Plasma proteomics of cancer risk in patients with chronic kidney disease

Background Patients with chronic kidney disease (CKD) have higher cancer risk and worse cancer outcomes than the general population. Although circulating proteins are important biomarkers of incident and recurrent cancer, CKD profoundly remodels the proteome. High‑throughput proteomic assays enable systematic interrogation of protein-phenotype associations and can disentangle CKD‑specific proteomic biomarkers and pathways of cancer risk. Methods We analyzed 4,091 plasma proteins measured by SomaScan (v4) at the year 1 visit of the Chronic Renal Insufficiency Cohort study (CRIC). The cohort included participants without prior cancer and with complete baseline covariates. For each protein, we fit Cox models for time-to-event (breast, colon, head & neck, leukemia, lung, lymphoma, prostate, and overall cancer). Fully adjusted models included age, sex, race, eGFR, smoking history, alcohol use, urine albumin‑to‑creatinine ratio, body mass index, diabetes, prior cardiovascular disease, and socioeconomic indicators. We reported Benjamini-Hochberg FDR‑significant proteins and flagged “nominal” associations (P < 0.001 and HR per SD ≥ 1.5 or ≤ 0.67). Downstream analyses assessed overlap between proteomic signatures across cancer types. Hallmark and KEGG enrichment analyses highlighted biological pathways associated with oncologic endpoints. Results Among the 2,975 participants who met study inclusion criteria (median age 59.5 years, 45.2% female), 395 developed cancers over a median of 12.5 years. Across cancer types, 78 proteins were FDR significant in demographics‑adjusted models. In fully adjusted models, 37 met the nominal significance threshold and 12 remained FDR‑significant, spanning colon, leukemia, lung, and prostate cancers. Many identified proteins were biologically plausible: higher soluble CD163-a marker of tumor‑associated macrophages-was associated with colon cancer; Gremlin‑2, a DAN‑family BMP antagonist implicated in hematopoietic dysregulation, was associated with leukemia; RAN‑binding protein 3, a regulator of nuclear export and TGF‑beta signaling, was associated with lung cancer; and cystatin C, a kidney filtration marker, was inversely associated with prostate cancer risk. Cancer types displayed distinct protein signatures, but pathway enrichment overlapped across outcomes, highlighting shared biological processes including epithelial-to-mesenchymal transition, KRAS signaling, inflammatory signaling, and cell adhesion. Conclusions Plasma proteomics revealed early circulating biomarkers associated with future cancer in CKD. Many identified proteins have established or mechanistically plausible roles in cancer biology. Protein-cancer associations were cancer type-specific; however, the proteomic signatures of different cancers converged on known oncogenic pathways. Identified proteins, if replicated, might facilitate earlier detection of cancer in CKD.