Blab AI
Quantum-Based AI Technique Enhances Neuroblastoma Outcome Prediction
In a breakthrough that blends the strange principles of quantum physics with the pressing needs of pediatric oncology, researchers have demonstrated that a quantum‑inspired algorithm can sharpen predictions of treatment success for children with neuroblastoma, the most common cancer in infants.
The study, published on 22 June 2026 in Applied Physics Letters (APL) Quantum, introduces a multitensor comparative spectral decomposition algorithm that parses high‑dimensional molecular data even when only a handful of patient samples are available. By treating each patient’s DNA and RNA profiles as a multi‑dimensional tensor, the method applies quantum concepts of entanglement and superposition to isolate linked patterns across tumor DNA, circulating tumor DNA, and tumor RNA.
Neuroblastoma develops when early nerve cells grow out of control. Patients are stratified into low, intermediate, and high‑risk groups, yet outcomes can vary wildly within each category. Traditional machine‑learning models struggle here because the number of genetic and transcriptomic variables—often in the millions or billions—far exceeds the modest sample sizes of 20 to 100 patients that clinical trials can provide. This imbalance hampers the models’ ability to learn reliable predictors.
Orly Alter of the University of Utah and her collaborators turned to the mathematics of quantum mechanics to overcome this barrier. Their spectral decomposition algorithm can dissect the full breadth of a patient’s molecular signature while remaining robust to small cohorts. Applying the method to publicly available neuroblastoma datasets, the team uncovered two novel predictors of overall survival that consistently outperformed established biomarkers derived from tumor or blood DNA alone. The markers maintained their predictive power across independent cohorts drawn from different hospitals and time periods, suggesting that the algorithm generalizes beyond the training data.
Beyond prognosis, the researchers validated the biological relevance of their findings experimentally. In separate studies on adult glioblastoma, the algorithm’s predictions of patient response and drug targets were confirmed using CRISPR‑Cas9 gene‑editing in pre‑clinical models. Because the algorithm is interpretable, researchers can trace each predictor back to specific genes or pathways, paving a clear path toward targeted therapies.
Alter’s work is being commercialized through Prism AI Therapeutics, a University of Utah spinoff in which she serves as chief scientific officer. The company applies the multitensor algorithm to help pharmaceutical and biotechnology firms identify patient subgroups most likely to benefit from clinical trials and to pinpoint genes that could be targeted to improve outcomes.
The authors emphasize that the algorithm is data‑agnostic and could be applied to other diseases—and even non‑medical domains such as sustainable energy. They also note that the ultimate goal is to use the method on a single patient’s data to recommend a personalized treatment plan.
The study was funded by the National Institutes of Health, the National Science Foundation, and several cancer‑focused foundations. While the algorithm shows promise, further work is needed to validate its performance in prospective clinical trials and to integrate it into routine diagnostic workflows.
In short, the quantum‑based multitensor approach offers a new way to extract actionable insights from complex, high‑dimensional data sets in small‑cohort studies. If its predictive accuracy is confirmed in larger, prospective trials, the method could become a valuable tool for precision oncology and drug development.
The study, published on 22 June 2026 in Applied Physics Letters (APL) Quantum, introduces a multitensor comparative spectral decomposition algorithm that parses high‑dimensional molecular data even when only a handful of patient samples are available. By treating each patient’s DNA and RNA profiles as a multi‑dimensional tensor, the method applies quantum concepts of entanglement and superposition to isolate linked patterns across tumor DNA, circulating tumor DNA, and tumor RNA.
Neuroblastoma develops when early nerve cells grow out of control. Patients are stratified into low, intermediate, and high‑risk groups, yet outcomes can vary wildly within each category. Traditional machine‑learning models struggle here because the number of genetic and transcriptomic variables—often in the millions or billions—far exceeds the modest sample sizes of 20 to 100 patients that clinical trials can provide. This imbalance hampers the models’ ability to learn reliable predictors.
Orly Alter of the University of Utah and her collaborators turned to the mathematics of quantum mechanics to overcome this barrier. Their spectral decomposition algorithm can dissect the full breadth of a patient’s molecular signature while remaining robust to small cohorts. Applying the method to publicly available neuroblastoma datasets, the team uncovered two novel predictors of overall survival that consistently outperformed established biomarkers derived from tumor or blood DNA alone. The markers maintained their predictive power across independent cohorts drawn from different hospitals and time periods, suggesting that the algorithm generalizes beyond the training data.
Beyond prognosis, the researchers validated the biological relevance of their findings experimentally. In separate studies on adult glioblastoma, the algorithm’s predictions of patient response and drug targets were confirmed using CRISPR‑Cas9 gene‑editing in pre‑clinical models. Because the algorithm is interpretable, researchers can trace each predictor back to specific genes or pathways, paving a clear path toward targeted therapies.
Alter’s work is being commercialized through Prism AI Therapeutics, a University of Utah spinoff in which she serves as chief scientific officer. The company applies the multitensor algorithm to help pharmaceutical and biotechnology firms identify patient subgroups most likely to benefit from clinical trials and to pinpoint genes that could be targeted to improve outcomes.
The authors emphasize that the algorithm is data‑agnostic and could be applied to other diseases—and even non‑medical domains such as sustainable energy. They also note that the ultimate goal is to use the method on a single patient’s data to recommend a personalized treatment plan.
The study was funded by the National Institutes of Health, the National Science Foundation, and several cancer‑focused foundations. While the algorithm shows promise, further work is needed to validate its performance in prospective clinical trials and to integrate it into routine diagnostic workflows.
In short, the quantum‑based multitensor approach offers a new way to extract actionable insights from complex, high‑dimensional data sets in small‑cohort studies. If its predictive accuracy is confirmed in larger, prospective trials, the method could become a valuable tool for precision oncology and drug development.
