Plasma Proteomics of Cancer Risk Among Patients with CKD

Patients with CKD have higher cancer risk and worse outcomes than the general population. Circulating proteins are important biomarkers of incident and recurrent cancer, but CKD profoundly remodels the proteome. High-throughput assays now allow systematic interrogation of protein-phenotype associations and may identify biomarkers of cancer risk in CKD.

We measured 4,091 plasma proteins using SomaScan v4 in patients without prior cancer at the year 1 visit of the Chronic Renal Insufficiency Cohort. For each protein, we fit Cox models evaluating incident overall and specific cancers (breast, colon, head and neck, leukemia, lung, lymphoma, prostate). Fully adjusted models included sociodemographic and clinical characteristics. To determine whether proteomic signatures were cancer-type-specific or reflected common oncologic signatures in CKD, we assessed overlap of findings across outcomes. Hallmark and Kyoto Encyclopedia of Genes and Genomes pathway analyses characterized enriched biological processes.

Among 2,975 eligible participants (mean age 58.5 years, 45% female), 395 developed cancers over a median of 5.5 years. 37 proteins were identified as nominal associations and 12 were significant after false discovery rate correction in fully adjusted models, spanning colon (e.g., CD163), leukemia (e.g., Gremlin-2), lung (e.g., RAN-binding protein 3), and prostate (e.g., cystatin C) cancers. Cancers had distinct protein signatures, but pathway enrichment overlapped, highlighting shared processes including epithelial-to-mesenchymal transition, KRAS signaling, and inflammatory signaling.

Plasma proteomics identified 12 proteins associated with future cancer in CKD. Several of these proteins have known or plausible roles in oncogenesis. Although associations were cancer-type-specific, signatures converged on known pathways. If validated, these plasma proteins could enable cancer early detection strategies in CKD.