Research ArticleOriginal Article

A novel recurrence-associated metabolic prognostic model for risk stratification and therapeutic response prediction in patients with stage I lung adenocarcinoma

Chengming Liu, Sihui Wang, Sufei Zheng, Xinfeng Wang, Jianbin Huang, Yuanyuan Lei, Shuangshuang Mao, Xiaoli Feng, Nan Sun and Jie He
Cancer Biology & Medicine August 2021, 18 (3) 734-749; DOI: https://doi.org/10.20892/j.issn.2095-3941.2020.0397

Article Figures & Data

Figures

  • Figure 1
    Figure 1

    Relationships between relapse and metabolic phenotype in stage I LUAD. (A) Significant enrichment of metabolic pathways in relapsed patients with stage I LUAD, compared with non-relapsed patients. NES: normalized enrichment score. (B) Volcano plot of 64 metabolism-related genes differentially expressed in stage I LUAD samples with and without relapsed tumors. (C) Functional analysis of 64 metabolism-related genes.

  • Figure 2
    Figure 2

    Development of RAMS for stage I LUAD in the training cohort. (A and B) The LASSO Cox proportional hazards regression model identified 23 genes most related to recurrence free survival (RFS). (C) Receiver operating characteristic (ROC) curve analysis of the RAMS for RFS. (D) Prognostic values of 2 selected genes using multivariate Cox proportional hazards regression analysis. (E) Kaplan-Meier survival curves of RFS for patients with stage I LUAD based on the RAMS. (F) Heat map of 2 gene expression profiles, risk score distributions, and recurrence status of each patient in the high and low risk groups. (G and H) Univariate (G) and multivariate (H) regression analyses of the associations between RAMS and clinical variables for the predictive value of RFS. (I) Performance was compared between RAMS and stages based on ROC curve analysis. (J) Time-dependent ROC curve analysis of the RAMS for overall survival (OS). (K) Kaplan-Meier survival curves of OS for patients with stage I LUAD, based on the RAMS. ***P < 0.001.

  • Figure 3
    Figure 3

    Validation of RAMS for stage I LUAD in the test cohort. (A) Receiver operating characteristic (ROC) curve analysis of the RAMS for recurrence free survival (RFS). (B) Kaplan-Meier survival curves of RFS for patients with stage I LUAD based on the RAMS. (C) Heat map of 2 gene expression profiles, risk score distributions, and recurrence status of each patient in the high and low risk groups. (D and E) Univariate (D) and multivariate (E) Cox regression analyses of the associations between RAMS and clinical variables for the predictive value of RFS. (F) The performance compared between RAMS and the stage was based on ROC curve analysis. (G) Time-dependent ROC curve analysis of the RAMS for overall survival (OS). (H) Kaplan-Meier survival curves of OS for patients with stage I LUAD based on the RAMS.

  • Figure 4
    Figure 4

    Validation of RAMS for stage I LUAD in the CICAMS cohort. (A) Receiver operating characteristic (ROC) curve analysis of the RAMS for recurrence free survival (RFS). (B) Kaplan-Meier survival curves of RFS for patients with stage I LUAD based on the RAMS. (C) Heat map of 2 gene expression profiles, risk score distributions, and recurrence status of each patient in the high and low risk groups. (D and E) Univariate (D) and multivariate (E) Cox regression analyses of the associations between the RAMS and clinical variables for the predictive value of RFS. (F) Nomogram to predict the 1-, 3-, and 5-year RFS of patients with stage I LUAD. (G) ROC curve analysis of the RAMS for overall survival (OS). (H) Kaplan-Meier survival curves of OS for patients with stage I LUAD based on the RAMS.

  • Figure 5
    Figure 5

    Associations between RAMS and immune response in stage I LUAD. (A) Significant enrichment of immune pathways between the high risk and low risk groups. NES: normalized enrichment score. (B) Heat map of the relationships between risk scores and 7 clusters of inflammatory metagenes. (C) Volcano plot of 4 clusters of inflammatory metagenes differentially enriched in the high and low risk groups. (D) Heat map of 4 clusters of inflammatory metagenes differentially enriched in the high and low risk groups. (E) Cross-correlogram based on Pearson’s correlation coefficient values between risk scores and 7 clusters of inflammatory metagenes.

  • Figure 6
    Figure 6

    Correlation between RAMS and immune cell infiltration or immune checkpoint profiles in stage I LUAD. (A) Differences in immune cell infiltration abundances between the high risk and low risk groups. (B) Cross-correlogram based on Pearson’s correlation coefficient values between risk scores and 22 tumor-infiltrated immune cells. (C) Heat map of immune checkpoint profiles in the high and low risk groups. (D) Differences in expressions of immune checkpoint molecules between high risk and low risk groups. (E) Cross-correlogram was based on Pearson’s correlation coefficient values between risk scores and expressions of immune checkpoint molecules. *P < 0.05; **P < 0.01; ***P < 0.001.

  • Figure 7
    Figure 7

    Relationships between RAMS and drug response in stage I LUAD. (A) Waterfall plot of TIDE prediction scores in the high and low risk groups. Red indicates a tumor that responded to therapy. Green indicates nonresponders. (B) A box plot evaluating immunotherapeutic responses between high risk and low risk patients using TIDE prediction scores. (C and D) Box plots evaluating responses to the chemotherapeutics docetaxel (C) and bortezomib (D) between high risk and low risk patients.

Supplementary Materials

Back to top

In this issue

Print
Download PDF
Email Article
Citation Tools

Jump to section

Keywords