Incidence and molecular subtype patterns of lung metastases in breast cancer: a systematic review and meta-analysis

Background Lung metastasis (LM) is a clinically significant manifestation of breast cancer (BC), yet its incidence across diverse populations remains insufficiently characterized. This study aimed to estimate the pooled reported proportion of LM and explore potential sources of heterogeneity. Methods This systematic review and meta-analysis was conducted following PRISMA guidelines and prospectively registered in PROSPERO (CRD420251275241). PubMed, Scopus, Embase, Web of Science, and Google Scholar were systematically searched from inception to September 30, 2026, without language restrictions. Two reviewers independently screened studies, extracted data, and assessed methodological quality using the Joanna Briggs Institute checklist. Pooled LM proportions were estimated using a random-effects model with restricted maximum likelihood (REML) estimation and logit-transformed proportions. Subgroup analyses, meta-regression, and sensitivity analyses were performed to explore heterogeneity and assess the robustness of the findings. Results Seventeen studies comprising 1,023,998 BC patients and 10,685 reported LM cases were included. The pooled reported LM proportion was 2.68% (95% CI: 1.40–5.07%), with substantial heterogeneity (I² = 99.87%). Subtype-specific pooled point estimates were 4.53% for HR−/HER2+, 3.86% for TNBC, 2.22% for HR+/HER2+, and 1.33% for HR+/HER2 − disease; however, the overall between-subgroup difference was not statistically significant ( p = 0.520). Subgroup analyses revealed significant differences across WHO regions ( p < 0.001), PM2.5 categories ( p = 0.002), and risk-of-bias classifications ( p = 0.012), whereas HDI, income level, study design, and publication year showed no significant between-group differences. The six primary moderators showed no significant associations in univariable meta-regression. Despite significant differences across PM2.5 categories in subgroup analyses, neither continuous PM2.5 nor AQI was significantly associated with LM proportion in the exploratory multivariable model (PM2.5: β = −0.272, p = 0.317; AQI: β = 0.108, p = 0.389). An inverse association between HDI and the reported LM proportion emerged only in the exploratory multivariable model (β = −12.924, p = 0.040) and should be interpreted cautiously given the limited number of studies and potential multicollinearity. Conclusion The pooled reported proportion of LM among BC patients was approximately 2.7%, with considerable heterogeneity across studies. Although geographical and environmental subgroup differences were observed, the inconsistency of study-level associations warrants cautious interpretation. Standardized metastatic-site reporting and improved epidemiological surveillance are essential for more reliable estimates. Clinical trial registration Not applicable.