Challenge
Multispecific therapeutic antibodies are designed to bind to multiple antigens simultaneously to improve the effectiveness of antibody-based immunotherapy for cancer. The detection of cancer cells co-expressing the target antigens is important to identify patients who are likely to respond to the drug. However, assessing co-expression using visual inspection of sequential single-plex IHCs or multiplex IHC can be challenging for pathologists.
Our client, a Top 10 biopharma company, needed to deploy an IHC assay for their Phase I multispecific trial, however, the pathologists had trouble with manual scoring because it required accurate, reproducible quantification of tumor cells co-expressing the 2 target proteins.
Solution
Our patent-pending AI-powered IHC algorithm is currently deployed in the Phase 1 trial to enable pathologists to accurately score co-expressing cells.
During the feasibility study, we helped the client decide whether to use two monoplex assays or a single duplex assay for detecting co-expression during trial enrollment (Figure 1). A multiplex IHC algorithm for color deconvolution and co-expression analysis was trained and then optimized and validated under design control with our in vitro diagnostic (IVD) partner.
Figure 1. AI-powered spatial scoring of co-expression and colocalization can optimize assay design and derisk multi-specific trials
Algorithm training was followed by a prevalence study for selecting the right tumor indications. Finally, the algorithm was deployed for prospective enrollment through integration with the IVD company’s pathology platform and their CLIA-certified lab (Figure 2).
Figure 2. Clinical deployment workflow a novel AI-powered duplex IHC assay for a Phase I trial through the IVD partner’s CLIA lab
Our three-way partnership with a Top 10 Biopharma and a leading IVD pathology company has established a pioneering, industry-leading framework for deploying AI-powered IHC assays for prospective trial enrollment.