Supplementary material: Challenges in conducting fractional polynomial and standard parametric network meta-analyses of immune checkpoint inhibitors for first-line advanced renal cell carcinoma

dataset

posted on 2024-04-15, 13:57 authored by Svenja Petersohn, Bradley McGregor, Sven Klijn, Jessica R. May, Flavia Ejzykowicz, Murat Kurt, Matthew Dyer, Bill Malcolm, Sebastien Branchoux, Katharina Nickel, Saby George, Sonja Kroep

These are peer-reviewed supplementary materials for the article 'Challenges in conducting fractional polynomial and standard parametric network meta-analyses of immune checkpoint inhibitors for first-line advanced renal cell carcinoma' published in the Journal of Comparative Effectiveness Research.

1. Feasibility assessment
1.1 Heterogeneity assessment
1.1.1 Items assessed for heterogeneity between trials
1.1.2 Overview of relevant baseline characteristics across trials
1.2 Assessment of the proportional hazards (PH) assumption
1.2.1 Criteria applied
1.2.2 Results
1.3 Network of evidence
2. Non-proportional hazards NMA outcomes
2.1 Model fitting
2.2 Model selection algorithm: face validity check of first- and second-order models
2.2.1 Time-varying HR plots versus trial hazards second order polynomial for OS
2.3 Model selection algorithm: Predictive accuracy against trial data
2.3.1 PFS
2.3.2 OS

Aim: Network meta-analyses (NMAs) increasingly feature time-varying hazards to account for nonproportional hazards between different drug classes. This paper outlines an algorithm for selecting clinically plausible fractional polynomial NMA models. Methods: The NMA of four immune checkpoint inhibitors (ICIs) + tyrosine kinase inhibitors (TKIs) and one TKI therapy for renal cell carcinoma (RCC) served as case study. Overall survival (OS) and progression free survival (PFS) data were reconstructed from the literature, 46 models were fitted. The algorithm entailed a-priori face validity criteria for survival and hazards, based on clinical expert input, and predictive accuracy against trial data. Selected models were compared with statistically best-fitting models. Results: Three valid PFS and two OS models were identified. All models overestimated PFS, the OS model featured crossing ICI + TKI versus TKI curves as per expert opinion. Conventionally selected models showed implausible survival. Conclusion: The selection algorithm considering face validity, predictive accuracy, and expert opinion improved the clinical plausibility of first-line RCC survival models.