Functional precision medicine has evolved significantly as a strategy to individualize treatment for patients with limited options. While ex vivo drug sensitivity testing has existed for decades, historical adoption was hindered by sample scarcity and inconsistent clinical correlation. However, recent real-world evidence of Optim.AI™ demonstrates a paradigm shift in managing hematological malignancies, particularly in complex cases of AML.
Clinical Methodology and Workflow
The Optim.AI™ platform utilizes a combinatorial approach to evaluate therapeutic efficacy. The workflow begins with tumor cell isolation from diverse sample types, including bone marrow and peripheral blood. Following isolation, patient cells undergo a 48-hour combinatorial drug treatment using a custom panel of up to 12 drugs.
Viability is subsequently quantified and analyzed via the Optim.AI™ algorithm to generate a prioritized list of treatment options. This rapid turnaround is critical for clinical utility; the study reported a median turnaround time of just 4 working days from sample receipt to report generation.
Performance Outcomes in a Clinical Laboratory
The study evaluated 169 samples, demonstrating high technical success rates. Of the samples received, 91% contained sufficient viable cells for processing. Within the viable cohorts, reports were successfully generated for 98% of AML samples and 96% of Non-Hodgkin’s Lymphoma (NHL) samples.
To assess clinical relevance, the platform used a Normalized Cell Viability (NCV) cut-off of 0.5. The results revealed:
-
High Concordance: 92% of 88 evaluated cases demonstrated concordance with prior lines of treatment ($p < 0.0001$).
-
Predictive Accuracy: The platform achieved an overall accuracy of 86% in predicting clinical responders.
-
Sensitivity and NPV: The analysis showed 100% sensitivity and a 100% Negative Predictive Value (NPV), indicating the platform’s strength in identifying ineffective therapies.
Clinical Utility in AML and Off-Label Combinations
Optim.AI™ identifies effective therapeutic options even when standard-of-care (SOC) therapies fail. The platform’s ability to identify “top-ranked off-label combinations” provides a vital roadmap for clinicians treating refractory patients.
For instance, Case Study 24-0079 involved a patient with MDS-AML who showed insufficient response to standard FLAG-Ida and Venetoclax regimens. Optim.AI™ guided the administration of a combination involving Bortezomib, Idarubicin, and Venetoclax, leading to a complete response and successful transplantation.
Conclusion
The real-world evidence of Optim.AI™ confirms its ability to predict treatment responses with high statistical significance. By providing actionable insights within a clinically relevant timeframe, the platform aids in both identifying intensive therapy options and guiding appropriate palliative care.