There is a recent emerging theory that suggests a cross-link between pathogens and cancer. In this context, we examined the association between the Mycobacterium tuberculosis (MTB) with its L-forms (MTB-L) and lung cancer. In the present study, we have optimized and applied a highly sensitive assay to detect the presence of MTB and MTB-L in 187 lung cancer samples and 39 samples of other cancer origins. By carefully controlling confounding factors, we have found that 62% of the lung cancer samples are MTB-L positive, while only 5.1% of the other cancer samples are MTB-L positive. Through generalized linear models and random forest models, we have further identified a set of clinical end-points that are strongly associated with MTB-L presence. Our finding provides the basis for future studies to investigate the underlying mechanism linking MTB-L infection to lung cancer development.