Long-Term Results of Postoperative Hypofractionated Accelerated Breast and Lymph Node Radiotherapy (HypoAR) with Hypofractionated Boost

Round 1

Reviewer 1 Report

Dear authors,
the presented study is an important contribution towards shorter, cheaper, and more convenient RT after breast cancer. The presented data are interesting, the applied statistical methods sound. However, I could not verify some cited values and computations in the current manuscript. As these are essential for the conclusions, these issues need to be clarified:

- secondary cancer risks are compared to Kirova et al. [25]. It is claimed secondary lung cancer + sarcoma risks in the present study (0.3%) to be much smaller than risks observed in [25]. However, I could not verify the cited "4.4%" sarcoma and lung cancer risk.  (just didn't find this number in [25]). Instead, about 0.6% of breast cancer patients suffered from sarcoma and lung cancer in [25]. How did this different number come about? Moreover, it should be noted that [25] includes treatments in the 80s, and also general population cancer rates were quite different back then. Therefore, while a risk of 0.3% within about a decade is certainly acceptable, a better comparison group would be necessary to judge whether this rate is above or below cancer rates expected from normal fractionation. Regarding lung cancer this is only possible if smoking prevalence is known. Moreover, not lung (1 case) but contralateral breast cancer (8 cases) was the most frequent second carcinoma in the present study. The most frequent endpoint may not be omitted when comparing secondary cancer risks.

- It is concluded an EQD2 of 50-52 Gy to be optimal. However, it is unclear how the authors calculate the EQD2 as some treatment time correction is implemented and this correction seems not to be consistent. In Table 2, the tumor quadrant has shorter treatment time (17 days) than the breast (22 days). How is this possible? The treatment time corrected EQD2 is also calculated for other trials. In the FAST trial [16] five fractions were applied, one per week. (Note that "five-fraction weekly" as stated in the present manuscript is misleading). If I apply the overall treatment time of 5 weeks, I do not obtain an EQD2 of 50-52 Gy. The same is true for the START A trial. Therefore, the authors should clearly substantiate the way they calculated EQD2 and tabulate the relevant numbers from different studies. If it is unclear how to calculate the time correction, it makes no sense to state that its result should be trusted (p11 l306).

Moreover, presentation needs to be improved:
-abstract: Follow-up time is an important information if stating late toxicities, in particular secondary cancer
- Historical context (30 year old!) should be skipped or cutted down in the introduction
- It is not necessary always to state the total number of patients in the text such as in "330/367".
- Table 1: What does PS stand for? Why is "extracapsular invasion" and "unknown(**)" bold? Why do counts in "Node involvement" do not add to 367?
- p4 l110: Ref [10] was not the first proposing the time correction.
- p4 l111: alpha/beta in the formula, not a/b
- Table 2: Why does the tumor quadrant obtain only 13 fractions. What's about the 3 remaining fractions to the breast?
Several blanks in the caption appear to be wrong.
- Tables 3,4: First lines could be skipped. "NoNone" in Tab 3, under "Pain"
- Fig. 1b: Different lines cannot be distinguished. Does p=0.01 refer to B vs other, or, instead, to C vs other?
- p7 l187, l189: Figs 1b, 1c are exchanged.
- p8 l208: "luminal type A were sign. ..better survival": But in Table 5, luminal types A,B are related with large HR? The same for l214.
- p8 l213: Why are Nx patients excluded here but not in Fig 1?
- p8 l213: "disease specific" but Fig2 shows "Overall survival"?
- Fig 2, caption: What is "disease specific overall survival"? Why is it abbreviated "LRFS"?

Author Response

REVIEWER 1

 

Comment 1.-

 secondary cancer risks are compared to Kirova et al. [25]. It is claimed secondary lung cancer + sarcoma risks in the present study (0.3%) to be much smaller than risks observed in [25]. However, I could not verify the cited "4.4%" sarcoma and lung cancer risk.  (just didn't find this number in [25]). Instead, about 0.6% of breast cancer patients suffered from sarcoma and lung cancer in [25]. How did this different number come about? Moreover, it should be noted that [25] includes treatments in the 80s, and also general population cancer rates were quite different back then. Therefore, while a risk of 0.3% within about a decade is certainly acceptable, a better comparison group would be necessary to judge whether this rate is above or below cancer rates expected from normal fractionation. Regarding lung cancer this is only possible if smoking prevalence is known. Moreover, not lung (1 case) but contralateral breast cancer (8 cases) was the most frequent second carcinoma in the present study. The most frequent endpoint may not be omitted when comparing secondary cancer risks.

Response:  

We now better report and present the analysis by Kirova et al. The percentages reported are now clearly presented. We revised the paragraph as follows: .” Indeed, in an analysis by Kirova et al, radiotherapy for breast cancer has been associated with increased risk for the development of secondary carcinomas, especially of the lung, and sarcomas [25]. Authors analysed 16.705 patients (13.472 treated with postoperative RT vs. 3.233 treated with surgery alone) with a median follow-up of 10.5 years. The incidence of second malignancies was 4.42% (596/13.472) in patients receiving radiotherapy vs. 3.5% (113/3.233) in patients receiving surgery alone [25]. In the Kirova series of patients 35 developed sarcomas, 27 of them considered to be radiation-induced (incidence 0.2%; 7.4-fold increased incidence compared to the non-RT group). The incidence of lung cancer was 0.4% (54/13.472) in the RT-group vs. 0.1% (4/3.233) in the non-RT group. In our study, within a median follow-up of 12 years, the incidence of lung cancer was 0.3%, and we recorded no case of sarcoma. Kirova et al reported that 52/58 women who developed secondary lung cancer had smoking history, but we have no such data available to report for our series. The overall incidence of neoplasia was 1.6%, which is lower than the recorded by Kirova et al., even than the non-RT group of patients”.  Moreover, we comment on the incidence of contralateral breast cancer as follows: “Regarding the risk of controlateral breast cancer, Kirova et al found no increase in the group of patients receiving RT. The risk was 8.2% (1.113/13.472) in the RT group vs. 7.1% (230/3.234) in the non-RT group of patients. In our series the risk of controlateral breast cancer was as low as 2.2%.

 

Comment 2

- It is concluded an EQD2 of 50-52 Gy to be optimal. However, it is unclear how the authors calculate the EQD2 as some treatment time correction is implemented and this correction seems not to be consistent. In Table 2, the tumor quadrant has shorter treatment time (17 days) than the breast (22 days). How is this possible?

Response:

We apologize for the mistake in the description of the Schedule B. The booster dose of 0,8Gy was given for the first 8 fractions (together with 2.7Gy directed to the whole breast) in 10 days (including weekend). Subsequently, RT continued to the whole breast-only for another 8 fractions of 2.7Gy in 12 days. Total treatment time 22 days for both the whole breast and the quadrant. The EQD2n,c and time corrected doses are reported in Table 2. (Τhe initially reported in the table analysis of the 3.5Gy x 13 scheme refers to a radical breast RT regimen for inoperable chemoresistant patients, not included in the current study).

 

Comment 3

The treatment time corrected EQD2 is also calculated for other trials. In the FAST trial [16] five fractions were applied, one per week. (Note that "five-fraction weekly" as stated in the present manuscript is misleading). If I apply the overall treatment time of 5 weeks, I do not obtain an EQD2 of 50-52 Gy. The same is true for the START A trial. Therefore, the authors should clearly substantiate the way they calculated EQD2 and tabulate the relevant numbers from different studies. If it is unclear how to calculate the time correction, it makes no sense to state that its result should be trusted (p11 l306).

 

Response:

We rephrased the sentence regarding the FAST trial in the discussion as follows: “The recently published 10-year follow-up of the ‘FAST trial’ on breast-only irradiation confirmed the safety and efficacy of a close to ultra-hypofractionation, five-fraction (one fraction per week) regimen delivering 5.7Gy/week [16].”

Moreover, we report in details the ΕQD2 and EQD2-T for the Canadian, START A and FAST trials showing that these deliver similar biological doses to our regimens. The paragraph has been rephrased as follows: “The recently published 10-year follow-up of the ‘FAST trial’ on breast-only irradiation confirmed the safety and efficacy of a close to ultra-hypofractionation, five-fraction (one fraction per week) regimen delivering 5.7Gy/week [16]. This regimen delivers 28.5Gy, thus an EQD2 of 46.07Gy in 29 days with minimal 2-day acceleration. The Canadian trial on breast-only irradiation also proved that an accelerated 22-day regimen (42.5Gy in 16 fractions) provides high efficacy, less acute toxicity, and improved quality of life compared to CRT [17]. This regimen delivers an EQD2 of 47.2Gy in 22 days, thus with an acceleration of 10 days, which gives an EQD2-T for the breast (λ=0.2Gy) of 49.2Gy. Similar results in terms of effectiveness and breast toxicity have been reported in the UK START A trial, where breast received 13 fractions of 3Gy [18]. This regimen delivers an EQD2 of 45.5Gy in 17 days. This regimen provides an acceleration of 14 days, so that the ΕQD2-T is estimated to 48.3Gy. All these schedules produce a similar toxicity to conventionally fractionated regimen delivering 50Gy to the breast. Our schedules A and B deliver a similar time-corrected biological dose of 47.35Gy and 50.24Gy, respectively, to the whole breast and as expected late breast toxicity was minimal”.

 

Comment 4

-abstract: Follow-up time is an important information if stating late toxicities, in particular secondary cancer

Response:

We now report the median follow-up (12 years) in the abstract.

 

Comment 5
- Historical context (30 year old!) should be skipped or cutted down in the introduction

Response:

the comments on this important historical paper have been cut down.

 

Comment 5
- It is not necessary always to state the total number of patients in the text such as in "330/367".

Response:

This has been corrected where appropriate.

 

Comment 6
- Table 1: What does PS stand for? Why is "extracapsular invasion" and "unknown(**)" bold? Why do counts in "Node involvement" do not add to 367?

Response:

‘PS’ has been corrected as ‘performance status’. The bold letters do not appear in our version. Moreover, we recalculated the ‘node involvement’ numbers (these are 367).

 

 

Comment 7
- p4 l110: Ref [10] was not the first proposing the time correction.

Response: The original study by Maciejewski has been added (Maciejewski B, Withers HR, Taylor JM, Hliniak A. Dose fractionation and regeneration in radiotherapy for cancer of the oral cavity and oropharynx: tumor dose-response and repopulation. Int J Radiat Oncol Biol Phys. 1989;16:831-43. doi: 10.1016/0360-3016(89)90503-8.)

 

Comment 8
- p4 l111: alpha/beta in the formula, not a/b

Response : Corrected

 

Comment 9
- Table 2: Why does the tumor quadrant obtain only 13 fractions. What's about the 3 remaining fractions to the breast?

Response:

As mentioned above, we apologize for the mistake in the description of schedule B. As mentioned in previous responses to the comments from the reviewer, the regimen included 16 fraction of 2.7Gy to the whole breast, the first 8 of them receiving a 0.8Gy booster dose to the quadrant. The radiobiological analysis of the schedule is reported in Table 2. The schedule was designed to deliver and EQD2-T of 50Gy to the breast and 62Gy to the quadrant.

 

Comment 10
Several blanks in the caption appear to be wrong.
- Tables 3,4: First lines could be skipped. "NoNone" in Tab 3, under "Pain"

Response:

we agree these could be skipped, although it is easier for the readers to see the percentage of the lack of symptomatology.

 

Comment 11
- Fig. 1b: Different lines cannot be distinguished. Does p=0.01 refer to B vs other, or, instead, to C vs other?
- p7 l187, l189: Figs 1b, 1c are exchanged.

Response:

These are now corrected in the text and in the Figure 1.

 

Comment 12
- p8 l208: "luminal type A were sign. ..better survival": But in Table 5, luminal types A,B are related with large HR? The same for l214.

Response:

The comparison is ‘other vs. Luminal A’. This has been now corrected in the table 5

 

Comment 13
- p8 l213: Why are Nx patients excluded here but not in Fig 1?

Response: We corrected the figure to include the Nx patients. These had a worse OS compared to N0 patients but the difference was not significant compared to other subgroups.

 

 

Comment 14
- p8 l213: "disease specific" but Fig2 shows "Overall survival"?

Response: the ‘overall survival’ curves are disease specific as mentioned in the legend of Figure 2

 

Comment 15
- Fig 2, caption: What is "disease specific overall survival"? Why is it abbreviated "LRFS"?

Response:

we deleted LRFS – These are disease specific overall survival curves (deaths from breast cancer taken into account only).

 

 

 

 

 

Author Response File: Author Response.doc

Reviewer 2 Report

This is an interesting paper describing hypofractionated RT to breast +/- nodes. 

1) I think the introduction could be modified slightly to reflect current evidence. The introduction establishes 50Gy in 25 fractions as the only standard treatment schedule for adjuvant breast cancers; however data (including the Canadian data cited later) has established 42.5Gy in 16 fractions as a safe and recommended alternative. Therefore the more interesting arm of this study is the 35Gy in 10 fraction arm, and the introduction might be more interesting if framed this way. 

2) modern data does not necessarily support a survival advantage to adjuvant breast RT in low risk patients, and I think this statement could be modified to be less concrete. For example, "may have a survival advantage."

3) It was unclear to me how patients were assigned to Scheme A or Scheme B. Was this based on phsyician preference? Some underlying different in the patients in each group? Table 1 could be broken down into Scheme A and Scheme B, with p values to show whether there are significant differences between the groups. 

4) In many places it the results, patients in Scheme A and Scheme B are analyzed together. These are 2 very different doses and should be analyzed separately.

5) A subset analysis of patients treated with breast only vs breast and nodal RT in each scheme would also be interesting. 

6) The rationale for boosting is divergent from that used in many institutions where only patients with young age or close/positive margin are prescibed a boost. Your recommendation that boost be used in a wider population seems tangential to this work and could be omitted. 

7) The conclusion does not reflect the data presented in the study.

 

Author Response

REVIEWER 2

 

Comment 1

I think the introduction could be modified slightly to reflect current evidence. The introduction establishes 50Gy in 25 fractions as the only standard treatment schedule for adjuvant breast cancers; however data (including the Canadian data cited later) has established 42.5Gy in 16 fractions as a safe and recommended alternative. Therefore the more interesting arm of this study is the 35Gy in 10 fraction arm, and the introduction might be more interesting if framed this way. 

Response:

The Canadian trial by Whelan et al has been added in the introduction, as a standard regimen used worldwide.

 

Comment 2

modern data does not necessarily support a survival advantage to adjuvant breast RT in low risk patients, and I think this statement could be modified to be less concrete. For example, "may have a survival advantage."

Response:

we corrected the statement in the introduction as follows: “Radiotherapy minimizes the risk of locoregional recurrence and decreases death rates in high risk patients [2]”.

 

 

Comment 3

It was unclear to me how patients were assigned to Scheme A or Scheme B. Was this based on phsyician preference? Some underlying different in the patients in each group? Table 1 could be broken down into Scheme A and Scheme B, with p values to show whether there are significant differences between the groups.

                Response:

Schedule B has been formulated to simulate the Canadian schedule with concomitant booster dose to the affected quadrant. The choice between schedule A and B was at the discretion of the physician. Schedule B was more cumbersome and it was soon abandoned.  These are now reported in the methods. As suggested by the reviewer, Table 1 is now broken down to report separately the distribution of patient parameters in schedule A vs. schedule B and p-values are reported. The only statistical difference was the treatment of Tis cancer patients more frequently with schedule B, and triple negative patients with schedule A.

 

4) In many places it the results, patients in Scheme A and Scheme B are analyzed together. These are 2 very different doses and should be analyzed separately.

Response:

Table 3 and 4 now report separately the toxicities. There were no significant differences. A trend for increased incidence of early dermatitis was noted in schedule B.

 

5) A subset analysis of patients treated with breast only vs breast and nodal RT in each scheme would also be interesting. 

Response:

As breast toxicities are related to breast irradiation and arm toxicities mainly to nodal irradiation this analysis is not reported.

 

Comment 6

The rationale for boosting is divergent from that used in many institutions where only patients with young age or close/positive margin are prescibed a boost. Your recommendation that boost be used in a wider population seems tangential to this work and could be omitted.

Response:

We rephrased the conclusion regarding the booster dose  as follows: “Although the use of booster radiation to the affected breast quadrant is not a standard practice in many institutions, the current study provides evidence that this can be safely applied in subgroups of patients with high risk, but its therapeutic value needs evaluation in randomized trials”. 

 

Comment 7

The conclusion does not reflect the data presented in the study.

Response:

the sentence with the suggestion that is not reflected by the data has been omitted from the conclusion.

 

 

Author Response File: Author Response.doc