Brightfield-Only AI Readouts for T-Cell Killing Assays: Label-Free, Segmentation-Free Hit Calling
November 2, 2026
Immune-cell therapeutics such as TCR-T cells and bispecific antibodies must show robust tumor-cell killing in co-culture assays. Conventional assays rely on fluorescent labels and segmentation-based image analysis, making them prone to phototoxicity, bleaching, segmentation errors and constant re-tuning of analysis parameters. We present a hands-free, scalable AI workflow that analyses brightfield images alone, with no fluorescent markers. Benchmarked against semi-manual, segmentation-based analysis of fluorescent images in a T-cell mediated killing assay, it performs well on phenotypically diverse cancer cells, removes manual adjustment steps and delivers results of equivalent consistency.