FDG-PET has proven valuable in cervical cancer staging1 and treatment response evaluation. PET-defined cervical tumor volume can predict progression-free and overall survival2. However access to 2-[fluorine-18]fluoro-2-deoxy-d-glucose (FDG) is limited in many developing countries where cervical cancer remains a leading cause of death for women. MR is more readily available in these regions such as Latin American and Southeast Asia. We set out to evaluate if MR can accurately assess tumor volume and predict clinical response.20 subjects with locally advanced cervical cancer were enrolled in this prospective study where a MR pelvis was obtained at baseline (exam 1), after 5 weeks of chemoradiation (exam 2) and 3 months after completion of chemoradiation (exam 3). Diffusion-weighted imaging was acquired with b-values of 0, 50, and 900 s/mm2 or 0, 900, and 1500 s/mm2, but only 0 and 900 s/mm2 were used to generate the apparent diffusion coefficient (ADC) for all cases. De-identified DICOM images were uploaded into a dedicated post-processing workstation. An abdominal radiology fellow contoured cervix and tumor boundaries on the ADC images for all 3 exams for each subject using a semi-automated algorithm. Cervix volume, tumor volume, mean ADC, and median ADC were then calculated for each of the 3 timepoints for the subjects. Values from the first timepoint were compared with the second and third timepoints. Comparison between groups was analyzed using student’s t-test.Complete data sets were available for 9 subjects, which were used for analysis. An additional subject was dropped as the tumor could not be identified on the second and third time points. The mean tumor volume decreased by 90.2% after 5 weeks of chemoradiation (exam 2) and by 97.9% at 3 months after completed chemoradiation (exam 3). The mean ADC value increased by 29% at exam 2 and by 50.2% at exam 3. Of the subjects analyzed, 3 (38%) had >99% resolution of the tumor evident on ADC. At baseline, the only value which discriminated these subjects from those with measurable tumor was the standard deviation (p=0.01). However, by the second time point, the absolute and change in the mean and median value were also discriminators (p≤0.02 for all). Standard deviation of the ADC value at baseline appears to differentiate subjects with complete tumor response (no measurable tumor during treatment scans). This suggests the heterogeneity of the original tumor may predict tumor’s sensitivity to chemoradiation. Evaluation is limited by the small number of subjects and robust response (>93% tumor volume reduction) in all. We expect to receive clinical outcome data and will evaluate tumor feature analysis which will clarify predictors for chemo/radio-sensitive tumors and more aggressive tumors.While PET has been shown to be optimal in assessing tumor volume and predicting treatment response, developing countries where cervical cancer remains prevalent are more likely to have access to MR than FDG-PET. Our prospective study shows that MRI can provide accurate baseline tumor volume and assess treatment response. The discriminatory value of the standard deviation of the ADC at baseline suggests that tumor heterogeneity may be predictive of response. This may support MR’s role in identifying more aggressive tumors and predicting chemo- or radio-resistant tumors.