An international team led by researchers from St. Jude Children's Research Hospital in the United States has developed an innovative AI algorithm named M-PACT, which can diagnose and classify pediatric brain tumors with an impressive accuracy of 92% through the analysis of spinal fluid alone, eliminating the need for invasive surgical procedures. The study was published in the British journal Nature Cancer, highlighting the use of this technology to analyze circulating tumor DNA found in cerebrospinal fluid. The approach employs deep learning neural networks to detect molecular patterns, offering a groundbreaking alternative to traditional diagnostic methods.

Traditionally, diagnosing brain tumors involves surgical biopsies of tumor tissue, a procedure that carries significant risks, especially for children. The new algorithm enables precise molecular identification of tumors and allows for monitoring the tumor's response to treatment, as well as detecting potential recurrences through what is known as 'liquid biopsy.' This less invasive method could significantly improve the quality of care for pediatric patients, reducing the physical and psychological toll of surgical interventions.

Researchers have trained and validated the model on hundreds of pediatric patient samples, demonstrating its high efficacy and potential impact on clinical practice. The advancement represents a significant leap forward in pediatric oncology, blending cutting-edge AI technology with critical healthcare challenges and opening pathways for more personalized and less invasive treatment options for children facing brain tumors.