Network Meta-Analysis of Survival Outcomes in the First-Line Treatment of Unresectable or Metastatic Melanoma: Scoping Review of Methodology
Author(s)
Andrei Lazarev, MS1, Daria Tolkacheva, MS1, Kirill Sapozhnikov, PhD1, Olga Mironenko, PhD2, Natalia Sableva, MS2.
1BIOCAD JSC, Saint Petersburg, Russian Federation, 2BIOCAD JSC, Moscow, Russian Federation.
1BIOCAD JSC, Saint Petersburg, Russian Federation, 2BIOCAD JSC, Moscow, Russian Federation.
OBJECTIVES: To review and compare methodological approaches used in network meta-analyses (NMAs) of the first-line treatments for metastatic or unresectable melanoma, with a particular emphasis on the synthesis models applied, the handling of time-to-event outcomes, and the use of individual patient data (IPD).
METHODS: A comprehensive literature search was conducted in Embase and PubMed databases in February 2025. Data on methodological approaches to NMA were extracted from the studies and analysed.
RESULTS: 7 NMAs published in 2016-2024 were identified and reviewed, including a total of 45 unique randomized controlled trials (RCTs). The most frequently used approach (6 studies) was the synthesis of aggregated published results (Hazard Ratios), IPD NMA with pseudo-IPD restored from Kaplan-Meier curves (with Guyot’s algorithm) was done only in 1 case. NMA was mainly performed in the Bayesian framework (6 studies), while the frequentist one was used in 1 study. Fixed effects (FE) and random effects (RE) models were used with equal frequency across the studies. In most cases, the choice between FE and RE was not clearly justified, but in 2 of the 4 studies with FE, they were chosen due to the small number of RCTs for direct comparison. Assumptions about proportional hazards were tested in 3 studies using visual inspection of complementary log-log survival plots and Schoenfeld residuals. Parametric and flexible survival models were used in only one study, where IPD NMA was done.
CONCLUSIONS: NMA is widely used to compare treatments for unresectable or metastatic melanoma, but most studies lack modern methodological approaches, including unclear model selection, absence of proportional hazards testing, and reliance on aggregated data, which limits the accurate modelling of time-dependent effects and interpretability of results. Future research should adopt IPD NMA with flexible survival models for more reliable results.
METHODS: A comprehensive literature search was conducted in Embase and PubMed databases in February 2025. Data on methodological approaches to NMA were extracted from the studies and analysed.
RESULTS: 7 NMAs published in 2016-2024 were identified and reviewed, including a total of 45 unique randomized controlled trials (RCTs). The most frequently used approach (6 studies) was the synthesis of aggregated published results (Hazard Ratios), IPD NMA with pseudo-IPD restored from Kaplan-Meier curves (with Guyot’s algorithm) was done only in 1 case. NMA was mainly performed in the Bayesian framework (6 studies), while the frequentist one was used in 1 study. Fixed effects (FE) and random effects (RE) models were used with equal frequency across the studies. In most cases, the choice between FE and RE was not clearly justified, but in 2 of the 4 studies with FE, they were chosen due to the small number of RCTs for direct comparison. Assumptions about proportional hazards were tested in 3 studies using visual inspection of complementary log-log survival plots and Schoenfeld residuals. Parametric and flexible survival models were used in only one study, where IPD NMA was done.
CONCLUSIONS: NMA is widely used to compare treatments for unresectable or metastatic melanoma, but most studies lack modern methodological approaches, including unclear model selection, absence of proportional hazards testing, and reliance on aggregated data, which limits the accurate modelling of time-dependent effects and interpretability of results. Future research should adopt IPD NMA with flexible survival models for more reliable results.
Conference/Value in Health Info
2025-11, ISPOR Europe 2025, Glasgow, Scotland
Value in Health, Volume 28, Issue S2
Code
SA72
Topic
Methodological & Statistical Research, Study Approaches
Topic Subcategory
Literature Review & Synthesis, Meta-Analysis & Indirect Comparisons
Disease
Oncology