Fragility index analysis of systemic therapy clinical trials in soft tissue sarcoma (STS).

Abstract

11581 Background: The fragility index (FI) measures the robustness of randomized clinical trials (RCT) that have positive results based on statistical p-values. It estimates the minimum number of events that are needed to reverse the trial results from positive to negative. Here, we perform FI analysis on RCTs evaluating systemic treatments for STS. Methods: A systematic search of the Medline and Embase databases for RCTs in adults with advanced STS (Jan 1998 to Dec 2023) was conducted. Gastrointestinal stromal tumor trials were excluded. The FI framework was adapted to allow the use of time-to-event (TTE) outcomes – progression-free survival (PFS) or overall survival (OS) as previously described (Desnoyers et al 2021). Survival tables were reconstructed from published data on positive TTE outcomes, using the Parmar Toolkit. Positive secondary endpoints were used only if the primary endpoint was negative. The number of additional events that would result in a non-significant effect for the hazard ratio (HR) of the positive endpoint of each trial (FI) was calculated and expressed as a proportion of the size of the experimental arm (fragility quotient; FQ). The number of censored patients – those who withdrew consent or were lost to follow-up (Cn) was noted. Results: Among 47 RCTs, 16 (8 phase II; 8 phase III) trials had positive outcomes. The primary endpoint was positive and used for FI analysis in 11/16 trials (68.8%). PFS was the most common positive outcome, evaluated in 13 trials (81.3%). The median FI was 6 (range 2 – 52), with FI < 10 observed in 11 trials (68.8%). Median FQ was 7% (range 1 – 59%) and 10 trials (62.5%) had FQ < 10%. Among 14 trials that reported data, Cn was ≥ FI in 7 trials (50%). Only 2 of 4 trials (50%) that led to regulatory approval had FI > 10 or FQ > 10%. Of the other 2, one drug approval was subsequently withdrawn (Table). Conclusions: Most positive clinical trials in STS were fragile and their outcomes may be confounded by the level of censoring. Real-world evaluation of approved systemic therapies and value scales such as ESMO Magnitude of Clinical Benefit scale or ASCO value framework are needed to confirm clinical benefit of systemic treatments in STS.[Table: see text]