Poster Presentation Clinical Oncology Society of Australia Annual Scientific Meeting 2026

Breast cancer recurrence in contemporary population-based studies: a systematic review and meta‑analysis of local, regional, locoregional and distant outcomes (145134)

Meghan Stephensen 1 , Susan Jordan 1 , Michelle Janossy-Taylor 2 , Rachel Tintera 2 , Nicole Mccarthy 3 , Marina Reeves 4
  1. School of Public Health, The University of Queensland, Brisbane, Queensland, Australia
  2. Medical School, The University of Queensland, Brisbane, Queensland, Australia
  3. Icon Cancer Centre (Wesley), Brisbane, Queensland, Australia
  4. Cancer and Palliative Care Outcomes Centre, Queensland University of Technology, Brisbane, Queensland, Australia

Aims

We aimed to update outdated estimates of population-based breast cancer recurrence risk by systematically reviewing and meta-analysing studies reporting local, regional, locoregional, distant and all breast cancer recurrences. We also aimed to describe how distant recurrence patterns vary by tumour grade, stage and subtype.

Methods

Following PRISMA guidelines (PROSPERO: CRD42024560649), we searched MEDLINE, Embase, Web of Science and CENTRAL (2005-2025). Eligible studies were population‑based (from national, state or regional cancer registries) and reported recurrence outcomes in adult women diagnosed with primary, non‑metastatic invasive breast cancer. Studies reported ≥1 recurrence endpoint, had ≥2 years follow‑up, and included diagnosis dates spanning 2005 and later to reflect contemporary treatment eras. Those with registry-derived subgroups, defined by age, stage or subtype only, were also eligible. Two reviewers independently screened studies, extracted data, and assessed risk of bias. Pooled recurrence risks were estimated using random-effects meta‑analysis.

Results

Of 2,150 studies, 23 studies were included. Pooled 5‑year cumulative incidence risks were 3.7% (95%CI:1.7-5.8; k=2) for local, 1.7% (95%CI:1.3-2.2; k=2) for regional, 9.1% (95%CI:7.8-10.3; k=7) for distant recurrence and 9.6% (95%CI:7.9-11.2; k=2) for all recurrences, noting high heterogeneity across pooled estimates (I²:87-99%). Triple‑negative and HER2-positive disease showed the highest 5‑year distant recurrence risks (19.8% [95%CI:13.8-25.8; k=5] and 19.2% [95%CI:14.6-23.9; k=3], respectively), whereas luminal subtypes were lower (HR+/HER2-:6.6% [95%CI:4.9-8.3; k=3]; HR+/HER2+:13.1% [95%CI:11.2-14.9; k=2]). Recurrence risks increased with higher stage and grade.

Conclusion

Across contemporary population‑based studies, pooled recurrence risks were lower than historical trial‑derived estimates. However, most included studies reported diagnosis periods spanning many years, with none providing recurrence rates for cohorts diagnosed solely in the era of modern targeted therapies (post-2015), limiting the extent to which contemporary treatment effects can be observed. More population‑based studies focused on modern treatment eras are needed to strengthen recurrence estimation, capture effects of targeted therapies and support survivorship planning.