Microbiota-Derived Extracellular Vesicles in Cancer Prognosis: A Systematic Review and Contextual Evidence Map of EV-Associated Microbial Signals

Background Microbiota-derived extracellular vesicles (EVs) are proposed as oncology biomarkers, but microbial vesicles and microbial signals in heterogeneous or host-EV preparations are not biologically equivalent.Methods We revised a prospectively registered systematic review of human cancer studies relating microbiota-derived EVs measured at or before prognostic baseline to time-to-event outcomes. Peer-review reruns covered PubMed, Scopus, Web of Science Core Collection, CENTRAL, and Embase. Biological provenance was re-adjudicated; QUIPS, REMARK-informed items, PROBAST where applicable, and GRADE-Prognosis informed appraisal.Results The update identified 9,862 records; 5,837 unique records remained after 4,025 duplicates were removed. Two P1 studies reported time-to-event associations. Fujikawa 2025, the only fully protocol-concordant study, found lower 5-year RFS with high versus low serum-EV BAF (66.5% vs 91.3%; p = 0.042). Uemura 2023 found shorter PFS (8 vs 20 months; p = 0.04) and OS (19 months vs not reached; p = 0.03) with high serum-EV Bacteroidia, but prognostic-baseline sampling time was unreported. Four P3 studies provided contextual outcome/treatment-response evidence. No study established treatment-specific predictive utility.Conclusions Evidence remains sparse and heterogeneous. Microbial-vesicle identity, baseline timing, independent prognostic value, and external validation must be demonstrated together before clinical use.