ESTRO 2021 Abstract Book

S501

ESTRO 2021

Conclusion This work will prospectively allow us to discriminate which patients are not meeting dose constraints with a 3D-CRT plan and would therefore benefit from an IMRT. This decision-making tool would impact our treatment planning times since it provides a better allocation of resources. We want to emphasize that our CNN model, which had been previously tested for internal validity, has shown excellent external validity for the first time. OC-0636 Utilising diffusion-weighted MRI for isotoxic dose escalated radiotherapy for glioblastoma Y. Pang 1 , M. Kosmin 2 , Z. Li 3,3 , X. Deng 3 , X. Li 3 , Y. Zhang 4 , B. Ding 4 , G. Royle 1 , S. Manolopoulos 1 1 University College London, Medical Physics And Bioengineering, London, United Kingdom; 2 UNIVERSITY COLLEGE LONDON HOSPITALS NHS FOUNDATION TRUST, Department Of Oncology, London, United Kingdom; 3 The 2nd Clinical Medical College (Shenzhen People’s Hospital) Of Jinan University, Department Of Radiation Oncology, Shenzhen, China; 4 University College London, Department Of Medical Physics And Biomedical Engineering, London, United Kingdom Purpose or Objective Glioblastoma (GBM) is the most common malignant primary brain tumour. Long term control is rarely achieved despite aggressive multimodality therapy, with local recurrence after radiotherapy (RT) the most common mode of failure. In standard RT practice, the prescription dose is applied uniformly across the target volume disregarding radiological tumour heterogeneity. We present a novel strategy for the use of diffusion- weighted (DW-) MRI to calculate the cellular density within the gross tumour volume (GTV) in order to facilitate dose escalation to a biological target volume (BTV) to improve tumour control probability (TCP). Materials and Methods The pre-treatment apparent diffusion coefficient (ADC) maps derived from DW-MRI of ten GBM patients treated with standard radical chemoradiotherapy (60Gy in 30 fractions with temozolomide 75mg/m 2 ) were used to calculate the local (per pixel) cellular density based on published data. ADC maps were registered to CT scans to match the GTVs which have been defined for the patients' standard treatment and clinically approved. Then, the cell density distributions within the GTVs were determined and a Poissonian linear quadratic (LQ) TCP model (α=0.12Gy -1 , β=0.015Gy -2 ) was used to calculate pixelated TCP maps from the derived cell density values. The prescription dose