Review Reports - Effectiveness of Contrast-Enhanced Spectral Mammography Following Contrast-Enhanced Computed Tomography in Breast Cancer Diagnosis

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

In the manuscript entitled “Effectiveness of Contrast-Enhanced Spectral Mammography Following Contrast-Enhanced Computed Tomography in Breast Cancer Diagnosis” the authors present relevant results. In spite of the offered results, there are some points that must be addressed before I recommend to accept the manuscript for publications in the journal diagnostics MDPI:

I recommend that you describe the contrast-administration protocol in greater detail so other centers can accurately reproduce the study. In addition to reporting the contrast agent, dose, and injection rate, please specify the exact temporal reference used (start versus end of injection), total injection duration, presence and volume of saline flush, total iodine load in mg I/kg, patient transfer time between rooms, and the precise timing of each projection acquisition (CC and MLO). Because your conclusions depend on contrast retention and washout kinetics, small differences in timing may significantly alter lesion enhancement; therefore, providing these parameters will clarify whether the observed diagnostic equivalence is due to the imaging technique or to residual intravascular contrast concentration.

I suggest that you replace the consensus visual scoring approach with an independently assessed interpretation and quantify inter-observer agreement. Each radiologist should evaluate the images separately, and you should report Cohen’s kappa (or weighted kappa) to demonstrate reliability. I also recommend clearly defining the criteria for each category of enhancement, conspicuity, and background parenchymal enhancement, ideally linking them to standardized descriptors such as BI-RADS terminology or measurable parameters (for example contrast-to-background ratio or lesion margin visibility). This will reduce subjectivity and allow other investigators to apply the same scoring system in a reproducible manner.

I recommend that you include a precise description of how tumor measurements and lesion detection were performed. Please indicate whether measurements were taken on low-energy, recombined, or subtraction images, which projection was selected, the calibration and magnification procedure of the workstation, and how non-mass enhancement was handled. In addition, clarify how imaging measurements were correlated with pathology, including whether surgical specimens were measured fresh or after fixation and whether tissue shrinkage was considered. Since tumor size estimation is a major clinical justification of the technique, transparent reporting of measurement methodology is essential for reproducibility and comparability with other studies.

I suggest strengthening the statistical analysis by including confidence intervals and adjusting for confounding variables. One group presents larger tumors, which may directly influence lesion detectability and therefore the apparent diagnostic performance. I recommend performing a multivariate analysis (e.g., logistic regression) controlling for tumor size and relevant clinical factors, and reporting sensitivity and specificity with 95% confidence intervals. Including receiver-operating-characteristic curves and a sample size justification or power calculation would further support your conclusions and demonstrate that the study can detect clinically meaningful differences between protocols.

I recommend clearly separating measured imaging outcomes from inferred clinical advantages. The discussion currently includes statements about workflow efficiency, cost reduction, and safety, but these variables were not directly evaluated. I suggest restricting the main conclusions to the diagnostic imaging findings and presenting workflow or safety benefits as potential implications that require prospective validation. Explicitly distinguishing evidence from interpretation will improve scientific clarity, prevent overstatement, and increase the credibility of the manuscript for reviewers and readers.

I suggest to add a paragraph before conclusion where possible application in future is mentioned, this can give the reader insight into the possible application (besides facilitate the identification of some structures or tissue in mammography) of the reported results, maybe helping automated systems based in A.I. for automatic pointing, recommended reference: Hernández-Vázquez, M.A.; Hernández-Rodríguez, Y.M.; Cortes-Rojas, F.D.; Bayareh-Mancilla, et. Al, Hybrid Feature Mammogram Analysis: Detecting and Localizing Microcalcifications Combining Gabor, Prewitt, GLCM Features, and Top Hat Filtering Enhanced with CNN Architecture. Diagnostics 2024, 14, 1691.

Author Response

“I recommend that you describe the contrast-administration protocol in greater detail so other centers can accurately reproduce the study. In addition to reporting the contrast agent, dose, and injection rate, please specify the exact temporal reference used (start versus end of injection), total injection duration, presence and volume of saline flush, total iodine load in mg I/kg, patient transfer time between rooms, and the precise timing of each projection acquisition (CC and MLO). Because your conclusions depend on contrast retention and washout kinetics, small differences in timing may significantly alter lesion enhancement; therefore, providing these parameters will clarify whether the observed diagnostic equivalence is due to the imaging technique or to residual intravascular contrast concentration”

We thank the reviewer for this valuable comment. We agree that detailed reporting of the contrast-administration protocol and temporal parameters is essential to ensure reproducibility and to clarify the potential impact of contrast timing within the early post-contrast phase on lesion enhancement.

The start time of the first CEM acquisition relative to contrast injection (T0) was recorded for each patient, and the minimum, maximum, and mean values were calculated. Because each projection is acquired within seconds and the protocol is fixed and sequential, second-level timing differences were not expected to materially influence enhancement assessment. All CEM acquisitions were completed within 10 minutes from T0 according to our standardized imaging protocol.In response to this comment, we have substantially expanded the CT and CEM protocol description in the Materials and Methods section (Section 2.2). Specifically, we have now provided: The exact temporal reference point (contrast injection start time defined as T0), total injection duration, presence and volume of saline flush, total iodine load expressed in mg I/kg, patient transfer time between CT and mammography units, exact interval between contrast injection and first CEM exposure in both groups. For clarity and reproducibility, we have additionally summarized all temporal and contrast-related parameters in a new table (Table 2), which presents a step-by-step timing sequence for both the CT/CEM and CEM-only groups.

To improve temporal clarity and avoid potential misinterpretation, the previously stated approximate total examination duration has been removed in materials and methods section and replaced with a more precise statement indicating that all CEM acquisitions were completed within 10 minutes from the start of contrast injection (T0) in both groups.

CEM was performed using a standardized four-view protocol (ipsilateral CC, contralateral CC, ipsilateral MLO, contralateral MLO) applied uniformly to all patients. Each projection consisted of one low-energy and one high-energy acquisition, with exposure times in the order of seconds per view, depending on automatic exposure control and breast thickness. Breast compression was released between projections according to standard clinical workflow. Because acquisition of each view occurs within seconds and projections are obtained sequentially without protocol variation, we did not record second-level start times for each individual exposure. However, we confirm that all four projections in both groups were consistently completed within 10 minutes from T0, ensuring that imaging occurred within the early post-contrast phase.

I suggest that you replace the consensus visual scoring approach with an independently assessed interpretation and quantify inter-observer agreement. Each radiologist should evaluate the images separately, and you should report Cohen’s kappa (or weighted kappa) to demonstrate reliability. I also recommend clearly defining the criteria for each category of enhancement, conspicuity, and background parenchymal enhancement, ideally linking them to standardized descriptors such as BI-RADS terminology or measurable parameters (for example contrast-to-background ratio or lesion margin visibility). This will reduce subjectivity and allow other investigators to apply the same scoring system in a reproducible manner.

While we agree that inter-observer agreement analysis can provide additional methodological strength, the primary aim of our study was protocol comparison rather than observer variability assessment. In this regard, consensus reading was intentionally chosen to reduce interpretative variability and to reflect routine clinical decision-making, particularly given the difference in experience levels between the two radiologists. The scoring system was applied uniformly across both groups under identical reading conditions, thereby minimizing systematic bias.

We acknowledge that independent assessment with kappa analysis could provide additional information regarding reproducibility and have noted this as a limitation in the revised manuscript.

Background parenchymal enhancement (BPE) was categorized according to BI-RADS terminology (minimal, mild, moderate, marked). As the BI-RADS atlas does not provide an ordinal classification for lesion enhancement intensity or conspicuity in CEM, these parameters were assessed using a structured 4-point visual scale reflecting increasing degrees of contrast uptake and lesion visibility.

To reduce subjectivity and improve transparency, we have now explicitly described the criteria for each scoring category in the revised Methods section. The same scoring framework was applied uniformly across all cases in both groups to ensure internal consistency.

I recommend that you include a precise description of how tumor measurements and lesion detection were performed. Please indicate whether measurements were taken on low-energy, recombined, or subtraction images, which projection was selected, the calibration and magnification procedure of the workstation, and how non-mass enhancement was handled. In addition, clarify how imaging measurements were correlated with pathology, including whether surgical specimens were measured fresh or after fixation and whether tissue shrinkage was considered. Since tumor size estimation is a major clinical justification of the technique, transparent reporting of measurement methodology is essential for reproducibility and comparability with other studies.

Tumor size estimation was based on the maximum linear extent of enhancement observed on recombined CEM images. When spiculated morphology or architectural distortion was present, the measurement incorporated the entire enhancing abnormality, including radiating extensions or the distorted breast parenchyma.

Measurements were performed on a dedicated CEM workstation using the built-in electronic caliper tool. Tumor dimensions were recorded in millimeters based on the system’s standard digital calibration. No additional magnification was applied during measurement.

Pathological measurements were obtained from routine histopathological evaluation of the surgical specimen. Only the invasive component size was considered for correlation with imaging findings. No correction factor for potential tissue shrinkage was applied.

I suggest strengthening the statistical analysis by including confidence intervals and adjusting for confounding variables. One group presents larger tumors, which may directly influence lesion detectability and therefore the apparent diagnostic performance. I recommend performing a multivariate analysis (e.g., logistic regression) controlling for tumor size and relevant clinical factors, and reporting sensitivity and specificity with 95% confidence intervals. Including receiver-operating-characteristic curves and a sample size justification or power calculation would further support your conclusions and demonstrate that the study can detect clinically meaningful differences between protocols.

Thank you for this valuable suggestion. To address potential confounding factors, we performed a multivariate logistic regression analysis including group (case vs. control) and primary tumor size as independent variables, with lesion detectability as the dependent variable. The analysis showed that neither group status (OR = 2.923, 95% CI: 0.527–16.211, p = 0.220) nor primary tumor size (OR = 1.052, 95% CI: 0.989–1.119, p = 0.107) had a statistically significant association with lesion detectability. Model fit was confirmed using the Hosmer–Lemeshow test (χ² = 10.173, p = 0.253). These results have now been added to the Results section of the revised manuscript..

I recommend clearly separating measured imaging outcomes from inferred clinical advantages. The discussion currently includes statements about workflow efficiency, cost reduction, and safety, but these variables were not directly evaluated. I suggest restricting the main conclusions to the diagnostic imaging findings and presenting workflow or safety benefits as potential implications that require prospective validation. Explicitly distinguishing evidence from interpretation will improve scientific clarity, prevent overstatement, and increase the credibility of the manuscript for reviewers and readers.

In accordance with this suggestion, the statement “while reducing overall contrast exposure and improving workflow efficiency” has been removed from the Introduction to ensure that the study aim reflects only the imaging parameters that were directly evaluated.

In addition, the Discussion and Conclusion sections have been revised to clearly distinguish measured imaging outcomes from potential practical implications. Statements regarding contrast utilization, workflow efficiency, cost, and safety have been moderated and explicitly presented as aspects not directly assessed in this study.

I suggest to add a paragraph before conclusion where possible application in future is mentioned, this can give the reader insight into the possible application (besides facilitate the identification of some structures or tissue in mammography) of the reported results, maybe helping automated systems based in A.I. for automatic pointing, recommended reference: Hernández-Vázquez, M.A.; Hernández-Rodríguez, Y.M.; Cortes-Rojas, F.D.; Bayareh-Mancilla, et. Al, Hybrid Feature Mammogram Analysis: Detecting and Localizing Microcalcifications Combining Gabor, Prewitt, GLCM Features, and Top Hat Filtering Enhanced with CNN Architecture. Diagnostics 2024, 14, 1691.

We thank the reviewer for this insightful suggestion. While we agree that artificial intelligence–based image analysis represents an important and evolving field in breast imaging, the present study was not designed to evaluate automated systems or feature extraction methods. To maintain focus and avoid overextending the scope of the manuscript, we have chosen not to expand the Discussion in this direction. However, we acknowledge that future research may explore the potential integration of sequential CT/CEM protocols into AI-assisted diagnostic frameworks.

We hope this clarifications adequately address the reviewer’s suggestions.

Reviewer 2 Report

Comments and Suggestions for Authors

This manuscript addresses a practical and clinically relevant question regarding the optimization of breast cancer staging workflows. The concept of "one-stop" staging using a single contrast bolus for both CT and CEM is innovative and has the potential to improve patient comfort and resource utilization. The authors successfully demonstrate the feasibility of this approach with a retrospective cohort. However, the study has notable limitations that need to be addressed. The significant difference in tumor size between the two groups is a major confounder; larger tumors in the experimental group may mask potential degradation in image quality due to the timing delay. Additionally, the restriction to only biopsy-proven malignancies prevents the assessment of specificity, which is critical for a diagnostic study. The authors should also consider providing objective quantitative metrics (CNR/SNR) to bolster their qualitative visual assessments. Addressing the BPE discrepancy between groups would also strengthen the physiological basis of the findings. Overall, this is a promising preliminary study that requires robust clarification of potential biases.

1. The study is retrospective with two groups (n=29 vs n=34). How were patients allocated to the CT/CEM group versus the standard CEM group? Was this based on physician preference, scheduling convenience, or clinical indication? The lack of randomization may introduce bias.

2. The text states that the CEM-only group used "the same injection protocol as that used for CT" (1.5 mL/kg). Standard CEM protocols often use a fixed dose (e.g., 1.5 mL/kg capped at a certain volume) or strictly 1.5 mL/kg. Please explicitly confirm that the iodine concentration (300 mg I/mL) and weight-based dosing were identical for both groups to ensure fair comparison.

3. The methods state that radiologists were blinded to clinical information. Were they also blinded to the group (CT/CEM vs. CEM-only)? Given the potential difference in image noise or enhancement timing, could the readers distinguish which protocol was used, potentially biasing the scoring?

4. The study only included patients with biopsy-proven breast cancer. Why were patients with benign or indeterminate lesions excluded? Including these would allow for the calculation of specificity, which is a key metric for any diagnostic test.

5. Table 4 shows a statistically significant difference in mean lesion diameter between groups (CT/CEM: 42.82 mm vs. CEM: 29.15 mm; p=0.008). Larger tumors are generally easier to visualize. How do the authors account for this confounding variable? Could the "comparable" conspicuity in the CT/CEM group be artificially inflated by the larger lesion size?

6. The results indicate that BPE significantly affected lesion conspicuity in the CT/CEM group (p=0.014) but not in the CEM-only group (p=0.061). Do the authors believe this is a physiological effect of the delayed timing in the CT/CEM group (washout phase of background tissue vs. tumor), or is it a statistical artifact due to sample size?

7. The results state "The primary lesion was detected in all patients (sensitivity 100%)." Since only biopsy-proven cancers were included, this study cannot assess false positives or true negatives. Would it be more accurate to frame this as "Lesion Detection Rate" rather than diagnostic sensitivity?

8. Did the authors analyze if the enhancement quality varied by histologic subtype (e.g., Invasive Lobular Carcinoma vs. Invasive Ductal Carcinoma)? ILC is known to have different enhancement patterns; were any of the false negatives or lower conspicuity scores associated with ILC?

9. The assessment of enhancement was purely qualitative (4-point visual scale). Did the authors consider measuring Signal-to-Noise Ratio (SNR) or Contrast-to-Noise Ratio (CNR) on the subtracted images to provide an objective comparison between the protocols?

Author Response

The study is retrospective with two groups (n=29 vs n=34). How were patients allocated to the CT/CEM group versus the standard CEM group? Was this based on physician preference, scheduling convenience, or clinical indication? The lack of randomization may introduce bias.

Thank you for this important comment. Patient allocation was not based on physician preference or disease severity. Both groups consisted of patients undergoing staging imaging at the time of initial diagnosis and prior to treatment, with the same clinical indication. In all cases, CT examinations were performed for staging purposes. The difference between groups was related to workflow and timing: in the study group, CT and CEM were performed consecutively using a single contrast bolus, whereas in the control group, CT and CEM were performed at different time points as part of standard practice. Therefore, although the study is retrospective and non-randomized, allocation was not driven by clinical factors that could systematically bias disease severity between groups. To address this concern, we have expanded this statement in Materials and Methods section and added the following statement to the Discussion section:

“Although the study was retrospective and non-randomized, both groups shared identical clinical staging indications, reducing the likelihood of severity-based allocation bias.”

The text states that the CEM-only group used "the same injection protocol as that used for CT" (1.5 mL/kg). Standard CEM protocols often use a fixed dose (e.g., 1.5 mL/kg capped at a certain volume) or strictly 1.5 mL/kg. Please explicitly confirm that the iodine concentration (300 mg I/mL) and weight-based dosing were identical for both groups to ensure fair comparison.

Thank you for this important clarification request. We confirm that both groups received an identical contrast administration protocol. In all patients, the same iodinated contrast agent with an iodine concentration of 300 mg I/mL was used, administered at a weight-based dose of 1.5 mL/kg body weight with a limit on maximum contrast volume of 120 cc).

To ensure transparency and allow fair comparison between groups, we have expanded the description of the contrast administration protocol in the Materials and Methods section in accordance with the reviewer’s suggestion. In addition, a new table (Table 2) has been added to summarize and clearly present the contrast-related technical parameters for both groups.

The methods state that radiologists were blinded to clinical information. Were they also blinded to the group (CT/CEM vs. CEM-only)? Given the potential difference in image noise or enhancement timing, could the readers distinguish which protocol was used, potentially biasing the scoring?

We confirm that the radiologists were blinded not only to all clinical and histopathologic information but also to the patient group (CT/CEM vs. CEM-only). To clarify this point, we have expanded the description in the Materials and Methods section (Section 2.2.3) to explicitly state that both radiologists were unaware of group allocation during image interpretation. Standardized CEM acquisition and post-processing protocols minimized any perceptible differences between CT/CEM and CEM-only images. As a result, it is unlikely that the readers could reliably identify the imaging protocol based on image characteristics alone, and therefore the scoring of lesion enhancement, conspicuity, and BPE was not biased by knowledge of the group.

The study only included patients with biopsy-proven breast cancer. Why were patients with benign or indeterminate lesions excluded? Including these would allow for the calculation of specificity, which is a key metric for any diagnostic test.

We appreciate the reviewer’s comment. The study focused on evaluating CEM immediately after contrast-enhanced CT in patients with biopsy-proven breast cancer. Both the CT/CEM and CEM-only groups included only cancer patients with identical clinical staging indications. The primary aim was to assess whether contrast timing affects lesion enhancement and conspicuity, rather than to determine specificity. This has been clarified in the materials methods section of the revised manuscript.

Table 4 shows a statistically significant difference in mean lesion diameter between groups (CT/CEM: 42.82 mm vs. CEM: 29.15 mm; p=0.008). Larger tumors are generally easier to visualize. How do the authors account for this confounding variable? Could the "comparable" conspicuity in the CT/CEM group be artificially inflated by the larger lesion size?

Thank you for this important observation. We agree that larger lesions may be easier to visualize and therefore could potentially influence lesion conspicuity. To account for this potential confounding factor, we performed a multivariate logistic regression analysis including group status (CT/CEM vs. CEM-only) and primary tumor size as independent variables, with lesion detectability as the dependent variable.

The analysis demonstrated that tumor size was not significantly associated with lesion detectability (OR = 1.052, 95% CI: 0.989–1.119, p = 0.107). Similarly, group status did not significantly influence lesion detectability (OR = 2.923, 95% CI: 0.527–16.211, p = 0.220). These findings suggest that the comparable lesion conspicuity observed between the groups cannot be explained solely by differences in tumor size. We have now included this multivariate analysis in the Results section and clarified this point in the Discussion section of the revised manuscript.

The results indicate that BPE significantly affected lesion conspicuity in the CT/CEM group (p=0.014) but not in the CEM-only group (p=0.061). Do the authors believe this is a physiological effect of the delayed timing in the CT/CEM group (washout phase of background tissue vs. tumor), or is it a statistical artifact due to sample size?

We thank the reviewer for this insightful comment. We believe that the observed effect is more likely attributable to physiological factors related to background parenchymal enhancement rather than solely a statistical artifact. In the CT/CEM group, CEM was performed immediately after CT within the early post-contrast phase, which may coincide with peak enhancement of background parenchyma and transiently reduce lesion-to-background contrast. This mechanism is consistent with the known influence of BPE on lesion conspicuity.

At the same time, we acknowledge that the difference in statistical significance between groups may partly reflect sample size limitations. Importantly, the overall pattern aligns with previously reported effects of BPE and does not suggest a systematic bias related to the imaging protocol.

To clarify this point, we have expanded the Discussion section accordingly.

The results state "The primary lesion was detected in all patients (sensitivity 100%)." Since only biopsy-proven cancers were included, this study cannot assess false positives or true negatives. Would it be more accurate to frame this as "Lesion Detection Rate" rather than diagnostic sensitivity?

We thank the reviewer for this important observation. As the study cohort consisted exclusively of patients with biopsy-proven breast cancer, true negative and false positive cases were not available for analysis. We agree that the term “sensitivity” may not accurately reflect the study design. Accordingly, we have revised the manuscript to remove references to sensitivity and specificity. In the Abstract, the statement has been modified to: “The primary lesion was identified in all patients in both groups.” Corresponding changes have also been made in the Statistical Analysis section.

Did the authors analyze if the enhancement quality varied by histologic subtype (e.g., Invasive Lobular Carcinoma vs. Invasive Ductal Carcinoma)? ILC is known to have different enhancement patterns; were any of the false negatives or lower conspicuity scores associated with ILC?

We thank the reviewer for this insightful suggestion. In our cohort, invasive lobular carcinoma (ILC) was present in 3 patients in the CT/CEM group and 4 patients in the CEM-only group. Given this limited number of cases, a statistically meaningful subgroup analysis according to histologic subtype was not feasible.

Importantly, no false-negative cases were observed in either group. While lower conspicuity scores may theoretically be associated with ILC due to its known diffuse growth pattern, the small number of ILC cases in our study precludes definitive conclusions. We have acknowledged this limitation in the revised Discussion section.

The assessment of enhancement was purely qualitative (4-point visual scale). Did the authors consider measuring Signal-to-Noise Ratio (SNR) or Contrast-to-Noise Ratio (CNR) on the subtracted images to provide an objective comparison between the protocols?

Quantitative measurements such as signal-to-noise ratio (SNR) or contrast-to-noise ratio (CNR) were not performed in the present study. As this was a retrospective analysis, our primary objective was to evaluate lesion visibility and enhancement characteristics in a manner reflecting routine clinical interpretation. Therefore, a qualitative 4-point visual scale assessed in consensus was used to simulate real-world diagnostic decision-making.

We acknowledge that objective quantitative parameters could provide additional technical insight and strengthen protocol comparison. This has now been recognized as a limitation in the revised Discussion section. Future prospective studies incorporating quantitative image analysis may further clarify potential differences between protocols.

We hope this clarifications adequately address the reviewer’s suggestions.

Reviewer 3 Report

Comments and Suggestions for Authors

This paper evaluates the diagnostic effectiveness of performing contrast-enhanced spectral mammography (CEM) immediately following a contrast-enhanced computed tomography (CT) scan using the same contrast bolus (CT/CEM). The study finds that this sequential approach provides diagnostic performance—including lesion visibility, enhancement, and conspicuity—comparable to standard CEM protocols that require a separate, additional contrast injection. Its main contribution is demonstrating a feasible workflow for breast cancer staging that reduces cumulative contrast exposure, patient risk, and healthcare costs without compromising diagnostic quality.

Paper Weaknesses

Small and Retrospective Sample Size: The study is limited by its retrospective design and relatively small cohort of 63 patients (29 in the CT/CEM group and 34 in the CEM group). This small sample size may limit the generalizability of the statistical findings.
Image interpretation was performed in consensus by two radiologists with markedly different experience levels, and interobserver variability was not assessed. This prevents evaluation of reproducibility. Additionally, inclusion of only biopsy-proven malignancies precludes assessment of specificity in a true diagnostic setting and may overestimate performance metrics.
There was a statistically significant difference in the mean diameter of primary lesions between the two groups (42.82 mm in CT/CEM vs. 29.15 mm in the standard group). Although the authors noted this, larger tumors are inherently easier to detect and enhance, which may have subtly biased the results in favor of the CT/CEM group's high sensitivity.
For institutions where the CT and mammography suites are not adjacent, what is the maximum "safe" time window you would recommend between the CT injection and CEM acquisition before diagnostic quality is expected to degrade?

The paper presents a highly practical and innovative clinical workflow that addresses patient safety and cost-effectiveness in breast cancer staging. The methodology is sound, and the results clearly support the feasibility of sequential imaging. However, the score is tempered by the small retrospective sample size and the significant difference in tumor sizes between the study groups, which introduces potential bias and limits the breadth of the findings. If the authors can address the impact of tumor size and provide more guidance on the "timing window" for broader institutional use, the paper's value would be significantly enhanced.

Author Response

RESPONSE TO REVIEWER

We sincerely thank the reviewer for the careful evaluation of our manuscript and for highlighting both the strengths and the limitations of our study. We appreciate the constructive suggestions, which have helped us improve the clarity and scientific rigor of the manuscript. Our responses to the individual comments are provided below.

Comment 1

Small and retrospective sample size.

Response:
We agree with the reviewer that the retrospective design and relatively small cohort represent important limitations of the study. This limitation has been acknowledged in the revised Discussion section. Our study should therefore be considered a preliminary feasibility analysis demonstrating the potential of a sequential CT/CEM workflow. Future prospective studies with larger patient populations will be necessary to confirm these findings and further validate the clinical utility of this approach.

Comment 2

Image interpretation was performed in consensus and interobserver variability was not assessed.

Response:
We appreciate this important observation. In the present study, image interpretation was performed in consensus by two radiologists to ensure consistent evaluation of lesion characteristics. However, we agree that the absence of an interobserver variability analysis limits the ability to assess reproducibility. This limitation has now been explicitly acknowledged in the Discussion section. Future studies including independent readings by multiple radiologists and assessment of interobserver agreement will be valuable to further validate the robustness of the findings.

We also acknowledge that the inclusion of only biopsy-proven malignant lesions limits the ability to evaluate specificity in a true screening or diagnostic setting. This point has also been clarified as a limitation in the revised manuscript.

Comment 3

Difference in tumor size between groups and potential bias.

Response:
We agree that larger tumors may be easier to visualize and could potentially influence lesion conspicuity. To address this potential confounding factor, we performed a multivariate logistic regression analysis including group status (CT/CEM vs. CEM-only) and primary tumor size as independent variables.

The analysis showed that tumor size was not significantly associated with lesion detectability (OR = 1.052, 95% CI: 0.989–1.119, p = 0.107). Similarly, group status did not significantly influence lesion detectability (OR = 2.923, 95% CI: 0.527–16.211, p = 0.220). These findings suggest that the comparable lesion conspicuity observed between the groups cannot be explained solely by differences in tumor size.

The results of this additional analysis have now been included in the Results section, and the potential impact of tumor size has been discussed in the Discussion section.

Comment 4

Timing window between CT injection and CEM acquisition.

Response:
We thank the reviewer for raising this clinically relevant question. In our study, CEM was performed immediately after contrast-enhanced CT, and the interval between CT acquisition and CEM imaging was typically within approximately 5–10 minutes. Within this time frame, lesion enhancement remained sufficient for reliable visualization.

Although our findings suggest that sequential imaging within this interval can provide adequate diagnostic quality, the optimal timing window has not yet been systematically investigated and may depend on several factors, including contrast agent pharmacokinetics and institutional workflow logistics. We have therefore clarified in the Discussion section that further prospective studies are needed to determine the maximum interval that preserves diagnostic image quality.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

Authors Have attend the comments and suggestions satisfactorly. Thus, I recomend to accept the manuscritp for publication.

Author Response

Thank you very much for your positive evaluation of our manuscript and for your time and consideration.

We appreciate your constructive comments, which helped us improve the quality of our work.

We are pleased that the revisions have been found satisfactory and look forward to the next steps in the publication process.

Reviewer 2 Report

Comments and Suggestions for Authors

The authors have completed the revisions in accordance with the review comments.

Author Response

Thank you very much for your positive evaluation of our manuscript and for your time and consideration. We appreciate the reviewer’s constructive comments, which helped us improve the quality of our work.

Reviewer 3 Report

Comments and Suggestions for Authors

Overall Review

This paper investigates the diagnostic effectiveness of performing contrast-enhanced spectral mammography (CEM) immediately following a contrast-enhanced computed tomography (CT) scan using a single contrast bolus, compared to standard CEM with a separate injection. The authors conducted a retrospective study on 63 women with biopsy-proven breast cancer, dividing them into a CT/CEM group (n=29) and a standard CEM group (n=34). The study found no significant differences between the two groups regarding primary lesion enhancement, lesion conspicuity, background parenchymal enhancement (BPE), or the detection of additional lesions. The main contribution of this paper is to demonstrate that sequential CT/CEM can offer diagnostic performance comparable to standard CEM while eliminating the need for an additional contrast injection.

Paper Weaknesses

There is a statistically significant difference in the mean primary lesion diameter between the CT/CEM group (42.82 mm) and the standard CEM group (29.15 mm). Although the authors performed a multivariate logistic regression to account for this, the substantial baseline disparity in tumor size could still heavily bias the subjective assessments of lesion enhancement and conspicuity. Larger tumors are inherently easier to detect, which could mask potential decreases in sensitivity associated with the CT/CEM protocol.
All CEM images were evaluated in consensus by two radiologists rather than independently. Consequently, inter-observer agreement was not assessed. For a study relying entirely on a subjective visual scoring system, failing to evaluate observer variability and reproducibility significantly limits the robustness of the findings.
Although the authors acknowledge this limitation, the relatively small cohort (n = 63) restricts statistical power and may reduce the reliability of subgroup analyses, particularly when evaluating lesion characteristics such as BPE effects or additional lesion detection.
The mean tumor diameter was significantly larger in the CT/CEM group than in the CEM-only group (p = 0.008). Larger tumors may inherently have higher enhancement and visibility, potentially biasing the comparison. Although a logistic regression analysis was performed, the study may still be underpowered to fully control this confounding factor.
Only patients with biopsy-proven malignancies were included, preventing evaluation of diagnostic specificity in distinguishing benign from malignant lesions. This limits the ability to assess the real-world diagnostic performance of the proposed protocol.
The manuscript discusses potential benefits such as reduced contrast exposure, lower cost, and improved workflow efficiency. However, these outcomes were not directly analyzed or quantified in the study and therefore remain speculative.

Author Response

We would like to thank the reviewer for the careful evaluation of our manuscript and for the constructive and insightful comments. We have revised the manuscript and edited the language accordingly, and our detailed responses are provided below.

There is a statistically significant difference in the mean primary lesion diameter between the CT/CEM group (42.82 mm) and the standard CEM group (29.15 mm). Although the authors performed a multivariate logistic regression to account for this, the substantial baseline disparity in tumor size could still heavily bias the subjective assessments of lesion enhancement and conspicuity. Larger tumors are inherently easier to detect, which could mask potential decreases in sensitivity associated with the CT/CEM protocol.

We thank the reviewer for this important observation. We agree that larger tumor size may influence lesion detectability and conspicuity. To address this potential confounding factor, we performed a multivariate logistic regression analysis including both group status and primary tumor size.

The analysis demonstrated that tumor size was not a statistically significant predictor of lesion detectability (OR = 1.052, 95% CI: 0.989–1.119, p = 0.107). However, we acknowledge that the baseline imbalance in tumor size may still introduce residual bias, particularly given the relatively small sample size.

Accordingly, we have revised the Discussion section to explicitly address this issue and emphasize that the findings should be interpreted with caution and validated in larger, more balanced cohorts

All CEM images were evaluated in consensus by two radiologists rather than independently. Consequently, inter-observer agreement was not assessed. For a study relying entirely on a subjective visual scoring system, failing to evaluate observer variability and reproducibility significantly limits the robustness of the findings.

We appreciate this important comment. In our study, image interpretation was performed in consensus to ensure consistent evaluation of lesion characteristics. However, we agree that the absence of interobserver variability analysis represents a limitation, particularly given the use of a subjective visual scoring system.

This limitation has now been clearly stated in the Discussion section, and we have emphasized that future studies incorporating independent readings and interobserver agreement analysis would be valuable to further assess reproducibility.

Although the authors acknowledge this limitation, the relatively small cohort (n = 63) restricts statistical power and may reduce the reliability of subgroup analyses, particularly when evaluating lesion characteristics such as BPE effects or additional lesion detection.

We agree that the relatively small sample size may limit statistical power, particularly in subgroup analyses such as BPE-related comparisons and additional lesion detection.

We have now explicitly addressed this limitation in the Discussion section, noting that some non-significant findings may be related to limited statistical power rather than true equivalence, and that larger prospective studies are needed.

The mean tumor diameter was significantly larger in the CT/CEM group than in the CEM-only group (p = 0.008). Larger tumors may inherently have higher enhancement and visibility, potentially biasing the comparison. Although a logistic regression analysis was performed, the study may still be underpowered to fully control this confounding factor.

We acknowledge this important point. While multivariate logistic regression did not demonstrate a statistically significant independent effect of tumor size, we recognize that the study may be underpowered to fully account for this confounding factor.

We have therefore revised the Discussion section to highlight that residual confounding cannot be completely excluded and that the findings should be interpreted cautiously.

Only patients with biopsy-proven malignancies were included, preventing evaluation of diagnostic specificity in distinguishing benign from malignant lesions. This limits the ability to assess the real-world diagnostic performance of the proposed protocol.

We agree with the reviewer that inclusion of only biopsy-proven malignant cases limits evaluation of diagnostic specificity and reduces generalizability to real-world clinical settings.

This limitation has now been explicitly clarified in the Discussion section, and we have emphasized the need for future studies including both benign and malignant lesions to better assess diagnostic performance.

The manuscript discusses potential benefits such as reduced contrast exposure, lower cost, and improved workflow efficiency. However, these outcomes were not directly analyzed or quantified in the study and therefore remain speculative.

We thank the reviewer for this important clarification. We agree that these outcomes were not directly measured in our study.

Accordingly, we have revised the relevant statements in the Discussion section to avoid overinterpretation. These aspects are now described as potential advantages, supported by previous literature, and clearly stated as not directly evaluated in the present study.

We believe that these revisions have strengthened the manuscript and improved its clarity and scientific rigor. We sincerely thank the reviewer for the valuable comments that contributed to improving our work.