Protein signatures outperforming individual biomarkers for improved clinical outcomes
The importance of protein biomarkers for delivering real-time biological insights is well-recognized. While research in this area initially relied on the use of individual biomarkers, these often lack the specificity needed for studying complex diseases driven by multiple pathological processes. Advances in multiplex protein analysis have enabled protein signatures to provide a more nuanced understanding of human biology. However, the growing reliance on protein biomarkers in clinical research highlights the need for a robust development pipeline to ensure reliable protein measurements capturing true biological signal.
White Paper
The following examples of protein biomarker signatures leading to clinical advancements were supported by Olink’s Proximity Extension Assay (PEA), a robust, highly sensitive and uniquely scalable multiplex immunoassay platform.
Protein signatures for early cancer detection and disease monitoring
Protein biomarker signatures have proven invaluable for early cancer detection in recent years. A study screened 3000 proteins in pancreatic ductal adenocarcinoma (PDAC) patients and controls, using machine learning (ML) to identify protein signatures linked to tumor progression, inflammation and matrix remodeling, outperforming CA19-9, the only FDA-approved PDAC biomarker (1). Similarly, another PDAC study utilizing multiplex protein analysis and ML identified a 6-protein signature model that effectively identified patients with disease progression during chemotherapy, outperforming both gene expression and CA19-9 (2). In a study aiming to uncover diagnostic biomarkers for pancreatic neuroendocrine tumors with proteomics, the top-performing 9-protein signature model for disease detection outperformed the standard biomarker CgA, even when CgA was excluded. This offers a new diagnostic tool for suspected neuroendocrine tumors and could be used in patient follow-up after radical resection (3).
Moving beyond NfL as a single biomarker in neurodegenerative disorders
Despite its relevance to neurodegeneration, neurofilament light polypeptide (NfL) lacks the specificity to be used as a single biomarker for diagnosing or monitoring conditions such as Alzheimer’s disease (AD) and multiple sclerosis (MS), as it is also implicated in unrelated conditions, such as cardiovascular disease, depression, anorexia, and schizophrenia (4,5).
In a prime example of protein signatures providing higher resolution for studying neurological disorders, Chitnis T. et al. screened over 1400 proteins in MS patients’ serum samples, developing an 18-protein assay for MS disease activity (MSDA). This test, validated for LDT use, offers >90% accuracy in patient monitoring, surpassing traditional biomarkers like NfL. (6).
The MSDA Test was clinically validated with improved performance compared to the top-performing single-protein model and can serve as a quantitative tool to enhance the care of MS patients.
In a study by del Campo et al., panels with 8- and 9-protein biomarkers were developed following CSF protein screening in AD and dementia patients, differentiating AD from controls and other dementia types with striking accuracy (7). Similarly, the group led by Nancy Ip identified a 21-protein model for early diagnosis and staging of AD using blood biomarkers, which significantly out-performed the classic ATN markers in identifying mild cognitive impairment (8).
Our findings demonstrate the feasibility of a blood-based biomarker assay for early screening and routine monitoring of pathological changes of AD. Moreover, the heterogeneity of AD progression between ethnic groups and individuals revealed by our assay emphasizes the importance of patient stratification and precision medicine for AD diagnostics and therapeutics.
Improving cardiovascular risk assessment with protein biomarker signatures
In addition to outperforming single protein biomarkers, protein signature models have shown superiority in disease risk assessment compared to traditional clinical risk scores, particularly in cardiovascular disease (CVD). A study with close to 800 patients identified a protein profile that enhances risk prediction for recurrent atherosclerotic CVD, with minimal benefit from adding clinical scores (9).
Single plasma risk markers have failed to robustly improve atherosclerotic cardiovascular disease risk scores to date. Using a panel of 50 proteins, we show a significant improvement in discrimination and clinical value attested by the net reclassification improvement in secondary prevention.
Another study assessing CVD risk prediction analyzed circulating protein biomarkers in patients with heart failure, identifying a 9-protein signature with improved primary endpoint risk prediction compared to existing scores (10). Finally, a study of almost 2000 patients with CVD risk factors used proteomics and ML to develop a risk score that far outperformed the Framingham clinical risk score in predicting mortality (11).
As the examples outlined above show, researchers are propelling the field toward precision medicine by leveraging multiplex protein analysis to uncover protein signatures, leading to new possibilities for early disease detection, treatment selection and patient monitoring across various pathologies.
White Paper
References
1. Athanasiou A, Kureshi N, Wittig A, et al. Biomarker Discovery for Early Detection of Pancreatic Ductal Adenocarcinoma (PDAC) Using Multiplex Proteomics Technology (2024) Journal of Proteome Research, DOI: 10.1021/acs.jproteome.4c00752
2. van Eijck CWF, Sabroso-Lasa S, Strijk GJ, et al. A liquid biomarker signature of inflammatory proteins accurately predicts early pancreatic cancer progression during FOLFIRINOX chemotherapy (2024) Neoplasia, DOI: 10.1016/j.neo.2024.100975
3. Thiis-Evensen E, Kjellman M, Knigge U, et al. Plasma protein biomarkers for the detection of pancreatic neuroendocrine tumours and differentiation from small intestinal neuroendocrine tumours. (2022) Journal of Neuroendocrinology, DOI:10.1111/jne.13176
4. Giacomucci G et al., Plasma neurofilament light chain as a biomarker of Alzheimer’s disease in Subjective Cognitive Decline and Mild Cognitive Impairment. (2022) J Neurol. doi: 10.1007/s00415-022-11055-5
5. Amrein M et al., Serum neurofilament light chain in functionally relevant coronary artery disease and adverse cardiovascular outcomes. (2023) Biomarkers. doi: 10.1080/1354750X.2023.2172211.
6. Chitnis T, Foley J, Ionete C, et al. Clinical validation of a multi-protein, serum-based assay for disease activity assessments in multiple sclerosis. (2023) Clinical Immunology. doi: 10.1016/j.clim.2023.109688
7. del Campo M, Peeters C, Johnson E, et al. CSF proteome profiling across the Alzheimer’s disease spectrum reflects the multifactorial nature of the disease and identifies specific biomarker panels. (2022) Nature Aging, DOI: 10.1038/s43587-022-00300-1
8. Jiang Y, Uhm H, Ip FC, et al. A blood-based multi-pathway biomarker assay for early detection and staging of Alzheimer’s disease across ethnic groups. (2024) Alzheimer’s & Dementia, DOI: 10.1002/alz.13676
9. Nurmohamed NS, Belo Pereira JP, et al. Targeted proteomics improves cardiovascular risk prediction in secondary prevention (2022) European Heart Journal, DOI: 10.1093/eurheartj/ehac055
10. Klimczak-Tomaniak D, de Bakker M, Bouwens E, et al. Dynamic personalized risk prediction in chronic heart failure patients: a longitudinal, clinical investigation of 92 biomarkers (Bio-SHiFT study). (2022) Scientific Reports, DOI: 10.1038/s41598-022-06698-3
11. Unterhuber M, Kresoja K-P, Rommel K-P, et al. Proteomics-Enabled Deep Learning Machine Algorithms Can Enhance Prediction of Mortality. (2021) Journal of the American College of Cardiology, DOI: 10.1016/j.jacc.2021.08.018
Want to know more about Olink/PEA?
Please us our contact form for any questions you may have about OIink’s products and services for next-generation protein biomarker research across a broad range if applications, at any scale of project.