This study combined two types of molecular data, proteomics (the study of proteins present in a cell) and transcriptomics (the study of gene activity measured by RNA levels), to identify a set of molecular signals that could predict survival in kidney cancer patients.
The cancer studied is called KIRC (Kidney renal clear cell carcinoma), which is the same as ccRCC, and is the most lethal form of kidney cancer. Despite advances in targeted therapy and immunotherapy, predicting which patients will have better or worse outcomes remains difficult.
By integrating two different levels of biological information, the research aimed to build a more powerful and accurate prognostic tool than either data type could provide alone.
The research team started with large-scale protein data from The Cancer Genome Atlas (TCGA) and integrated it with RNA expression data to identify proteins whose levels correlated with patient survival. Multiple statistical filters were applied to reduce thousands of candidates to a manageable set.
After analysis, 7 prognostic proteins were identified as the core signature: ACC1, P21, P70S6K_pT389, SMAC, BRAF_pS445, MIG6, and PEA15. Each of these proteins plays a role in cancer-relevant pathways including cell cycle regulation, programmed cell death, and tumor suppression.
Researchers validated the proteins using two independent methods: the Human Protein Atlas (HPA) database, which documents protein levels across human tissues, and qRT-PCR experiments in four kidney cell lines (HK-2, ACHN, 786-O, and 769-P), where they confirmed the gene expression levels matched predictions.
Using the 7-protein signature, patients were divided into high-risk and low-risk groups. Patients in the high-risk group had significantly worse survival compared to those in the low-risk group, with this difference holding up in training, testing, and external validation datasets.
The model's performance was measured by the 5-year overall survival ROC AUC, which reached 0.810 in the training set, 0.775 in the test set, and 0.788 in the combined cohort. An AUC closer to 1.0 means better predictive accuracy, and values above 0.75 are generally considered clinically meaningful.
Compared to using clinical variables alone (such as tumor stage and patient age), adding the protein signature improved prediction, demonstrating that molecular information adds valuable insight beyond what doctors can observe through standard clinical assessments.
The high-risk patient group identified by the signature showed higher immune scores, meaning their tumors were more heavily infiltrated by immune cells. Surprisingly, this did not translate into better outcomes, suggesting a dysfunctional immune response in these tumors.
High-risk patients also showed greater expression of immune checkpoint genes, including PD-L1 and related molecules that cancer cells use to suppress immune attack. This makes the tumor resistant to natural immune defense mechanisms.
This finding has direct clinical implications: patients classified as high-risk by the signature may be better candidates for immune checkpoint inhibitor therapies such as pembrolizumab or nivolumab, which block the immune-suppressing signals that protect tumors.
ACC1 is an enzyme involved in fatty acid synthesis, a pathway cancer cells rely on for rapid growth. P21 is a cell cycle regulator that normally halts cell division when DNA is damaged, acting as a brake on uncontrolled growth.
P70S6K_pT389 is an activated form of a protein that promotes cell growth and protein production. SMAC promotes programmed cell death and normally counters cancer cell survival. BRAF_pS445 is an activated cancer-driving protein.
MIG6 suppresses growth signals from a receptor called EGFR, and PEA15 has complex roles in both promoting and inhibiting cancer depending on context. Together, this group reflects multiple cancer hallmarks including abnormal growth, survival, and metabolism.
This prognostic signature could help doctors identify, at the time of diagnosis, which kidney cancer patients are at highest risk of poor outcomes. This information would help prioritize closer monitoring and more aggressive treatment for those patients.
The combination of proteomics and transcriptomics makes this approach more robust than single-platform studies. Proteins are the actual functional molecules in cells, so measuring them alongside gene activity gives a more complete picture of tumor biology.
With further validation in prospective clinical studies, this 7-protein signature could be incorporated into clinical decision-making tools, ultimately helping guide more personalized treatment strategies for patients with kidney clear cell carcinoma.