Visual Evaluation of Ultrafast MRI in the Assessment of Residual Breast Cancer after Neoadjuvant Systemic Therapy: A Preliminary Study Association with Subtype

Round 1

Reviewer 1 Report

This retrospective study of 55 breast MRIs performed after neoadjuvant therapy compared four acquisitions for correlation between imaging and pathological findings. The authors find that the final acquisition of an ultrafast sequence outperforms higher resolutions acquisitions performed at later time points after contrast administration. They further find associations across tumor subtypes.

1. The main strength of ultrafast MRI is its improved temporal resolution. However, this study uses only a single timepoint – the final acquisition. As such, the main difference in the ultrafast image versus the ‘early’ image is its timing and thicker slices. The authors should comment on what factors of the ultrafast acquisition are leading to the improved detection of residual tumor in the absence of any dynamic analysis.
2. The title ‘Ultrafast dynamic contrast-enhanced MRI…’ is misleading as no dynamic analysis is performed.
3. It appears that the UF measures the smallest lesion, followed by ‘Early’, and then ‘Delayed.’ As these are acquired in that order is lesion size simply a function of time after contrast agent injection, with larger lesion sizes at later time points due to contrast agent extravasation?
4. The text states that 23 lesions achieved pCR but Tables 2 and 3 indicates 33 lesions were pCR.
5. The statement that ‘Use of UF-DCE MRI to assess response to treatment has not been established’ is not supported. Previous studies have used ‘ultrafast’ MRI to investigate response to therapy, among them the manuscript’s citation #24 and doi 10.1007/s00330-021-08530-4.
6. Please report the qualifications of the readers, e.g., board-certified breast radiologist for XXX years.
7. The timing of the DCE acquisition is difficult to follow. A flowchart demonstrating the timing of the acquisitions would aid the reader.
8. What was the imaging plane for acquisition? It appears to be axial from the images, but this should be specified in the manuscript.
9. I would not refer to a kappa of 0.81 as ‘almost perfect.’ Simply state the kappa value and allow the reader to assess whether this kappa level is sufficient.
10. The title of Table 3 should be more descriptive, such as ‘Diagnostic performances for detected residual lesion across two readers.’
11. The results section does not refer to Figures 2, 5, or 6. These should be introduced in the text.
12. For Figure 6, please provide the calculated lesion size for each acquisition.

Author Response

This retrospective study of 55 breast MRIs performed after neoadjuvant therapy compared four acquisitions for correlation between imaging and pathological findings. The authors find that the final acquisition of an ultrafast sequence outperforms higher resolutions acquisitions performed at later time points after contrast administration. They further find associations across tumor subtypes.   1. The main strength of ultrafast MRI is its improved temporal resolution. However, this study uses only a single timepoint – the final acquisition. As such, the main difference in the ultrafast image versus the ‘early’ image is its timing and thicker slices. The authors should comment on what factors of the ultrafast acquisition are leading to the improved detection of residual tumor in the absence of any dynamic analysis.  

Improved temporal resolution leads to more the capture of contrast effects at more pinpoint times. While conventional DCE MRI reflects averaged contrast effects from 60 to 120 seconds after contrast injection, the final phase of our UF-DCE MRI reflects contrast effects only within 56 to 60 seconds after injection. This may be the reason why Uf-DCE MRI yields the improved detection of residual cancer. The following sentence was added in the 1st paragraph of the discussion section.    

“The final phase of UF-DCE MRI may be an appropriate time to capture the enhancement of residual invasive cancer.”  

2. The title ‘Ultrafast dynamic contrast-enhanced MRI…’ is misleading as no dynamic analysis is performed.  

We appreciate the reviewer’s comment. The title was corrected as follows.   “Visual evaluation of ultrafast MRI for evaluation of residual breast cancer after neoadjuvant systemic therapy: a preliminary study association with subtype”  

3. It appears that the UF measures the smallest lesion, followed by ‘Early’, and then ‘Delayed.’ As these are acquired in that order is lesion size simply a function of time after contrast agent injection, with larger lesion sizes at later time points due to contrast agent extravasation?  

We agree that enhanced lesion becomes larger at later points due to contrast agents’ extravasation. However, in the post NAC status, residual tumor is enhanced weaker than before, which overlaps slow enhancement of post-treatment inflammation or scar and causes overestimation of residual tumor. Hypothesizing that UF-DCE MRI is more sensitive to the hypervascular tissues and more accurate in depicting viable tumor, the current study assessed the diagnostic performance of UF-DCE in diagnosing pCR. The following sentences were inserted in the introduction.  

“In the post NAC status, residual tumor is enhanced weaker than before, which overlaps slow enhancement of post-treatment inflammation or scar and causes overestimation of residual tumor. We hypothesized that UF-DCE MRI may perform better in distinguishing rapidly-enhancing viable cancer tissue from gradually-enhancing scar or inflammatory tissue after NST compared with conventional DCE MRI”  

4. The text states that 23 lesions achieved pCR but Tables 2 and 3 indicates 33 lesions were pCR.  

Thank you for pointing this out. We’ve corrected the number in the text.  

5. The statement that ‘Use of UF-DCE MRI to assess response to treatment has not been established’ is not supported. Previous studies have used ‘ultrafast’ MRI to investigate response to therapy, among them the manuscript’s citation #24 and doi 10.1007/s00330-021-08530-4.  

We appreciate the reviewer’s suggestion. The sentence was modified with the references.  

“Recently, some studies investigated the use of UF-DCE MRI to assess breast cancer response to NST. Kim et al. found the association of UF-DCE MRI-derived parameters with pCR among triple-negative breast cancer [16]. According to one study evaluating the diagnostic performance of UF-DCE MRI to detect residual cancer after NST, UF-DCE MRI was more specific than conventional DCE MRI [17]. Image findings after NST may differ among breast cancer subtypes, which may affect the diagnostic performance of UF-DCE MRI.”    

6. Please report the qualifications of the readers, e.g., board-certified breast radiologist for XXX years.  

Thank you for the comment. I’ve added that explanation.  

7. The timing of the DCE acquisition is difficult to follow. A flowchart demonstrating the timing of the acquisitions would aid the reader.  

We appreciate the reviewer’s comment. We’ve added a figure demonstrating the timing of each acquisition as Figure 3.  

8. What was the imaging plane for acquisition? It appears to be axial from the images, but this should be specified in the manuscript.  

The imaging plane is specified in Table 1.  

9. I would not refer to a kappa of 0.81 as ‘almost perfect.’ Simply state the kappa value and allow the reader to assess whether this kappa level is sufficient.  

Thank you for the comment. We’ve rephrased the sentence as follows.  
“The inter-reader agreement for presence or absence of a residual lesion was better for UF (kappa = 0.81) than for Early, Delayed and HR (kappa = 0.64, 0.68 and 0.52, respectively).”  

10. The title of Table 3 should be more descriptive, such as ‘Diagnostic performances for detected residual lesion across two readers.’  

We appreciated the reviewer’s comment. The title has been modified as such.  

11. The results section does not refer to Figures 2, 5, or 6. These should be introduced in the text.  

The indications for the figures (figure numbers have been changed to Figure 4, 7, and 8 as some figures are added) are in Page5, Line189 and Page 9 Line 231.  

12. For Figure 6, please provide the calculated lesion size for each acquisition.  

Thank you for the suggestion. We added the lesion size in the figure caption.  

 

Reviewer 2 Report

This manuscript is a study on the usefulness of ultrafast MRI for the evaluation of reponse to neoadjuvant chemotherapy. This is very interesting and well-written.

I would like to make a minor comment in the Introduction section. Please revise the sentence  in page 2, line 64, 'Use of UF-DCE MRI to assess response to treatment has not been established.' There are two recent studies to show the usefulness of ultrafast MRI to assess response to neoadjuvant chemotherapy. Please add these references and revise the sentence. 

- European Radiology 2022. doi: 10.1007/s00330-021-08530-4.

Ultrafast dynamic contrast.enhanced breast MRI: association with pathologic complete response in neoadjuvant treatment of breast cancer

- Jpn J Radiol 2021 Aug;39(8):791-801.  

Value of ultrafast and standard dynamic contrast-enhanced magnetic resonance imaging in the evaluation of the presence and extension of residual disease after neoadjuvant chemotherapy in breast cancer

Author Response

This manuscript is a study on the usefulness of ultrafast MRI for the evaluation of reponse to neoadjuvant chemotherapy. This is very interesting and well-written.

I would like to make a minor comment in the Introduction section. Please revise the sentence in page 2, line 64, 'Use of UF-DCE MRI to assess response to treatment has not been established.' There are two recent studies to show the usefulness of ultrafast MRI to assess response to neoadjuvant chemotherapy. Please add these references and revise the sentence. 

- European Radiology 2022. doi: 10.1007/s00330-021-08530-4.

Ultrafast dynamic contrast.enhanced breast MRI: association with pathologic complete response in neoadjuvant treatment of breast cancer

- Jpn J Radiol 2021 Aug;39(8):791-801.  

Value of ultrafast and standard dynamic contrast-enhanced magnetic resonance imaging in the evaluation of the presence and extension of residual disease after neoadjuvant chemotherapy in breast cancer

We appreciate the reviewer’s suggestion. The sentence was modified with the references.

“Recently, some studies investigated the use of UF-DCE MRI to assess breast cancer response to NST. Kim et al. found the association of UF-DCE MRI-derived parameters with pCR among triple-negative breast cancer [16]. According to one study evaluating the diagnostic performance of UF-DCE MRI to detect residual cancer after NST, UF-DCE MRI was more specific than conventional DCE MRI [17]. Image findings after NST may differ among breast cancer subtypes, which may affect the diagnostic performance of UF-DCE MRI.”

Reviewer 3 Report

I would like to congratulate the authors for their interesting manuscript.

 

Here, I have made a few suggestions that, in my opinion, could help improve the overall quality of the manuscript.

 

In particular, the authors may consider augmenting the number of each category of breast cancer subtype; in fact, the principat subgroup of breast cancer who undergoing NACT/NST are the HER2+ and Triple negative (TNBC). In this paper, the number of these hystopathological subgroup are, in my opinion, too small. I suggest, for a future study, to improve the number of the these two important categories of breast cancer. 

In relation of the limited numer of partecipants, I suggest to add in the title or in the study limitations, that it is a "preliminary study" and it will be necessary for further studies to increase the sample size. 

Author Response

I would like to congratulate the authors for their interesting manuscript. Here, I have made a few suggestions that, in my opinion, could help improve the overall quality of the manuscript.    In particular, the authors may consider augmenting the number of each category of breast cancer subtype; in fact, the principat subgroup of breast cancer who undergoing NACT/NST are the HER2+ and Triple negative (TNBC). In this paper, the number of these hystopathological subgroup are, in my opinion, too small. I suggest, for a future study, to improve the number of the these two important categories of breast cancer.    In relation of the limited numer of partecipants, I suggest to add in the title or in the study limitations, that it is a "preliminary study" and it will be necessary for further studies to increase the sample size.  

We appreciate the reviewer’s comment. The title was corrected as follows.  

“Visual evaluation of ultrafast MRI for evaluation of residual breast cancer after neoadjuvant systemic therapy: a preliminary study association with subtype”  

The following sentence was added in the limitation section.  

“Future studies including a sufficient number of cases are needed to clarify the use of UF-DCE MRI in the evaluation of treatment response in each subtype, especially TNBC and HER2+BC.”

 

Reviewer 4 Report

This manuscript by Honda et al. reports the ultrafast dynamic contrast-enhanced MRI for evaluation of residual breast cancer after neoadjuvant systemic therapy: association with subtype. The results show that UF-DCE MRI demonstrated higher AUC and specificity for more accurate detection of residual cancer and visualization of tumor extent than conventional DCE MRI. This is an interesting study, I recommend a minor revision.

1. It is suggested to add more discussion in INTRODUCTION to enrich the background.

2. I suggest authors to add a scheme describing the main content of this study at the beginning of the manuscript.

3. The resolution of MRI images should be improved.

Author Response

This manuscript by Honda et al. reports the ultrafast dynamic contrast-enhanced MRI for evaluation of residual breast cancer after neoadjuvant systemic therapy: association with subtype. The results show that UF-DCE MRI demonstrated higher AUC and specificity for more accurate detection of residual cancer and visualization of tumor extent than conventional DCE MRI. This is an interesting study, I recommend a minor revision.

1. It is suggested to add more discussion in INTRODUCTION to enrich the background.

We appreciate the reviewer’s suggestion. The 3rd to 5th paragraphs were modified as follows.

“Because pCR can only be confirmed after surgery, cCR is used as an alternative for preoperative treatment planning. Although a biopsy is often useful, it sometimes fails to sample the viable portion of a heterogeneous tumor, whereas DCE MRI can evaluate the entire residual lesion. Meanwhile, performance of MRI for prediction of pCR has unsatisfactory accuracy of around 74% [8]. Furthermore, enhancing scar or inflammatory tissue is sometimes misdiagnosed as residual cancer [9]. The improved accuracy of MRI in predicting pCR may lead to the establishment of non-surgical treatment options for breast cancer patients.”

“In the post NAC status, residual tumor is enhanced weaker than before, which overlaps slow enhancement of post-treatment inflammation or scar and causes overestimation of residual tumor. We hypothesized that UF-DCE MRI may perform better in distinguishing rapidly-enhancing viable cancer tissue from gradually-enhancing scar or inflammatory tissue after NST compared with conventional DCE MRI”

“Recently, some studies investigated the use of UF-DCE MRI to assess breast cancer response to NST. Kim et al. found the association of UF-DCE MRI-derived parameters with pCR among triple-negative breast cancer [16]. According to one study evaluating the diagnostic performance of UF-DCE MRI to detect residual cancer after NST, UF-DCE MRI was more specific than conventional DCE MRI [17]. Image findings after NST may differ among breast cancer subtypes, which may affect the diagnostic performance of UF-DCE MRI.”

2. I suggest authors to add a scheme describing the main content of this study at the beginning of the manuscript.

We appreciate the reviewer’s comment. We’ve added a scheme explaining our study hypothesis and design as Figure 1. 

3. The resolution of MRI images should be improved.

Thank you for the comment. The MR images in Figures 7 and 8 have been changed to images with improved resolution.