Transcriptional patterns reveal tumor histologic heterogeneity and immunotherapy response in lung adenocarcinoma

Tumoral heterogeneity has proven to be a leading cause of difference in prognosis and acquired drug resistance. High intratumor heterogeneity often means poor clinical response and prognosis. Histopathological subtypes suggest tumor heterogeneity evolved during the progression of lung adenocarcinoma, but the exploration of its molecular mechanisms remains limited. In this work, we first verified that transcriptional patterns of a set of differentially expressed genes profoundly revealed the histologic progression of lung adenocarcinoma. Next, a predictive model based on the transcriptional patterns was established to accurately distinguish histologic subtypes. Two crucial genes were identified and used to construct a tumor heterogeneous scoring model (L2SITH) to stratify patients, and we found that patients with low heterogeneity score had better prognosis. Low L2SITH scores implied low tumor purity and beneficial tumor microenvironment. Moreover, L2SITH effectively identified cohorts with better responses to anti-PD-1 immunotherapy.