Summary:
In oncology clinical trials, a treatment’s fate often hinges on a single question: did the tumor shrink? Standardized frameworks like RECIST and RANO were built to answer that question consistently. Blinded Independent Central Review (BICR) adds a further safeguard, with two independent radiologists reading the same images blind to treatment assignment, and a third adjudicator stepping in when they disagree. Yet even within this carefully designed system, reader variability persists, showing up across response criteria, cancer types, and imaging modalities, and quietly shaping trial outcomes. This webinar digs into why that variability exists, what’s being done about it, and where the field may be headed next.

3 Key Questions We Will Explore:

1. Tackling variability head-on: How is MERIT approaching reader variability in practice? What mitigation strategies are actually being implemented today?

2. The AI question: Can artificial intelligence read oncology images in this context? And more importantly, can it help readers reach higher agreement, not just replace them?

3. A persistent problem, or a solvable one? Is reader variability an unavoidable byproduct of BICR’s structure and the inherent ambiguity of disease presentation, or could emerging technology and smarter mitigation strategies eventually close the gap?