Milan 2.

Methods	Study design: randomised clinical trial Country: Italy Study period: 1996 to 2000 (trial entry period) Inclusion criteria: women with primary operable breast cancer ≤ 2 cm in mammographic diameter, clinically negative axillary nodes, aged 65 to 80 years Exclusion criteria: synchronous bilateral breast cancer, distant metastases at diagnosis, history of other malignancy (except basal cell carcinoma or intraepithelial cervical cancer) Length of follow‐up: ALND: median (range) = 150 (125‐175) months. No ALND: median (range) = 149 (124‐174) months
Participants	No. in trial arm: ALND: N = 109; no ALDN: N = 110 Age: ALND: median (range) = 70 (65‐80 ) years; no ALND: median (range) = 70 (65‐80 ) years Stage distribution: ALDN: T1a, N = 2; T1b, N = 30; T1c, N = 69; T2, N = 8. No ALDN: T1a, N = 6; T1b, N = 44; T1c, N = 52; T2, N = 8 Proportion node positive: ALDN: 25/109. No ALDN: not reported, but 2/110 (1.8%) required delayed axillary dissection for overt axillary disease during follow‐up Pathological type of breast cancer: ALDN: Infiltrating ductal carcinoma, N = 60; infiltrating lobular carcinoma, N = 20; other infiltrating carcinoma, N = 29. No ALDN: infiltrating ductal carcinoma, N = 61; infiltrating lobular carcinoma, N = 19; other infiltrating carcinoma, N = 30
Interventions	Quadrantectomy plus axillary dissection (all 3 Berg levels removed) vs quadrantectomy alone
Outcomes	Overall mortality, breast cancer mortality, breast events (ipsilateral tumour recurrence, contralateral breast cancer, distant metastases)
Axillary node surgery	Minimum no. nodes to be removed according to protocol: not reported Nodes removed axillary dissection arm: not reported Nodes removed no axillary dissection arm: not reported Method of node pathological analysis: not reported Further treatment for node‐positive cases: no
Radiotherapy	RT ALND arm: postoperative RT to residual breast within 4 weeks of surgery. Axillary, supraclavicular and internal nodes were NOT irradiated, but RT fields used typically included the lower part of level I of the axilla. 50 Gy over 5 weeks, with a supplemental boost of 10 Gy to the tumour bed RT no ALND arm: postoperative RT to residual breast within 4 weeks of surgery. Axillary, supraclavicular and internal nodes were NOT irradiated, but RT fields used typically included the lower part of level I of the axilla. 50 Gy over 5 weeks, with a supplemental boost of 10 Gy to the tumour bed RT same in all trial arms? yes
Hormone and chemotherapy	All women were prescribed 10 mg tamoxifen twice daily after surgery for 5 years. 15% discontinued tamoxifen owing to side effects.
Notes	Baseline differences? possible excess of stage T1c in axillary dissection arm ‐ Table 5 (page 3, Martelli et al 2005). No P values were reported. Intention‐to‐treat analyses? yes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation list was reported (page 242, Martelli et al 2005), but it was not reported how this list was derived.
Allocation concealment (selection bias)	Low risk	Allocation was performed by calling data centre manager at study centre (page 2, Martelli et al 2005).
Blinding of outcome assessment (detection bias) Disease control in the axilla	Unclear risk	This was not reported.
Blinding of outcome assessment (detection bias) Breast cancer recurrence	Unclear risk	This was not reported.
Blinding of outcome assessment (detection bias) Short term adverse events	Unclear risk	Outcome was not reported.
Blinding of outcome assessment (detection bias) Long term adverse events	Unclear risk	Outcome was not reported.
Incomplete outcome data (attrition bias) Survival	Unclear risk	14 participants were excluded from analysis for protocol violation. It is unclear to which group they were randomised.
Incomplete outcome data (attrition bias) Axillary recurrence	Unclear risk	14 participants were excluded from analysis for protocol violation. It is unclear to which group they were randomised.
Incomplete outcome data (attrition bias) Breast cancer recurrence	Unclear risk	14 participants were excluded from analysis for protocol violation. It is unclear to which group they were randomised.
Incomplete outcome data (attrition bias) Short term adverse events	Unclear risk	Outcome was not reported.
Incomplete outcome data (attrition bias) Long term adverse events	Unclear risk	Outcome was not reported.
Selective reporting (reporting bias)	Unclear risk	Short‐term and long‐term adverse events were not reported.