Yatiri Bio Achieves 100% Accuracy in Predicting AML Drug Response Using AI-Driven Biomarker Platform
核心洞察
Yatiri Bio (搜索)'s AI-driven ProteoCharts™ platform demonstrated 100% concordance between biomarker-guided predictions and clinical outcomes in a blinded study of Foghorn Therapeutics' FHD-286 for relapsed/refractory AML and MDS patients.
The validation study analyzed eight blinded pre-treatment blood samples from a Phase 1 trial, correctly identifying patients who achieved complete or partial responses versus those with stable disease or treatment failure.
Following the successful validation, Yatiri Bio (搜索) and Foghorn Therapeutics signed a collaboration agreement worth over $40 million in clinical and sales milestones to advance FHD-286 development using biomarker-guided patient selection.
Yatiri Bio (搜索) has achieved a significant milestone in precision oncology by demonstrating 100% accuracy in predicting patient responses to Foghorn Therapeutics' FHD-286 (camibirstat) using its AI-driven proteomic platform. The validation study represents a breakthrough in oncology patient stratification, with potential to transform how clinical trials are designed and conducted.
Breakthrough Validation Study Results
In a blinded retrospective clinical validation study, Yatiri Bio (搜索)'s ProteoCharts™ platform analyzed eight pre-treatment peripheral blood mononuclear cell (PBMC) samples from patients with relapsed or refractory acute myeloid leukemia (R/R AML) or myelodysplastic syndrome (MDS) enrolled in Foghorn's Phase 1 trial (ClinicalTrials.gov identifier: NCT04891757). The platform achieved 100% concordance between biomarker-guided predictions and observed clinical responses.
The AI platform correctly identified patients who achieved complete or partial responses (CR/PR) and accurately classified patients who experienced stable disease or treatment failure. Notably, all predictions were generated exclusively from pre-dose peripheral blood samples, without access to additional patient information, including mutational status, dosing, trial arm assignment, or clinical outcomes.
Advanced AI-Driven Platform Technology
Yatiri Bio (搜索)'s AML ProteoCharts™ represents a molecularly defined disease landscape that integrates unbiased proteomic profiles from AML patient samples with associated clinical metadata. The platform conducts drug testing on proprietary patient-derived cellular models with proteomic profiles of the drug-perturbed systems. These patient-derived cellular models capture patient heterogeneity beyond other commercially available cell lines, allowing for highly refined patient clusters.
Using deep neural networks, the company developed an AML model specifically evaluated with FHD-286. This approach identified a distinct set of biomarkers for FHD-286 that accurately stratify patients according to therapeutic sensitivity.
Target and Mechanism of Action
FHD-286 is a novel, potent, and selective inhibitor of the ATPase subunits of the BAF complex (搜索), specifically targeting SMARCA4 (搜索) and SMARCA2 (搜索). Inhibition of this complex has demonstrated cytotoxic activity in selected AML cell lines in preclinical models and has been associated with clinical responses in the R/R AML and MDS Phase 1 trial.
Strategic Partnership and Commercial Implications
Following the successful validation study, Yatiri Bio (搜索) and Foghorn Therapeutics have signed a collaboration agreement for Foghorn to supply material for the further clinical development of FHD-286. The terms of the agreement include clinical and sales milestones valued at over $40 million.
"Our mission is to match the right patients to the right therapies before clinical trials begin," said Dr. Jackson. "As the biotech industry becomes more capital disciplined, eliminating clinical ambiguity through more precise patient stratification is not optional, it is essential."
Broader Platform Expansion Plans
Yatiri Bio (搜索) plans to expand its ProteoCharts™ platform across hematologic and solid tumors in 2026, including MDS, ovarian cancer, prostate, breast, and colorectal cancer (CRC). The company is actively advancing partnerships with biopharmaceutical companies seeking to de-risk clinical development by integrating biomarker-based patient selection into early-stage trials.
The validation study results suggest that patient selection based on predictive biomarkers could have significantly altered the observed response rate in the original trial, highlighting the potential for improved clinical outcomes through precision medicine approaches. This breakthrough validates the platform's reliability for patient stratification and its potential to inform targeted clinical development across multiple cancer types.