The landscape and prognostic value of tumor-infiltrating immune cells in gastric cancer

• Published in: PeerJ (San Francisco, CA) Vol. 7; p. e7993
• Main Authors: Li, Linhai, Ouyang, Yiming, Wang, Wenrong, Hou, Dezhi, Zhu, Yu
• Format: Journal Article
• Language: English
• Published: United States PeerJ. Ltd 10-12-2019; PeerJ, Inc; PeerJ Inc
• Subjects: Algorithms; Analysis; Biochemistry; Bioinformatics; Cancer; Cancer genetics; Cancer research; Cancer staging; Cancer therapies; CD4 antigen; CD8 antigen; Datasets; Dendritic cells; Gastric cancer; Gastroenterology and Hepatology; Gene expression; Genes; Genetic aspects; Genomes; Genomics; Health aspects; Immunological memory; Immunology; Lymphocytes; Lymphocytes T; Macrophages; Malignancy; Medical prognosis; Memory cells; Mortality; Multivariate analysis; Nomogram; Oncology; Patients; Prognosis; Proteins; Regression analysis; Risk assessment; Stomach cancer; Survival analysis; T cells; Tumor-infiltrating immune cells; Tumor-infiltrating lymphocytes; Tumors; Tumor-infiltrating immune cells; Gastric cancer; Nomogram
• ISSN: 2167-8359; 2167-8359
• DOI: 10.7717/peerj.7993

Gastric cancer (GC) is the fourth most frequently diagnosed malignancy and the second leading cause of cancer-associated mortality worldwide. The tumor microenvironment, especially tumor-infiltrating immune cells (TIICs), exhibits crucial roles both in promoting and inhibiting cancer growth. The aim of the present study was to evaluate the landscape of TIICs and develop a prognostic nomogram in GC.
A gene expression profile obtained from a dataset from The Cancer Genome Atlas (TCGA) was used to quantify the proportion of 22 TIICs in GC by the CIBERSORT algorithm. LASSO regression analysis and multivariate Cox regression were applied to select the best survival-related TIICs and develop an immunoscore formula. Based on the immunoscore and clinical information, a prognostic nomogram was built, and the predictive accuracy of it was evaluated by the area under the curve (AUC) of the receiver operating characteristic curve (ROC) and the calibration plot. Furthermore, the nomogram was validated by data from the International Cancer Genome Consortium (ICGC) dataset.
In the GC samples, macrophages (25.3%), resting memory CD4 T cells (16.2%) and CD8 T cells (9.7%) were the most abundant among 22 TIICs. Seven TIICs were filtered out and used to develop an immunoscore formula. The AUC of the prognostic nomogram in the TCGA set was 0.772, similar to that in the ICGC set (0.730) and whole set (0.748), and significantly superior to that of TNM staging alone (0.591). The calibration plot demonstrated an outstanding consistency between the prediction and actual observation. Survival analysis revealed that patients with GC in the high-immunoscore group exhibited a poor clinical outcome. The result of multivariate analysis revealed that the immunoscore was an independent prognostic factor.
The immunoscore could be used to reinforce the clinical outcome prediction ability of the TNM staging system and provide a convenient tool for risk assessment and treatment selection for patients with GC.