Identifying Plasma Proteins Associated with Risk of Solid Cancers: A 25-Year Prospective Analysis of 4,712 Circulating Proteins in the ARIC Study
Wang, Z.; Burk, V.; Huang, Z.; Zahed, H.; Muller, D.; Yarmolinsky, J.; Lee, M. A.; Joshu, C.; Lin, Z.; Prizment, A.; Butler, K. R.; Couper, D.; Smith-Byrne, K.; Kolijn, M.; Vermeulen, R. C. H.; Riboli, E.; Gunter, M.; Coresh, J.; Chatterjee, N.; Platz, E. A.
Show abstract
This study investigated potential pre-diagnostic proteomic risk markers for 7 solid cancers independent of known risk factors within a multi-center, prospective cohort study. Using the SomaScan(R) 5K assay, we analyzed 4,712 unique plasma proteins (4,955 aptamers) in the Atherosclerosis Risk in Communities (ARIC) study among 9,391 middle-aged and older Black (23%) and White (77%) men and women. Over a maximum follow-up of 25.9 years, incident cases included 136 bladder, 271 colorectal, 96 kidney, 22 liver, 416 lung, 88 pancreatic, and 588 prostate cancers. After false discovery rate (FDR) correction, we identified 144 unique protein-cancer associations in common risk-factor adjusted models, and 41 protein-cancer associations in both common and cancer site-specific risk-factor adjusted models. Associations included several novel circulating proteins related to liver (33 proteins) and lung (4 proteins) cancer risk, and confirmed previously established proteins associated with kidney (HAVCR1 and MMP7) and prostate (KLK3 and ACP3) cancer risk. External validation in the European Prospective Investigation into Cancer and Nutrition cohort (SomaScan 7K) confirmed that the majority of FDR-significant proteins showed consistent effect directions and nominal significance, with the proportion of confirmed proteins varying between 75% and 100% depending on the cancer site. Time-lagged analysis demonstrated that 90% of the identified cancer-associated proteins are markers for long-term cancer risk, with observed associations more than 5 years pre-diagnosis after multiple-testing correction. These findings underscore the potential of circulating proteomic markers beyond known risk factors for elucidating etiologic mechanisms and improving risk stratification across cancers.
Matching journals
The top 7 journals account for 50% of the predicted probability mass.