Researchers at King’s College London have successfully employed artificial intelligence (AI) to predict the survival rates of patients diagnosed with glioblastoma, an aggressive and often fatal form of brain cancer. Glioblastoma patients typically have a survival time of 12 to 18 months, with only 25% surviving beyond one year and a mere 5% surviving beyond five years. The AI system was trained on a dataset comprising 10,000 MRI scans of various cancer types and demonstrated its effectiveness by analyzing a single MRI scan. The primary objective of this AI application is to determine, post-radiotherapy, whether a patient will survive at least eight months. This prediction allows doctors to explore alternative treatments for patients with a poor prognosis.
Glioblastomas are considered the most lethal type of brain cancer, and their fast-growing cells invade nearby brain tissue, making them challenging to remove. The new AI-driven approach aims to streamline the assessment of patient outcomes, eliminating the need for multiple scans and expediting survival predictions. The method is particularly valuable in identifying patients who may benefit from early second-line treatment or enrollment in clinical trials based on their initial treatment response. Dr. Thomas Booth, involved in the research, emphasized that the AI provides an immediate and accurate prediction, empowering clinicians to guide patients in making informed decisions about their treatment. This breakthrough represents a significant advancement in cancer research, leveraging cutting-edge AI to enhance the prediction and understanding of survival rates for patients with glioblastoma.