On April 18, 2023 Notable Labs, Inc. ("Notable"), a clinical stage therapeutic platform company developing predictive precision medicines for cancer patients, reported clinical data regarding its Predictive Precision Medicine Platform (PPMP) at the 2023 American Association for Cancer Research (AACR) (Free AACR Whitepaper) Annual Meeting being held in Orlando, Florida from April 14-19, 2023 (Press release, Notable Labs, APR 18, 2023, View Sourcenbl-validation-aacr-2023/" target="_blank" title="View Sourcenbl-validation-aacr-2023/" rel="nofollow">View Source [SID1234630249]).

Schedule your 30 min Free 1stOncology Demo!
Discover why more than 1,500 members use 1stOncology™ to excel in:

Early/Late Stage Pipeline Development - Target Scouting - Clinical Biomarkers - Indication Selection & Expansion - BD&L Contacts - Conference Reports - Combinatorial Drug Settings - Companion Diagnostics - Drug Repositioning - First-in-class Analysis - Competitive Analysis - Deals & Licensing

                  Schedule Your 30 min Free Demo!

"Together with our collaborators at Washington University, today we are reporting on the fourth successful validation study of our Predictive Precision Medicine Platform (PPMP)," said Thomas Bock, M.D., Chief Executive Officer of Notable. "The study assessed our platform’s accuracy in predicting whether a patient will clinically respond to their induction chemotherapy for acute myeloid leukemia (AML). We are excited about these clinical results as they not only corroborate, but expand upon, those of our three other validation trials. Using a specially designed approach to training our machine learning algorithm, PPMP achieved 100% accuracy in its predictions for response to venetoclax plus decitabine (VenDec). That is, all patients predicted to respond clinically actually did while those patients predicted not to respond, did not. These results add further validation and promise to Notable’s strategy of de-risking precision medicines and developing them selectively in patients predicted to clinically respond."

Abstract title: Predictive precision medicine platform accurately predicts individual patient response to AML treatments to maximize outcomes.

This study assessed the capacity of Notable’s Predictive Precision Medicine Platform to accurately predict newly diagnosed AML patients’ response to treatment with cytarabine plus idarubicin (7+3) or VenDec. Employing two different training methods, the predictive algorithm assessed pre-induction blood samples from 31 patients, 18 of whom received 7+3 and 13 of whom received VenDec. The "original" training approach bases predictions on the number of live blasts remaining after treatment with the induction therapies, while the enhanced PPMP approach employs a novel machine learning method and is explicitly designed to maximize the accuracy of predictions for VenDec. The study assessed the correlations between predictive and actual outcomes using four metrics: positive predictive value (PPV/predictive precision, the proportion of predicted responders who actually responded), the negative predictive value (NPV, the proportion of predicted non-responders who, in fact, did not respond), the area under the curve (AUC) of the Receiver Operating Characteristic (ROC) curve (the probability that the predictor correctly ranks a randomly chosen responder higher than randomly chosen nonresponder), and accuracy (the proportion of correct predictions out of all predictions).

For both 7+3 and VenDec, PPMP trained using the original method achieved a PPV of 100%, i.e., all predicted responders actually responded. A PPV of 100% indicates, for example, that a clinical trial selectively enrolling predicted clinical responders would result in a 100% response rate. The NPV was 67% for 7+3 and 57% for VenDec (i.e., 67% of patients on 7+3 and 57% on VenDec who were predicted not to respond, did not). Trained using the original method AUC was 0.91 for 7+3 and 0.81 for VenDec, and the accuracy of this approach was 94% for 7+3 and 77% for VenDec.

To further increase the accuracy on VenDec, a machine learning algorithm integrated the behavior of malignant and non-malignant cell populations and examined responses to the therapeutics along multiple biological dimensions. This novel enhanced method resulted in 100% PPV, 100% NPV and 100% accuracy on VenDec. These compelling results highlight the platform’s potential as a tool for guiding the identification of, the decision-making regarding, and the clinical development of optimal AML therapies for the individual patient. Additional meeting information can be found on the AACR (Free AACR Whitepaper) website, View Source The poster will be available on the Company’s website at View Source shortly after the event.