AI Driven Early Cancer Detection with Deep Plasma Proteomics

Artificial intelligence is transforming how researchers interpret biological data, creating new opportunities for biomarker discovery and the early detection of cancer. However, the power of AI is only as strong as the data behind it, making consistency, quality, and scalability more important than ever.

In this webinar, leading experts will discuss how deep plasma proteomics and AI-driven data analysis are working together to uncover disease-associated patterns at increasingly early stages. Discover how standardized, high-throughput proteomic workflows generate the robust datasets needed to power reliable AI models, deliver meaningful biological insights, and accelerate the translation of scientific innovation into real-world clinical impact.

Machine Learning and Deep Plasma Proteomics for Breast Cancer Screening and Biomarker Discovery

• Alec Horrmann, Astrin Biosciences

Proteome-guided liquid biopsy tests hold immense promise for the future of early cancer detection. Here, we analyzed the plasma proteome from over 1,200 biobanked samples consisting of healthy women and women with breast cancer. A classifier was trained on 845 women (466 healthy and 379 with newly diagnosed, treatment-naïve breast cancer) and validated on 397 women (195 healthy and 202 with breast cancer) from the same collection cohort. All plasma samples were processed in an automated, blinded manner coupled with semi-quantitative, label-free mass spectrometry-based analysis, resulting in held-out performance of 92.3% specificity, 92.6% sensitivity, and an AUC of 0.975. Observed sensitivity remained high across all breast cancer stages and pathological and molecular subtypes, albeit with small sample sizes for some subtypes. Gene set enrichment analyses identified epithelial-to-mesenchymal transition and PI3K-AKT signaling as enriched in the breast cancer samples, suggesting the potential to identify cancer-related proteins in early-stage patients. Overall, our proteomic classifier demonstrates high sensitivity and specificity and is a favorable supplemental test post-mammogram.

Using Deep Plasma Proteomics and Adaptive Machine Learning to Improve Adnexal Mass Classification

• Johannes B. Müller-Reif, Project Group Leader at Department of Proteomics and Signal Transduction at Max Planck Institute of Biochemistry

Adnexal masses often present a diagnostic challenge, as current approaches relying on imaging and serum biomarkers can struggle to accurately distinguish benign from malignant disease. In this webinar, Johannes B. Müller-Reif will showcase how deep plasma proteomics combined with adaptive machine learning can reveal biological signatures associated with malignancy, enabling improved classification of adnexal masses from a single blood sample. The findings demonstrate the value of high-throughput plasma proteomics as a complementary tool to ultrasound, supporting more accurate patient triage and clinical decision-making.

Speaker: Sindisiwe Buthelezi (Senior Researcher at the Council for Scientific and Industrial Research (CSIR), South Africa)

Speaker: Pieter Langerhorst, Junior Group Leader at Sanquin, The Netherlands