- Email: [email protected]
Omission of radiotherapy in elderly patients with early breast cancer: 15-Year results of a prospective non-randomised trial
European Journal of Cancer (2015) xxx, xxx– xxx
Available at www.sciencedirect.com
ScienceDirect journal homepage: www.ejcancer.com
Omission of radiotherapy in elderly patients with early breast cancer: 15-Year results of a prospective non-randomised trial Gabriele Martelli a,⇑, Patrizia Boracchi b, Eleonora Guzzetti a, Giuseppe Marano b, Laura Lozza c, Roberto Agresti a, Cristina Ferraris a, Domenico Piromalli a, Marco Greco a a
Breast Unit, Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy Department of Clinical Science and Community Health, University of Milan, Milan, Italy c Radiotherapy Unit, Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy b
Received 21 February 2015; received in revised form 6 April 2015; accepted 24 April 2015
KEYWORDS Breast cancer Radiotherapy Elderly patients Breast-conserving surgery Quadrantectomy Tamoxifen
Abstract Background: Whether radiotherapy (RT) is beneficial in elderly (P70 years) patients undergoing conservative surgery for early breast cancer has long been controversial. Recent randomised trials show that most elderly patients do not benefit from RT. We started a prospective non-randomised trial to address this issue in 1987 and now present results for the 627 consecutive pT1/2cN0 patients recruited, and treated by conservative surgery (quadrantectomy) and tamoxifen, and assigned non-randomly to RT or no RT. Methods: We used multivariate competing risks models to estimate 15-crude cumulative incidence (CCI) of ipsilateral breast tumour recurrence (IBTR), distant metastasis and breast cancer mortality. The models incorporated a propensity score as a measure of probability of receiving RT based on baseline characteristics, to account for the lack of randomisation. Results: For pT1 patients, 15-year CCIs of IBTR, distant metastasis and breast cancer death were indistinguishable in the RT and no RT groups. For pT2 patients, 15-year CCI of IBTR was much higher in those not given RT (14.6% versus 0.8%, p = 0.004), although breast cancer mortality and distant metastasis did not differ significantly between RT and no RT. Conclusions: Consistent with the findings of recent randomised trials, our long-term data indicate that most elderly, ER-positive patients with pT1 cN0 breast cancer treated by quadrantectomy do not benefit from RT. The 14.6% CCI of IBTR in our pT2 patients is an additional finding not presented in the trials and suggests that RT should be administered to elderly patients with pT2 disease. Ó 2015 Elsevier Ltd. All rights reserved.
⇑ Corresponding author at: Fondazione IRCCS Istituto Nazionale dei Tumori, Via Venezian 1, 20133 Milan, Italy. Fax: +39 02 2367430.
E-mail address: [email protected] (G. Martelli). http://dx.doi.org/10.1016/j.ejca.2015.04.018 0959-8049/Ó 2015 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Martelli G. et al., Omission of radiotherapy in elderly patients with early breast cancer: 15-Year results of a prospective non-randomised trial, Eur J Cancer (2015), http://dx.doi.org/10.1016/j.ejca.2015.04.018
2
G. Martelli et al. / European Journal of Cancer xxx (2015) xxx–xxx
1. Introduction The meta-analysis of 10,801 breast cancer patients from 17 randomised trials published by the Early Breast Cancer Trialists’ Collaborative Group in 2011 [1], showed that radiotherapy (RT) after breast-conserving surgery not only reduces the risk of breast cancer recurrence but also improves overall survival. However, although the incidence of breast cancer increases with advancing age [2] – and around 40% of breast cancers occur in women over 65 years – most of the randomised trials that assessed postoperative breast RT excluded patients over 70 years [3–6] and it was unclear for many years whether RT after breast-conserving surgery is beneficial in elderly patients. An early trial carried out at our Institute [3], which recruited patients with tumour 62.5 cm, found in an ad hoc analysis, that the incidence of cancer recurrence in the no RT arm declined markedly with age, and the authors suggested that RT may not be necessary in older women. By contrast, another early trial, published in 1992 [4], did not identify a subgroup of patients at low risk of local recurrence who could be spared RT; in particular, for women over 50 years with tumour up to 2 cm who did not receive RT, the recurrence rate at 7 years was 22% [5]. In 2013, the long-term results of the CALGB 9343 randomised trial were published [7]. This trial recruited women aged 70 years or older with T1N0 ER-positive breast cancer, treated with lumpectomy and tamoxifen, and randomised to either postoperative RT or no RT. Fewer local relapses occurred in the RT arm (2% versus 9%) without, however, affecting breast cancer mortality, distant disease-free survival, or breast conservation, after a median follow-up of 12.6 years. In 1987 we initiated a prospective non-randomised study to investigate whether breast RT can be safely avoided in elderly patients (P70 years) undergoing conservative surgery (quadrantectomy) and prescribed tamoxifen. Here we present 15-year results of that study. Whether or not RT was given depended on patient preference and the treating surgeon’s opinion. No patients were formally contraindicated for RT. We used a propensity score with multivariable analyses to adjust for the imbalance in baseline characteristics between the two groups.
resection margins, multifocal lesions, synchronous bilateral breast cancer, distant metastasis at diagnosis, previous cancer at another site, or contraindication for RT, were excluded. 2.2. Treatment Two hundred and seven patients received postoperative RT, 420 patients did not. RT was administered to the residual breast using a cobalt unit or 6-MeV photons to deliver 50 Gy (daily target dose 2 Gy) in the two opposing tangential fields. In all cases, a 10 Gy boost was administered to the tumour bed. The irradiation aimed to treat the breast only, with no attempt to include the axillary, supraclavicular or internal mammary nodes in the irradiation field. Axillary dissection was performed in 134/430 (31%) patients with pT1 disease, and 33/197 (17%) patients with pT2 disease. After surgery, regardless of hormone receptor status, all women were prescribed 20 mg tamoxifen daily for at least 3 years. 2.3. Hormone receptor status Hormone receptor status was determined by the dextran-coated charcoal technique [8]. Cancers with oestrogen receptor (ER) concentration above 10 fmol/mg cytosol protein were considered ER positive; those with progesterone receptor (PgR) above 25 fmol/mg were considered PgR positive. Cancers with receptor content below these values were considered receptor negative. 2.4. Follow-up Patients were seen every 6 months for the first 5 years and annually thereafter. Mammography and chest X-ray were performed once a year and bone scan every 2 years. Gynaecological examination with pelvic ultrasound was performed annually since patients were receiving tamoxifen. Disease status or cause of death was ascertained from clinical records, by consulting general practitioners or from mortality databases. Median follow-up was 188.1 months in the no RT group and 208.9 months in the RT group. No patients were lost to follow-up.
2. Patients and methods 2.5. Statistical methods 2.1. Patients Of 1451 consecutive breast cancer patients P70 years presenting at the Milan National Cancer Institute between January 1987 and December 1992, we evaluated 627 consecutive cN0 patients (430 pT1; 197 pT2 [63 cm]) who underwent conservative surgery with or without postoperative RT. Patients with involved
During follow-up possible events were ipsilateral breast tumour recurrence (IBTR), axillary relapse, distant metastasis, contralateral breast cancer, new primary at non-breast site, breast cancer death and death for any other cause. These events compete with each other to occur first, so competing risk methods were used to estimate crude cumulative incidences (CCIs) of events of
Please cite this article in press as: Martelli G. et al., Omission of radiotherapy in elderly patients with early breast cancer: 15-Year results of a prospective non-randomised trial, Eur J Cancer (2015), http://dx.doi.org/10.1016/j.ejca.2015.04.018
G. Martelli et al. / European Journal of Cancer xxx (2015) xxx–xxx
interest [9] (study end-points: IBTR, distant metastasis and breast cancer mortality) in the RT and no RT groups, taking account of the interaction of pT. The Fine and Gray [10] models were also used to estimate adjusted subdistribution hazard ratios (HRs) with 95% confidence intervals (CIs) for events, with P values obtained from the Wald test. The covariates entered in the Fine and Gray models were: RT (1 = yes, 0 = no), pT (1 = pT2, 0 = pT1), the interaction between these two covariates and a propensity score. The propensity score [11] was introduced because RT was not assigned randomly, and covariates likely to have influenced whether RT was given need to be included as confounders in the models. The propensity score is the probability that a patient will receive the treatment (RT) based on baseline characteristics. In the present case we estimated the probability of receiving RT in relation to patient age, receptor status (ER, PGR coded as 1 = positive and 0 = negative), axillary surgery (coded as 1 = yes, 0 = no) and histology (reference category ductal carcinoma). Since treatments and recruitment criteria did not change over the study period, year of diagnosis was not included in the propensity score. Data on other factors possibly influencing whether RT was given were not available. It has been shown that use of a propensity score based on potential confounders is equivalent to including the individual covariates in the model, and has the advantage of limiting the number of covariates included. The results of propensity score model are reported as odds ratios (ORs) with 95% CIs. The analyses were carried out with R software [12] and the cmprsk package [13].
3
3. Results Clinical characteristics of 627 consecutive patients according to whether or not they received RT are shown in Table 1. ER and PgR receptor status was available for 600 patients: 542 were ER positive, 377 were ER and PgR positive and 54 were ER and PgR negative. Invasive ductal carcinoma was the commonest histologic type (65.9%) followed by invasive lobular carcinoma (22.5%). Table 2 shows 15-year CCIs of first events and deaths by treatment group. Eight women experienced IBTR as first event in the pT1/RT group, and 29 in the pT1/no RT group. All patients who developed IBTR were treated by conservative surgery not mastectomy. Of the eight pT1/RT patients who had IBTR, one developed distant metastases and died of breast cancer, five died of unrelated conditions and the remaining two were alive at latest check-up. Of the 29 pT1/no RT patients who had IBTR, four developed distant metastases and died of their disease, 14 died of unrelated conditions (three other malignancies, 11 non-neoplastic conditions) and the remaining 11 were alive at latest check-up. Among pT2 patients, one in the RT group and 11 in the no RT group developed IBTR. The pT2/RT patient was disease-free and alive at latest follow-up; of the 11 pT2/no RT patients, two developed distant metastases and died of their disease, four died of unrelated conditions and the remaining five were alive at latest check-up. Fifteen-year CCIs of IBTR were 8.1% in both pT1/RT and pT1/no RT patients (Fig. 1); and 0.8%
Table 1 Characteristics of 627 consecutive elderly (P70 years) early, cN0 breast cancer patients, according to whether or not they received radiotherapy (RT). No radiotherapy (RT) (N 420)
RT (N 207)
Overall (N 627)
Median age in years (range)
76 (70–91)
75 (70–88)
76 (70–91)
Axillary surgery, N (%) No Yes
333 (79.3%) 87 (20.7%)
127 (61.4%) 80 (38.6%)
460 (73.4%) 167 (26.6%)
Tumour size, N (%) pT1 pT2
344 (81.9%) 76 (18.1%)
86 (41.5%) 121 (58.5%)
430 (68.6%) 197 (31.4%)
Oestrogen receptor (ER), N (%) ER-positive ER-negative Not assessed
373 (88.8%) 31 (7.4%) 16 (3.8%)
169 (81.6%) 27 (13.0%) 11 (5.3%)
542 (86.4%) 58 (9.3%) 27 (4.3%)
Progesterone receptor (PGR), N (%) PgR-positive PgR-negative Not assessed
265 (63.1%) 139 (33.1%) 16 (3.8%)
116 (56.0%) 80 (38.6%) 11 (5.3%)
381 (60.8%) 219 (34.9%) 27 (4.3%)
Histology, N (%) Invasive ductal carcinoma Invasive lobular carcinoma Other invasive carcinoma
276 (65.7%) 90 (21.4%) 54 (12.9%)
137 (66.2%) 51 (24.6%) 19 (9.2%)
413 (65.9%) 141 (22.5%) 73 (11.6%)
Please cite this article in press as: Martelli G. et al., Omission of radiotherapy in elderly patients with early breast cancer: 15-Year results of a prospective non-randomised trial, Eur J Cancer (2015), http://dx.doi.org/10.1016/j.ejca.2015.04.018
4
G. Martelli et al. / European Journal of Cancer xxx (2015) xxx–xxx
Table 2 Numbers of adverse events and estimates of crude cumulative incidences (CCIs) at 15 years and 95% confidence intervals (CIs) in 627 consecutive elderly (P70 years) early cN0 breast cancer patients, according to whether or not radiotherapy (RT) was given.
First event Ipsilateral breast tumour recurrence (IBTR) pT1 pT2 Axillary disease Distant metastasis pT1 pT2 Contralateral breast cancer Other primary cancer Death Breast cancer pT1 pT2 Other cancer Other cause Total deaths
No radiotherapy (RT)
RT
N events
N events
Crude cumulative incidence (CCI) (95% confidence interval (CI))
Overall CCI (95% CI)
N events
CCI (95% CI)
40 29 11 14 37 24 13 6 23
9.2 (6.3, 12.1) 8.1 (5.1, 11.0) 14.6 (6.5, 22.7) 3.1 (1.4, 4.8) 8.9 (6.1, 11.6) 7.0 (4.3, 9.7) 17.2 (8.6, 25.7) 1.4 (0.3, 2.6) 5.7 (3.4, 7.9)
9 8 1 11 27 8 19 6 8
3.9 (1.2, 6.5) 8.1 (2.3, 14.0) 0.8 (0.0, 2.5) 5.3 (2.3, 8.4) 13.1 (8.5, 17.7) 9.3 (3.1, 15.5) 15.7 (9.2, 22.3) 2.9 (0.6, 5.1) 3.9 (1.2, 6.5)
49 37 12 25 64 32 32 12 31
7.4 (5.4, 9.5) 8.0 (5.4, 10.6) 6.1 (2.8, 9.5) 3.8 (2.3, 5.3) 10.2 (7.9, 12.6) 7.5 (5.0, 10.0) 16.3 (11.1, 21.5) 1.9 (0.8, 3.0) 5.0 (3.3, 6.7)
54 36 18 18 244 316
12.6 (9.4, 15.8) 10.5 (7.2, 13.8) 22.5 (13.0, 32.0) 4.1 (2.2, 6.0) 56.2 (51.2, 61.2)
34 10 24 7 114 155
16.6 (11.5, 21.7) 11.6 (4.8, 18.5) 20.2 (12.8, 27.2) 3.4 (0.9, 5.9) 48.0 (41.2, 54.9)
88 46 34 25 358 471
13.9 (11.2, 16.6) 10.7 (7.7, 13.6) 21.0 (15.2, 26.7) 3.9 (2.3, 5.4) 53.5 (49.4, 57.5)
Fig. 1. Estimated crude cumulative incidences (CCIs) of ipsilateral breast tumour recurrence (IBTR) (A) and breast cancer mortality (B) in pT1 patients. Thin line: no radiotherapy (RT); thick line: RT.
and 14.6% in pT2/RT and pT2/no RT patients, respectively (Fig. 2). Fifteen-year CCIs of breast cancer death were 11.6% in pT1/RT, 10.5% in pT1/no RT (Fig. 1), 20.2% in pT2/RT and 22.5% in pT2/no RT (Fig. 2). Considering the 296 pT1 patients who did not undergo axillary dissection, none of the 22 RT and 12 of the 274 no RT patients developed axillary disease (15-year CCI 0.0% and 4.0% respectively). None of the 134 pT1 patients who underwent axillary dissection had an axillary relapse. As regards factors influencing whether a patient received RT, patient age significantly influenced RT assignment, with older patients more likely not to receive RT (OR 0.94, 95% CI 0.90–0.98), as did axillary dissection, with those receiving axillary dissection more
likely to receive RT (OR 1.88, 95% CI 1.26–2.80). Concerning oestrogen receptor status, ER-positive (OR 0.55, 95% CI 0.30–1.03) and PgR-positive (OR 0.93, 95% CI 0.62–1.38) patients were less likely to receive RT. Patients with lobular carcinoma tended to receive RT more often than those with ductal carcinoma (OR 1.21, 95% CI 0.79–1.85). Separate analysis of lobular carcinomas showed that pT1 cases did not benefit from RT, while no RT pT2 patients had a significantly greater incidence of IBTR than RT pT2 patients (data not presented). The effect of RT on IBTR was found to differ significantly (p interaction 0.009) between pT1 and pT2 cases (Table 3). Specifically, Fine and Gray modelling showed that RT was significantly protective against IBTR in
Please cite this article in press as: Martelli G. et al., Omission of radiotherapy in elderly patients with early breast cancer: 15-Year results of a prospective non-randomised trial, Eur J Cancer (2015), http://dx.doi.org/10.1016/j.ejca.2015.04.018
G. Martelli et al. / European Journal of Cancer xxx (2015) xxx–xxx
5
Fig. 2. Estimated crude cumulative incidences (CCIs) of ipsilateral breast tumour recurrence (IBTR) (A) and breast cancer mortality (B) in pT2 patients. Thin line: no radiotherapy (RT); thick line: RT.
those with pT2 disease (subdistribution HR for IBTR 0.05, 95% CI: 0.01–0.39 compared to no RT; p = 0.004, Wald test), but had no effect on IBTR in pT1 patients (HR 0.90, 95% CI: 0.41–2.02; p = 0.806). By contrast, the effects of RT on distant metastasis and breast cancer mortality did not depend on pT status, with P values for interaction not significant (p = 0.806 and 0.784 respectively). Furthermore there was no evidence that giving RT had a protective effect against either of these end-points (p = 0.814 and p = 0.498, respectively, Table 3). We did not analyse the effect of ER status on IBRT as there were too few events and too few patients with ER-negative disease.
Table 3 Adjusted estimates of subdistribution hazard ratios (HRs), produced by Fine and Gray models, for end-points. P values are given by the Wald test. HRs for ipsilateral breast tumour recurrence (IBTR) according to whether or not radiotherapy (RT) given (no RT reference) are separate for pT1 and pT2 patients. HRs are also given for a 10% increase in propensity score which is the estimated probability of receiving RT according to patient age and whether or not axillary surgery was given. Hazard ratio (HR) (95% confidence interval (CI)) Ipsilateral breast tumour recurrence (IBTR) Radiotherapy (RT) versus no RT pT1 0.90 (0.41, 2.02) pT2 0.05 (0.01, 0.39) For 10% increase in propensity score 1.19 (0.71, 1.47)
P value
0.806 0.004 0.806
Distant metastasis RT versus No RT pT2 versus pT1 For 10% increase in propensity score
0.94 (0.54, 1.62) 2.63 (1.52, 4.55) 1.35 (1.13, 1.63)
0.814 <0.001 0.001
Breast cancer mortality RT versus No RT pT2 versus pT1 For 10% increase in propensity score
0.85 (0.53, 1.36) 2.43 (1.52, 3.89) 1.24 (1.06, 1.45)
0.498 0.001 0.008
4. Discussion In this prospective non-randomised study on elderly cN0 breast cancer patients treated by conservative surgery and adjuvant tamoxifen, for those with pT1 disease, 15-year CCIs of IBTR, distant metastasis and breast cancer death were indistinguishable in the RT and no RT groups (Table 2). By contrast, for patients with pT2 disease, the 15-year CCI of IBTR was much higher in those not given RT (14.6% versus 0.8%), although breast cancer mortality (RT 20.2%, no RT 22.5%, and distant metastasis (RT 15.7%, no RT 17.2%) did not differ significantly (Table 2). Outcomes for lobular carcinoma cases were closely similar to those of the whole series: RT was associated with significantly reduced CCI of IBTR only in the pT2 subgroup. These findings are consistent with those of the CALGB 9343 trial [7] which recruited women of P70 years with ER-positive, T1 breast cancer: although there were more local relapses in the no RT arm (9% versus 2%), breast cancer mortality, distant disease-free survival and breast conservation did not differ significantly in the RT and no RT arms, after a median follow-up of 12.6 years. Our findings are more similar to those of the recent Italian multicentric randomised trial [14], which found no significant difference in IBTR between the RT and no RT arms. The trial recruited patients aged 55–75 years with breast cancer up to 2.5 cm, treated by modified quadrantectomy (1 cm margins) and randomly allocated to RT versus no RT. After a median follow-up of 108 months, 12 (3.4%) IBTRs occurred in the RT arm against 16 (4.4%) in the no RT arm. Overall survival was 81.4% in the RT arm and 83.7% the no RT arm, with no significant differences in IBTR, distant disease-free survival or breast cancer-specific death between the arms. Over 80% of patients received hormonal therapy.
Please cite this article in press as: Martelli G. et al., Omission of radiotherapy in elderly patients with early breast cancer: 15-Year results of a prospective non-randomised trial, Eur J Cancer (2015), http://dx.doi.org/10.1016/j.ejca.2015.04.018
6
G. Martelli et al. / European Journal of Cancer xxx (2015) xxx–xxx
It is noteworthy that one of these trials [7] only recruited pT1 patients, while in the other [14], 11% of patients had small (up to 2.5 cm) pT2 disease. The novelty of our study is that tumour size (pT status) interacted significantly with the relation between RT and IBTR, and the 15-year CCI of IBTR was much higher in the no RT than RT group in pT2 patients. RT probably made a significant difference in elderly pT2 patients because the local relapse rate is intrinsically higher in this group, probably in relation to shorter distance between resection margin and tumour margin, or greater number of disease foci distant from the main tumour mass, and RT is able to ‘sterilise’ these foci – as it does in younger patients, who in general should always receive RT. In their 12-year follow-up [15] of the Milan III trial conducted at our Institute [3] Veronesi et al. also found that, while IBTR was significantly higher in the no RT arm (10-year CCI 23.5% versus 5.8% in the RT arm), for women over 65 years there were no differences in IBTR or other outcomes between the two arms. By contrast some North American studies [16,17] found larger differences (>10%) in local failure rate when RT is not given to older patients with early ER-positive breast cancer treated conservatively. Possible explanations are that under half the women (total 160 patients) in the Fisher et al. three-arm trial [16] were P60 years, while in the Fyles et al. trial [17] the analysis was based on the whole cohort, 58% of whom were 50–69 years. It is also likely that differences in surgical technique may have contributed: quadrantectomy, the operation generally used in Italy, is a more extensive excision than the lumpectomy generally used in North America, and has greater probability of removing all macroscopic disease foci. Even with RT, local recurrence rates are lower in patients who receive quadrantectomy than those who receive lumpectomy [18]. As shown initially by the trial of Fisher et al. [19] long-term tamoxifen prolongs the disease-free interval of postmenopausal patients with ER-positive disease. Tamoxifen almost certainly contributed to the low IBTR rates in our series, 86% of whom were ER-positive. By delaying IBTR onset, our elderly patients were more likely to die of competing events than experience relapse or die of breast cancer. Thus, only 88 (19% of total deaths) of our patients died of breast cancer compared to 383 for other causes. Among the 296 women in the present study with pT1 disease who did not undergo axillary dissection (patient preference or treating surgeon’s opinion), none who received RT (CCI: 0.0%) and 12 who did not receive RT (CCI: 4.0%) developed axillary disease after 15 years. The lack of axillary disease in the RT group could be due in part to the RT itself which may have included the first level of the axilla. However a recent prospective Italian study [20] reported that the axilla does not receive a therapeutic dose when the treatment plan aims to irradiate
the breast only. We therefore feel that the contribution of RT to axillary control is minimal. In our randomised study with 5-year follow-up on older patients with T1N0 breast cancer randomised to conservative surgery with or without axillary dissection, followed by RT and tamoxifen [21], we found that women with high grade cancer had more locoregional events than those with low-intermediate grade disease, with no adverse effect on breast cancer mortality. High grade has been associated with increased risk of local failure in other studies [1,22]. We have no data on tumour grade for the present study as grade was not available in 1987, when the study started. Nevertheless based on other data [1,21,22] we suggest that elderly patients with early ER-positive breast cancer should receive postoperative RT if they have high grade disease. The main limitation of this prospective non-randomised study is the possibility of bias in selecting patients for RT versus no RT. Thus, although all patients were eligible for RT, it is likely that those in poorer general health tended not to receive RT. Performance status was not evaluated. However overall status is likely to have influenced the decision to perform RT and since it was not included in the propensity score is a possible source of bias. Grade and presence of lymphovascular invasion were not assessed, they were unknown to the treating physician, could not have influenced the RT decision and are not a source of bias. As regards extensive intraductal component, this may have been available to physicians in a few cases, and may have influenced the RT decision. However this information was not included in the database since it was not determined routinely. Since it was not included in the propensity score it too is a possible source of bias. Cases with involved margins were excluded prospectively and the quadrantectomy performed during the study period was more ample than it is today (2 cm margins were the norm). There was surely some variation, but even if the margin was only 1 cm it is unlikely to have influenced the RT decision. To conclude, data from the present prospective non-randomised study and randomised trials [7,14] strongly suggest that in most elderly patients with pT1 cN0 ER-positive breast cancer treated by conservative surgery and tamoxifen, the contribution of RT to disease control is minimal and its omission does not impact on breast cancer mortality. However data from the present study suggest that patients with pT2 disease should receive RT to limit the number of local recurrences. Conflict of interest statement None declared.
Please cite this article in press as: Martelli G. et al., Omission of radiotherapy in elderly patients with early breast cancer: 15-Year results of a prospective non-randomised trial, Eur J Cancer (2015), http://dx.doi.org/10.1016/j.ejca.2015.04.018
G. Martelli et al. / European Journal of Cancer xxx (2015) xxx–xxx
Acknowledgement The authors thank Don Ward for help with the English. References [1] Darby S, McGale P, Correa C, Taylor C, Arriagada R, Clarke M, et al. Effect of radiotherapy after breast-conserving surgery on 10year recurrence and 15-year breast cancer death: meta-analysis of individual patient data for 10,801 women in 17 randomized trials. Lancet 2011;378:1707–16. [2] Ferlay J, Shin HR, Bray F, Forman D, Mathers C, Parkin DM. Estimates of worldwide burden of cancer in 2008: GLOBOCAN 2008. Int J Cancer 2010;15(127):2893–917. [3] Veronesi U, Luini A, Del Vecchio M, Greco M, Galimberti V, Merson M, et al. Radiotherapy after breast-preserving surgery in women with localized cancer of the breast. N Engl J Med 1993;328:1587–91. [4] Clark RM, McCulloch PB, Levine MN, Lipa M, Wilkinson RH, Mahoney LJ, et al. Randomized clinical trial to assess the effectiveness of breast irradiation following lumpectomy and axillary dissection for node-negative breast cancer. J Natl Cancer Inst 1992;84(9):683–9. [5] Clark RM, Whelan T, Levine M, Roberts R, Willan A, McCulloch P, et al. Randomized clinical trial of breast irradiation following lumpectomy and axillary dissection for node-negative breast cancer: an update. Ontario Clinical Oncology Group. J Natl Cancer Inst 1996;88:1659–64. [6] Fisher B, Anderson S, Redmond CK, Wolmark N, Wickerham DL, Cronin WM. Reanalysis after 12 years of follow-up in a randomized clinical trial comparing total mastectomy with lumpectomy with or without irradiation in the treatment of breast cancer. N Engl J Med 1995;333:1456–61. [7] Hughes KS, Schnaper LA, Bellon JR, Cirrincione CT, Berry DA, McCormick B, et al. Lumpectomy plus tamoxifen with or without irradiation in women age 70 years or older with early breast cancer: long-term follow-up of CALGB 9343. J Clin Oncol 2013;31:2382–7. [8] Ronchi E, Granata G, Brivio M, Coradini D, Miodini P, Di Fronzo G. A double-labeling assay for simultaneous estimation and characterization of estrogen and progesterone receptors using radioiodinated estradiol and tritiated Org 2058. Tumori 1986;72(3):251–7. [9] Kalbfleisch JD, Prentice RL. The statistical analysis of failure time data. 2nd ed. Hoboken, New Jersey: John Wiley & Sons; 2011. [10] Fine JP, Gray RJ. A proportional hazards model for the subdistribution of a competing risk. J Am Stat Assoc 1999;94:496–509.
7
[11] D’Agostino RB. Propensity score methods for bias reduction in the comparison of a treatment to a non-randomized control group. Stat Med 1998;17:2265–81. [12] R Development Core Team. R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0, [accessed 11.14]. [13] Gray B. Subdistribution analysis of competing risks, 2.2-7 [accessed 11.14]. [14] Tinterri C, Gatzemeier W, Costa A, Gentilini MA, Zanini V, Regolo L, et al. Breast-conservative surgery with and without radiotherapy in patients aged 55–75 years with early-stage breast cancer: a prospective, randomized, multicenter trial analysis after 108 months of median follow-up. Ann Surg Oncol 2014;21:408–15. [15] Veronesi U, Marubini E, Mariani L, Galimberti V, Luini A, Veronesi P, et al. Radiotherapy after breast-conserving surgery in small breast carcinoma: long-term results of a randomized trial. Ann Oncol 2001;12:997–1003. [16] Fisher B, Bryant J, Dignam JJ, Wickerham DL, Mamounas EP, Fisher ER, et al. National Surgical Adjuvant Breast and Bowel Project. Tamoxifen, radiation therapy, or both for prevention of ipsilateral breast tumor recurrence after lumpectomy in women with invasive breast cancers of one centimeter or less. J Clin Oncol 2002;20:4141–9. [17] Fyles AW, McCready DR, Manchul LA, Trudeau ME, Merante P, Pintilie M, et al. Tamoxifen with or without breast irradiation in women 50 years of age or older with early breast cancer. N Engl J Med 2004;351:963–70. [18] Veronesi U, Volterrani F, Luini A, Saccozzi R, Del Vecchio M, Zucali R, et al. Quadrantectomy versus lumpectomy for small size breast cancer. Eur J Cancer 1990;26:671–3. [19] Fisher B, Costantino J, Redmond C, Poisson R, Bowman D, Couture J, et al. A randomized clinical trial evaluating tamoxifen in the treatment of patients with node-negative breast cancer who have estrogen-receptor-positive tumors. N Engl J Med 1989;320:479–84. [20] Aristei C, Chionne F, Marsella AR, Alessandro M, Rulli A, Lemmi A, et al. Evaluation of level I and II axillary nodes included in the standard breast tangential field and calculation of the administered dose: results of a prospective study. Int J Radiat Oncol Biol Phys 2001;51:69–73. [21] Martelli G, Boracchi P, De Palo M, Pilotti S, Oriana S, Zucali R, et al. A randomized trial comparing axillary dissection to no axillary dissection in older patients with T1N0 breast cancer: results after 5 years of follow-up. Ann Surg 2005;242:1–6. [22] Albert JM, Pan IW, Shih YC, Jiang J, Buchholz TA, Giordano SH, et al. Effectiveness of radiation for prevention of mastectomy in older breast cancer patients treated with conservative surgery. Cancer 2012;118:4642–51.
Please cite this article in press as: Martelli G. et al., Omission of radiotherapy in elderly patients with early breast cancer: 15-Year results of a prospective non-randomised trial, Eur J Cancer (2015), http://dx.doi.org/10.1016/j.ejca.2015.04.018
Available at www.sciencedirect.com
ScienceDirect journal homepage: www.ejcancer.com
Omission of radiotherapy in elderly patients with early breast cancer: 15-Year results of a prospective non-randomised trial Gabriele Martelli a,⇑, Patrizia Boracchi b, Eleonora Guzzetti a, Giuseppe Marano b, Laura Lozza c, Roberto Agresti a, Cristina Ferraris a, Domenico Piromalli a, Marco Greco a a
Breast Unit, Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy Department of Clinical Science and Community Health, University of Milan, Milan, Italy c Radiotherapy Unit, Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy b
Received 21 February 2015; received in revised form 6 April 2015; accepted 24 April 2015
KEYWORDS Breast cancer Radiotherapy Elderly patients Breast-conserving surgery Quadrantectomy Tamoxifen
Abstract Background: Whether radiotherapy (RT) is beneficial in elderly (P70 years) patients undergoing conservative surgery for early breast cancer has long been controversial. Recent randomised trials show that most elderly patients do not benefit from RT. We started a prospective non-randomised trial to address this issue in 1987 and now present results for the 627 consecutive pT1/2cN0 patients recruited, and treated by conservative surgery (quadrantectomy) and tamoxifen, and assigned non-randomly to RT or no RT. Methods: We used multivariate competing risks models to estimate 15-crude cumulative incidence (CCI) of ipsilateral breast tumour recurrence (IBTR), distant metastasis and breast cancer mortality. The models incorporated a propensity score as a measure of probability of receiving RT based on baseline characteristics, to account for the lack of randomisation. Results: For pT1 patients, 15-year CCIs of IBTR, distant metastasis and breast cancer death were indistinguishable in the RT and no RT groups. For pT2 patients, 15-year CCI of IBTR was much higher in those not given RT (14.6% versus 0.8%, p = 0.004), although breast cancer mortality and distant metastasis did not differ significantly between RT and no RT. Conclusions: Consistent with the findings of recent randomised trials, our long-term data indicate that most elderly, ER-positive patients with pT1 cN0 breast cancer treated by quadrantectomy do not benefit from RT. The 14.6% CCI of IBTR in our pT2 patients is an additional finding not presented in the trials and suggests that RT should be administered to elderly patients with pT2 disease. Ó 2015 Elsevier Ltd. All rights reserved.
⇑ Corresponding author at: Fondazione IRCCS Istituto Nazionale dei Tumori, Via Venezian 1, 20133 Milan, Italy. Fax: +39 02 2367430.
E-mail address: [email protected] (G. Martelli). http://dx.doi.org/10.1016/j.ejca.2015.04.018 0959-8049/Ó 2015 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Martelli G. et al., Omission of radiotherapy in elderly patients with early breast cancer: 15-Year results of a prospective non-randomised trial, Eur J Cancer (2015), http://dx.doi.org/10.1016/j.ejca.2015.04.018
2
G. Martelli et al. / European Journal of Cancer xxx (2015) xxx–xxx
1. Introduction The meta-analysis of 10,801 breast cancer patients from 17 randomised trials published by the Early Breast Cancer Trialists’ Collaborative Group in 2011 [1], showed that radiotherapy (RT) after breast-conserving surgery not only reduces the risk of breast cancer recurrence but also improves overall survival. However, although the incidence of breast cancer increases with advancing age [2] – and around 40% of breast cancers occur in women over 65 years – most of the randomised trials that assessed postoperative breast RT excluded patients over 70 years [3–6] and it was unclear for many years whether RT after breast-conserving surgery is beneficial in elderly patients. An early trial carried out at our Institute [3], which recruited patients with tumour 62.5 cm, found in an ad hoc analysis, that the incidence of cancer recurrence in the no RT arm declined markedly with age, and the authors suggested that RT may not be necessary in older women. By contrast, another early trial, published in 1992 [4], did not identify a subgroup of patients at low risk of local recurrence who could be spared RT; in particular, for women over 50 years with tumour up to 2 cm who did not receive RT, the recurrence rate at 7 years was 22% [5]. In 2013, the long-term results of the CALGB 9343 randomised trial were published [7]. This trial recruited women aged 70 years or older with T1N0 ER-positive breast cancer, treated with lumpectomy and tamoxifen, and randomised to either postoperative RT or no RT. Fewer local relapses occurred in the RT arm (2% versus 9%) without, however, affecting breast cancer mortality, distant disease-free survival, or breast conservation, after a median follow-up of 12.6 years. In 1987 we initiated a prospective non-randomised study to investigate whether breast RT can be safely avoided in elderly patients (P70 years) undergoing conservative surgery (quadrantectomy) and prescribed tamoxifen. Here we present 15-year results of that study. Whether or not RT was given depended on patient preference and the treating surgeon’s opinion. No patients were formally contraindicated for RT. We used a propensity score with multivariable analyses to adjust for the imbalance in baseline characteristics between the two groups.
resection margins, multifocal lesions, synchronous bilateral breast cancer, distant metastasis at diagnosis, previous cancer at another site, or contraindication for RT, were excluded. 2.2. Treatment Two hundred and seven patients received postoperative RT, 420 patients did not. RT was administered to the residual breast using a cobalt unit or 6-MeV photons to deliver 50 Gy (daily target dose 2 Gy) in the two opposing tangential fields. In all cases, a 10 Gy boost was administered to the tumour bed. The irradiation aimed to treat the breast only, with no attempt to include the axillary, supraclavicular or internal mammary nodes in the irradiation field. Axillary dissection was performed in 134/430 (31%) patients with pT1 disease, and 33/197 (17%) patients with pT2 disease. After surgery, regardless of hormone receptor status, all women were prescribed 20 mg tamoxifen daily for at least 3 years. 2.3. Hormone receptor status Hormone receptor status was determined by the dextran-coated charcoal technique [8]. Cancers with oestrogen receptor (ER) concentration above 10 fmol/mg cytosol protein were considered ER positive; those with progesterone receptor (PgR) above 25 fmol/mg were considered PgR positive. Cancers with receptor content below these values were considered receptor negative. 2.4. Follow-up Patients were seen every 6 months for the first 5 years and annually thereafter. Mammography and chest X-ray were performed once a year and bone scan every 2 years. Gynaecological examination with pelvic ultrasound was performed annually since patients were receiving tamoxifen. Disease status or cause of death was ascertained from clinical records, by consulting general practitioners or from mortality databases. Median follow-up was 188.1 months in the no RT group and 208.9 months in the RT group. No patients were lost to follow-up.
2. Patients and methods 2.5. Statistical methods 2.1. Patients Of 1451 consecutive breast cancer patients P70 years presenting at the Milan National Cancer Institute between January 1987 and December 1992, we evaluated 627 consecutive cN0 patients (430 pT1; 197 pT2 [63 cm]) who underwent conservative surgery with or without postoperative RT. Patients with involved
During follow-up possible events were ipsilateral breast tumour recurrence (IBTR), axillary relapse, distant metastasis, contralateral breast cancer, new primary at non-breast site, breast cancer death and death for any other cause. These events compete with each other to occur first, so competing risk methods were used to estimate crude cumulative incidences (CCIs) of events of
Please cite this article in press as: Martelli G. et al., Omission of radiotherapy in elderly patients with early breast cancer: 15-Year results of a prospective non-randomised trial, Eur J Cancer (2015), http://dx.doi.org/10.1016/j.ejca.2015.04.018
G. Martelli et al. / European Journal of Cancer xxx (2015) xxx–xxx
interest [9] (study end-points: IBTR, distant metastasis and breast cancer mortality) in the RT and no RT groups, taking account of the interaction of pT. The Fine and Gray [10] models were also used to estimate adjusted subdistribution hazard ratios (HRs) with 95% confidence intervals (CIs) for events, with P values obtained from the Wald test. The covariates entered in the Fine and Gray models were: RT (1 = yes, 0 = no), pT (1 = pT2, 0 = pT1), the interaction between these two covariates and a propensity score. The propensity score [11] was introduced because RT was not assigned randomly, and covariates likely to have influenced whether RT was given need to be included as confounders in the models. The propensity score is the probability that a patient will receive the treatment (RT) based on baseline characteristics. In the present case we estimated the probability of receiving RT in relation to patient age, receptor status (ER, PGR coded as 1 = positive and 0 = negative), axillary surgery (coded as 1 = yes, 0 = no) and histology (reference category ductal carcinoma). Since treatments and recruitment criteria did not change over the study period, year of diagnosis was not included in the propensity score. Data on other factors possibly influencing whether RT was given were not available. It has been shown that use of a propensity score based on potential confounders is equivalent to including the individual covariates in the model, and has the advantage of limiting the number of covariates included. The results of propensity score model are reported as odds ratios (ORs) with 95% CIs. The analyses were carried out with R software [12] and the cmprsk package [13].
3
3. Results Clinical characteristics of 627 consecutive patients according to whether or not they received RT are shown in Table 1. ER and PgR receptor status was available for 600 patients: 542 were ER positive, 377 were ER and PgR positive and 54 were ER and PgR negative. Invasive ductal carcinoma was the commonest histologic type (65.9%) followed by invasive lobular carcinoma (22.5%). Table 2 shows 15-year CCIs of first events and deaths by treatment group. Eight women experienced IBTR as first event in the pT1/RT group, and 29 in the pT1/no RT group. All patients who developed IBTR were treated by conservative surgery not mastectomy. Of the eight pT1/RT patients who had IBTR, one developed distant metastases and died of breast cancer, five died of unrelated conditions and the remaining two were alive at latest check-up. Of the 29 pT1/no RT patients who had IBTR, four developed distant metastases and died of their disease, 14 died of unrelated conditions (three other malignancies, 11 non-neoplastic conditions) and the remaining 11 were alive at latest check-up. Among pT2 patients, one in the RT group and 11 in the no RT group developed IBTR. The pT2/RT patient was disease-free and alive at latest follow-up; of the 11 pT2/no RT patients, two developed distant metastases and died of their disease, four died of unrelated conditions and the remaining five were alive at latest check-up. Fifteen-year CCIs of IBTR were 8.1% in both pT1/RT and pT1/no RT patients (Fig. 1); and 0.8%
Table 1 Characteristics of 627 consecutive elderly (P70 years) early, cN0 breast cancer patients, according to whether or not they received radiotherapy (RT). No radiotherapy (RT) (N 420)
RT (N 207)
Overall (N 627)
Median age in years (range)
76 (70–91)
75 (70–88)
76 (70–91)
Axillary surgery, N (%) No Yes
333 (79.3%) 87 (20.7%)
127 (61.4%) 80 (38.6%)
460 (73.4%) 167 (26.6%)
Tumour size, N (%) pT1 pT2
344 (81.9%) 76 (18.1%)
86 (41.5%) 121 (58.5%)
430 (68.6%) 197 (31.4%)
Oestrogen receptor (ER), N (%) ER-positive ER-negative Not assessed
373 (88.8%) 31 (7.4%) 16 (3.8%)
169 (81.6%) 27 (13.0%) 11 (5.3%)
542 (86.4%) 58 (9.3%) 27 (4.3%)
Progesterone receptor (PGR), N (%) PgR-positive PgR-negative Not assessed
265 (63.1%) 139 (33.1%) 16 (3.8%)
116 (56.0%) 80 (38.6%) 11 (5.3%)
381 (60.8%) 219 (34.9%) 27 (4.3%)
Histology, N (%) Invasive ductal carcinoma Invasive lobular carcinoma Other invasive carcinoma
276 (65.7%) 90 (21.4%) 54 (12.9%)
137 (66.2%) 51 (24.6%) 19 (9.2%)
413 (65.9%) 141 (22.5%) 73 (11.6%)
Please cite this article in press as: Martelli G. et al., Omission of radiotherapy in elderly patients with early breast cancer: 15-Year results of a prospective non-randomised trial, Eur J Cancer (2015), http://dx.doi.org/10.1016/j.ejca.2015.04.018
4
G. Martelli et al. / European Journal of Cancer xxx (2015) xxx–xxx
Table 2 Numbers of adverse events and estimates of crude cumulative incidences (CCIs) at 15 years and 95% confidence intervals (CIs) in 627 consecutive elderly (P70 years) early cN0 breast cancer patients, according to whether or not radiotherapy (RT) was given.
First event Ipsilateral breast tumour recurrence (IBTR) pT1 pT2 Axillary disease Distant metastasis pT1 pT2 Contralateral breast cancer Other primary cancer Death Breast cancer pT1 pT2 Other cancer Other cause Total deaths
No radiotherapy (RT)
RT
N events
N events
Crude cumulative incidence (CCI) (95% confidence interval (CI))
Overall CCI (95% CI)
N events
CCI (95% CI)
40 29 11 14 37 24 13 6 23
9.2 (6.3, 12.1) 8.1 (5.1, 11.0) 14.6 (6.5, 22.7) 3.1 (1.4, 4.8) 8.9 (6.1, 11.6) 7.0 (4.3, 9.7) 17.2 (8.6, 25.7) 1.4 (0.3, 2.6) 5.7 (3.4, 7.9)
9 8 1 11 27 8 19 6 8
3.9 (1.2, 6.5) 8.1 (2.3, 14.0) 0.8 (0.0, 2.5) 5.3 (2.3, 8.4) 13.1 (8.5, 17.7) 9.3 (3.1, 15.5) 15.7 (9.2, 22.3) 2.9 (0.6, 5.1) 3.9 (1.2, 6.5)
49 37 12 25 64 32 32 12 31
7.4 (5.4, 9.5) 8.0 (5.4, 10.6) 6.1 (2.8, 9.5) 3.8 (2.3, 5.3) 10.2 (7.9, 12.6) 7.5 (5.0, 10.0) 16.3 (11.1, 21.5) 1.9 (0.8, 3.0) 5.0 (3.3, 6.7)
54 36 18 18 244 316
12.6 (9.4, 15.8) 10.5 (7.2, 13.8) 22.5 (13.0, 32.0) 4.1 (2.2, 6.0) 56.2 (51.2, 61.2)
34 10 24 7 114 155
16.6 (11.5, 21.7) 11.6 (4.8, 18.5) 20.2 (12.8, 27.2) 3.4 (0.9, 5.9) 48.0 (41.2, 54.9)
88 46 34 25 358 471
13.9 (11.2, 16.6) 10.7 (7.7, 13.6) 21.0 (15.2, 26.7) 3.9 (2.3, 5.4) 53.5 (49.4, 57.5)
Fig. 1. Estimated crude cumulative incidences (CCIs) of ipsilateral breast tumour recurrence (IBTR) (A) and breast cancer mortality (B) in pT1 patients. Thin line: no radiotherapy (RT); thick line: RT.
and 14.6% in pT2/RT and pT2/no RT patients, respectively (Fig. 2). Fifteen-year CCIs of breast cancer death were 11.6% in pT1/RT, 10.5% in pT1/no RT (Fig. 1), 20.2% in pT2/RT and 22.5% in pT2/no RT (Fig. 2). Considering the 296 pT1 patients who did not undergo axillary dissection, none of the 22 RT and 12 of the 274 no RT patients developed axillary disease (15-year CCI 0.0% and 4.0% respectively). None of the 134 pT1 patients who underwent axillary dissection had an axillary relapse. As regards factors influencing whether a patient received RT, patient age significantly influenced RT assignment, with older patients more likely not to receive RT (OR 0.94, 95% CI 0.90–0.98), as did axillary dissection, with those receiving axillary dissection more
likely to receive RT (OR 1.88, 95% CI 1.26–2.80). Concerning oestrogen receptor status, ER-positive (OR 0.55, 95% CI 0.30–1.03) and PgR-positive (OR 0.93, 95% CI 0.62–1.38) patients were less likely to receive RT. Patients with lobular carcinoma tended to receive RT more often than those with ductal carcinoma (OR 1.21, 95% CI 0.79–1.85). Separate analysis of lobular carcinomas showed that pT1 cases did not benefit from RT, while no RT pT2 patients had a significantly greater incidence of IBTR than RT pT2 patients (data not presented). The effect of RT on IBTR was found to differ significantly (p interaction 0.009) between pT1 and pT2 cases (Table 3). Specifically, Fine and Gray modelling showed that RT was significantly protective against IBTR in
Please cite this article in press as: Martelli G. et al., Omission of radiotherapy in elderly patients with early breast cancer: 15-Year results of a prospective non-randomised trial, Eur J Cancer (2015), http://dx.doi.org/10.1016/j.ejca.2015.04.018
G. Martelli et al. / European Journal of Cancer xxx (2015) xxx–xxx
5
Fig. 2. Estimated crude cumulative incidences (CCIs) of ipsilateral breast tumour recurrence (IBTR) (A) and breast cancer mortality (B) in pT2 patients. Thin line: no radiotherapy (RT); thick line: RT.
those with pT2 disease (subdistribution HR for IBTR 0.05, 95% CI: 0.01–0.39 compared to no RT; p = 0.004, Wald test), but had no effect on IBTR in pT1 patients (HR 0.90, 95% CI: 0.41–2.02; p = 0.806). By contrast, the effects of RT on distant metastasis and breast cancer mortality did not depend on pT status, with P values for interaction not significant (p = 0.806 and 0.784 respectively). Furthermore there was no evidence that giving RT had a protective effect against either of these end-points (p = 0.814 and p = 0.498, respectively, Table 3). We did not analyse the effect of ER status on IBRT as there were too few events and too few patients with ER-negative disease.
Table 3 Adjusted estimates of subdistribution hazard ratios (HRs), produced by Fine and Gray models, for end-points. P values are given by the Wald test. HRs for ipsilateral breast tumour recurrence (IBTR) according to whether or not radiotherapy (RT) given (no RT reference) are separate for pT1 and pT2 patients. HRs are also given for a 10% increase in propensity score which is the estimated probability of receiving RT according to patient age and whether or not axillary surgery was given. Hazard ratio (HR) (95% confidence interval (CI)) Ipsilateral breast tumour recurrence (IBTR) Radiotherapy (RT) versus no RT pT1 0.90 (0.41, 2.02) pT2 0.05 (0.01, 0.39) For 10% increase in propensity score 1.19 (0.71, 1.47)
P value
0.806 0.004 0.806
Distant metastasis RT versus No RT pT2 versus pT1 For 10% increase in propensity score
0.94 (0.54, 1.62) 2.63 (1.52, 4.55) 1.35 (1.13, 1.63)
0.814 <0.001 0.001
Breast cancer mortality RT versus No RT pT2 versus pT1 For 10% increase in propensity score
0.85 (0.53, 1.36) 2.43 (1.52, 3.89) 1.24 (1.06, 1.45)
0.498 0.001 0.008
4. Discussion In this prospective non-randomised study on elderly cN0 breast cancer patients treated by conservative surgery and adjuvant tamoxifen, for those with pT1 disease, 15-year CCIs of IBTR, distant metastasis and breast cancer death were indistinguishable in the RT and no RT groups (Table 2). By contrast, for patients with pT2 disease, the 15-year CCI of IBTR was much higher in those not given RT (14.6% versus 0.8%), although breast cancer mortality (RT 20.2%, no RT 22.5%, and distant metastasis (RT 15.7%, no RT 17.2%) did not differ significantly (Table 2). Outcomes for lobular carcinoma cases were closely similar to those of the whole series: RT was associated with significantly reduced CCI of IBTR only in the pT2 subgroup. These findings are consistent with those of the CALGB 9343 trial [7] which recruited women of P70 years with ER-positive, T1 breast cancer: although there were more local relapses in the no RT arm (9% versus 2%), breast cancer mortality, distant disease-free survival and breast conservation did not differ significantly in the RT and no RT arms, after a median follow-up of 12.6 years. Our findings are more similar to those of the recent Italian multicentric randomised trial [14], which found no significant difference in IBTR between the RT and no RT arms. The trial recruited patients aged 55–75 years with breast cancer up to 2.5 cm, treated by modified quadrantectomy (1 cm margins) and randomly allocated to RT versus no RT. After a median follow-up of 108 months, 12 (3.4%) IBTRs occurred in the RT arm against 16 (4.4%) in the no RT arm. Overall survival was 81.4% in the RT arm and 83.7% the no RT arm, with no significant differences in IBTR, distant disease-free survival or breast cancer-specific death between the arms. Over 80% of patients received hormonal therapy.
Please cite this article in press as: Martelli G. et al., Omission of radiotherapy in elderly patients with early breast cancer: 15-Year results of a prospective non-randomised trial, Eur J Cancer (2015), http://dx.doi.org/10.1016/j.ejca.2015.04.018
6
G. Martelli et al. / European Journal of Cancer xxx (2015) xxx–xxx
It is noteworthy that one of these trials [7] only recruited pT1 patients, while in the other [14], 11% of patients had small (up to 2.5 cm) pT2 disease. The novelty of our study is that tumour size (pT status) interacted significantly with the relation between RT and IBTR, and the 15-year CCI of IBTR was much higher in the no RT than RT group in pT2 patients. RT probably made a significant difference in elderly pT2 patients because the local relapse rate is intrinsically higher in this group, probably in relation to shorter distance between resection margin and tumour margin, or greater number of disease foci distant from the main tumour mass, and RT is able to ‘sterilise’ these foci – as it does in younger patients, who in general should always receive RT. In their 12-year follow-up [15] of the Milan III trial conducted at our Institute [3] Veronesi et al. also found that, while IBTR was significantly higher in the no RT arm (10-year CCI 23.5% versus 5.8% in the RT arm), for women over 65 years there were no differences in IBTR or other outcomes between the two arms. By contrast some North American studies [16,17] found larger differences (>10%) in local failure rate when RT is not given to older patients with early ER-positive breast cancer treated conservatively. Possible explanations are that under half the women (total 160 patients) in the Fisher et al. three-arm trial [16] were P60 years, while in the Fyles et al. trial [17] the analysis was based on the whole cohort, 58% of whom were 50–69 years. It is also likely that differences in surgical technique may have contributed: quadrantectomy, the operation generally used in Italy, is a more extensive excision than the lumpectomy generally used in North America, and has greater probability of removing all macroscopic disease foci. Even with RT, local recurrence rates are lower in patients who receive quadrantectomy than those who receive lumpectomy [18]. As shown initially by the trial of Fisher et al. [19] long-term tamoxifen prolongs the disease-free interval of postmenopausal patients with ER-positive disease. Tamoxifen almost certainly contributed to the low IBTR rates in our series, 86% of whom were ER-positive. By delaying IBTR onset, our elderly patients were more likely to die of competing events than experience relapse or die of breast cancer. Thus, only 88 (19% of total deaths) of our patients died of breast cancer compared to 383 for other causes. Among the 296 women in the present study with pT1 disease who did not undergo axillary dissection (patient preference or treating surgeon’s opinion), none who received RT (CCI: 0.0%) and 12 who did not receive RT (CCI: 4.0%) developed axillary disease after 15 years. The lack of axillary disease in the RT group could be due in part to the RT itself which may have included the first level of the axilla. However a recent prospective Italian study [20] reported that the axilla does not receive a therapeutic dose when the treatment plan aims to irradiate
the breast only. We therefore feel that the contribution of RT to axillary control is minimal. In our randomised study with 5-year follow-up on older patients with T1N0 breast cancer randomised to conservative surgery with or without axillary dissection, followed by RT and tamoxifen [21], we found that women with high grade cancer had more locoregional events than those with low-intermediate grade disease, with no adverse effect on breast cancer mortality. High grade has been associated with increased risk of local failure in other studies [1,22]. We have no data on tumour grade for the present study as grade was not available in 1987, when the study started. Nevertheless based on other data [1,21,22] we suggest that elderly patients with early ER-positive breast cancer should receive postoperative RT if they have high grade disease. The main limitation of this prospective non-randomised study is the possibility of bias in selecting patients for RT versus no RT. Thus, although all patients were eligible for RT, it is likely that those in poorer general health tended not to receive RT. Performance status was not evaluated. However overall status is likely to have influenced the decision to perform RT and since it was not included in the propensity score is a possible source of bias. Grade and presence of lymphovascular invasion were not assessed, they were unknown to the treating physician, could not have influenced the RT decision and are not a source of bias. As regards extensive intraductal component, this may have been available to physicians in a few cases, and may have influenced the RT decision. However this information was not included in the database since it was not determined routinely. Since it was not included in the propensity score it too is a possible source of bias. Cases with involved margins were excluded prospectively and the quadrantectomy performed during the study period was more ample than it is today (2 cm margins were the norm). There was surely some variation, but even if the margin was only 1 cm it is unlikely to have influenced the RT decision. To conclude, data from the present prospective non-randomised study and randomised trials [7,14] strongly suggest that in most elderly patients with pT1 cN0 ER-positive breast cancer treated by conservative surgery and tamoxifen, the contribution of RT to disease control is minimal and its omission does not impact on breast cancer mortality. However data from the present study suggest that patients with pT2 disease should receive RT to limit the number of local recurrences. Conflict of interest statement None declared.
Please cite this article in press as: Martelli G. et al., Omission of radiotherapy in elderly patients with early breast cancer: 15-Year results of a prospective non-randomised trial, Eur J Cancer (2015), http://dx.doi.org/10.1016/j.ejca.2015.04.018
G. Martelli et al. / European Journal of Cancer xxx (2015) xxx–xxx
Acknowledgement The authors thank Don Ward for help with the English. References [1] Darby S, McGale P, Correa C, Taylor C, Arriagada R, Clarke M, et al. Effect of radiotherapy after breast-conserving surgery on 10year recurrence and 15-year breast cancer death: meta-analysis of individual patient data for 10,801 women in 17 randomized trials. Lancet 2011;378:1707–16. [2] Ferlay J, Shin HR, Bray F, Forman D, Mathers C, Parkin DM. Estimates of worldwide burden of cancer in 2008: GLOBOCAN 2008. Int J Cancer 2010;15(127):2893–917. [3] Veronesi U, Luini A, Del Vecchio M, Greco M, Galimberti V, Merson M, et al. Radiotherapy after breast-preserving surgery in women with localized cancer of the breast. N Engl J Med 1993;328:1587–91. [4] Clark RM, McCulloch PB, Levine MN, Lipa M, Wilkinson RH, Mahoney LJ, et al. Randomized clinical trial to assess the effectiveness of breast irradiation following lumpectomy and axillary dissection for node-negative breast cancer. J Natl Cancer Inst 1992;84(9):683–9. [5] Clark RM, Whelan T, Levine M, Roberts R, Willan A, McCulloch P, et al. Randomized clinical trial of breast irradiation following lumpectomy and axillary dissection for node-negative breast cancer: an update. Ontario Clinical Oncology Group. J Natl Cancer Inst 1996;88:1659–64. [6] Fisher B, Anderson S, Redmond CK, Wolmark N, Wickerham DL, Cronin WM. Reanalysis after 12 years of follow-up in a randomized clinical trial comparing total mastectomy with lumpectomy with or without irradiation in the treatment of breast cancer. N Engl J Med 1995;333:1456–61. [7] Hughes KS, Schnaper LA, Bellon JR, Cirrincione CT, Berry DA, McCormick B, et al. Lumpectomy plus tamoxifen with or without irradiation in women age 70 years or older with early breast cancer: long-term follow-up of CALGB 9343. J Clin Oncol 2013;31:2382–7. [8] Ronchi E, Granata G, Brivio M, Coradini D, Miodini P, Di Fronzo G. A double-labeling assay for simultaneous estimation and characterization of estrogen and progesterone receptors using radioiodinated estradiol and tritiated Org 2058. Tumori 1986;72(3):251–7. [9] Kalbfleisch JD, Prentice RL. The statistical analysis of failure time data. 2nd ed. Hoboken, New Jersey: John Wiley & Sons; 2011. [10] Fine JP, Gray RJ. A proportional hazards model for the subdistribution of a competing risk. J Am Stat Assoc 1999;94:496–509.
7
[11] D’Agostino RB. Propensity score methods for bias reduction in the comparison of a treatment to a non-randomized control group. Stat Med 1998;17:2265–81. [12] R Development Core Team. R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0,
Please cite this article in press as: Martelli G. et al., Omission of radiotherapy in elderly patients with early breast cancer: 15-Year results of a prospective non-randomised trial, Eur J Cancer (2015), http://dx.doi.org/10.1016/j.ejca.2015.04.018