Caution in interpreting comparative effectiveness of breast localization techniques

Correspondence to: Janhavi Venkataraman (email: janhavi.Venkataraman@hcahealthcare.co.uk)

The London Breast Institute

Princess Grace Hospital

42-52 Nottingham Place

London

W1U 5NY

UK

BJS Open, https://doi.org/10.1093/bjsopen/zraf153, published 01 January 2026

We read with interest the multicentre real-world study by Corsi et al. comparing preoperative localization techniques for non-palpable breast lesions¹. While the authors present valuable observational data, several methodological and interpretative issues limit causal inference and clinical applicability.

Marked baseline imbalances were evident between groups, including age, menopausal status, lesion biology, and morphology. Patients undergoing wire-guided localization (WGL) were older and more frequently postmenopausal, and the proportion of malignant (B5) lesions differed substantially between techniques (96.3% in WGL versus 70.8% in radio-guided (ROLL)). Microcalcifications—identified by the authors as an independent predictor of margin positivity—were over-represented in the ROLL cohort. These differences are likely to influence margin status and resection extent, yet were not adequately addressed through adjusted or propensity-based analyses.

Each participating centre employed only its preferred localization technique, introducing confounding by centre-specific expertise, surgical practice, and pathology workflows. Although hierarchical modelling was undertaken, technique effects cannot be reliably disentangled from institutional experience.

Interpretation of the calculated resection ratio (CRR) also warrants caution. The assumption that lower CRR reflects superior performance is questionable, as higher CRR may indicate more conservative excision with reduced removal of healthy tissue. Notably, randomized data have previously reported larger excision volumes with ROLL compared with WGL³. Without validation against re-excision rates, cosmetic outcomes, or patient-reported measures, the clinical relevance of CRR remains uncertain.

Localization failure rates and technical success were not reported, despite being key determinants of real-world effectiveness. Important technique-specific limitations were also omitted, including MRI artefacts associated with magnetic seed localization following neoadjuvant systemic therapy² and potential interference of ROLL with sentinel lymph node identification, particularly for upper outer quadrant tumours. The absence of cost-effectiveness analysis further limits translational relevance, despite evidence that patients, surgeons, and radiologists prefer wire-free localization approaches⁴.

Finally, the retrospective design spanning 2016–2024 introduces potential temporal confounding and learning-curve effects, particularly for newer technologies.

In summary, significant selection bias, centre-specific confounding, lack of technical success metrics, and uncertain interpretation of CRR restrict conclusions regarding comparative effectiveness. Prospective studies with standardized selection criteria and validated, clinically meaningful endpoints are needed.