Abstract
BACKGROUND: Proteins have an integral role in cancer aetiology and inform cancer detection, treatment, and prognosis. While published data from prospective proteomic studies identifying early detection cancer risk markers have rapidly increased in the past five years, the landscape of evidence remains unclear. We aim to synthesise the evidence on circulating proteins and cancer risk. METHODS: We conducted a systematic review and meta-analysis. We searched Embase and Medline up to December 2023 with reference-list screening and hand-searching up to August 2025. Prospective cohort studies were eligible if they used multiplex panels, included adults without cancer diagnosis at baseline, and reported associations between baseline circulating proteins and risk of incident cancer. Based on highest global incidence, we included oesophageal, stomach, colorectal, liver, lung, breast, cervical, prostate, bladder, thyroid cancer. We conducted exome-protein-score and Mendelian-randomisation analyses and integrated exome-gene-burden results for protein associations that passed false-discovery-rate correction to assess protein roles in cancer aetiology. FINDINGS: Of 4949 articles, we included 26 unique studies comprising 84,129 participants and 14,326 cases. The studies profiled 2434 unique proteins and reported 19,130 total protein-cancer-associations. We conducted 3448 meta-analyses and detected 216 associations (131 significant only in the meta-analyses) that passed false-discovery-rate correction for stomach (n = 2), colorectal (n = 27), lung (n = 172), breast (n = 14), and prostate (n = 1) cancer. No significant associations were observed for bladder cancer. Meta-analyses were not possible for oesophageal, liver, cervical, or thyroid cancer, due to limited data. Supporting genetic evidence was found for 39 protein-cancer-associations. INTERPRETATION: We identified 131 protein-cancer-associations that reached statistical significance in the meta-analysis, of which some may be cancer markers and others may have aetiological roles, indicated by supporting genetic analyses, including ITGA11 and lung cancer. Our findings highlight the need for large, diverse, and mature prospective proteomic cohort studies of cancer risk to ensure the equity and generalisability of insights into cancer risk. FUNDING: AD is funded by the University of Oxford. KSB is funded by Cancer Research UK and UK Research Institute. RCT is funded by Cancer Research UK. The funders had no role in study design, data collection, analysis, decision to publish, or manuscript preparation.