The Absolute Effect / Evidence / SOUND
SOUND
In women with breast cancer ≤2 cm and negative preoperative axillary ultrasonography, can axillary surgery be omitted when nodal pathology would not change postoperative treatment?
Trial registry: NCT02167490 · One underlying study; follow-up publications consolidated.
Design, population and eligibility
Multicentre, open-label, phase III randomized non-inferiority trial; 1:1 allocation.
1,463 women enrolled; 1,405 included in the intention-to-treat analysis. Median age 60 years; median tumour size 1.1 cm; 87.8% ER-positive/HER2-negative.
Breast cancer ≤2 cm; clinically node-negative with negative preoperative axillary ultrasound; candidates for breast-conserving treatment. Exclusion/selection criteria and nodal imaging details are in the protocol.
Median 5.7 years for disease assessment.
Treatment and comparator
Sentinel lymph-node biopsy (SLNB), followed by adjuvant treatment according to standard clinicopathological factors.
No axillary surgery; breast surgery and adjuvant management otherwise based on available clinical and tumour information.
Primary endpoint: Five-year distant disease-free survival (DDFS), tested for non-inferiority.
Other endpoints: Distant recurrence, axillary recurrence, DFS, OS and adjuvant treatment recommendations.
Key efficacy results
Five-year DDFS was 97.7% with SLNB and 98.0% without axillary surgery; HR 0.84 (90% CI 0.45–1.54); non-inferiority P=0.02. The ordinary superiority log-rank P value was 0.67.
Locoregional relapses were 12/708 (1.7%) versus 11/697 (1.6%); distant metastases 13 (1.8%) versus 14 (2.0%); deaths 21 (3.0%) versus 18 (2.6%).
In the SLNB arm, 97/708 (13.7%) had positive nodes, illustrating the limited nodal yield in this selected ultrasound-negative population.
Source-reported outcome visualization
Five-year distant disease-free survival · 5 years
Kaplan–Meier estimates reported in the primary publication. Similar observed rates support the prespecified non-inferiority result; this is not evidence of superiority.
Absolute effect
The observed five-year DDFS difference was +0.3 percentage points for omission (98.0% versus 97.7%). The trial's conclusion rests on its prespecified non-inferiority analysis, not on a clinically meaningful superiority effect.
NNT: Not assigned: the comparison was designed for non-inferiority and the observed difference is small; do not interpret the reciprocal of this difference as a treatment NNT.
NNH: Not quantifiable from a harmonized fixed-time adverse-event endpoint in the primary abstract.
Toxicity, harms and treatment burden
SLNB adds an operation and risks of seroma, sensory symptoms, restricted shoulder movement and lymphedema; the trial's principal rationale was to avoid morbidity when nodal information would not change management.
Arm-specific grade 3–5 adverse-event and treatment-discontinuation rates are not consistently reported in the primary abstract; not estimated here.
The trial does not test omission when nodal status is needed to determine systemic therapy or regional nodal irradiation.
Quality of life and patient-reported outcomes
The trial assessed treatment recommendations and patient-relevant consequences, but a numerical QoL effect is not reproduced here because it is not established in the primary abstract. Do not infer a QoL benefit solely from avoiding surgery.
Subgroups, maturity and limitations
Most participants had small, hormone-receptor-positive/HER2-negative cancers; generalizability to larger, biologically aggressive or clinically suspicious disease is limited.
The median follow-up of 5.7 years is adequate for the prespecified five-year primary endpoint but not a lifetime recurrence assessment. The planned non-inferiority margin was 2.5 percentage points for five-year DDFS. The trial was registered after 492 patients had already been randomized; the investigators reported that no interim data were examined before registration. The treatment-recommendation analysis was not powered to detect small differences. Radiotherapy was not standardized: all available RT options were allowed, and 114 patients (16.3%) in the no-axillary-surgery group received ELIOT as a full dose or intraoperative boost; this is relevant when interpreting the combined surgical and local-treatment strategy.
The intervention is a strategy of omitting axillary staging only when missing nodal information does not alter postoperative treatment decisions.
Clinical interpretation
Current-practice context · 1 November 2026 (full guideline freshness review pending): Consider contemporary de-escalation guidance and eligibility criteria. The result is most applicable to selected low-risk patients receiving breast-conserving treatment; do not extrapolate to clinically node-positive disease or situations in which nodal status changes RT or systemic treatment.
Primary sources and registry
Source hierarchy: primary randomized report and prespecified follow-up first; registry and current guidance are supplementary. Values not reported in the cited primary sources are explicitly marked unavailable rather than imputed.