Katherine Louise Dextraze MS, Presenter: Nothing to Disclose

Florian Maier, Abstract Co-Author: Nothing to Disclose

Judy Un Chong Ahrar MD, Abstract Co-Author: Nothing to Disclose

Yvette Teniente, Abstract Co-Author: Nothing to Disclose

Kamran Ahrar MD, Abstract Co-Author: Nothing to Disclose

R. Jason Stafford PhD, Abstract Co-Author: Nothing to Disclose

A statistical model was investigated to quantify the extent of damage within the kidney parenchyma based on tissue position with respect to the iceball surface as visualized on images during the MRI-guided cryoablation procedure.

A retrospective study of 20 patients cases was performed in order to statistically correlate the lack of perfusion seen on periprocedural contrast enhanced T1 post-treatment images with the iceball signal deficit seen on MRI-guided cryoblation monitoring images. Manual land-mark based registration and manual segmentation were performed on the data sets prior to analysis. In order to reduce variability in the segmentations, repeated segmentation trials to submitted to a truth-estimation scheme. Automated measurements of the distance between the iceball surface and the perfusion deficit edge were made and logistic regression model was fit to these measurements using original MATLAB scripts. The Kolmogorov-Smirnov test was applied to the Pearson residuals of the logistic regression model to assess goodness-of-fit of the model to the data. Measurements were restricted to renal parenchyma, where reliable registration could be applied.

Using 20 patient cases and over 600 data points, the perfusion loss likelihood of renal parenchyma within the iceball was described by a unique logistic regression curve, where the parameters are alpha = -0.45 and beta = 0.79. From this curve, it was determined that tissue is 50% likely to lose perfusion at 0.57mm within the iceball, while perfusion loss is 95% likely at 4.28 mm within the iceball edge. The Kolmogorov-Smirnov test for goodness-of-fit confirmed that the logistic regression model reported here describes the observed data appropriately.  

Through a retrospective study of 20 patient cases, the relationship between likelihood of perfusion loss in renal parenchyma and distance within iceball was statistically quantified. From the statistical model, the margin for 95% perfusion loss likelihood was found to be 4.28mm within the iceball, which agrees the clinically accepted 3-5mm margin that is estimated during the procedure.

The statistical model presented here could serve effectively as a quantitative approach to assessing treatment progress during the MRI-guided cryoablation procedure, rather than relying on visual estimation.

Dextraze, K, Maier, F, Ahrar, J, Teniente, Y, Ahrar, K, Stafford, R, A Statistical Model of the Relationship between Iceball and Perfusion Deficit Visualized during MRI-guided Cryoablation.  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14005821.html