Unveiling non-small cell lung cancer treatment effect heterogeneity: a comparative analysis of statistical methods Journal Article
| Authors: | Lavery, J. A.; Chen, Y.; Panageas, K. S.; Wang, Y. J. |
| Article Title: | Unveiling non-small cell lung cancer treatment effect heterogeneity: a comparative analysis of statistical methods |
| Abstract: | Background For patients with advanced non-small cell lung cancer lacking targetable genomic alterations, the impact of clinicogenomic characteristics on the effectiveness of combining chemotherapy with immunotherapy is unclear.Methods We evaluated 4 statistical methods for detecting heterogeneous treatment effects related to clinical factors, including programmed death-ligand 1 expression, tumor mutation burden, and stage at diagnosis, using the American Association for Cancer Research Project Genomics Evidence Neoplasia Exchange BioPharma Collaborative dataset supplemented with institutional data collected under the same data curation model. A 2-sided P value of no more than .05 was used to denote statistical significance for all analyses.Results The mixture model revealed 2 latent subgroups: in one subgroup, there was no meaningful treatment effect, with average progression-free survival (PFS) only 5% longer with immunotherapy alone (95% confidence interval [CI] = -19% to 35%); in the second subgroup, immunotherapy alone was associated with a 35% decrease in average PFS (95% CI = -59% to 2%), corresponding to a ratio in treatment effects of 1.62 (95% CI = 1.02 to 2.57). There was a marginal association between lower tumor mutation burden levels and membership in the subgroup with improved PFS following receipt of chemoimmunotherapy. The causal survival forest highlighted the importance of tumor mutation burden (variable importance ranking: 1) and programmed death-ligand 1 (variable importance ranking: 3) when assessing heterogeneity. In contrast, the accelerated failure time and Cox proportional hazards models did not detect any statistically significant heterogeneous treatment effects. In simulations, the mixture model identified heterogeneous treatment effects more frequently than other methods, especially with weak covariate relationships, demonstrating its utility for informing personalized treatment approaches.Conclusions The application of novel statistical methods to large scale clinico-genomic databases offers an opportunity to more accurately identify heterogeneous treatment effects in some settings as compared to traditional statistical methods. Applying such methods to the AACR Project GENIE BPC non-small cell lung cancer data indicated a potential association between decreasing tumor mutation burden and improved outcomes with chemoimmunotherapy as compared to immunotherapy alone. |
| Keywords: | immunotherapy; tumor mutational burden; subgroup identification |
| Journal Title: | JNCI: Journal of the National Cancer Institute |
| ISSN: | 0027-8874 |
| Publisher: | Oxford University Press |
| Publication status: | Online ahead of print |
| Date Published: | 2025-01-01 |
| Online Publication Date: | 2025-01-01 |
| Language: | English |
| ACCESSION: | WOS:001548838900001 |
| DOI: | 10.1093/jnci/djaf176 |
| PROVIDER: | wos |
| Notes: | Article; Early Access -- Source: Wos |
Altmetric
Citation Impact
BMJ Impact Analytics
Related MSK Work