Meta-analysis of trials of radiotherapy in DCIS Early Breast Cancer Trialists’ Collaborative Group (EBCTCG) The world’s breast cancer trialists have shared their data every 5 years 1985, 1990, 1995, 2000, 2005/6 Many MODERATE gains in long-term survival have been identified Together MODERATE gains from various treatment improvements (and a moderate contribution from breast screening) have almost halved UK/US breast cancer mortality at ages 35-69 2 EBCTCG 5th Cycle EBCTCG 5th Cycle 3 4 EBCTCG 5th Cycle EBCTCG 5th Cycle • Trials starting before 1 Jan 2000 • Events up to 30 Sept 2006 • Included trials of BCS±RT in DCIS for the first time – ie, trials of RT versus the same management but without RT 5 EBCTCG 5th Cycle Randomized trials of BCS±RT in DCIS Entry Women randomised Median followup NSABP B-17 1985-90 818 EORTC 10853 1986-96 SweDCIS UK/ANZ DCIS† Study Surgery* Negative margins required % with boost 16.5 y WLE (37% ad) Yes 9% 1010 10.4 y WLE (20% ad) Yes 5% 1987-99 1067 8.4 y Sect res (17% ad) No 2% 1990-98 1030 4.8 y WLE (No ad) Yes NR‡ 1999-06 636 N/A WLE (No ad) Yes NR‡ Data available: Data unavailable:§ RTOG 9804 * WLE: wide local excision; ad: axillary dissection; Sect res: sector resection † 2x2 factorial design: ±RT±Tam ‡ Boost not recommended § Some patients still taking tamoxifen Median follow-up for all trials with data available: 8.9 y 6 EBCTCG 5th Cycle Randomized trials of BCS±RT in DCIS Eligibility of randomized patients for EBCTCG meta-analysis Eligibility status NSABP B-17 EORTC 10853 Swedish BCCG UK/ANZ Total Eligible* 798 918 1,011 1,002 3,729 Ineligible† 20 92 56 28 196 Total 818 1,010 1,067 1,030 3,925 *Taking into account all information available to EBCTCG † Benign lesion only; microinvasion, invasion, Paget's disease, or other cancer present at randomization; or other study-specific eligibility violation 7 EBCTCG 5th Cycle Randomized trials of BCS±RT in DCIS Distribution of women by calendar year of randomisation Calendar year NSABP B-17 EORTC 10853 Swedish BCCG UK/ANZ Total 1985-89 71% 17% 6% - 21% 1990-94 29% 60% 51% 46% 47% 1995-99 - 23% 43% 54% 32% Total 100% 100% 100% 100% 100% 8 EBCTCG 5th Cycle Main outcome analysed: • Ipsilateral breast recurrence (either recurrence of CIS or occurrence of invasive cancer) 9 EBCTCG 5th Cycle DCIS: BCS + RT vs. BCS Ips. BREAST RECURRENCE (CIS & Inv) 10 EBCTCG 5th Cycle DCIS: BCS + RT vs. BCS Ips. BREAST RECURRENCE (CIS & Inv) 3729 women 11 EBCTCG 5th Cycle DCIS: BCS + RT vs. BCS Ips. BREAST RECURRENCE (CIS & Inv) Ratio of annual event rates by patient and tumour characteristics 12 DCIS: BCS + RT vs. BCS Ips. BREAST RECURRENCE (CIS & Inv) Age at diagnosis 13 EBCTCG 5th Cycle DCIS: BCS + RT vs. BCS Ips. BREAST RECURRENCE (CIS & Inv) Extent of breast-conserving surgery 14 EBCTCG 5th Cycle DCIS: BCS + RT vs. BCS Ips. BREAST RECURRENCE (CIS & Inv) Use of tamoxifen (in both treatment arms) 15 EBCTCG 5th Cycle DCIS: BCS + RT vs. BCS Ips. BREAST RECURRENCE (CIS & Inv) Method of DCIS detection 16 EBCTCG 5th Cycle DCIS: BCS + RT vs. BCS Ips. BREAST RECURRENCE (CIS & Inv) Margin status 17 EBCTCG 5th Cycle DCIS: BCS + RT vs. BCS Ips. BREAST RECURRENCE (CIS & Inv) Focality 18 EBCTCG 5th Cycle DCIS: BCS + RT vs. BCS Ips. BREAST RECURRENCE (CIS & Inv) Histological grade 19 EBCTCG 5th Cycle DCIS: BCS + RT vs. BCS Ips. BREAST RECURRENCE (CIS & Inv) Nuclear grade 20 EBCTCG 5th Cycle DCIS: BCS + RT vs. BCS Ips. BREAST RECURRENCE (CIS & Inv) Comedonecrosis 21 EBCTCG 5th Cycle DCIS: BCS + RT vs. BCS Ips. BREAST RECURRENCE (CIS & Inv) Architecture 22 EBCTCG 5th Cycle DCIS: BCS + RT vs. BCS Ips. BREAST RECURRENCE (CIS & Inv) Clinical primary tumour size 23 EBCTCG 5th Cycle DCIS: BCS + RT vs. BCS Ips. BREAST RECURRENCE (CIS & Inv) Pathological primary tumour size 24 EBCTCG 5th Cycle DCIS: BCS + RT vs. BCS Ips. BREAST RECURRENCE (CIS & Inv) among women with 1-20 mm tumours and negative margin status 25 EBCTCG 5th Cycle DCIS: BCS + RT vs. BCS Ips. BREAST RECURRENCE (CIS & Inv) Ratio of annual event rates by age and histological grade 26 DCIS: BCS + RT vs. BCS Ips. BREAST RECURRENCE (CIS & Inv) Ratio of annual event rates by age and comedonecrosis 27 EBCTCG 5th Cycle DCIS: BCS + RT vs. BCS 3729 women Contralateral breast events Regional or distant events 28 EBCTCG 5th Cycle DCIS: BCS + RT vs. BCS 3729 women Any breast event * *ie ipsilateral or contralateral DCIS, ipsilateral or contralateral invasive cancer, regional or distant metastasis, or death certified as due to breast cancer 29 EBCTCG 5th Cycle DCIS: BCS + RT vs. BCS 3729 women Mortality with a breast event Mortality without a breast event Mortality from all causes 30 EBCTCG 5th Cycle DCIS: BCS + RT vs. BCS 3729 women Heart disease mortality Non-breast primary cancer incidence 31 EBCTCG 5th Cycle EBCTCG analysis of radiotherapy in DCIS Conclusions • In these trials, radiotherapy halved the recurrence rate of either DCIS or invasive cancer in the ipsilateral breast. After 5 years the absolute gain was about 10% (18% vs. 8%) and after 10 years it was about 15% (28% vs. 13%). • The absolute gain in ‘low risk’ women was as large as in other other women. • The proportional reduction in the ipsilateral recurrence rate was greater in women aged 50+ years than in younger women. • Radiotherapy had no significant effect on mortality from breast cancer, at least to year 10. • In these data, radiotherapy had no significant adverse effect on cancers other than breast cancer, or on mortality from non-breast cancer causes, including heart disease. EBCTCG 5th Cycle 32