Urine Metabolic Profiling
for Rapid Lung Cancer Screening:
A Strategy Combining Rh-Doped SrTiO3‑Assisted Laser
Desorption/Ionization Mass Spectrometry and Machine Learning
posted on 2024-02-28, 06:44authored byKe Jia, Yawei Wang, Lixia Jiang, Mi Lai, Wenlan Liu, Liping Wang, Huihui Liu, Xiaohua Cao, Yuze Li, Zongxiu Nie
Lung cancer ranks among the cancers with the highest
global incidence
rates and mortality. Swift and extensive screening is crucial for
the early-stage diagnosis of lung cancer. Laser desorption/ionization
mass spectrometry (LDI-MS) possesses clear advantages over traditional
analytical methods for large-scale analysis due to its unique features,
such as simple sample processing, rapid speed, and high-throughput
performance. As n-type semiconductors, titanate-based perovskite materials
can generate charge carriers under ultraviolet light irradiation,
providing the capability for use as an LDI-MS substrate. In this study,
we employ Rh-doped SrTiO3 (STO/Rh)-assisted LDI-MS combined
with machine learning to establish a method for urine-based lung cancer
screening. We directly analyzed urine metabolites from lung cancer
patients (LCs), pneumonia patients (PNs), and healthy controls (HCs)
without employing any pretreatment. Through the integration of machine
learning, LCs are successfully distinguished from HCs and PNs, achieving
impressive area under the curve (AUC) values of 0.940 for LCs vs HCs
and 0.864 for LCs vs PNs. Furthermore, we identified 10 metabolites
with significantly altered levels in LCs, leading to the discovery
of related pathways through metabolic enrichment analysis. These results
suggest the potential of this method for rapidly distinguishing LCs
in clinical applications and promoting precision medicine.