Assessing lung cancer progression and survival with infrared spectroscopy of blood serum

Abstract Infrared molecular fingerprinting has been identified as a new minimally invasive technological tool for disease diagnosis. While the utility of cross-molecular infrared fingerprints of serum and plasma for in vitro cancer diagnostics has been recently demonstrated, their potential for stratifying and predicting the prognosis of lung cancer remained unexplored. This study investigates the capability of this approach to predict survival and stratify lung cancer patients. Molecular fingerprinting through vibrational spectroscopy is employed to probe lung cancer. Fourier-transform infrared (FTIR) spectroscopy is applied to blood sera from 160 therapy-naive lung cancer patients, who were followed for up to 4 years. Machine learning is then utilized to evaluate the prognostic utility of this new approach. Additionally, a case-control study involving 501 individuals is analyzed to investigate the relationship between FTIR spectra and disease progression. Overall, we establish a strong correlation between the infrared fingerprints and disease progression, specifically in terms of tumor stage. Furthermore, we demonstrate that infrared fingerprinting provides insights into patient survival at performance levels comparable to those of tumor stage and relevant blood-based biomarkers. Identifying the combined capacity of infrared fingerprinting to complement primary lung cancer diagnostics and to assist in the assessment of lung cancer survival represents the first proof-of-concept study underscoring the potential of this profiling platform. This may provide new avenues for the development of tailored, personalized treatment decision-making.