Related Experiment Video
Updated: Jan 7, 2026
Orthotopic Transplantation of Syngeneic Lung Adenocarcinoma Cells to Study PD-L1 Expression
Published on: January 19, 2019
Advancing Prognosis Prediction and Immunotherapy Efficacy in Lung Adenocarcinoma Through Machine Learning: Novel
Shan Li1,2, Wenhang Zhou3, Chen Hu4
1Department of Laboratory Medicine, Northern Jiangsu People's Hospital, Yangzhou, Jiangsu, China, yzsbh.com.
Introduction:
Anoikis, a type of programmed cell death induced by detachment from the extracellular matrix (ECM), is crucial in cancer progression. Resistance to anoikis often correlates with enhanced invasion, metastasis, treatment resistance, and tumor recurrence. However, no research has systematically explored anoikis-regulated tumor microenvironment (TME) in lung adenocarcinoma (LUAD).
Methods:
We used single-cell RNA sequencing (scRNA-seq) and spatial transcriptome RNA sequencing (stRNA-seq) analyses to reveal the subtype of anoikis-related epithelial cells, demonstrating its spatial location characteristics. With the maker genes of prognostic significance, we depicted the molecular landscapes of anoikis regulator patterns in RNA-seq data. We developed the anoikis-related signature (Anoikis.Sig) by integrating 10 machine learning (ML) algorithms to accurately predict prognosis in LUAD. Based on the median risk score computed by Anoikis.Sig, patients were divided into high- and low-risk groups. We employed extensive analysis between two risk groups, in terms of clinic implications, immune microenvironment, somatic mutations, immunotherapy, chemotherapy, and single-cell landscape. Finally, we verified the prognosis value of two Anoikis.Sig model genes.
Results:
By integrative analysis of scRNA-seq and stRNA-seq datasets, we defined diverse function subtypes of anoikis-related epithelial cells and investigated their spatial regulator patterns. Through its marker genes and leave-one-out cross-validation, we utilized the RSF algorithm to develop the Anoikis.Sig with a superior predictive ability, outperformed other LUAD signatures and clinical indicators. We categorized LUAD patients into high- and low-risk groups, which demonstrated the low-risk group had a better survival outcome, an ample immune infiltration, a distinct mutational landscape, and response to immunotherapy. ScRNA-seq analysis revealed biologically intercellular disparities delineated by Anoikis.Sig. qRT-PCR validated the prognostic value of two model genes of Anoikis.Sig in LUAD.
Conclusion:
Through multiomics analyses and ML algorithms, we succeeded in establishing the Anoikis.Sig to efficiently predict prognosis in Anoikis.Sig, which delineated molecular landscapes of anoikis regulator patterns and clinical applications of Anoikis.Sig.
