aiPTO Announces DELPHAI AI Model for Heterogeneous Tumor Drug Responses
Published in Artificial Intelligence, Bioinformatics, Drugs development & discovery.
October 8, 2026 — aiPTO TechBio announced DELPHAI, an AI model designed to predict how heterogeneous cell populations respond to drug perturbations. The company is a spin-out of the German Cancer Research Center with a focus on brain cancer.
DELPHAI incorporates a mechanism called fitness gating to model cell-specific survival alongside changes in gene expression. The aim is to estimate which populations a treatment depletes and how the composition of a tumor may change, rather than assuming every measured cell survives.
aiPTO reports that DELPHAI ranked first for recovery of differentially expressed genes in two benchmarking frameworks and first overall in scPerturBench’s out-of-distribution setting. The company is training the model on drug-perturbation data from its Individualized Patient Tumor Organoid platform and testing predictions back in those biological models.
The cited research is a preprint. Benchmark rankings do not establish that AI-guided treatment improves patient survival or reliably selects an individual’s best therapy. The announcement concerns a preclinical research platform; clinical utility requires further validation.
Source: aiPTO TechBio announcement, October 8, 2026. Research: DELPHAI preprint.