Assessment of Treatment Response of Nasopharyngeal Cancer Using Simultaneous 18F-FDG-PET and MRI

The overarching goal of this study is to develop PET/MR techniques for accurate assessment of
treatment response during and immediately after chemoradiation therapy. The central
hypothesis is that the GMR measured using a simultaneous PET/MR scanner can more accurately
detect residual tumor than conventional SUV measures from PET alone. It is important to note
that SUV depends on both tumor metabolic rate and tracer delivery, which makes the
interpretation of SUV challenging. For instance, inflammatory tissue can have high SUV due to
increased vascularity and vascular permeability and cannot be easily differentiated from
tumor based on the SUV. Investigators hypothesize that inflammatory tissue will have lower
GMR than residual tumor that contain highly proliferating cells with increased expression of
glucose transporters (GLUT). Measuring GMR accurately will improve the specificity of PET
while maintaining the high sensitivity of PET for detection of residual tumor. In order to
test our hypothesis, investigators propose to conduct dynamic PET and MRI scans with NPC
patients who are undergoing a conventional two-stage chemoradiation therapy at our
institution; the first stage for 7-week chemoradiation therapy followed by the second stage
for 3-month chemotherapy. A combination of PET/CT and nasopharynx MRI is currently obtained
before the initiation of treatment and 3 months after completion of treatment to assess
treatment response.

This study proposes to introduce PET/MR scans at the time of these exams (scan #1 for
pre-treatment & scan #4 for 3 months after completion) and to add two additional PET/MR scans
in between them; one immediately after the first stage of treatment (scan #2) and another one
immediately after the second stage (scan #3). A primary clinical endpoint of this study is
the treatment response assessed at 3 months after completion of treatment. A secondary
endpoint is 6 month follow-up exam. Complete responder will be determined based on clinical
and imaging assessment of residual tumor size at each endpoint.

It is hoped that preliminary data obtained from this study will be useful in planning larger
studies to formally investigate the utility of GMR for detection of residual tumor and
prediction of treatment response.