1057P - Blinded independent central review versus local investigator assessment of progression-free survival in randomized controlled trials of immunotherapy in advanced cancers: A systematic review and meta-analysis

Background

Progression free survival (PFS) in randomized controlled trials (RCTs) investigating the role of immunotherapy (IO) is adopted for approvals of oncology drugs. Assessment and interpretation of PFS data by investigators might be inaccurate in RCTs with open label design. Thus, we explored potential differences between blinded independent central review (BICR) and local investigator assessment of PFS in trials of IO in advanced cancers.

Methods

We systematically reviewed articles of RCTs testing IO in advanced solid tumors, published in Pubmed-indexed journals from 01/2010 to 12/2023. For each RCT reporting results for both BICR and local investigator assessment of PFS, we collected: i) The number of patients at risk; ii) PFS results by BICR and iii) by local investigators. We calculated a discrepancy index (DI) between BICR and investigator Hazard Ratios. Finally, an overall DI and relative confidence interval was calculated using a fixed model weighted for variance.

Results

Of the 141 RCTs testing IO in advanced cancers, only 32 (22.6%) reported both BICR and investigator PFS data, including 17,054 patients. PFS was the only primary endpoint or a co-primary endpoint in 19/32 (59.4%) and 9/32 (28.2%) trials, respectively. The study design was open label or double-blind in 17/32 (53.1%) and 15/32 (46.9%) RCTs, respectively. The overall DI was 1.07 (95% CI 1.01-1.13; I²=0, p=0.02), revealing a statistically significant difference between BICR and local investigator assessment of PFS, with a more optimistic analysis of results in favour of local investigator. Of note, in the subgroup of 17 open label trials the overall DI was 1.09 (95% CI 1.02 – 1.17, I²=0, p=0.02), while in the 15 double-blind RCTs the overall DI was 1.03 (95% CI 0.95 – 1.12, I²=0, p=0.51), revealing a more optimistic interpretation of PFS results by local investigators in open label RCTs.

Conclusions

This was the first study reporting a statistically significant difference between BICR and local investigator assessment of PFS in trial of IO in cancer. These results suggest that the double assessment is strongly recommended in RCTs testing IO, especially in open label trials.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

The authors.

Funding

Has not received any funding.

Disclosure

All authors have declared no conflicts of interest.