AI-Powered Cancer Relapse Tool Validated
CareDx announced positive clinical validation results for AlloHeme™, an AI-driven surveillance solution for AML and MDS patients after cell therapy. The tool, which uses next-generation sequencing data, reportedly detected cancer relapse earlier than standard monitoring methods. The results highlight the growing application of AI analytics in post-treatment monitoring and its potential to influence in-process analytics for cell therapies.
- Data from the ACROBAT study, a prospective trial across 11 U.S. centers, showed AlloHeme detected relapse a median of 41 days earlier than clinical methods. The study involved 198 patients with AML or MDS who had undergone allogeneic hematopoietic cell transplant. - AlloHeme demonstrated 85% sensitivity and 92% specificity in detecting relapse, with patients testing positive at 6 months post-transplant having a 12-fold higher risk of relapse. This performance is a significant improvement over standard methods like short tandem repeat (STR)-PCR analysis, which has a detection limit of 1-5% compared to AlloHeme's 0.2%. - Unlike marker-specific methods, AlloHeme is a "tumor-naive" surveillance tool, meaning it does not require identifiable genetic targets to work. This makes it universally applicable for AML and MDS patients, a key advantage in diseases known for tumor heterogeneity. - The underlying technology is based on next-generation sequencing (NGS) to monitor chimerism, which is the presence of two different sets of genetic material (donor and recipient) after a transplant. An AI-driven algorithm analyzes longitudinal data to predict relapse when it detects an increase in the recipient's own residual diseased cells. - Cancer relapse is the leading cause of death for AML and MDS patients after hematopoietic cell transplantation, with 2-year relapse rates between 30% and 45%. Early detection provides a critical window for preemptive interventions, such as tapering immunosuppression or administering donor lymphocyte infusions (DLI). - The integration of NGS-based diagnostic data into a manufacturing context requires a robust Laboratory Information Management System (LIMS). A LIMS designed for genomics can automate workflows, track samples from collection to analysis, and integrate with sequencing instruments, which is crucial for managing the large datasets generated and ensuring data integrity for electronic batch records. - This application of AI in post-transplant monitoring parallels the growing use of digital twins in cell therapy manufacturing. Digital twins—virtual replicas of manufacturing processes—use real-time data to optimize bioreactor conditions, predict equipment failure, and de-risk scale-up, reflecting a broader industry trend toward Pharma 4.0. - CareDx plans for CLIA laboratory readiness in 2026 and a U.S. commercial launch in 2027, with a focus on approximately 200 bone marrow transplant centers. This positions them in the growing cell therapy market, which is estimated to have included 20,000 patients undergoing allogeneic HCT and CAR-T therapy in 2025.
