Round 1

Reviewer 1 Report (Previous Reviewer 2)

Comments and Suggestions for Authors

A major revision of the manuscript is necessary before the manuscript is recommended for publication in Biomedicines.

Author Response

Response to Reviewer Comments

Manuscript ID: Biomedicines-4082160 | Coil-only intracranial aneurysm embolisation – local DRLs (low-dose protocol)

 

Dear Reviewer,

We sincerely thank you for your detailed and constructive review. We have revised the manuscript to address each point and to ensure that our conclusions remain strictly aligned with the study design (a single-centre retrospective dose audit aimed at establishing local DRLs). Below we provide a point-by-point response; reviewer comments are reproduced in summary form for readability.

 

Reviewer #1

Comment 1. Lack of controlled comparison between low-dose and standard protocols

Reviewer comment: The central claim relies on comparisons with historical literature rather than an internally controlled comparison against a conventional frame-rate protocol (e.g., 7.5–15 fps) on the same system; the SNR analysis is based on a small subset and is not linked to diagnostic or clinical endpoints.

 

Response: We agree that an internal controlled comparison would be the most rigorous approach to isolate the independent contribution of frame-rate reduction. However, such a comparison is not feasible in our centre: throughout the entire study period, procedures on this system were performed exclusively using our routine low-dose workflow and conventional frame rates were not used in clinical practice. In addition, a prospective protocol-comparison study would require intentionally increasing acquisition parameters in a subgroup solely for research purposes, thereby exposing patients to additional radiation without direct clinical benefit, which we did not consider ethically justifiable. Accordingly, we have revised the manuscript to avoid causal wording ("protocol X causes dose reduction") and to frame the results as a retrospective description of dose distributions and local DRLs achieved in routine practice on a contemporary biplane platform. Regarding image quality, the retrospective SNR calculation (n=15) was added in response to the reviewer request as a limited quantitative illustration under the routine protocol; we explicitly state that it is supportive information only and does not constitute a comprehensive diagnostic-performance or outcomes validation.

 

Changes made in the manuscript:

• Revised Abstract/Discussion/Conclusions to avoid causal language and to state explicitly that no internal standard-protocol comparison was performed.
• Expanded the Limitations section to explain why a contemporaneous control group was not available and why a prospective dose-escalation comparison was not undertaken.
• Clarified the purpose and limitations of the SNR analysis (small sample; supportive indicator only; not linked to clinical endpoints).

 

Comment 2. Inadequate handling of confounding variables in dose stratification

Reviewer comment: Stratified analyses (sex, rupture status, anterior/posterior circulation) are presented without adjustment for key confounders (e.g., aneurysm size/morphology, vascular tortuosity, procedural complexity, patient anthropometrics). Multivariable regression or propensity-score matching is recommended.

 

Response: We agree that multiple anatomical and procedural factors can influence fluoroscopy time, the number of acquisitions, and ultimately DAP and Ka,r. However, the primary objective of this manuscript is to establish local DRLs and describe the distribution of routinely reported dose metrics in a large real-world cohort, not to model independent determinants of dose or to draw causal inferences from subgroup differences. In our retrospective dataset, detailed aneurysm morphology and procedural complexity metrics were not captured in a sufficiently structured and complete manner across all cases to support robust multivariable modelling without introducing substantial classification uncertainty. In addition, head-region anthropometric measures that best represent cranial attenuation (e.g., head circumference or AP/LAT head dimensions) were not systematically available. Therefore, we retained subgroup analyses as descriptive/exploratory benchmarking only and strengthened the manuscript by explicitly stating their unadjusted nature and the potential for residual confounding. We also outline this as a priority for follow-up work with standardized extraction/collection of anatomical and complexity parameters.

 

Changes made in the manuscript:

• Explicitly labelled subgroup analyses as descriptive/unadjusted and not intended for causal inference.
• Expanded Limitations to note the lack of structured head anthropometrics, aneurysm morphology variables, and procedural complexity metrics across the retrospective dataset.
• Added a statement in the Discussion that observed subgroup differences may reflect residual confounding and should be interpreted cautiously.

 

Comment 3. Overgeneralization of single-operator, single-centre findings

Reviewer comment: All procedures were performed by a single operator; operator-dependent factors may drive the low doses. The manuscript should acknowledge a best-case scenario and the need for multi-operator validation.

 

Response: We fully agree. The single-centre, single-operator setting reduces within-study variability, but it limits external validity and reproducibility in less specialized settings. We now state explicitly that the reported DRLs may reflect a highly specialised practice environment (a potential "best-case" scenario) rather than a population-level benchmark. We also emphasise that multicentre and multi-operator validation is required to assess robustness, transferability, and learning-curve effects.

 

Changes made in the manuscript:

• Added explicit statements in the Discussion and Limitations that the DRLs may reflect a best-case specialised setting and are not intended as population-level benchmarks.
• Added a forward-looking statement on the need for multi-operator/multicentre validation and learning-curve assessment.

 

Comment 4. Incomplete image-quality assessment and lack of clinical validation

Reviewer comment: Image quality is assessed only via retrospective SNR from DSA, without CNR, blinded multi-reader assessment, or clinical endpoints (e.g., residual aneurysm, thromboembolic events).

 

Response: We agree that comprehensive evaluation of diagnostic image quality would ideally include additional quantitative metrics (e.g., CNR), blinded multi-reader assessment using standardized criteria, and linkage to clinical endpoints. The present work, however, was designed as a DRL-focused dose audit and does not contain prospectively collected image-quality scores or outcome endpoints. We have therefore clarified that the SNR assessment was provided solely as limited supportive information on request and should not be interpreted as a standalone validation of diagnostic performance or procedural safety. We also note that system performance is routinely monitored through a structured QC programme and regular service, which supports stability of imaging parameters but does not replace a dedicated research-grade image-quality study.

 

Changes made in the manuscript:

• Expanded Limitations to explicitly state that CNR, blinded multi-reader scoring, and clinical endpoints were not assessed.
• Revised Discussion to avoid overinterpretation of SNR and to keep conclusions within the DRL/dose-audit scope.
• Added clarifying text on routine QC and maintenance as a background assurance of stable system performance.

 

Comment 5. Methodological issues in statistical analysis and data reporting

Reviewer comment: Parametric and non-parametric tests appear to be used inconsistently and without justification (normality/variance testing). There are inconsistencies between narrative statements and p-values in tables; effect sizes and 95% confidence intervals are missing.

 

Response: We thank the Reviewer for this careful assessment. We have revised the Statistical Analysis section to describe the decision pathway explicitly. All continuous variables were assessed for distributional assumptions using the Shapiro–Wilk test and visual inspection. For dose metrics that did not meet normality assumptions (as is common for DAP and Ka,r), we used non-parametric tests (Mann–Whitney U for independent groups; Wilcoxon for paired data where applicable). For variables approximating normality (including after log10 transformation where appropriate), we used parametric testing and applied the Welch correction when variance heterogeneity was present. We additionally audited all p-values and revised the text to clearly distinguish overall comparisons from stratified subgroup comparisons, eliminating any apparent contradictions. Finally, given that the principal outputs are dose distributions and percentiles (P75 as DRL), we present subgroup p-values as exploratory/descriptive and avoid over-interpretation; we state this explicitly.

 

Changes made in the manuscript:

• Expanded the Methods (Statistics) to document normality testing (Shapiro–Wilk), use of log10 transformation where appropriate, and test selection (parametric vs non-parametric; Welch correction when needed).
• Cross-checked and harmonized all p-values and narrative statements across text, tables, and figures; clarified overall vs stratified comparisons.
• Added explicit wording that subgroup comparisons are exploratory/descriptive within a DRL-focused dose audit.

 

Comment 6. Claims regarding safety and broad applicability of ultra-low frame rates

Reviewer comment: The manuscript advocates routine adoption of ultra-low frame rates without adequate discussion of risks and training/QA pathways; applicability to complex cases is untested (coil-only only).

 

Response: We agree that our conclusions must be framed in proportion to the retrospective DRL-focused design. We have therefore tempered the manuscript language to avoid implying outcome-validated safety or universal applicability. Our analysis is limited to coil-only procedures, and we now state explicitly that the presented DRLs should not be extrapolated to more complex interventions (e.g., stent-assisted coiling, balloon assistance, flow diversion). Within our routine coil-only workflow, ultra-low frame rates (fluoroscopy and DSA) were operationally sufficient and did not require systematic escalation; however, the present study was not designed to formally quantify event-detection sensitivity, learning-curve effects, or clinical outcomes. Accordingly, we added a consolidated Limitations subsection and implementation considerations emphasising that any broader adoption should be accompanied by structured training, centre-specific protocol governance, and routine QC/service.

 

Changes made in the manuscript:

• Tempered wording in Abstract/Discussion/Conclusions to avoid claims of outcome-validated safety or universal applicability.
• Restricted applicability statements to coil-only procedures and stated explicitly that DRLs should not be extrapolated to complex techniques.
• Added a consolidated Limitations subsection including the absence of learning-curve quantification and a note on training/QA needs for implementation.

 

Sincerely,

Joanna Sowa, on behalf of all authors

Reviewer 2 Report (New Reviewer)

Comments and Suggestions for Authors

This retrospective single-center study analyzes radiation exposure metrics in 245 coil-only endovascular treatments of intracranial aneurysms performed between 2018 and 2024, to establish local diagnostic reference levels (DRLs) and typical values for dose–area product (DAP), reference air kerma (Ka,r), fluoroscopy time (FT), and number of DSA frames. By employing an ultra–low-dose imaging protocol, —the authors report radiation dose values that are several-fold lower than those previously published in the literature. Subgroup analyses stratified by sex, aneurysm rupture status, vascular territory, and imaging configuration (single-plane vs biplane) further contextualize dose variability and procedural complexity. Overall, the study proposes local DRLs that may serve as practical benchmarks for radiation optimization and quality assurance in coil-only aneurysm embolization.

This work is novel in its use of imaging technology to quantify the real-world impact of an aggressively optimized low-frame-rate strategy applied consistently across a large, well-defined cohort of coil-only aneurysm procedures. However, there are several comments:

1- The principal limitation is the single-center, retrospective design, which inherently limits external validity.

2-The reported low radiation doses are closely tied to a highly experienced operator and prolonged familiarity with low-frame-rate imaging; thus, reproducibility in less specialized settings or among less experienced operators remains uncertain.

3-The lack of a contemporaneous control group using standard frame-rate protocols prevents direct causal attribution of dose reduction solely to the low-dose strategy. While subgroup analyses are extensive, some comparisons are underpowered, reducing the robustness of statistical inferences.

4- Finally, the reliance on a single angiographic system and vendor-specific dose-reduction algorithms constrains generalizability across different platforms, and multicenter validation would be required before extrapolating these DRLs as wider reference standards

 

Recommendations:

While some limitations are implicitly acknowledged in the Discussion (e.g., small subgroup sizes such as posterior circulation UIA cases, single-center experience, and the need for larger cohorts to confirm findings), these points are scattered and descriptive rather than consolidated. There is no dedicated limitation subsection that systematically outlines methodological, statistical, or generalizability constraints. Please include

 

Author Response

Thank you for this constructive and detailed comment. We fully agree that the single-centre, retrospective nature of the study limits external validity and that the low dose levels observed in our cohort may be influenced by operator-related and system-related factors. In the revised manuscript, we have therefore strengthened and consolidated the limitations into a dedicated Limitations subsection that systematically addresses methodological, statistical, and generalisability constraints.

Specifically, we have added/clarified that:

1. Study design: This is a single-centre, retrospective dose audit and DRL analysis, which inherently limits external validity and does not allow causal inference.

2. Operator dependence: All procedures were performed by one highly experienced neurointerventionalist with long-standing familiarity with low-frame-rate workflows; therefore, reproducibility in less specialised settings and among less experienced operators remains uncertain.

3. No contemporaneous control group / underpowered subgroups: We did not have a contemporaneous control group using standard frame-rate protocols on the same system, which prevents direct causal attribution of dose reduction solely to the low-dose strategy or frame-rate reduction. We also explicitly state that some subgroup comparisons involved small sample sizes (e.g., posterior circulation subgroups) and should be interpreted as exploratory rather than robust inferential findings.

4. Single platform / vendor-specific technologies: All procedures were performed on a single biplane angiography platform with vendor-specific dose-reduction and dose-optimisation features, which limits generalisability across systems and configurations; therefore, multicentre and multi-operator validation is required before extrapolating these DRLs as broader reference standards.

We believe these revisions directly address the Reviewer’s recommendation by presenting a clear, consolidated limitations section and by ensuring that our conclusions are restricted to the scope of retrospective dose audit and local DRL establishment.

Round 2

Reviewer 1 Report (Previous Reviewer 2)

Comments and Suggestions for Authors

This manuscript is recommended for acceptance.

Reviewer 2 Report (New Reviewer)

Comments and Suggestions for Authors

Accept in present form

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The authors investigated a local DRL for aneurysm embolization procedures using coil-only techniques under low-dose imaging protocols.

 

Major:

In the Results section, both the t-test and the Mann–Whitney test are used. Does this imply that the authors selected the tests based on the presence or absence of normality? If so, the statistical methodology should be described in the Materials & Methods section.

 

The Results section is excessively long, redundant, and difficult to read. The authors should present the findings more concisely.

 

Is the comparison with previous studies truly valid? Can the authors legitimately claim a statistical difference? The results are highly dependent on operator experience, as well as aneurysm size, morphology, orientation, and the condition of the parent vessel, making it difficult to judge whether such comparisons are appropriate. Moreover, device performance has improved over the years, so it may also be necessary to examine what type of equipment was used and in which manufacturing year.

 

The comparison between single-center and multi-center settings is also a significant limitation. If multiple operators are involved in a multi-center study, the variability becomes even more difficult to control.

 

The authors should also consider evaluating the operators’ years of experience.

 

This study lacks an evaluation of image quality. Image quality and low-dose radiation (radiation exposure) are in a trade-off relationship, and it is not possible to determine whether the image quality at the authors’ institution is comparable to that of previous studies. Even if subjective, an assessment of image quality is necessary.

 

What exactly was evaluated in the post-procedural CT scan?

 

Minor:

In Figure 9, aren’t the labels for Male RIA and Female UIA reversed?

The spelling of “patient number” in Table 2 is incorrect.

 

Reviewer 2 Report

Comments and Suggestions for Authors

A major revision of the manuscript is necessary before the manuscript is recommended for publication in Biomedicines.

Comments for author File: Comments.pdf

Comments on the Quality of English Language

The English could be improved to more clearly express the research.