Akoya Biosciences (AKYA) investor relations material

AI strategy and clinical application

• AI-powered profiling of trillions of immune cell data points identified key predictors of cancer recurrence risk, focusing on liver cancer.

• Developed a structured approach to reduce massive spatial data to a clinically actionable five-gene immune scoring system, achieving superior predictive accuracy (AUC 0.82) over traditional prognostic factors.

• AI models were designed to be explainable and human-in-the-loop, allowing pathologists to validate and interpret results at the single-cell level.

• The immune scoring system stratifies patients into high and low recurrence risk, guiding potential adjuvant treatment decisions.

• The approach enables more efficient clinical trial recruitment by predicting biomarker positivity from standard H&E slides, reducing resource use.

Validation and mechanistic insights

• Findings were validated across RNA, protein, ex vivo, and in vivo models, confirming the role of NK cells and their interaction with CD8 T cells in preventing recurrence.

• The five-gene panel was packaged into a multiplex IHC/IF assay for clinical translation.

• Ex vivo and spatial analyses demonstrated that SPON2-positive NK cells, in proximity to interferon gamma-positive CD8 T cells, are critical for tumor cell killing.

• Knockdown experiments confirmed the mechanistic importance of NK cell-specific SPON2 in tumor infiltration and prognosis.

Technological innovation and future directions

• Introduced H&E 2.0/3.0 and virtual staining, leveraging AI to predict biomarker status from routine pathology slides.

• Developed cloud-based platforms for spatial data visualization and automated quantification.

• Advanced spatial mass spectrometry (deep visual proteomics) enabled validation and discovery of novel protein markers, including unannotated proteins (spatial dark proteomics).

• Ongoing work aims to automate image quantification and reporting, generalizing the approach for broader clinical use.

• Efforts are underway to apply these AI-driven methods to global health challenges, including projects in Africa and Southeast Asia.