EP1.04-03 Immune Cell Filtration as a Biomarker for the Diagnosis and Prognosis of Lung Adenocarcinoma

Abstract

Since tumor-infiltrating immune cells provides meaningfully information of prognosis in lung adenocarcinoma, we aimed to construct a novel prognostic immune model on the basis of a systematic assessment of the immune landscape calculated from cancer transcriptomes of lung adenocarcinoma patients. We used an advanced algorithm, which named “Cell type Identification By Estimating Relative Subsets Of RNA Transcripts (CIBERSORT)”, to estimate the 22 immune cell types from public datasets. The selection operator model and least absolute shrinkage and random forest method were then applied to develop immune scores for tumor diagnosis and prognosis. 355 lung adenocarcinoma patients and 204 normal controls were obtained to develop a diagnostic model and the diagnostic immune risk score (dIRS) suggested high sensitivity and specificity in both the training sets (AUC= 0.93, P<0.01) and validation sets (AUC=0.89). A prognostic immune score (pIRS) was also established and served as an independent prognostic factor for overall-free survival, which showed better prognostic value than TNM stage. Additionally, by integrating the pIRS with clinical information in a complete nomogram, the result suggested higher accuracy of recurrence risk prediction with well-calibrated curves. In summary, we studied the potential application of immune cells in cancer diagnosis, prognosis and treatment. The proposed diagnostic and prognostic model (dIRS and pIRS) might provide integrative and meaningful signatures for precision medicine and personal management of lung adenocarcinoma patients.