RT Journal Article SR Electronic T1 Multi-parametric FDG PET/MRI as an Early Predictor of Response to Neoadjuvant Chemotherapy in Patients wit Epithelial Ovarian Cancer JF Journal of Nuclear Medicine JO J Nucl Med FD Society of Nuclear Medicine SP 397 OP 397 VO 59 IS supplement 1 A1 Franceschi, Ana A1 Pothuri, Bhavana A1 Frey, Melissa A1 Chandarana, Hersh A1 Jackson, Kimberly A1 Friedman, Kent YR 2018 UL http://jnm.snmjournals.org/content/59/supplement_1/397.abstract AB 397Purpose: There is limited data regarding how many cycles of chemotherapy are optimal prior to debulking surgery in metastatic ovarian cancer. Furthermore, early identification of non-responders would prompt discontinuation of chemotherapy and earlier surgical management. The purpose of our study was to investigate the performance of FDG PET, dynamic contrast-enhanced (DCE) and intra-voxel incoherent motion (IVIM) MR as early predictors of treatment response in ovarian cancer. Parametric images of molecular diffusion restriction (D), tissue perfusion (D[asterisk]), vascular volume fraction (F), blood->interstitium constant of transfer (Ktrans), interstitum->plasma constant of transfer (Kep), extravascular/extracellular volume % (Ve) and plasma volume % (Ve) were investigated along with routine measures of SUV and ADC. Materials & Methods: Five subjects with a new diagnosis of epithelial ovarian cancer enrolled in the study. All subjects underwent 3 cycles of standardized chemotherapy followed by cytoreduction (debulking surgery). FDG PET/MR including DCE and IVIM was performed at baseline (T1), after cycle 1 (T2) and after cycle 3 (T3) of chemotherapy. Final responses were categorized at T3 by RECIST 1.1. Olea 3.0 software was used to generate parametric images from the multi-B-value DWI and DCE-MR datasets at all three timepoints. Parametric DICOM images were then co-registered to anatomical datasets using MIMvista and fusion was manually adjusted to optimize co-registration of tumor lesions across the multiple datasets. VOIs were manually drawn on clearly visible solid tumor deposits on PET, DCE-MR and DWI MR images. The parametric images derived from IVIM and DCE-MR at T2 were analyzed as early predictors of final response. Results: Five subjects completed FDG PET and IVIM-MR, three of which underwent DCE-MR. All subjects were partial responders by RECIST at T3. SUV values were only available for 4/5 patients due to technical difficulties and DCE-MR was only available for 3/5. All 5 subjects had good IVIM data. At T2, the SUVmax decreased on average by -39% across all subjects (p<0.001) and the SUVmean decreased on average by -43% across all subjects (p<0.001). At T2, the ADCmean increased on average by +25% across all subjects (p<0.05). At T2, the molecular diffusion restriction (D) increased on average by +43% across all subjects, approaching statistical significance (p=0.058). Furthermore, D[asterisk], F, Kep, Ktrans, and Vp increased in some subjects and decreased in others, without any recognizable pattern. Ve decreased in 3/3 patients, however, not reaching statistical significance. Conclusions: In this current FDG PET/MR study of ovarian cancer, SUVmax and ADCmean values obtained after one cycle of chemotherapy were consistently associated with partial anatomical treatment responses at end of therapy. These findings are in agreement with pre-existing literature studying the value of SUV and ADC in early treatment response assessment. Only one of seven advanced perfusion/diffusion metrics (D; molecular diffusion restriction) was reliably associated with treatment response. This finding that D is associated with treatment response is not surprising given that it is based on ADC without the contribution of intravascular diffusion. Our current small dataset does not yet demonstrate the value of the remaining analyzed advanced DCE-MR and DWI parameters. Further study is required to determine the utility of DCE- and IVIM-derived parameters in early response assessment. Voxelwise correlative studies and other advanced data processing methods are underway to determine if these advanced quantitative parameters may provide further information in the early assessment of chemotherapy treatment response.