Review Reports - Contrast-Enhanced Mammography vs. Breast MRI for Assessing Neoadjuvant Chemotherapy Response: A Prospective Clinical Comparison Study

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The authors assessed tumour response using CEM and MRI in 74 breast cancers receiving neoadjuvant chemotherapy with CEM being comparable or slightly better for assessing tumor size vs pathology and for CPR. It is vital that time from imaging to surgery is stated. Some of the over prediction of MRI may be accounted for by further tumor shrinkage between imaging and surgery. Overall this is an important study showing that CEM can be used as an alternative

Specific comments

‘The mean residual tumor size was 11.7 mm.’ -what was the mean tumor size before treatment?
Please state the median time and IQR from imaging to surgery
Why was 1cm used for accuracy? Why not 0.5cm? please also provide results for 0.5cm
Predicting the response to neoadjuvant chemotherapy. Can the addition of tomosynthesis improve the accuracy of contrast-enhanced spectral mammography? A comparison with breast MRI.
A number of reports comparing CEM and MRI have been overlooked in the discussion :PMID: 37399083 Free PMC article.PMID: 36754936 Free PMC article. PMID: 35831526; AJR Am J Roentgenol. 2022 Dec;219(6):884-894. doi: 10.2214/AJR.22.27756. Epub 2022 Jun 22.
‘This study showed that in patients receiving NAC, CEM was superior to MRI in predicting which patients had a complete pathological response. – This difference in sensitivity is not significant please make this clear -chi square p=0.24
Please assess over prediction of MRI based on time to surgery

Author Response

Response: We thank the reviewer for this important comment and for the positive evaluation of our study. The median interval between post-NAC imaging and surgery was approximately 2 weeks, and this information has now been added to the Methods and Results sections. Although some degree of additional tumor regression may theoretically occur during this interval, both CEM and MRI were performed within the same timeframe prior to surgery. Therefore, any interval tumor shrinkage would be expected to affect both modalities similarly, and is unlikely to fully explain the observed differences in overestimation rates. This point has been clarified in the revised manuscript.

Specific comments

‘The mean residual tumor size was 11.7 mm.’ -what was the mean tumor size before treatment?

Response: Thank you for this comment. Although baseline tumor sizes were presented in Table 2, we have now explicitly added the mean pre-treatment tumor sizes to the Results section for clarity.

Please state the median time and IQR from imaging to surgery

Response: We thank the reviewer for this important comment. The median time from post-NAC imaging to surgery was 14 days (IQR: 10–18 days). This information has now been added to the Methods section of the revised manuscript.

Why was 1cm used for accuracy? Why not 0.5cm? please also provide results for 0.5cm

Response: Thank you for this important methodological question. The ±1 cm threshold was selected based on prior literature evaluating imaging-pathology concordance in the neoadjuvant setting. In response to your suggestion, we additionally evaluated a stricter ±0.5 cm threshold.

Using the ±0.5 cm criterion, accuracy decreased for both modalities (CEM: 68.1%; MRI: 58.3%), and the paired comparison remained statistically non-significant (exact McNemar p = 0.118). These results have now been added to the manuscript and are presented in Supplementary Table S1.

Predicting the response to neoadjuvant chemotherapy. Can the addition of tomosynthesis improve the accuracy of contrast-enhanced spectral mammography? A comparison with breast MRI.

Response: We thank the reviewer for this insightful suggestion. The potential addition of digital breast tomosynthesis to contrast-enhanced mammography may indeed improve lesion characterization and response assessment. However, the present study was designed to compare CEM and MRI, and tomosynthesis was not incorporated into our imaging protocol.

We agree that future prospective studies evaluating CEM combined with tomosynthesis would be valuable, and this point has now been acknowledged in the Discussion section as a potential direction for further research.

A number of reports comparing CEM and MRI have been overlooked in the discussion :PMID: 37399083 Free PMC article.PMID: 36754936 Free PMC article. PMID: 35831526; AJR Am J Roentgenol. 2022 Dec;219(6):884-894. doi: 10.2214/AJR.22.27756. Epub 2022 Jun 22.

Response: We thank the reviewer for highlighting these important and recent studies. The suggested references (PMID: 37399083; 36754936; 35831526) have now been incorporated into the Discussion section and cited where appropriate to strengthen the contextualization of our findings within the current literature.

‘This study showed that in patients receiving NAC, CEM was superior to MRI in predicting which patients had a complete pathological response. – This difference in sensitivity is not significant please make this clear -chi square p=0.24

Response: Thank you for this important clarification. We agree that the difference in sensitivity between CEM and MRI did not reach statistical significance (χ² p = 0.24). Accordingly, we have revised the Abstract, Results, Discussion, and Conclusion sections to clearly state that the higher sensitivity observed for CEM was numerical and not statistically significant. The manuscript no longer describes CEM as superior, but rather as demonstrating comparable performance with numerically higher sensitivity.

Please assess over prediction of MRI based on time to surgery

Response: Thank you for this important comment. The median interval between post-NAC imaging and surgery was approximately 14 days (IQR: 10–18 days). Both CEM and MRI were performed within the same timeframe prior to surgery; therefore, any additional tumor regression during this interval would be expected to affect both modalities similarly.

Although MRI showed a higher overestimation rate than CEM (15.3% vs 6.9%), we did not observe a correlation between imaging-to-surgery interval and overestimation. Given the relatively short and comparable interval for both modalities, time to surgery is unlikely to have selectively influenced MRI overestimation in our cohort. This interpretation has been clarified in the Discussion s

Reviewer 2 Report

Comments and Suggestions for Authors

This is a simple, well done, and important study that supports previously published data and should make more clinicians comfortable performing CEM for patients who are unable to get an MRI. A short discussion on the findings of lower specificity of CEM vs. MRI in this study compared to other studies would strengthen the paper. Comparison of accuracy of the imaging modalities between pre and postmenopausal patients, by tumor stage, and receptor status would also strengthen the impact of the paper

Author Response

Response: Thank you for this valuable comment. In our study, CEM demonstrated slightly lower specificity compared with MRI (70.6% vs 74.5%), which differs from some previous reports showing comparable or higher specificity for CEM. This discrepancy may be related to differences in patient selection, tumor biology distribution, and sample size. In addition, small residual microscopic tumor foci with limited contrast enhancement may contribute to response misclassification and influence specificity estimates. This point has now been expanded in the Discussion section with comparison to previous studies.

Comparison of accuracy of the imaging modalities between pre and postmenopausal patients, by tumor stage, and receptor status would also strengthen the impact of the paper

Response: Thank you for this valuable suggestion. In response to your comment, we performed additional subgroup analyses stratified by menopausal status, tumor stage, and molecular receptor subtype.

Across these subgroups, CEM showed numerically higher sensitivity in several categories, while specificity was generally comparable between CEM and MRI, without statistically significant differences in paired comparisons. These findings have been incorporated into the Results section and are presented in Supplementary Table S2. Relevant interpretation has also been added to the Discussion section.

Reviewer 3 Report

Comments and Suggestions for Authors

I agree with the conclusion of this paper CEM is becoming a very useful tool in breast imaging, it gives almost the same information as MRI at a lower cost and is more easily available. It is a good alternative to MRI.

I have some questions for the authors: I. notice that 74% of the cases had mastectomies in spite of the fact that tumors where relatively small. Did you see any change in practice to more breast conserving surgery. Besides assessing the response to treatment CME or MRI help decide if the patient is eligible for breast conserving surgery.

The PCR was 31% but you had 25.7% of the cases as Luminal A. We know PCR in this group is less. Was there any difference in sensitivity or specificity between MRI and CEM in this specific group.

Author Response

Comment: I agree with the conclusion of this paper CEM is becoming a very useful tool in breast imaging, it gives almost the same information as MRI at a lower cost and is more easily available. It is a good alternative to MRI.

Response: Thank you for your positive and supportive comment. We appreciate your recognition of the clinical value of CEM as a practical and accessible alternative to MRI.

Comment: I have some questions for the authors: I. notice that 74% of the cases had mastectomies in spite of the fact that tumors where relatively small. Did you see any change in practice to more breast conserving surgery. Besides assessing the response to treatment CME or MRI help decide if the patient is eligible for breast conserving surgery.

Response: Thank you for this important comment. Although post-NAC residual tumor sizes were relatively small in many patients, a considerable proportion initially presented with stage II–III disease, multifocal involvement, or extensive tumor bed areas, which influenced the decision toward mastectomy. In addition, 12.2% of patients had confirmed BRCA mutations, which may favor mastectomy. Patient preference and institutional surgical practices during the study period also contributed. Both CEM and MRI were used as part of routine surgical planning; however, the present study was designed to evaluate imaging accuracy rather than to assess changes in surgical decision-making or rates of breast-conserving surgery. This point has been clarified in the Discussion section.

Comment: The PCR was 31% but you had 25.7% of the cases as Luminal A. We know PCR in this group is less. Was there any difference in sensitivity or specificity between MRI and CEM in this specific group.

Response: Thank you for this important comment. Luminal A tumors represented 25.7% of our cohort, while the overall pCR rate was 31.1%. As expected, no pathological complete response was observed among Luminal A patients.

In subgroup analysis limited to Luminal A tumors, no difference was observed between CEM and MRI in specificity, with both modalities demonstrating identical specificity rates (84.2%). Sensitivity could not be meaningfully calculated because no Luminal A patient achieved pCR.

These findings are consistent with the known lower pCR rates in Luminal A tumors and suggest comparable performance of CEM and MRI in ruling out complete response within this subgroup. The manuscript has been revised accordingly in the Results and Discussion sections.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The authors have responded well to reviewer comments