Drug Response Profiling with AI Identifies Optimal Drug Combinations for Children with Relapsed ALL
Key Insights
A large-scale European study across 17 countries used drug response profiling (DRP) and AI to analyze drug efficacy in 340 children with acute lymphoblastic leukemia (search), generating approximately 135,000 individual data records.
Around two-thirds of patients who received DRP-based therapy responded, a notable success given the treatment-resistant nature of relapsed ALL.
Tyrosine kinase inhibitors emerged as promising agents, and DRP delivered actionable results within one to two weeks to guide clinical decision-making.
A landmark study led by researchers at the University Children's Hospital Zurich (search) has demonstrated that drug response profiling (DRP) combined with artificial intelligence can identify the most effective drug combinations for children with acute lymphoblastic leukemia (search) (ALL), particularly those who have experienced treatment failure or relapse. The study, which analyzed data collected from 2016 to 2023 across 17 European countries, represents one of the most ambitious efforts to bring functional precision oncology into routine clinical decision-making for pediatric leukemia.
The research, led by Fabio Steffen, head of the functional precision oncology division at the University Children's Hospital Zurich (search), encompassed data from 340 patients. The logistical undertaking required samples and data from hospitals to be recorded, delivered, analyzed, and fed back to treating physicians virtually in real time, forming the basis for clinical decisions.
How Drug Response Profiling Works
The DRP platform integrates high-throughput laboratory automation with machine learning to evaluate how individual drugs and drug combinations affect patient-derived cancer cells. A pipetting robot injects drugs in the form of tiny droplets onto prepared tumor samples. After 72 hours, a fluorescence microscope captures images of the cells, and machine learning algorithms quantify how many cancer cells have been destroyed by each medication.
In total, the study generated approximately 135,000 individual data records, yielding several key findings. Notably, various tyrosine kinase inhibitors emerged as promising drugs for targeting cancer cells. Steffen offered a vivid analogy for their mechanism: they operate like a police officer who brings traffic back under control at a defective traffic light that is only displaying green because of the disease, thereby stopping leukemia cancer cells from multiplying.
Clinical Impact and Response Rates
One of the critical practical advantages of DRP is its turnaround time. The study confirmed that drug response profiling could deliver results within one to two weeks, enabling timely therapy modifications. Around two-thirds of patients who received DRP-based therapy responded to it. "This rate can be deemed a success when you consider this disease is very resistant to treatment," said Steffen.
The researchers caution against drawing premature conclusions given the heterogeneity of therapies and timescales involved. However, for individual drugs, Steffen noted that "thanks to drug response profiling, we're seeing the first indications of an actual improvement in the clinical benefit," particularly when a drug is used as bridging therapy to minimize the number of leukemia cells in the bone marrow ahead of subsequent stem cell treatment or immunotherapy.
Functional DRP Twins and Pattern Recognition
An intriguing finding from the study involved what the researchers term "functional DRP twins"—patients whose cells react to a treatment in a very similar way, even though their genetic characteristics do not necessarily suggest such similarity. This observation underscores the added value of functional profiling beyond genomic data alone in guiding therapy selection.
The Broader Context of Pediatric ALL Treatment
Acute lymphoblastic leukemia (search) is the most common form of cancer in children, accounting for approximately 60 cases per year in Switzerland. While cure rates today reach around 90 percent in children, outcomes depend heavily on the child's age and the specific ALL subtype. Conventional chemotherapies, though effective, carry significant toxicity and long-term effects. According to Jean-Pierre Bourquin, pediatric oncologist and head of the National Center of Competence in Research (NCCR) Children & Cancer, around one-fifth of patients require long-term medical care after their leukemia has been cured due to issues such as bone damage and learning or concentration disorders.
"It's no longer about how we can cure the disease, but rather how we can treat it as gently as possible and avoid serious long-term effects," Bourquin stated. He added that many young patients could undergo less intensive treatment if it were possible to predict more accurately who would respond well to a therapy.
Expanding the Platform and Building Infrastructure
Following the success of this research, the University Children's Hospital Zurich (search) is expanding the DRP program to include other types of cancer, such as brain tumors and bone tumors. To make this method routine in the future, the team is collaborating with the BioVisionCenter (search) at the University of Zurich to devise methods for processing and using vast amounts of data on a terabyte scale more easily and efficiently.
Steffen expressed confidence in the approach: "The combination of data-driven precision oncology and AI presents huge potential, and drug response profiling is an important piece of the jigsaw for managing complex oncology cases."