Glioblastoma cancer: comparing the effectiveness of 3D conformal radiation therapy and volumetric modulated radiotherapy - an artificial intelligence-based survival prediction

This study aimed to evaluate the survival efficacy of different radiotherapy strategies in patients with glioblastoma (GBM). Furthermore, by utilizing various artificial intelligence algorithms and machine learning models, including neural networks, logistic regression, and decision trees, among others, the possibility of obtaining outcome predictions for a specific point in time was explored. The study considered data from radiotherapy treatments for patients affected by GBM. Eligible data included patients treated with 3D conformal radiotherapy or intensity-modulated radiotherapy, who reported overall survival and progression-free survival. The impact of different radiotherapy modalities on survival was evaluated through direct comparisons of the available data. In the second part of the study, the possibility of using artificial intelligence to predict the survival status of patients after a specific period following the end of radiation treatment was explored. To test our hypothesis, we used data from new patients and asked the machine learning models with the best fit to our data to predict survival for these new patients. A total of 30 elderly GBM patients treated with modern radiotherapy strategies were examined, showing a better overall survival when volumetric modulated radiotherapy (VMAT) was used compared to the 3D conformal radiation therapy technique. The artificial intelligence algorithm was asked to predict the survival status of three new patients. The neural network method, compared to the others used, is the one that responded correctly in 100% of the cases submitted. In second place was the decision tree method, which responded correctly in 67% of the cases. Our results suggested VMAT as a standard radiotherapy modality with potentially superior survival outcomes for selected patients with GBM.