Microbial Oncology: A Critical Appraisal of Microbe-associated Carcinogenesis from Established Infectious Causes to the Contested Tumour Microbiome

Infectious agents remain among the few modifiable causes of human cancer for which prevention has already altered population-level incidence, yet the intellectual boundaries of the field have shifted substantially. Classical infectious carcinogenesis, built on a small set of viruses, bacteria and parasites with documented aetiological roles, now coexists with a far broader proposition that resident microbial communities within tumours and at distant mucosal sites shape tumour initiation, progression and treatment response. This review examines how securely that broader proposition is supported, and where the evidence remains provisional. Literature was identified through structured searching of biomedical and multidisciplinary scholarly sources, supplemented by backward and forward citation tracking and by targeted retrieval of corrections and retractions, with appraisal focused on study design, biological specificity, contamination control, replication and the distance between mechanistic plausibility and demonstrated clinical effect. Several conclusions are well supported: oncogenic viruses operate largely through sustained expression of viral oncoproteins and genomic integration; Helicobacter pylori eradication reduces gastric cancer incidence, although the absolute benefit depends strongly on baseline risk and mucosal stage at treatment; and vaccination against hepatitis B virus and human papillomavirus has produced measurable declines in attributable malignancy. Evidence for bacterial contributions to colorectal carcinogenesis is mechanistically rich but complicated by unresolved taxonomy, inconsistent detection thresholds and limited prospective data. The pan-cancer intratumoural microbiome literature has been materially destabilised by demonstrated analytical errors and a high-profile retraction, which narrowed rather than abolished the field's claims. Microbiota-based prediction of immunotherapy response replicates only partially across cohorts, and randomised faecal microbiota transplantation evidence remains small and early phase. The central unresolved problems concern causal attribution in low-biomass settings, strain-level rather than genus-level resolution, and the near-absence of prospective interventional designs capable of distinguishing microbial drivers from microbial consequences of established tumours. Progress will depend less on additional descriptive surveys than on contamination-controlled, strain-resolved and prospectively designed studies.