Apparent diffusion coefficient benchmarking and inter-scanner variability; preparing for biological image guided adaptive radiotherapy treatments.

Introduction: Biological Image Guided Adaptive Radiotherapy Treatments (BIGART) rely on the ability to utilise quantitative imaging biomarkers, qMRI. Our aim is to evaluate the performance of MRI scanners in terms of apparent diffusion coefficient (ADC) values, a form of qMRI, and compare with reference data, as a prelude for BIGART.

Methods: ADC values for various materials were measured by two MRI scanners (Siemens SOLA 1.5T) using a dedicated phantom developed by CaliberMRI. The data acquisition followed the protocol developed by QIBA and the data analysis carried out using the QCAL-MR® software platform.

Results: The scanners produced ADC values for different materials that had a mean deviation of less than 4 % from those certified by the National Institute of Standards and Technology (NIST, USA), and were within 3.6 % between each scanner (inter-scanner variability).

Conclusion: It is feasible to benchmark the performance of MRI scanners vis-à-vis qMRI data, such as those inferred from DWI, i.e. ADC. Moreover, our investigation showed that the ADC values produced by two MRI scanners conform to the QIBA brain profile, thus offering the assurance that "A measured change in the ADC of a brain lesion of 11 % or larger indicates that a true change has occurred with 95 % confidence".⁶ IMPLICATIONS FOR PRACTICE: Our work encourages the development of multicentre BIGART trial protocols that utilise qMRI biomarkers, like ADC, for the treatment of de novo GBM or other tumours. These qMRI biomarkers can be utilised for the personalisation of radiotherapy treatments, either upfront or during the course of radiotherapy, by adapting the treatment plan to the patient's response.