The invention provides a
lung cancer noninvasive screening
system and method based on expiration microbiomics, and belongs to the technical field of biomedical detection.The method comprises the steps that a sample is obtained through an expiration collector, a
bacteria abundance matrix is generated through a UMI-5R-16S sequencing technology, and the low-
biomass expiration
microorganism detection rate is remarkably increased; the method comprises the following steps of: constructing three types of hierarchical
random forest models: M1 model: healthy human vs.
Early lung cancer (I-II stage), and outputting early
risk probability based on markers such as Ochrobactrum and the like; m2 model: healthy human vs. Middle and
advanced stage lung cancer (III-IV stage), and
advanced stage risk probability is output based on
Burkholderia and other markers; m3 model: healthy human vs. Full
lung cancer (including
morning and
evening), outputting a comprehensive risk
score based on Dietzia and other markers; and M3, primarily screening positive samples, inputting the positive samples into M1 and M2, and performing staging evaluation, wherein a higher
risk probability is taken as a criterion during conflict. According to the invention, the total
lung cancer primary screening sensitivity is improved, and the
early detection rate is increased; the
microorganism detection rate of an expiration sample is high; and the single detection cost is lower.