Is There a Role for Risk-Reducing Bilateral Breast Surgery in BRCA1/2 Ovarian Cancer Survivors? An Observational Study

Round 1

Reviewer 1 Report

I find that the subject of this trial is highly relevant and of special interest. RRBBS is controversial in BRCA1/2-OC survivors not only due to the high relapse rate of OC, but also due the lack of data suggesting improved OS in patients undergoing RRBBS vs patients choosing for follow up (there is significant decrease of the risk to have an invasive or in situ BC, but not proven OS benefit); literature is controversial and even more when regarding the subgroup of BRCA1/2-OC survivors.

Text is well written and in proficient English language.

The authors consider  the incidence of BC after BRCA1/2-OC and  the role of RRBBS  as the primary outcomes of this trial.  I think that this trial fails to promptly address these two primary points:

• The incidence of BC after BRCA1/2-OC   from a 69 patient retrospective cohort can probably give a hint but still nine patients (14.3%) developing BC is only indicative and could be significantly different in a bigger cohort.  For example in a  recent trial (Safra et al, 2021, DOI: 10.1016/j.ygyno.2021.06.009)  including  502 OC  patients with BRCA1/2 mutation, BC cancer incidence was 6.2% with a median time to BC diagnosis after OC of 3.8 years (vs 7.7 years in the trial under review).
• To be noted that according to the authors  (line 72) “clinical data from six of BRCA1/2-OC women was not sufficiently accurate and updated  to establish if they developed or not BC” limiting further the population to 63 patients   
• Furthermore, and  given that the trial included patients from January 2000 to August 2022,  the follow up varies significantly (from less than one and up to 22 years)  and some patients have a very limited follow up making the interpretation of the results of such a small cohort even more difficult.  
• Additionally, the authors state that  ‘In this cohort of 63 patients, the risk of metachronous BC in the two years after the diagnosis of OC was 3.2% whereas this risk in the five years after OC was 7.9% and in the 10 years after OC was 11.1%” but there is nowhere clearly stated to what absolute number of patients these percentages are referred to.  (I guess 11.1% refers to the total population of 63 patients, if I assume that they have a minimum follow up of 2 years, but what is the referral follow up population for the 3.2% and 7.9% is not clear)
• The trials the authors refer to in the discussion section also present  a lower risk of developing a subsequent BC in OC patients with BRCA-associated what is in accordance with these results but numbers are also limited to mostly small populations :  ‘Vencken PM et al’ included 116 OC patients and BRCA,  ‘Domchek SM et al’ included  164 patients,  and “Gangi et al”  included 135 patients (still al three trials are bigger than the cohort of the trial under review); while  ‘McGee J et al’ included  a much bigger population of 509 BRCA mutation carriers with ovarian cancer and gives already a more robust conclusion: Twenty(3.9%) of the 509 patients developed BC in 10 years (the actuarial risk was 7.8%) and more importantly already suggests  that:  “neither preventive mastectomy nor MRI screening is warranted, except for those who have survived ovarian cancer without recurrence for ten years and for those with early stage ovarian cancer.” (what is also in accordance with the data of  Safra et al)
•  
• The role of RRBBS  (second endpoint) is not possible to be addressed from a trial with this design and such a small population  and the question “should BRCA1/2-OC survivors unergo RRBBS ?” remains without an answer from this 69 patient cohort: Only three patients (4.3%) performed RRBBS at a mean age of 58 years and nine patients (14.3%) developed BC with a mean age of 56.2 years. I don’t think that these numbers can be interpreted in any way  and certainly not to draw conclusions.  Furthermore data are more robust from the bigger trials of Safra at al  and  McGee et al  and already point to some conclusions on the role of RRBBS for this population.  
• In the trial of Safra at al  in the  502 patients with OC and BRCA, no deaths from breast cancer were recorded. Overall survival did not differ statistically between patients with an ovarian cancer diagnosis only and those diagnosed with breast cancer after ovarian cancer. The authors concluded that “The low incidence of breast cancer after ovarian cancer in women carrying BRCA1/2 mutations suggests that routine breast surveillance, rather than risk-reducing surgical interventions, may be sufficient in ovarian cancer survivors.” These conclusions  are also mostly in accordance with the suggestions of the trials the authors refer to in the discussion (especially the bigger trial of Mcgee): RRBBS is probably not needed in this population. This is why I do not see additional  input from the trial under review, regarding their second primary point.

For these reasons I don’t think that this small retrospective cohort analysis adds to our knowledge or answers adequately the two primary endpoints and I  will reject this trial from publication. I  would suggest to the authors to extend their cohort  in an multicenter setting If possible in order to achieve a more interpretable result from a robust population or even attempt an metanalysis including their data and the data  of the relevant trials.

Author Response

Dear Reviewer, 

We are very grateful for your input regarding our manuscript entitled “Is there a role for risk-reducing bilateral breast surgery in BRCA1/2 ovarian cancer survivors?”.

Please see below the specific answers to your comments:

We report the characterization of a cohort of 69 BRCA1/2-OC survivors to address the incidence of BC in this subpopulation. Although we are aware regarding the limitations to draw general conclusions from a cohort of 69 patients, we report a significant mean follow up duration of 7.4 years since OC diagnosis, ranging from 1 to 22 years (data added in the revised version of the manuscript). Indeed, our study collects retrospective data regarding OC diagnosis, but women are invited for prospective follow up after genetic testing. This follow up is maintained until death, lost to follow up or consent withdrawal, whichever occurs first. This is now clearly stated in the manuscript, where we added follow-up data in five of these women with previously missing information. Indeed, we now have complete follow up data from 68 patients. One patient dropped out from the follow up program more than two years before her death, so she is not considered when assessing BC (or other cancer type) risk. In the revised version of the manuscript, we also include additional information regarding FIGO stage, mean follow up period, mean age at death (and mean time to death after ovarian cancer diagnosis), cause of death in BRCA1/2-OC patients with BC.

We also describe our experience with RRBBS in this subpopulation and discuss it considering what is reported in previous studies. Since, the number of BRCA1/2-OC survivors who undergo RRBBS is generally small, we consider that our data adds information on this important issue, since this is a routine question in clinical practice for healthcare professionals that manage an increasing number of identified BRCA1/2 ovarian cancer patients. Our BRCA1/2-OC survivors who undergo RRBBS is also small but more exhaustive data form our population shows that reducing risk mastectomy was more frequent in the context of a BC diagnosis than in the context of reducing bilateral BC risk.

Regarding the discussion of our data in comparison with previously reported papers, it is also important to remark that patterns of BRCA1/2 pathogenic variants in Portugal disclose a higher prevalence of BRCA2 mutations. This finding highly contrasts with other published cohorts and only further follow up will allow us to conclude if OC survival as well as BC incidence are related to this BRCA1/2 pattern. At this time, although our results are generally in accordance with those previously reported, they reinforce the potential conclusion that BC incidence in this population is lower than in non-affected BRCA1/2 carriers and competing risks of OC mortality against the risk of BC that should be individually addressed, particularly in longer disease-free survivors.

We agree that a metanalysis including all data available would be of interest. However, we also remark on possible limitations of such an analysis since different inclusion criteria were defined for different studies. We are committed to further inclusion of OC patients in our long term follow up and will be happy to participate in collaborative studies that could allow clinicians to have evidence-based recommendations concerning BC risk and the role of RRBBS in BRCA1/2-OC survivors.

Please see attached the second version of our manuscript.

Reviewer 2 Report

Oliveira et al aimed to evaluate in a single-center retrospective study the incidence of breast carcinoma in ovarian carcinoma patients with BRCA 1 or 2 germline mutations. Women who had other cancer(s) before the presentation of ovarian carcinoma were excluded.

The research question is relevant to clinical practice, with the study adding data to current literature focusing on BRCA germline mutation carriers who had first presentation of ovarian carcinoma.

Clarifications are needed on several aspects of this manuscript:

(1)Information on duration of follow-up is not clearly defined in this cohort. What were the median and range of follow-up times among the patients studied?

(2) The reported risks of breast carcinoma appear to be based on the total cohort size including a mix of women receiving and not receiving prophylactic bilateral mastectomy. What was the incidence of breast carcinoma among those not having received prophylactic bilateral mastectomy?

(3) The researchers first stated that the cohort included 69 patients (Table 1) but later mentioned that data from six women “was not sufficiently accurate and updated” (line 72) to establish if they developed breast cancer or not. Was the exclusion of the 6 patients due to missing information on breast carcinoma status, or other clinical parameters related to study inclusion criteria?

(4) The data presented in Table 1 is confusing regarding the FIGO stage distribution of ovarian carcinoma. There appears significant missing/unknown data (n=45) and the percentages listed in the table and text also do not correctly correspond to the listed numbers over the reported cohort size of 69 in the table. Clarifications are on those numbers are needed.

(5) All-cause mortality data were reported. What were the cancer-specific mortality, data on recurrence(s) and clinical treatment status in this cohort, which are pertinent to evaluating cancer risks and related mortality in BRCA germline mutation carriers who first presented with ovarian carcinoma?

Author Response

Dear Reviewer, 

We are very grateful for your input regarding our manuscript entitled “Is there a role for risk-reducing bilateral breast surgery in BRCA1/2 ovarian cancer survivors?”.

Please see below the specific answers to your comments:

1. Information on duration of follow-up is not clearly defined in this cohort. What were the median and range of follow-up times among the patients studied?

Information on follow up is now included in the second version of the manuscript. BRCA1/2-OC women were followed-up in the Familial Risk Clinic of IPO Lisboa from January 2000 to August 2022 with a mean follow up duration since OC diagnosis of 7.4 years, ranging from 1 to 22 years. We also reinforce that even data before genetic testing was collected retrospectively, after testing women consented on prospective long-term follow up and this is maintained until consent withdrawal, death or lost to follow up.

2. The reported risks of breast carcinoma appear to be based on the total cohort size including a mix of women receiving and not receiving prophylactic bilateral mastectomy. What was the incidence of breast carcinoma among those not having received prophylactic bilateral mastectomy?

We reported a BC risk of 13.2% in a cohort of 68 BRCA1/2-OC survivors (retrieved data regarding the follow up of five women previously excluded because of missing data is now included). In this cohort, three women who had undergone RRBBS were included. Additionally, five other patients that underwent bilateral mastectomy in a context of unilateral BC were included. Although the inclusion of these women could potentially decrease the overall incidence of BC in our cohort, as these women still have a remaining BC risk and need surveillance, we decided to consider them and add a note in the manuscript regarding this issue.

3. The researchers first stated that the cohort included 69 patients (Table 1) but later mentioned that data from six women “was not sufficiently accurate and updated” (line 72) to establish if they developed breast cancer or not. Was the exclusion of the 6 patients due to missing information on breast carcinoma status, or other clinical parameters related to study inclusion criteria?

Our cohort includes 69 women, from 63 different families, with ovarian cancer and a confirmed BRCA1/2 variant.  Characterization of this cohort is provided. All patients met the described inclusion criteria. In the first version of the manuscript, missing follow up data for the two years before the cutoff date (August 2022) excluded six patients from BC risk assessment. In this second version, as reported to the Editor, we made an effort to contact or access medical records of these patients, gathering follow up data in five of these women. In conclusion, we now have complete follow up data from 68 patients. One patient dropped out from the follow up program more than two years before her death, so she is not considered when assessing BC (or other cancer type) risk in this cohort.

4. The data presented in Table 1 is confusing regarding the FIGO stage distribution of ovarian carcinoma. There appears significant missing/unknown data (n=45) and the percentages listed in the table and text also do not correctly correspond to the listed numbers over the reported cohort size of 69 in the table. Clarifications are on those numbers are needed.

Our cohort comprises 69 BRCA1/2-OC survivors although data on FIGO stage is not accessible for all of them. In the second version of the manuscript, we clearly improved the number of patients with available data regarding FIGO stage. We now have data from 61 patients and all percentages were calculated considering this number of patients of whom data is available. The same principle was applied to data regarding histology of ovarian cancer (n=58).

5. All-cause mortality data were reported. What were the cancer-specific mortality, data on recurrence(s) and clinical treatment status in this cohort, which are pertinent to evaluating cancer risks and related mortality in BRCA germline mutation carriers who first presented with ovarian carcinoma?

We had provided data on all-cause mortality in the first version of this manuscript.  We now include additional information, namely mean age at death and mean time after ovarian cancer diagnosis. Cancer-specific mortality data was not evaluated for the entire cohort. However, cause of death of BRCA1/2-OC women with BC was added in the second version. Three out of the nine BRCA1/2-OC women with BC died at a mean age of 54 years [range: 50-61 years] and at a mean time of 7 years after the OC diagnosis [range: 5-9 years]. Considering data added, we can now state that the cause of death of these three patients was unrelated to BC.

Please see attached the second version of our manuscript.

Round 2

Reviewer 1 Report

I find the answers of the authors regarding my comments on  the missing data,  the follow up period satisfactory. The cohort population and the incidence of BC remains very low for further commentary.

I also find the major review of the disscussion, with al the additional comments on the results of the other trials and the updated literature pointing at the right dirrection. The adjusted focus of the primary points in the BRCA2  subgroup  is at least clever in order to include an innovative point to the existing data.

 I still believe that these data are very limited to add evidence for the BRCA population ) and i would again suggest to consider a metanalysis with your data and those from the existing litarature; still with focus in the BRCA2 subgroup these data can be considered as additional knowllege. On the other hand the higher incidence of BRCA2  (vs BRCA1) in your trial can  also be biassed from the small sample

 I will leave the final decision for publication to the editors

Author Response

We are most grateful for the additional input regarding our manuscript entitled “Is there a role for risk-reducing bilateral breast surgery in BRCA1/2 ovarian cancer survivors?”.

Regarding a potential bias of the mutational BRCA1/2 pattern of our cohort due to low numbers, please allow us to inform of the following:

• in a previous paper, from 2007, even if we exclude the variant c.156_157insAlu (previously described by Teugels et al and found to be a founder variant of Portuguese origin), BRCA2 variants were already diagnosed more frequently in our population than BRCA1 (https://ascopubs.org/doi/full/10.1200/JCO.2006.06.9443)
• in a more recent exploratory study (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8426025/), we disclosed our general molecular data until 2018 and for the whole group of patients BRCA2 variants, even if excluding the c.156_157insAlu are more frequently diagnosed than BRCA1 variants in the Portuguese population.
• specifically regarding ovarian cancer, in 2021 we presented data from our Ovarian BRCA1/2 patients in the ESMO congress (https://oncologypro.esmo.org/meeting-resources/esmo-congress-2021?event_resources_filter_form%5Bareas%5D%5B%5D=924&event_resources_filter_form%5Bformat%5D%5B%5D=ePoster&event_resources_filter_form%5Bsearch%5D=Pedro%20Meireles). In that communication BRCA2 already contributed to more than half of BRCA1/2 OC cases. At that time, although not significant, a trend to a better OS for BRCA2-OC patients was observed.

Please see attached the third version of our manuscript.