Multi-parametric FDG PET/MRI as an early predictor of response to neoadjuvant chemotherapy in patients with epithelial ovarian cancer.

Abstract

5569 Background: For patients with ovarian cancer undergoing neoadjuvant chemotherapy, the effectiveness of treatment is not evaluable by conventional methods until all or much of the treatment has been given. The purpose of this study is to investigate the performance of FDG PET, dynamic contrast-enhanced (DCE) and intra-voxel incoherent motion (IVIM) MR as early predictors of treatment response. Methods: Subjects with a new diagnosis of epithelial ovarian cancer underwent 3 cycles of standardized chemotherapy followed by cytoreduction. FDG PET/MR including DCE and IVIM was performed at baseline (T0), after cycle 1 (T1) and after cycle 3 (T2) of chemotherapy. Final responses were categorized at T2 by RECIST 1.1. Image volumes at T1 were analyzed as predictors of final response. Parametric images of molecular diffusion restriction (D), tissue perfusion (D*), 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. Results: Nine subjects were enrolled, 8 were responders by RECIST at T2 and one had stable disease. At T0 the mean, min, and max SUVmax of dominant tumor deposits was 11.5, 6.3, 19.0, respectively. Mean, min, and max values were 1.0, 0.75 and 1.63 for ADCmean and 0.62, 0.30, 0.96 for ADCmin. At T1, ADCmean increased in 8 subjects by +0.22% (s.d. +/- 13%) and decreased by -3% in one subject. ADCmin increased in 8 subjects by +21% (s.d. +/-11%) and decreased by -23% in one subject. D increased for 8 subjects (average +29% s.d. +/- 13%) and decreased by -10% in one. D*, F, Kep, Ktrans, Ve and Vp had no recognizable pattern. At T2, SUVmax, SUVmin, and ADCmean maintained their change direction across all subjects with measurable lesions. The only subject with a complete response at T2 had the highest ADCmin and ADCmean change at +45% after one cycle of chemotherapy (T1). The subject with stable disease at T2 had no significant difference in changes amongst all metrics. Conclusions: FDG PET/MR SUVmax and ADCmean values obtained after one cycle of neoadjuvant chemotherapy were consistently associated with partial anatomical treatment responses after three cycles. Molecular diffusion restriction also was reliably associated with treatment response. Future studies evaluating FDG PET/MR in platinum-resistant patients may allow for early discontinuation of ineffective and toxic treatment.