![]() |
![]() |
||||||||||||
|
02420_2 OPTIMAL ANALYSIS OF MRI DATA TO QUANTITATE THE DISTRIBUTION OF A VAGINAL PRODUCT Barnhart, Kurt* Introduction: MRI can be used to safely and acutely visualize and quantitate spread of a potential microbicide formulation. We have used MRI to characterize the spread of a number of vehicles and active compounds including Gynol II, 1.0% C31Gg, 6% Cellulose Sulfate, Replens and KY jelly; at a variety ofvolumes. By combining the experience of these trials we have been able to analyze MRI as a technique to compare gels and optimize experimental conditions. This abstract summarizes some of these conclusions. Methods: Raw data from 5 experimental protocols and greater than 30 women were combined. When possible, data were aggregated. We explored a variety of analytic techniques, statistical methods and experimental designs to determine the optimal use of MRI to quantiate intra-vaginal spread of a potential microbicide. Results: MRI has outstanding intra-person validity and reproducibility. Using this data the best characterization of intra-vaginal spread of a vaginal gel is a linear relationship with time (linear mixed model). The slope of the curve is dependant on the specific gel, the volume, time and the baseline vaginal dimensions. The baseline vaginal dimensions of length below the cervix and transverse measurement at the flexion may be dependant on weight and gravity respectively. Spread is best characterized in two distinct anatomical categories: the upper and lower vagina, using the pelvic diaphragm to delineate the two. Individual gels do not spread equally in both the upper and lower vagina. Conclusions: MRI can be reproducibly used to assess and compare intra-vaginal spread of a microbicide. Given the complex anatomic shape of the human vagina a simple linear measurement in one plane is inadequate to characterize spread. A summary measurement of multiple assessments of both the upper and lower vagina best characterizes vagina coverage. Paired study designs provide the most efficient comparison of two products or volumes. This research was funded by CONRAD Dr. Kurt T. Barnhart |
|||||||||||||