Salt Lake City (LabNews Media LLC) – A new AI method based on quantum mechanics principles can derive more precise prognoses and therapy recommendations for cancer patients from complex molecular data, even with small patient groups. This is shown by a study from the University of Utah, published in the journal APL Quantum.
The team around Orly Alter developed algorithms based on the quantum mechanical concepts of superposition and entanglement. These so-called “multitensor comparative spectral decompositions” break down multiple layers of molecular data – including tumor and blood DNA as well as tumor RNA – into coherent patterns.
When applied to neuroblastoma data from 71 patients, the researchers were able to derive new predictors for the children's life expectancy that surpassed conventional biomarkers. The predictions could be successfully validated in independent patient groups. Furthermore, they provided interpretable insights into disease mechanisms and potential targets for new therapies.
“It’s about much more than just a single gene – everything happening in the patient’s cells plays a role,” explained Orly Alter. The method makes it possible to extract relevant information from different data layers, such as from the patients’ blood in addition to the tumor.
The technique has already been successfully applied to adult glioblastomas and experimentally validated with CRISPR-Cas9. Alter sees this as an important step towards precision medicine: “Ultimately, this is the ultimate personalized medicine – you have an individual person and can derive a treatment from their data.”
The researchers have incorporated the method into the spin-off company Prism AI Therapeutics to support biotech and pharmaceutical companies in drug development and the selection of suitable study participants.
https://pubs.aip.org/aip/apq/article/3/2/026116/3395875/Quantum-mechanics-based-multitensor-AI-ML-uniquely
