ESTRO 38 Abstract book

S240 ESTRO 38

quantitative imaging measure, particularly in multisite trials. This motivated the Quantitative Imaging Biomarker Alliance (QIBA), which initially formed in 2007, to write one of their earliest profiles on DCE-MRI. The DCE-MRI Quantification QIBA Profile places requirements on the acquisition device, the actors (technologists, radiologists, physicists), reconstruction software and image analysis tools involved in image acquisition, image data reconstruction, and image analysis. AS QIBA recognized that the clinical and research environment for these quantitative imaging tools is ever evolving, the Profile sits online as a living document and there are current ongoing revisions to address the increasing use of 3.0T magnets and parallel imaging. The Profile requirements are determined based on published literature on repeatability and reproducibility. Despite a long history of interest by the oncology community in using DCE-MRI as a quantitative imaging biomarker, the recent systemic review of clinical literature by the DCE-MRI Committee within QIBA discovered a very limited number of publications on repeatability and reproducibility data. In the era of precision medicine, the community-at-large is starting to understand that variability in measurement limits our ability to quantitatively characterize tumors or evaluate treatment response. Only through standardized processes of image acquisition, quality assessment, processing and analysis will there be the potential for collaborative, quantitative investigation. As part of this standardization process, the community will need to come together to determine shared nomenclature and define reporting standards that ensure enough transparency to enable cross-validation as well as supportive data collection, including repeatability and reproducibility data. SP-0467 Quality assurance for quantitative MRI in a multicenter trial P. Van Houdt 1 1 The Netherlands Cancer Institute, Radiation Oncology, Amsterdam, The Netherlands Abstract text Quantitative MRI (qMRI) is promising as a biomarker for prediction of treatment outcome and for treatment response monitoring. However, current evidence is mainly built on small patient cohorts. Large, multi-center studies are necessary to investigate the clinical application of qMRI. However, these studies typically involve a wide variety in MRI systems, with different vendors, field strengths, and generations. As a result novel sequences may be available in one institute, but not in another. One approach to reach consistency would be to use standardized MRI protocols in all centers. The drawback is, however, that this will force us to design the sequences for the oldest system. Therefore, we have built a framework where we deal with this variety by optimizing the trial sequences on each system individually. In this way all institutes are free to choose the sequence to their preference and system possibilities. We have set up a quality assurance (QA) procedure using calibration phantoms to assess consistency of the trial sequences between institutes. This includes the measurement of benchmark sequences to investigate whether deviations between institutes result from protocol differences in the trial sequences or from system variations. These benchmark sequences are well known reference standards that are available on all systems using identical parameter settings. In this talk, we will illustrate how we applied this framework in the IQ-EMBRACE trial and for treatment response monitoring studies using MR-linac systems. IQ- EMBRACE (sub-study of the EMBRACE-II trial) is a large, multi-center trial in which patients with cervical cancer will undergo an MRI exam prior to radio(chemo)therapy to investigate qMRI as a potential biomarker for treatment

skull base or retroperitoneum occupied by vulnerable anatomy. Therefore, modifications have been introduced by some groups in the interest of dose reduction to spare late morbidity. For example, a post-operative “boost” of 10-16 Gy to CTV1 to complete the 66 Gy total dose prescription was traditional in pre-operative radiotherapy but has generally been omitted in recent years when resection margins are not involved. Moreover some centres now also omit this “boost” when the margins are involved if this was a “planned positive” margin in a very small region juxtaposed to a vulnerable structure. Some colleagues, e.g. in Scandinavia and France, have also restricted post-operative radiotherapy courses to approximately 50 Gy in 25 fractions or equivalent rather than delivering an additional boost to CTV1 for a total of 66 Gy if the resection margins are uninvolved. The concept of restricting the dose to 50 Gy pre-operatively and post-operatively is currently being tested in the Canadian 50/50 randomized trial (NCT02565498). Differences in RT response of different sarcoma histological subtypes exists and has been shown to convert to improved local control in myxoid liposarcoma, with further potential for treatment with lower pre-op RT doses (36 Gy) in this pathological subtype currently under evaluation in the Doremy study (NCT02106312). There is also evidence that RT volume adaptation may be facilitated based on on-treatment response during pre- operative RT evident in patients treated in the first clinical trial of IMRT in extremity (STS NCT00188175). With daily CBCTs online, we have seen incremental changes in volume, and a 1 cm change affects dosimetry and coverage. This may be facilitated by implementation of MRgRT systems, with eventual potential generation of new IMRT plans for patients while still on treatment. Additional possibilities for RT dose reduction with shorter overall treatment time seem possible using RT regimens combined with chemotherapy or targeted agents. Finally, RT doses may be reduced through volume restriction by modification of RT delivery including use of particle beam, intra-operative RT, or pathological response following intratumoral injection of hafnium oxide nanoparticles with RT. SP-0466 A Critical Look of Quantitative Dynamic Contrast Enhanced MRI: From QIBA guidelines to Clinical Implementation C. Chung 1 1 MD Anderson Cancer Center, Radiation Oncology, Houston, USA Abstract text The promise of dynamic contrast-enhanced MRI (DCE-MRI) as a prognostic and predictive biomarker in solid malignancies is not a new concept. There have been efforts over several decades to investigate the methods and clinical value of DCE-MRI in oncological assessment. Many studies from individual institutions have suggested that DCE-MRI can characterize tumor angiogenesis and serial measurements may reflect changes in tumor vascular physiology in response to treatment or with further tumor progression. In Radiation Oncology, DCE- MRI has frequently been investigated in conjunction with tumor hypoxia, which has been associated with greater tumor aggressiveness and metastatic potential. In Medical Oncology, DCE-MRI was of particular interest when anti- angiogenic and anti-vascular therapies were discovered, but remain of interest in the context of drug delivery and early biomarkers of response. Despite strong interest and promising single site studies, there have been great challenges in using DCE-MRI as a Symposium: Quantitative Imaging for Radiation Oncology