The Impact of Dense Breasts on the Stage of Breast Cancer at Diagnosis: A Review and Options for Supplemental Screening

Round 1

Reviewer 1 Report

This is a timely and comprehensive review of supplemental screening options for women with dense breasts. The readership should find this to be a valuable reference. The paper is well written overall, but several areas of revision are suggested. First, I found that the intended message about the stage at diagnosis became buried in other details. Perhaps a summary table could help emphasize the intended communication of differing breast cancer detection stage, with the information presented by both modality and density. Also, it seems that this paper is intended for a specific audience. It is suggested that the author revise some aspects to make the message internationally applicable. For example, the suggestion that screening guidelines must decrease the interval between screening mammograms should be given more context such that the reference case is understood. Also, risk is mentioned multiple times, but with little mention of what risk models are applied, and what constitutes high/moderate/low risk. Additional specific comments follow.

Abstract, line 20, I find it to be overly general to suggest that “screening guidelines must decrease the interval between screening mammograms” – Although I appreciate the point about equality, I’m sure that the benefit from adjustments to the screening interval would depend on the local standard interval. As such, I recommend qualifying this statement.

Pg 3, line 56, suggested revision, “The more important risk [with respect to screening performance] is the masking of cancers”

Pg 7, lines 154 – 157, I suggest removing this quote from ref 46. It appears that it might have been a placeholder given the relevant citations have been included.

Pg 8, lines 76, 84, 98: Please be consistent about the identification of the breast density assessment method, especially when this is important to interpret quantitative results and to compare them between studies. E.g., when reporting on the Bertrand et al study (ref 14) no explanation was given for the 51% value (measured by a computer-assisted area-based method). Ref 9 is based on a visual quantification of the Wolfe breast density pattern. The method should also be identified for the Eriksson et al study. Also, be sure to identify whether BI-RADS 4^th ed or 5^th ed was used in a study for instances when the particular edition is important for comparisons.

Fig 6. can alternate DBT images be obtained? The DBT included appears to show the first or last slice only, at the skin line for each instance.

Pg 27, add some text to introduce Fig 7. It currently follows a quote about the ACR guidelines recommended use of MRI, but it doesn’t directly illustrate this.

Pg 27, lines 359 to 365, this quote is actually from ref 103. Also, including the entire quote seems unnecessary. Was this meant as a placeholder??

Pg 31, CEM subtraction images are also sometimes referred to as dual-energy (DE)

Fig 8, while it is agreed that HE images are not valuable for interpretation, could new images for Fig 8C be obtained with contrast adjusted to show some parenchyma? These images are rather unfair as they look like image segmentations/masks. I appreciate that processed HE images aren’t available, but a notation could be made that they are ‘For Processing’ format to set the expectation that they will appear quite different to conventional DM.

Pg 34, line 501 Since this sentence currently repeats the ACR recommendations that were stated on pg 27, I would recommend adding a discussion of the latest NCCN guidelines (version 1.2021), which recommend whole breast US and also suggest CEM as an alternative to MRI.

Pg 35, line 519, since it’s now available, you could update this to add a reference to the ACR BI-RADS CEM supplement https://www.acr.org/-/media/ACR/Files/RADS/BI-RADS/BIRADS_CEM_2022.pdf

Pg 39, line 612, for the availability of MRI it is suggested to reference OECD stats: https://data.oecd.org/healtheqt/magnetic-resonance-imaging-mri-units.htm

Pg 39, lines 618-619, suggest revisions to the CEM statement, “CEM has sensitivity approaching that of MRI, [and] uses standard mammography equipment and is becoming more widely-available. It requires injection, but uses non-ionic [iodinated] contrast; not gadolinium”

Pg 39, line 637, apart from the discussion around differences in interval cancers rates between screening programs with different screening intervals, little was suggested about screening intervals. To further support the proposal to decrease the screening interval for women with dense breasts I would suggest adding some direct discussion of this topic in the main text. Also, please clarify whether you are suggesting decreasing the mammographic screening interval, or decreasing the screening interval using any modality? Is there a minimum interval at which no additional benefit is derived?

Minor point – note that some acronyms (such as RCTs) are introduced several times. This could be cleaned up.

Author Response

Responses to Reviewer 1

Thanks for your thoughtful comments and suggestions.

Re: Perhaps a summary table could help emphasize the intended communication of differing breast cancer detection stage, with the information presented by both modality and density.

I have referred to a table from densebreast-info.org, which summarizes summarizes the ICDR in various breast densities, the additional recalls, and the impact on the IC rate of DBT, US, MBI, CEM, MRI and AB MRI.

Re: Abstract, line 20, I find it to be overly general to suggest that “screening guidelines must decrease the interval between screening mammograms” – Although I appreciate the point about equality, I’m sure that the benefit from adjustments to the screening interval would depend on the local standard interval. As such, I recommend qualifying this statement.

In the abstract, I have changed the sentence to: “To provide equity, breast cancer screening guidelines should be created with the goal of maximizing mortality reduction and allowing less aggressive therapy, which may include decreasing the interval between screening mammograms and recommending consideration of supplemental screening for women with dense breasts.” And in the second-to-last paragraph, I have added: “Guidelines for breast cancer screening in the USA vary considerably, depending on the issuing organization. There is still a moratorium on the United States Preventive Services Task Force guidelines that recommend starting mammography screening at age 50.”

Pg 3, line 56, suggested revision, “The more important risk [with respect to screening performance] is the masking of cancers” – I’ve added that phrase

Pg 7, lines 154 – 157, I suggest removing this quote from ref 46. It appears that it might have been a placeholder given the relevant citations have been included. –

I’ve removed the quote, and changed the sentence to: “Summarizing these, 94% of the cancers were invasive, and 70% of them were 1 cm or smaller in size. Ninety percent were stage 0 or stage I. And 90.5% of the women with sonographically-detected cancers had either heterogeneously dense or extremely dense parenchyma[46].”

Pg 8, lines 76, 84, 98: Please be consistent about the identification of the breast density assessment method, especially when this is important to interpret quantitative results and to compare them between studies. E.g., when reporting on the Bertrand et al study (ref 14) no explanation was given for the 51% value (measured by a computer-assisted area-based method). Ref 9 is based on a visual quantification of the Wolfe breast density pattern. The method should also be identified for the Eriksson et al study. Also, be sure to identify whether BI-RADS 4^th ed or 5^th ed was used in a study for instances when the particular edition is important for comparisons.

– rather than be consistent (because there’s inconsistency among the references cited), I’ve taken your suggestion and included specifics. I’ve changed the first sentence to: “The other risk for women with dense breasts is an increased risk of getting breast cancer, but the magnitude of increased risk varies, possibly in part because of the different ways of measuring density.”

- I’ve changed the 3^rd sentence to: “Boyd et al, using a semi-quantitative method called interactive thresholding to determine breast density found that…”

- I’ve changed the next sentence to: “Using a five-category scale of percentage breast density, McCormack and dos Santos Silva, in a meta-analysis of 42 studies using three different density-grading methods, found…”

Fig 6. can alternate DBT images be obtained? The DBT included appears to show the first or last slice only, at the skin line for each instance.

I agree that they look blurry, but I assume they’ll look less blurry if they’re published in a smaller size image. They were given to me by a colleague, so can’t switch them. Because the DBT studies were negative, it wouldn’t matter what slice was used. Ideally, showing the entire cine of the DBT would be best, but clearly that’s not possible in a static document. If you prefer, I can omit the DBT images from the figure, and just include the US images.

Pg 27, add some text to introduce Fig 7. It currently follows a quote about the ACR guidelines recommended use of MRI, but it doesn’t directly illustrate this.

- I’ve added to the text: Figure 7 shows a tiny cancer detected on MRI, and subsequently on directed US, in a 52-year-old high risk patient, whose 2D screening mammogram showed BIRADS C density and was negative.”

Pg 27, lines 359 to 365, this quote is actually from ref 103. Also, including the entire quote seems unnecessary. Was this meant as a placeholder??

- I’ve removed the quotes and paraphrased them as: “the American College of Radiology now recommends MRI for women with genetics-based increased risk (and their untested first-degree relatives), with a calculated lifetime risk of 20% or more, women with a history of chest or mantle radiation therapy at a young age, women with dense tissue who’ve been treated for breast cancer, or those diagnosed by age 50. They recommend consideration of MRI surveillance for women diagnosed with breast cancer or atypia at biopsy, especially if they have other risk factors are. They suggest that ultrasound be considered for those who qualify for but cannot undergo MRI. And they recommend that all women, especially black women and those of Ashkenazi Jewish descent, should be evaluated for breast cancer risk no later than age 30, to identify those at higher risk so that they can benefit from supplemental screening” and changed the citation to reference 103.

Pg 31, CEM subtraction images are also sometimes referred to as dual-energy (DE).

With respect: that is incorrect. “dual energy” refers to the technique, of obtaining 2 exposures, one low energy and one high. Before contradicting you, I checked with Dr. Maxine Jochelson (the author of one of the papers I cited) who concurs.

Fig 8, while it is agreed that HE images are not valuable for interpretation, could new images for Fig 8C be obtained with contrast adjusted to show some parenchyma? These images are rather unfair as they look like image segmentations/masks. I appreciate that processed HE images aren’t available, but a notation could be made that they are ‘For Processing’ format to set the expectation that they will appear quite different to conventional DM.

I obtained these images from Dr. Anat Kornecki, the radiologist in London Ontario who introduced CEM to Canada. It’s my understanding that she doesn’t adjust contrast on HE images, since they are not used for interpretation. It is notable that Dr. Jochelson did not include any HE images in her State-of-the-Art paper. If you prefer that I leave them out, I have no problem doing so.

Pg 34, line 501 Since this sentence currently repeats the ACR recommendations that were stated on pg 27, I would recommend adding a discussion of the latest NCCN guidelines (version 1.2021), which recommend whole breast US and also suggest CEM as an alternative to MRI.

I’ve included NCCN in that paragraph. I read the website: https://www.nccn.org/professionals/physician_gls/pdf/breast-screening.pdf and found these statements pertaining to US and CEM:

“Handheld or automated ultrasound can increase cancer detection rates in individuals with dense breast tissue, but may increase recall and increase benign breast biopsies.”

“Contrast-enhanced mammography is also an emerging efficacious option for higher risk breast cancer screening”

However I cannot find a statement that specifically suggests CEM as an alternative to MRI.

So I’ve worded it as follows: “In addition to the ACR recommendations for screening high-risk women, the National Comprehensive Cancer Network (NCCN) now states that whole breast US can increase cancer detection rates in individuals with dense breast tissue, but may increase recall and increase benign breast biopsies, and that contrast-enhanced mammography is also an emerging efficacious option for higher risk breast cancer screening.” I’ve added a reference.

Pg 35, line 519, since it’s now available, you could update this to add a reference to the ACR BI-RADS CEM supplement https://www.acr.org/-/media/ACR/Files/RADS/BI-RADS/BIRADS_CEM_2022.pdf

I have, and added a reference.

Pg 39, line 612, for the availability of MRI it is suggested to reference OECD stats: https://data.oecd.org/healtheqt/magnetic-resonance-imaging-mri-units.htm

added

Pg 39, lines 618-619, suggest revisions to the CEM statement, “CEM has sensitivity approaching that of MRI, [and] uses standard mammography equipment and is becoming more widely-available. It requires injection, but uses non-ionic [iodinated] contrast; not gadolinium”

changed

Pg 39, line 637, apart from the discussion around differences in interval cancers rates between screening programs with different screening intervals, little was suggested about screening intervals. To further support the proposal to decrease the screening interval for women with dense breasts I would suggest adding some direct discussion of this topic in the main text. Also, please clarify whether you are suggesting decreasing the mammographic screening interval, or decreasing the screening interval using any modality? Is there a minimum interval at which no additional benefit is derived?

I have changed the first sentence in the paragraph (now starting line 768) to: “Guidelines for breast cancer screening in the USA vary considerably, depending on the issuing organization. There is still a moratorium on the United States Preventive Services Task Force guidelines that recommend starting mammography screening at age 50. Most screening in the USA is ad hoc, whereas elsewhere it is largely performed in organized programs.”

And I have changed the final sentence to (line 778): “Until these studies are completed and to provide equity, organized screening programs that screen less often than annually should consider decreasing the screening interval for mammography for women at higher-than-averge risk, and consider issuing  recommendations supplemental screening for women at increased risk, and for women with dense breasts.”

Minor point – note that some acronyms (such as RCTs) are introduced several times. This could be cleaned up.

Done, and at the request of another reviewer, a table was inserted at the beginning of the paper, just after the key words, which unfortunately changed all the line numbers.

Reviewer 2 Report

A very well written review! It was a pleasure to read it, even if a little to long...

My only suggestion would be to try to shorten it to a reasonable number of pages.

Author Response

Thank-you. I realize it is long, but the journal lists no limit to the length of the manuscript or number of references. 

Reviewer 3 Report

General Comments

This review article is well written and comprehensive, and covers an important topic: imaging of breasts that are mammographically dense. The author also includes a superb set of figures, some her own, and others shared by fellow clinicians. I am confident that this paper will add significantly to the field of breast imaging once it is published.

Given that the author covered all the different imaging modalities, I was surprised that she did not include papers that explore the benefits of dual-modality imaging, in particular the combination of mammography and ultrasound in a single platform. I refer here to the work of the groups based in Michigan, USA, Heidelberg, Germany, and Cape Town, South Africa.

There is extensive use of acronyms throughout the manuscript (e.g. ACR, BI-RADS, DBT, etc.) and so I would recommend the inclusion of a table at the end (or beginning) where each is defined.

Specific Comments

page 2 line 16   Define RCT the first time it is introduced

page 13 line 289   ... posteriorly (arrows).

page 13 line 292   ... facilities had DBT units and ...

page 13 line 297    ... in six Canadian cities [80].

page 29 line 398   ... there is no CPT code for abbreviated MRI.

page 30 line 432    ... also applies for average-risk women.

page 39 line 604    ... So its use in initial screening ...

page 50 line 1093   Reference 153 appears to be incomplete

Author Response

Thank-you for your thoughtful comments and suggestions. Shocking how many typos you caught, even after I read it dozens of times!

Re the comment:

Given that the author covered all the different imaging modalities, I was surprised that she did not include papers that explore the benefits of dual-modality imaging, in particular the combination of mammography and ultrasound in a single platform. I refer here to the work of the groups based in Michigan, USA, Heidelberg, Germany, and Cape Town, South Africa.

I was unable to find references from Michigan, USA, Heidelberg, Germany, and Cape Town, South Africa, but added a comparison chart from Kuhl et al, 2010, and from DBI

Re the comment: There is extensive use of acronyms throughout the manuscript (e.g. ACR, BI-RADS, DBT, etc.) and so I would recommend the inclusion of a table at the end (or beginning) where each is defined.

I’ve created a Table and inserted it under the abstract and keywords, but I defer to the editor/reviewer whether it is necessary, given the changes made to the manuscript as follows:

In addition to page 2 line 16   Define RCT the first time it is introduced:

I had it in the “Background” line 34, but have I added (RCTs) to line 16

I’ve added addition al expansions of the acronyms in the text, and deleted some redundancies including:

page 2, lines 46-47, have added American College of Radiology Breast Imaging and Data System

For Interval cancers (ICs), this was already in p. 5, line 108

Have added “estrogen receptor (ER) and Progesterone receptor (PR) on p. 6 lines 115, 116

Ultrasound (US), digital breast tomosynthesis (DBT), magnetic resonance imaging (MRI), contrast-enhanced mammography (CEM) and molecular breast imaging (MBI) were already on p. 6, lines 137-138

Removed “randomized controlled trials” from p. 6 line 140, but left “RCTs”

Removed “randomized controlled trials” from p. 10 line 267, but left “RCTs”

p. 7 line 161 added “ductal carcinoma in situ”

incremental cancer detection rate (ICDR) already in p. 7 line 165

added “American College of Radiology Imaging Network” to p. 7 line 169

added “positive predictive value” (PPV) to p. 7 line 173

removed “positive predictive value” from p. 7, line 187

added “the Japan Strategic Anti-cancer Randomized Trial (J-START),” to p. 9, line 225

added “United States Food and Drug Administration” (FDA) to p.11, line 284

added synthetic “two-dimensional” (2D), p. 11 line 285

added “mediolateral oblique” p. 13, line 301

added “craniocaudal” p. 20, line 335

I did not expand “DENSE,” in the DENSE trial, p. 28, line 396, since it is really not an acronym, and is explained in the cited reference

quality-adjusted life-years (QALYs), already in p. 29 line 418

European Society of Breast Imaging (EUSOBI) already in p. 29 line 423

Added “computed tomography” (CT) p. 30 line 478

Added “intravenous” (IV) p. 30 line 480

Added “kilovoltage peak” (kVp) p. 30 line 481

Changed LE to low energy p. 31 line 491

p. 40, line 632, changed “FFDM” to 2D mammography to avoid introducing an unnecessary acronym

specific comments:

page 2 line 16   Define RCT the first time it is introduced - added

page 13 line 289   ... posteriorly (arrows).- this is now line 309, have added the “s

page 13 line 292   ... facilities had DBT units and ... ”- now line 313, have added the B

page 13 line 297    ... in six Canadian cities [80]. Now line 318, have added the s

page 29 line 398   ... there is no CPT code for abbreviated MRI. Now line 421, have added “no”

page 30 line 432    ... also applies for average-risk women. Now line 458, have removed “to”

page 39 line 604    ... So its use in initial screening ... Now line 640 – removed the apostrophe

page 50 line 1093   Reference 153 appears to be incomplete – added https://www.theabr.org/diagnostic-radiology/initial-certification/core-exam

Round 2

Reviewer 1 Report

Thank you for the comprehensive response to my comments and suggested revisions. I'm happy for this to be published. There are only a few minor comments for me to respond to.

1. For the NCCN guidelines, the update that mentions CEM (and US) is under “BSCR-2” and “BSCR-3” and reads, “Consider contrast-enhanced mammography or whole breast ultrasound for those who qualify for but cannot undergo MRI. I will leave it to the author's discrection of whether or not to revise further based on this info.

2. You are quite right that “dual energy” refers to the technique. The problem was my poor wording as my comment was intended to suggest another alternative wording for “CEM subtraction images” as they are sometimes called “DE subtraction images.” Anyhow, I have no problem if you leave this out.

3. It seems reasonable to leave out the HE images for Fig 8 as the "blank" looking images aren't entirely representative. Agreed that it's unusual to publish these, so they won't be missed.

Author Response

Point 1. I've made the change, p. 35-36, lines 597-609.

Point 2. Thanks, I'll leave it out.

Point 3. I've removed the HE images, adjusted the labelling and the figure legend. I just received the pathology from the mastectomy, so added that to the legend.

FYI: I was refused permission from JCO to use the figure from the Kuhl paper (!), so had to remove that figure, but included the information shown in the graphic, in the text.