Abstract

We present a blood-based optical platform built on vibrational spectroscopy, which directly leverages a quantum property of matter: molecular vibrational energy levels are quantized. When light interacts with plasma, photon–molecule coupling produces a spectrum shaped by discrete vibrational modes, bond strengths, and local chemical environments. In complex biofluids, like plasma, this yields a high-dimensional “molecular fingerprint”, an integrated readout of millions of overlapping vibrational contributions from circulating molecules and EV cargo, without requiring prior knowledge of specific mutations or single biomarkers.

Cancer alters not only genomes, but the entire molecular environment of the body: metabolites, lipids, proteins, and extracellular vesicles (EVs) that coordinate signaling, immune response, and tissue remodeling. Conventional liquid biopsies largely target sparse, sometimes undetectable analytes (e.g., tumor DNA early in disease), while imaging can be ambiguous or insensitive at small lesion sizes and in post-treatment tissue environments. These gaps are particularly acute in head and neck cancer surveillance, where recurrence may be asymptomatic until late, and indeterminate imaging complicate decision-making.

To translate this quantum-resolved spectral structure into an actionable tool, we apply machine-learning classifiers to identify latent pattern shifts associated with disease states and trajectories. Our studies in head and neck cancer indicate the potential for highly specific detection (with early datasets showing no false positives), supporting a design goal of minimizing unnecessary anxiety and downstream testing in survivorship care.

By treating blood as an optically interrogable quantum ensemble of molecular vibrations, this approach aims to solve problems that single-analyte assays and imaging struggle to address: earlier recurrence signals, improved surveillance triage, and complementary assessment when radiographic findings are equivocal. More broadly, the same physics-driven measurement framework is extensible across diseases and therapeutic contexts where the systemic molecular environment, not one biomarker, determines clinical reality.

@ 2026 allumbra, Inc.