Research Paper Volume 13, Issue 9 pp 12691—12709

Immune subgroup analysis for non-small cell lung cancer may be a good choice for evaluating therapeutic efficacy and prognosis

Subgroup prediction model based on binomial logistic regression and prognostic efficacy of immune subgroups. (A) Distinguishing the predictive power of cluster 2 from the other two groups. (B) Distinguishing the predictive power of cluster 1 from cluster 3. The upper part is ROC-AUC, and the lower part is PR-AUC. (C) In all NSCLC samples, immune subgroups cannot significantly distinguish the overall survival of patients. (D) In the GEO independent verification set, the immune subgroup can significantly distinguish the prognosis of patients. Among patients with lung squamous cell carcinoma (E) or those older than 60 years (F), immune subgroups can significantly distinguish the prognosis of the patients.

Figure 2. Subgroup prediction model based on binomial logistic regression and prognostic efficacy of immune subgroups. (A) Distinguishing the predictive power of cluster 2 from the other two groups. (B) Distinguishing the predictive power of cluster 1 from cluster 3. The upper part is ROC-AUC, and the lower part is PR-AUC. (C) In all NSCLC samples, immune subgroups cannot significantly distinguish the overall survival of patients. (D) In the GEO independent verification set, the immune subgroup can significantly distinguish the prognosis of patients. Among patients with lung squamous cell carcinoma (E) or those older than 60 years (F), immune subgroups can significantly distinguish the prognosis of the patients.