Obesity and Breast Cancer: Interaction or Interference with the Response to Therapy?

Round 1

Reviewer 1 Report

The aim of this study was to investigate a cut-off of BMI with significant prognostic impact, in a cohort of stage I-II BC patients on systemic adjuvant therapy with aromatase inhibitors.

I have some comments:

Subjects and Methods section:

Were the women's various gynecological histories recorded, including the use of oral contraceptives?

Regarding “A further cohort of 116 female patients with radically resected colorectal cancer (CRC), undergoing adjuvant chemotherapy was chosen to represent the "Non-breast" control group”. Why was the reason that the authors choose this control group? Since the obesity is considered a risk factor for colorectal cancer.

Regarding “In addition, data on variables routinely collected at the moment of hormone therapy or chemotherapy start were collected to analyse differences between patient subgroups”. Please include which datas were important for their analysis.

Besides, Were the women's various gynecological histories recorded, including the use of oral contraceptives, hormonal replacement therapy or if used metformin due to diabetes?

The authors asked about the obesity history of the patients? That is, from what age do they have obesity?.

How did you calculate and what was the statistical power of your study?

The study had confounding factors?

Results section:

In spite that the authors in Table 1 present the status of the menopause (pre-or-postmenopausal), the all general characteristics must present divided in premenopausal or postmenopausal status. The latter is since is not adequate mix both group as that different the effect of the obesity in the breast cancer (as the authors refer in their introduction).

Why only document the impact of BMI on 2-years disease-free survival (2yDFS)? Why no 3 years?

If is important the presence of obesity in the poorest prognosis in terms of DFS, the authors only documented the BMI at the moment of the diagnosis of breast cancer? Or the BMI presented in which momento was documented? Do the authors take into account the changes in BMI during the two years of follow-up?

Discussion section:

What were the limitation of their study?

Conclusions:

Regarding “With a view to developing predictive models of risk of relapse, it could be useful to add BMI, with a cut-off of 29 kg/m2, among the criteria to be taken into consideration”

They must be cautious in their conclusions, since the authors do not take into account the change in BMI throughout the follow-up, which is very important to be sure that their cut-off point finding is realistic.

Author Response

Reviewer #1

The aim of this study was to investigate a cut-off of BMI with significant prognostic impact, in a cohort of stage I-II BC patients on systemic adjuvant therapy with aromatase inhibitors.

I have some comments:

Subjects and Methods section:

Were the women's various gynecological histories recorded, including the use of oral contraceptives?

Yes, all anamnestic aspects were recorded. Forty-four women reported use of either hormone replacement therapy (n=10) or oral contraceptives (n=34), but they were evenly distributed among the different groups with no inter-group statistical difference (p=0.96). These data have been added to the text. Thirty-seven patients had a prior history of ovarian stimulation, but they were excluded from the analysis.

Regarding “A further cohort of 116 female patients with radically resected colorectal cancer (CRC), undergoing adjuvant chemotherapy was chosen to represent the "Non-breast" control group”. Why was the reason that the authors choose this control group? Since the obesity is considered a risk factor for colorectal cancer.

We thank the Reviewer for her/his question that might give us the opportunity of better explain the reason to include adjuvant-treated CRC patients as comparator group. Indeed, as the Reviewer correctly pointed out, the obesity is considered a risk factor for colorectal cancer, and this is exactly the reason why we chose this group which is at high risk but who is not under aromatase inhibition. Our results go in this direction, in that they do not demonstrate significant differences in survival.

Regarding “In addition, data on variables routinely collected at the moment of hormone therapy or chemotherapy start were collected to analyse differences between patient subgroups”. Please include which data were important for their analysis.

We thank the Reviewer for her/his comment. Among the variables commonly registered at baseline, prior to therapy start, in order to assess hematological, hepatic and renal function, plus the lipid and mineral balance of AI-treated BC patients, we reported the significant ones. These data have been added to the text.

Besides, Were the women's various gynecological histories recorded, including the use of oral contraceptives, hormonal replacement therapy or if used metformin due to diabetes?

As stated in response to the first part of the question, yes, it was recorded (see first answer). Regarding the use of anti-diabetic drugs, only 42 patients were using oral hypoglycemic medications, mainly consisting of sulfonylureas.

The authors asked about the obesity history of the patients? That is, from what age do they have obesity?

Unfortunately previous history of obesity, i.e. when the obesity developed in individual patients, has not been recorded, however the BMI was available also at the time of the first mammogram, which was at a median time of 4 months before treatment start, and no significant BMI change was observed in this period

How did you calculate and what was the statistical power of your study?

No formal sample size and power calculation was performed since it was a retrospective study including all consecutive patients with the defined inclusion criteria.

The study had confounding factors?

Among commonly considered confounding factors when evaluating BC patients (age at diagnosis, presence of specific genes, family history of breast cancer, reproductive history, pathological tumour and nodal stage, tumour grade, molecular subtype, type of surgery/axillary surgery, and the receipt of chemotherapy or endocrine therapy), we have evaluated the majority of them. Only the presence of specific genes was not included in the statistical analysis. We might, therefore, consider genetic variables as potential confounders which could be evaluated in future studies.

 

Results section:

In spite that the authors in Table 1 present the status of the menopause (pre-or-postmenopausal), the all general characteristics must present divided in premenopausal or postmenopausal status. The latter is since is not adequate mix both group as that different the effect of the obesity in the breast cancer (as the authors refer in their introduction).

We guess that the observation is that a separate analysis should performed for pre and post-menopausal patients. Although we do agree that the effect of obesity might vary according to the menopausal status, we must consider that menopause is induced by chemotherapy and also by the use of LHRH analogues, we therefore emphasize that our results apply mainly to post-menopausal women. This has been better highlighted in the text.

Why only document the impact of BMI on 2-years disease-free survival (2yDFS)? Why no 3 years?

The choice to use a DFS interval of 2 years was derived from the fact that, generally, patients with triple-negative carcinomas have a 5-year prognosis that is more unfavourable as compared to luminal-like cancers. Indeed, it has been shown that after adjuvant/neo-adjuvant chemotherapy, the 5-year overall survival (OS) for TNBC turns is around 64% compared to other molecular subtypes (85%) (Saleh RR et al, 2021, Ref# 51). In addition, the five-year disease-free survival (5yDFS) for TNBC is 57% compared to 72% (Saleh RR et al, 2021, Ref# 51). Due to these evidences, we thought that a longer evaluation might have been masked by the cancer aggressiveness in such population. Furthermore, the results of an exploratory analysis from the ATAC Trial (Sestak I et al, 2010, Ref#24) demonstrated that the best period to evaluate the disease recurrence is at 2 years.

If is important the presence of obesity in the poorest prognosis in terms of DFS, the authors only documented the BMI at the moment of the diagnosis of breast cancer? Or the BMI presented in which momento was documented? Do the authors take into account the changes in BMI during the two years of follow-up?

We thank the Reviewer for her/his comment. Indeed, the BMI was recorded at time of first visit and monitored during the 2-years follow-up and up to the fifth. The registered modification were slight and not such as to assign the patients to a different BMI group. These data have been added to the text.

 

Discussion section:

What were the limitation of their study?

Limitations of the study have been added to the text.

Conclusions:

Regarding “With a view to developing predictive models of risk of relapse, it could be useful to add BMI, with a cut-off of 29 kg/m2, among the criteria to be taken into consideration”

They must be cautious in their conclusions, since the authors do not take into account the change in BMI throughout the follow-up, which is very important to be sure that their cut-off point finding is realistic.

We strongly agree with the Reviewer and added this consideration among the conclusions.

Author Response File: Author Response.pdf

Reviewer 2 Report

Listed below are minor edits to the manuscript:

1.     Please follow a consistent pattern when referring to 2-year Disease Free Survival, either choosing 2yDFS or 2-year DFS 

2.     Greater than or lesser than can be depicted in a better way than < vs>  used in the manuscript.

3.     Full form of BMI (Body Mass Index) can be included on the line 15 in its first instance and can be referred as BMI later in the article.

4.     Line 191 : Discussion can start from next page making the section more clear.

Figures:

 

Figure2: Labels on the axis and the overall graph is unclear, resolution and font size can be increased. 

Figure3: Labels on the axis is unclear, resolution and font size can be increased, more information about the graph can be added 

Figure5: The title of the graph can be better along with resolution and font sizes of labels and titles.

Author Response

Reviewer #2

Listed below are minor edits to the manuscript:

1. Please follow a consistent pattern when referring to 2-year Disease Free Survival, either choosing 2yDFS or 2-year DFS

Emended as requested.

2. Greater than or lesser than can be depicted in a better way than < vs> used in the manuscript.

Emended as requested.

3. Full form of BMI (Body Mass Index) can be included on the line 15 in its first instance and can be referred as BMI later in the article.

Emended as requested.

4. Line 191 : Discussion can start from next page making the section more clear.

Unfortunately the Journal format does not allow to insert page breaks within the manuscript.

Figures:

Figure2: Labels on the axis and the overall graph is unclear, resolution and font size can be increased.

Emended as requested

Figure3: Labels on the axis is unclear, resolution and font size can be increased, more information about the graph can be added

Emended as requested

Figure5: The title of the graph can be better along with resolution and font sizes of labels and titles.

Emended as requested

Author Response File: Author Response.pdf

Reviewer 3 Report

Riondino et al highlight a long-known association between BCa patients treated with AI and obesity.  Their result suggests that patients with BMI>29 are at increased risk. 

I have a couple of concerns that should be addressed via data or commentary in the discussion.  First, the use of CRC as a negative control does not accurately reflect the molecular hallmarks driving each cancer.  A better control would be PCa as both are hormone driven adenocarcinomas. 

Second, the authors should discuss the disparity between BMI and A1C levels, especially since white and Asian women have different risk profiles.  As noted in the text, increased level of adipose tissue contributes to an elevation in estrogen production, thus leading to significant levels of circulating estrogen. Since many of these obese patients have a higher level of IGF and raised insulin resistance by A1C, this could activate the tumor cell survival pathways.  Numerous studies have highlighted the association between patients having high insulin levels have with higher breast malignancy incidence and, importantly, mortality

Also, the need for further studies with comparable size subgroups and larger cohorts and uniform neoadjuvant regimes would be advantages to this study. 

The graphics in Figure 2 (bottom) were very hard to read and the data in Figure 3 should be presented with better legend to describe the curve(s).  

Author Response

Reviewer #3

Riondino et al highlight a long-known association between BCa patients treated with AI and obesity.  Their result suggests that patients with BMI>29 are at increased risk.

I have a couple of concerns that should be addressed via data or commentary in the discussion.  First, the use of CRC as a negative control does not accurately reflect the molecular hallmarks driving each cancer.  A better control would be PCa as both are hormone driven adenocarcinomas.

We thank the Reviewer for her/his question that might give us the opportunity of better explain the reason to include adjuvant-treated CRC patients as comparator group. Indeed, as the Reviewer correctly pointed out, prostate cancer would have been the best comparator as both are hormone driven adenocarcinomas. However, our aim was to assess obesity interference with aromatase treatment efficacy in a woman population. Since obesity is considered a risk factor for colorectal cancer, we chose CRC women who are at higher risk due to their overweight/obese status, but who are not under aromatase inhibition and not strongly influenced by hormones in the pathogenesis.

Second, the authors should discuss the disparity between BMI and A1C levels, especially since white and Asian women have different risk profiles.  As noted in the text, increased level of adipose tissue contributes to an elevation in estrogen production, thus leading to significant levels of circulating estrogen. Since many of these obese patients have a higher level of IGF and raised insulin resistance by A1C, this could activate the tumour cell survival pathways.  Numerous studies have highlighted the association between patients having high insulin levels have with higher breast malignancy incidence and, importantly, mortality

We strongly agree with the Reviewer’s comment. Indeed, we added this concern among the study limitations. However, the mean glycaemic level of the obese BC women was 119 mg/dL and only for 64 patients HA1C values were available (mean 5.7%), with pathological values (mean 6.2%) only in 42 patients. When considering the two subgroups, with BMI greater or lesser than 29, we had a mean HA1c of 6% in 62 patients with BMI greater than 29, compared to 5.6% in patients with BMI lesser than 29.

Also, the need for further studies with comparable size subgroups and larger cohorts and uniform neoadjuvant regimes would be advantages to this study.

We thank the Reviewer for her/his interesting suggestion. The group of women under adjuvant treatment were chosen based on criteria depicted in Fig 1. Neoadjuvant treatment was allowed but many neoadjuvant-treated patients were not included because they did not meet the criteria. This will surely represent the topic of a further study.

The graphics in Figure 2 (bottom) were very hard to read and the data in Figure 3 should be presented with better legend to describe the curve(s). 

Emended as requested

Author Response File: Author Response.pdf