Automated fluorescent multiplex immunohistochemistry image analysis ～Applied to anticancer drug research and development (R&D) and spatial analysis of the tumor microenvironment.～

Challenge

Fluorescent multiplex Immunohistochemistry (IHC, a technique that uses fluorescent dye-labeled antibodies to simultaneously detect multiple molecules expressed in cancer cells, immune cells, and other tissues on a pathological specimen) enables the measurement of expression levels of multiple molecules at the single-cell level, allowing for detailed cell classification.

Daiichi Sankyo uses LPIXEL’s AI technology to automate the analysis of fluorescent immunostained tissue images focusing on Antibody-Drug Conjugate (ADC) target molecules and immune cells.

Traditionally, skilled researchers relied on conventional (non-AI-based) analysis software that required manually setting and verifying analysis parameters through visual inspection. This presented significant challenges, including lengthy analysis times and results that were dependent on the analyst’s skill and experience.

LPIXEL’s Solution – AI Technology and Impact

The automation of fluorescent multiplex IHC image analysis was achieved through the following two-step process:

Recognition of individual cell nuclei within the tissue

Classification and quantification of cell types based on marker expression in the detected cells

Analyses were conducted across multiple cancer types, including breast, colorectal, and pancreatic cancers.

Results demonstrated that the AI achieved cell nucleus recognition accuracy and quantitative cell classification performance comparable to conventional methods.

Going forward, LPIXEL will expand the range of cancer types supported and integrate analyses with other experimental data, such as spatial omics, to realize quantitative spatial analysis of the tumor microenvironment.