Timing of radiotherapy and chemotherapy after breast-conserving surgery for node-positive breast cancer: long term results from IBCSG Trials VI and VII

Abstract

Purpose

The aims of the present study were to update the previous report from two randomized clinical trials, now with median follow-up of 16 years, to analyze the effect of radiotherapy timing on local failure and disease-free survival.

Methods and Materials

From July 1986 to April 1993 the International Breast Cancer Study Group (IBCSG) Trial VI randomly assigned 1475 pre/perimenopausal women with node-positive breast cancer to receive 3 or 6 cycles of initial chemotherapy (CT). IBCSG Trial VII randomly assigned 1212 postmenopausal women with node-positive breast cancer to receive tamoxifen for 5 years, or tamoxifen for 5 years with three early cycles of initial CT. For patients who received breast-conserving surgery (BCS), radiotherapy (RT) was delayed until initial CT was completed; 4 or 7 months after BCS for Trial VI and 2 or 4 months for Trial VII. We compared RT-timing groups among 433 patients on Trial VI and 285 patients on Trial VII who received BCS plus RT. Endpoints were local failure, regional/distant failure, and disease-free survival (DFS).

Results

Among pre/perimenopausal patients there were no significant differences in disease-related outcomes. 15-year DFS was 48.2% in the group allocated 3 months initial CT and 44.9% in the group allocated 6 months initial CT (HR=1.12; 95% CI:0.87–1.45). Among postmenopausal patients, 15-year DFS was 46.1% in the no-initial-CT group and 43.3% in the group allocated 3 months initial CT (HR=1.11; 95% CI:0.82–1.51). Corresponding HRs for local failures were 0.94 (95% CI: 0.61–1.46) in Trial VI and 1.51 (95% CI:0.77–2.97) in Trial VII. For regional/distant failures, the respective HRs were 1.15 (95% CI:0.80–1.63) and 1.08 (95% CI:0.69–1.68).

Conclusions

This study confirms that, after more than 15 years of follow-up, it is reasonable to delay radiotherapy until after the completion of standard chemotherapy.

Summary

This study looks at the 16-year results of two clinical trials that required radiotherapy after surgery and chemotherapy for patients who received breast-conserving surgery. Patients received radiotherapy after initial chemotherapy within 2 months, 4 months, and 7 months after surgery. Our results confirm no increased risk of local failures in patients delaying their radiotherapy until after systemic treatment.

Introduction

Postoperative radiotherapy to the breast after breast-conserving surgery increases local control rates and long term survival (1). The optimal timing of this common radiotherapy is still not known. To start postoperative adjuvant treatment as soon as reasonably possible seems logical since the aim of the adjuvant treatment is to eradicate remaining clonogenic cells after surgery, and the probability of success would theoretically be better if the time for these cells to proliferate after surgery were shorter (2). However, the risk of waiting before initiating adjuvant radiotherapy differs among cancers and cancer subtypes (3).

Several retrospective breast cancer studies on the effect of delay of radiotherapy have yielded variable results (4–8). Often the time to start irradiation is not independent of the primary tumor characteristics, making interpretations difficult. Generally an increased risk for local failure with longer delay between surgery and radiotherapy was seen in univariate analysis, but this effect was not observed after adjusting for prognostic factors in multivariate analysis. In a report from the International Breast Cancer Study Group (IBCSG), no increased risk was observed with a delay of up to 20 weeks for patients with hormone receptor-positive breast cancer receiving endocrine therapy (9).

Three review articles indicate an increased risk for local failures with a longer delay than 8 weeks between surgery and radiotherapy (3, 10, 11). These reviews, however, include a variety of different study designs and various breast cancer subtypes.

A previous IBCSG publication reported the effect of delaying radiotherapy in patients randomized to different chemotherapy regimens (12). In this study, no significant difference was observed in 4-year local failure rate among pre/peri-menopausal patients starting radiotherapy after 3 or 6 courses of chemotherapy or among postmenopausal patients starting radiotherapy within 3 months after surgery or after 3 courses of chemotherapy. An advantage with this study was that the delay of radiotherapy was predefined by the study protocol and independent of primary tumor characteristics. However, the median follow-up was only 4 years and the power too low to exclude clinically relevant differences in outcome.

The aims of the present study were to update the previous report (12) with a median follow-up of 16 years, to analyze the effect of radiotherapy timing on local failure rates and disease-free survival, and to further study possible modifying effects of estrogen-receptor status, nodal status and age.

Patients and Methods

The clinical trial designs and treatment details have been previously published (12,13,14). In brief, patients were recruited to IBCSG Trial VI (pre/peri-menopausal) and IBCSG Trial VII (postmenopausal) from July 1986 to April 1993 and were randomized to the treatment groups shown in Table 1. A total of 2687 patients meeting eligibility criteria were enrolled in the two trials (1475 in Trial VI and 1212 in Trial VII). Of these, 718 (27%) elected to undergo breast-conserving surgery and radiation therapy subsequent to chemotherapy (as required by protocol in both trials). Thus, the differing chemotherapy schedules (Table 1) resulted in four radiotherapy-timing groups: pre/peri-menopausal patients assigned to receive radiation therapy beginning 4 months from surgery (3 months initial CT then radiotherapy, CT×3→RT); pre/peri-menopausal patients assigned to receive radiation therapy beginning 7 months from surgery (6 months initial CT then radiotherapy, CT×6→RT); postmenopausal patients assigned to receive radiation therapy within 2 months after surgery (no initial CT, radiotherapy after surgery); and postmenopausal patients assigned to receive radiation therapy at 4 months (3 months initial CT then radiotherapy, CT×3→RT). The protocols were approved by the relevant ethics committees as required by the governing bodies in each country at the time.

Table 1.

Treatment Groups and Patient Characteristics in Trials VI and VII

	Pre/perimenopausal (Trial VI)		Postmenopausal (Trial VII)
	CMF×3	CMF×6	No Initial CMF	CMF × 3
Randomized	725	750	614	598
Treatment Assignment
CT×3→RT	360
CT×3→RT→CT×1 at mos 6, 9, 12	364
CT×6→RT		375
CT×6→RT→CT×1 at months 9, 12, 15		375
Tamoxifen + RT			306
Tamoxifen + RT→CT×1 mos 9, 12, 15			308
Tamoxifen+CT×3→RT				302
Tamoxifen+CT×3→RT→CT×1 mos 9, 12, 15				296
Breast-conserving surgery+RT (analytic cohort)	213	220	146	139
Age median (range)	44 (27–56)	44 (26–56)	60 (45–76)	60 (46–76)
Age <40 (%)	51 (24)	50 (23)
Age ≥40 (%)	162 (76)	170 (77)
Age <65 (%)			112 (77)	108 (78)
Age ≥65 (%)			34 (23)	31 (22)
ER Negative* (%)	54 (25)	64 (29)	30 (21)	22 (16)
ER Positive (%)	159 (75)	156 (71)	116 (79)	117 (84)
1–3 Nodes positive (%)	165 (77)	164 (77)	111 (76)	98 (71)
≥4 Nodes positive (%)	48 (23)	56 (25)	35 (24)	41 (29)

^*<10 fmol/mg cytosol protein

According to the protocol, radiotherapy was to be given by two opposing tangential fields encompassing the entire breast parenchyma. None of the regional lymph node areas were specifically included in the target volume. However, no special attempt was made to exclude part of the internal mammary or lower axillary lymph nodes if these interfered with the irradiation of the breast. A planned dose of 50 Gy, in a specification point in the middle of the breast along the central axis of the two opposing tangential fields, was prescribed. The radiotherapy was given with daily 2 Gy fractions 5 times a week. A boost was not prescribed according to the protocol, but was left to discretion of each radiotherapist.

We compared outcomes and patterns of failure for the radiation-therapy groups based on intention to treat. Outcomes were disease-free survival (DFS) and time to recurrence. Disease-free survival was defined as the time from randomization until the first occurrence of invasive breast cancer in any site (including ipsilateral breast, contralateral breast, regional, distant), second (non-breast) primary tumor, or death from any cause. Time to recurrence was defined as the time from randomization to diagnosis of recurrent breast cancer or, if confirmed later, when it was first suspected. Site of recurrence was defined as local, regional, or distant. Recurrence sites were considered to be simultaneous if they were diagnosed within 2 months of one another.

We also investigated the timing of RT in an as-treated analysis. This analysis provided results very similar to the intention-to-treat analysis (data not shown) and did not alter our conclusions in any way.

We analyzed DFS using the Kaplan-Meier product limit method and the log-rank test to compare the radiotherapy groups. We assessed time to recurrence according to type of recurrence using methods for competing risks analysis (15). We used proportional-hazards regression and competing-risks regression to evaluate the impact of covariates on outcome (16). Covariates were estrogen-receptor status (defined as negative for <10 fmol/mg cytosol protein, otherwise positive), number of positive lymph nodes involved (1–3 or ≥4), and age (grouped as <40 vs. ≥40 among pre/peri-menopausal patients and <65 vs. ≥65 among postmenopausal patients). A two-sided p-value less than 0.05 was considered statistically significant. Statistical analysis was performed using SAS version 9.3 and R version 3.0.0.

Results

The characteristics of the 718 patients included in this analysis have been published previously (12). In brief, among the 433 pre/peri-menopausal patients, 213 were in the CT×3→RT group and 220 were in the CT×6→RT group; the median age was 44 years; 73% had estrogen receptor-positive disease; 24% had ≥4 positive lymph nodes; and 96% received radiotherapy. Among the 285 postmenopausal patients, 146 were allocated no initial CT and 139 were in the CT×3→RT group; the median age was 60 years; 82% had estrogen receptor-positive disease; 27% had ≥4 positive lymph nodes; and 95% received radiotherapy.

The median follow-up for the entire cohort was 16 years (range, 0.3–21.3). Table 2 shows the numbers of events observed. A total of 398 patients had a DFS event. There were a total of 114 local failures (with or without other type of failure) and 188 distant failures (with or without regional failure). Figure 1 illustrates the cumulative incidence of local failures (with or without other events). Among pre/peri-menopausal patients, the 15-year cumulative incidence was 18.8±2.7% in the CT×3→RT group and 18.5±2.7% in the CT×6→RT group (HR = 0.94; 95% CI: 0.61–1.46). Among postmenopausal patients, 15-year cumulative incidence was 9.0±2.4% in the no-initial-CT group and 14.3±3.1% in the CT×3→RT group (HR = 1.51; 95% CI: 0.77–2.97). Figure 2 illustrates the cumulative incidence of regional/distant (regional or distant or both) failures. Among pre/peri-menopausal patients, the 15-year cumulative incidence was 26.4±3.1% in the CT×3→RT group and 29.7±3.1% in the CT×6→RT group (HR = 1.15; 95% CI: 0.80–1.63). Among postmenopausal patients, 15-year cumulative incidence was 24.7±3.6% in the no-initial-CT group and 27.3±3.9% in the CT×3→RT group (HR = 1.08; 95% CI: 0.69–1.68). Figure 3 shows no significant differences in DFS when comparing the radiotherapy groups. Among pre/peri-menopausal patients, 15-year DFS was 48.2±3.5% (±standard error) in the CT×3→RT group and 44.9±3.5% in the CT×6→RT group (hazard ratio [HR] = 1.12; 95% confidence interval [CI]: 0.87–1.45). Among postmenopausal patients, 15-year DFS was 46.1±4.3% in the no-initial-CT group and 43.3±4.4% in the CT×3→RT group (HR = 1.11; 95% CI: 0.82–1.51).

Table 2.

Numbers of Events Observed

	Pre/perimenopausal		Postmenopausal
	CT×3→RT	CT×6→RT	No initial CT, RT after surgery	CT×3→RT	Total
Number of patients	213	220	146	139	718
Number of DFS events	112	122	80	84	398
Local only	13	9	1	5	28
Local ± other events	41	39	14	20	114
Regional only	3	4	2	1	10
Distant ± regional	54	62	35	37	188
Distant or regional	57	66	37	38	198
Contralateral breast	7	4	3	3	17
Second malignancy, non-breast	7	11	11	12	41
Death without recurrence	0	2	15	11	28

Abbreviations: CT, chemotherapy; CT×3, 3 months initial chemotherapy; CT×6, 6 months initial chemotherapy; DFS, disease-free survival; RT radiotherapy

Figure 1.

Cumulative Incidence of Local Failures for (A) Pre/peri-menopausal Patients and (B) Postmenopausal Patients

Figure 2.

Cumulative Incidence of Regional/Distant Failures for (A) Pre/peri-menopausal Patients and (B) Postmenopausal Patients

Figure 3.

Disease-Free Survival for (A) Pre/peri-menopausal Patients and (B) Postmenopausal Patients

We performed multivariable analyses of DFS, local failure and regional/distant failure (Table 3). After adjustment for nodal status, estrogen-receptor status and age, radiotherapy timing was not significantly associated with outcome. We also evaluated interaction effects in the multivariable models. In all cases, the radiotherapy-timing effect was not significantly modified by nodal status, estrogen-receptor status, or age. For DFS in the pre/peri-menopausal group, the interaction p-values were 0.32, 0.89 and 0.56 for the modifying effect of nodal status, estrogen-receptor status, and age, respectively. For DFS in the postmenopausal group, the respective interaction p-values were 0.12, 0.15 and 0.96.

Table 3.

Multivariable Analysis of DFS, Local Failure, and Regional or Distant Failure

	Local ± Other		Regional or Distant Failure		Disease-Free Survival
Factor	Hazard Ratio (95% CI)	P-value	Hazard Ratio (95% CI)	P-value	Hazard Ratio (95% CI)	P-value
Pre/peri-menopausal patients
Treatment
CT×3→RT	1.00		1.00		1.00
CT×6→RT	0.95 (0.61–1.48)	0.82	1.11 (0.78–1.59)	0.57	1.09 (0.84–1.41)	0.50
Nodal status
1–3 positive nodes	1.00		1.00		1.00
≥4 positive nodes	1.09 (0.65–1.82)	0.74	1.96 (1.34–2.87)	0.0005	1.74 (1.31–2.30)	0.0001
ER status
Negative	1.00		1.00		1.00
Positive	0.80 (0.50–1.28)	0.35	0.96 (0.61–1.48)	0.84	0.87 (0.65–1.16)	0.34
Age
< 40 years	1.00		1.00		1.00
≥ 40 years	0.50 (0.32–0.80)	0.003	0.76 (0.50–1.16)	0.21	0.63 (0.47–0.84)	0.002
Postmenopausal patients
Treatment
No initial CT	1.00		1.00		1.00
CT×3→RT	1.53 (0.77–3.01)	0.22	1.04 (0.66–1.64)	0.87	1.07 (0.79–1.46)	0.65
Nodal status
1–3 positive nodes	1.00		1.00		1.00
≥4 positive nodes	1.35 (0.65–2.78)	0.42	2.05 (1.28–3.27)	0.003	1.83 (1.32–2.54)	0.0003
ER status
Negative	1.00		1.00		1.00
Positive	0.70 (0.31–1.58)	0.39	0.87 (0.49–1.58)	0.66	0.69 (0.48–1.01)	0.06
Age
< 65 years	1.00		1.00		1.00
≥ 65 years	1.49 (0.70–3.15)	0.30	0.92 (0.52–1.64)	0.78	1.42 (1.00–2.03)	0.05

Abbreviations: CT, chemotherapy; ER, estrogen receptor; CT×3, 3 months initial chemotherapy; CT×6, 6 months initial chemotherapy; RT, radiotherapy

Discussion

With a median follow-up of 16 years and with greater statistical power we detected no significantly increased risk of local failures in patients delaying their radiotherapy until after systemic treatment. In multivariable analysis age < 40 years was significantly associated with increased risk for local recurrence and nodal status, age <40 years (premenopausal) and >65 years (postmenopausal) were significantly associated with decreased disease-free survival. However, no modifying effect of nodal status, estrogen receptor status or age on the risk of delaying radiotherapy in the different radiotherapy-timing groups could be observed.

This is in line with a randomized trial in which 244 patients were randomized between 12 weeks of chemotherapy (CAMFP-cyclophosphamide, doxorubicin, methotrexate, fluorouracil, prednisone) before or after radiotherapy (17, 18). With a median follow-up of 135 months 11% of the patients randomized to CT-first versus 10% in patients having radiotherapy first developed a local recurrence. The corresponding rates for distant failures at 10 years were 35% and 36% in the two treatment arms, respectively (17). The authors concluded that the low power of the study could not exclude clinically relevant differences in outcome between the arms.

In a previous up-date of trial VI, disease free survival in patients receiving 3 or 6 courses of CMF were equal. Three courses were shown to be adequate for patients at least 40 years old and with estrogen positive tumors (19).

In the present study the possible preventive effect of longer duration of adjuvant chemotherapy on risk for local recurrence could counteract a possible increased risk of delaying the start of radiotherapy and the nature of this study could not sort out these two opposing effects On the other hand, one can hypothesize that systemic treatment might be reduced by giving radiotherapy earlier, as there was no difference at all between the two treatment arms regarding all endpoints (Figures 1–3). However, for patient care it is important to consider the sum of these effects in an unbiased way. An advantage of this trial was the independence between primary tumor characteristics and delay of starting radiotherapy. The main limitation of the current study is the relatively low total number of patients, so no firm conclusions can be drawn for different subgroups of patients. For instance in the postmenopausal group the 15 year local failure rate was 9.0% versus 14.3% in patients starting radiotherapy within 2 months or after 4 months from surgery (HR = 1.51; 95% CI: 0.77–2.97).

Nevertheless, this study further confirms that it is reasonable to delay radiotherapy until after the completion of standard chemotherapy even if small differences in local failure rates in specific subgroups still cannot be excluded.