This study investigates the use of chromatin organization in peripheral blood mononuclear cells (PBMCs) from liquid biopsies as biomarkers for cancer diagnosis and treatment evaluation. The researchers hypothesized that chromatin organization, integrating secretome signals with gene expression, provides efficient biomarkers. Using a combined chromatin imaging and machine learning pipeline, they classified healthy and pan-tumor patients, and further distinguished between three tumor groups (meningioma, glioma, head and neck) with up to 89% accuracy. The pipeline also monitored treatment effects of proton therapy, demonstrating its potential for cancer diagnostics and therapy evaluation.

Kiran Challa, Daniel Paysan, Dominic Leiser, Nadia Sauder, Damien C. Weber, G. V. Shivashankar