8].

Comment 3.2: “Both the Conclusions (line 274) and Future Directions (line 320) are labeled as Section 4.”

Response: Thank you- a typo after splitting the sections per recommendations from the first round of revisions. Now corrected on page 10.

Comment 3.3: “The Methods describe a "narrative review with structured literature search", yet no study-selection numbers (identified/screened/included), no exclusion tally, and no risk-of-bias consideration are provided. If the authors intend a narrative framing, align with SANRA reporting standards.”

Response: We thank the reviewer for this important suggestion. We agree that a narrative review should be reported transparently and in a manner aligned with accepted standards for narrative evidence synthesis. In response, we revised the Methods section to more clearly describe the review as a narrative review with a structured literature search and to note that the manuscript was prepared with attention to key SANRA domains, including explicit statement of aims, transparent description of the literature search, appropriate referencing, and balanced evidence synthesis. [Page 2] We believe this clarification improves the methodological transparency and reporting of the manuscript while preserving its intended narrative design.

Comment 3.4: “Table 1 (hormonal therapy) and Table 2 (NSAIDs) list only two studies each, omitting influential series (e.g., Nishida meloxicam 2020). Table 4 conflates TKIs and γ-secretase inhibitors despite being labeled "targeted therapies"; consider splitting, or retitling and adding a separating row.”

Response: We thank the reviewer for this comment, highlighting an opportunity to add to the key articles we highlight. We reviewed the cited literature and were not able to identify a Nishida first author meloxicam publication from 2020. However, there is a relevant Nishida meloxicam study: a 2010 pilot study, which we have now included in Table 2. We also agree that the organization of Table 4 benefits from clearer distinction between therapeutic classes, and we therefore revised the table to add separate labels demarcating the tyrosine kinase inhibitors and the γ-secretase inhibitors. We appreciate this suggestion, which improved the clarity and structure of the manuscript.

Comment 3.5: “ The statement that NSAID use followed observation of COX-2 overexpression is anachronistic, Tsukada 1992 sulindac data (ref [20]) predates that rationale. The doxorubicin mechanism description (line 199, "topoisomerase inhibitor … believed to function via DNA intercalation") conflates two mechanisms imprecisely; topoisomerase II poisoning is the primary accepted action. The review also omits emerging agents (tegavivint/β-catenin-directed therapies) and real-world comparative sorafenib/pazopanib data (Noujaim 2024, already in the reference list as [30] but not discussed).

 

Response: We thank the reviewer for this careful and constructive comment. We agree that these points improve both the historical precision and scientific clarity of the review. In response, we revised the NSAID section to avoid implying that NSAID use began only after recognition of COX-2 overexpression, and instead clarified that NSAIDs were explored earlier as relatively well-tolerated noncytotoxic agents, with later biologic studies providing additional mechanistic rationale for their continued investigation. [Page 4] We also revised the doxorubicin discussion to more accurately describe its multifaceted mechanism of action, including topoisomerase II–associated cytotoxicity while acknowledging emerging data supporting additional topoisomerase II–independent intercalative effects during DNA replication. [Page 6] Finally, we expanded the discussion of future and contemporary therapies to include emerging β-catenin-directed strategies such as tegavivint and added mention of the real-world comparative sorafenib/pazopanib data from Noujaim et al. [ Page 7 & Page 10]

Round 3

Reviewer 3 Report

Comments and Suggestions for Authors

1. Round 2 flagged that the Gounder phase III sorafenib trial was incorrectly cited as [28] (Azzarelli MTX/vinblastine) at two locations. Line 250 has been corrected to [34], but line 253, supporting the key efficacy figures of "81% vs 36% 2-year PFS"  still cites [28]. Reference [28] is the Azzarelli MTX/vinblastine paper, not the sorafenib trial. Given that this exact error was specifically raised in the prior round and the author response stated it had been corrected, the residual artifact warrants careful re-checking.
2. The author response indicated removal of "the Penel and Kasper citations after the sentence discussing sorafenib." However, the sentence still contains [37] (Penel imatinib) and [39] (Kasper imatinib)." The reference numbers shifted because of newly inserted citations, but the cited studies are unchanged. These two imatinib trials do not support the specific claim about sorafenib's first-line TKI status and should be removed as previously committed.
3. The revised description is better but still places topoisomerase II–associated DNA damage and DNA intercalation on roughly equal footing. The accepted hierarchy is that topoisomerase II poisoning is the primary cytotoxic mechanism, with intercalation as a contributing or secondary effect.

Author Response

Comment 3.1: “Round 2 flagged that the Gounder phase III sorafenib trial was incorrectly cited as [28] (Azzarelli MTX/vinblastine) at two locations. Line 250 has been corrected to [34], but line 253, supporting the key efficacy figures of "81% vs 36% 2-year PFS"  still cites [28]. Reference [28] is the Azzarelli MTX/vinblastine paper, not the sorafenib trial. Given that this exact error was specifically raised in the prior round and the author response stated it had been corrected, the residual artifact warrants careful re-checking.”

Response: Thank you for identifying this citation error- it was indeed indicated in round two that there was an incorrect citation, and only one of the locations was corrected in the previous round of edits. The second location is now corrected. [Page 8]

Comment 3.2: “The author response indicated removal of "the Penel and Kasper citations after the sentence discussing sorafenib." However, the sentence still contains [37] (Penel imatinib) and [39] (Kasper imatinib)." The reference numbers shifted because of newly inserted citations, but the cited studies are unchanged. These two imatinib trials do not support the specific claim about sorafenib's first-line TKI status and should be removed as previously committed.”

Response: We agree with the reviewer, the transition to sorafenib on line 245 does not need the imatinib citations as “context” for the claim- they were not correctly removed in the prior round, but they are now both removed from line 245. [Page 8]

Comment 3.3: “The revised description is better but still places topoisomerase II–associated DNA damage and DNA intercalation on roughly equal footing. The accepted hierarchy is that topoisomerase II poisoning is the primary cytotoxic mechanism, with intercalation as a contributing or secondary effect.

Response: We appreciate the thoughtful comments on the current knowledge pertaining to the mechanism of action of doxorubicin. On page six, to address this area of ongoing investigation, we wrote: “Doxorubicin is an anthracycline with multifaceted cytotoxic activity, classically linked to topoisomerase II–associated DNA damage, but also mediated through DNA intercalation. More recent data suggest that, during DNA replication, doxorubicin may additionally impair fork progression through a topoisomerase II–independent intercalative mechanism”. We aimed to highlight that the classically accepted mechanism of action is topoisomerase II poisoning, however, recent studies, including the cited 2022 publication, have identified distinct mechanistic pathways that function independently of topoisonmerase II. We hoped to include this recent literature in the manuscript to highlight the ongoing evolution in the space. We have made the following modification to further emphasize that the topoisomerase II mechanism is the primary mechanism: “ Doxorubicin is an anthracycline with multifaceted cytotoxic activity, predominantly linked to topoisomerase II–associated DNA damage, but also mediated through DNA intercalation.” [Page 6]