advanced, locally advanced, and relapsed NSCLC
Conditions
Interventions
None listed
Sponsors
Yokohama City University Hospital
Eligibility
Sex/Gender
All
Inclusion criteria
Inclusion criteria: Study selection Articles written in English that present a randomised controlled trial (RCT) evaluating systemic immune-therapy for advanced, locally advanced, and relapsed NSCLC are eligible for inclusion. Conference abstracts are not accepted. Patients Patients with operable NSCLC are evaluated, irrespective of pathological subtype or driver mutation, provided they are considered candidates for systemic immune-therapy by the original study authors. Treatment This study focuses on ICI-based systemic immuno-therapy such as ICI monotherapy, dual ICI therapy, and chemoimmunotherapy. Absence or presence of previous systemic therapy is not considered. However, subgroup analysis focusing on first-line treatment and on second- or later-line treatment are expected.
Exclusion criteria
Exclusion criteria: Conference abstracts are not accepted. Systematic therapy without ICI is not allowed. Multimodal treatment combined with radiotherapy is excluded. Studies utilising regimens containing obsolete cytotoxic agents, such as mitomycin C and vindesine, are also excluded as they do not represent current standard-of-care.
Design outcomes
Primary
| Measure | Time frame |
|---|---|
| The surrogacy of HR of PFS for HR of OS. To eliminate potential bias arising from arbitrary label-/arm-assignment, an exhaustive label-assignment method is developed. For N RCTs, all 2^N possible label-assignment patterns are evaluated. STE, correlation coefficient, and P-values are determined as the median values across all 2^N sign-permutations. Surrogacy is evaluated using the weighted Peason's correlation coefficient (r). According to the generic inverse variance method, the weight assigned to each study is determined by the inverse variance of the natural log HR of survival, where the variance is the squared standard error. The correlation is interpreted as follows: no correlation (|r| < 0.2), weak (0.2 < |r| < 0.4), moderate (0.4 < |r| < 0.6), strong (0.6 < |r| < 0.8), very strong (0.8 < |r| < 0.9), or excellent (0.9 < |r|). Statistical significance is determined by a Z-statistic by dividing the random-effects meta-regression slope coefficient by its standard error, which inherently incorporates both within-trial sampling variances and between-trial heterogeneity (tau^2). The corresponding P-value is derived from this Z-statistic. A bivariate random-effects meta-analytic model is utilized to jointly model treatment effects on both endpoints and to estimate their variance-covariance structure. From this joint framework, the trial-level surrogate regression line and its 95% prediction interval (PI) are derived. The Surrogate threshold effect (STE) is defined as the threshold where this 95% PI crosses the line of no effect (HR = 1.0). This structural model-based approach is prioritized over ordinary univariate mixed-effects meta-regression to ensure a more robust estimation of the surrogacy relationship. | — |
Countries
Japan
Contacts
Public ContactNobuyuki Horita
Yokohama City University Hospital Chemotherapy Center
Outcome results
None listed