1 Elec Introd The mo cancer ex vivo ment, c in vivo models may re ticity p the def Metho The stu went an based o The ex 3D mo employ model tration face of image. and ou this app the ima an elas modulu mined additio allows the d compu process illustra Result To eva T2w sl 46.7% rigid a after n On m relative after no 93.3±0 Discu We ou intensit MRE. which Refer [1] Koz [2] Dre [3] Cru A M ctrical and Compu duction ost accurate image probability maps o images can be ea change in volume MRI. Biomechan s are assigned arb esult in inaccurate parameters to the p formation maps an ods udy was approved n in vivo pre-op T on transperineal a x vivo post-op pros odel of the prostat y a model-based re onto the field of v algorithm [5] that f the model. Final Rather than defor utside its surface. proach, each voxe age is associated w sticity value (Youn us) that was de by the MRE scan on, this appro the forces that dr deformations to uted on the entire v sing step, we com ates the registration ts aluate registration lices. The results a initially, 78.4% a alignment, and 86 non-rigid registrat mid-gland slices, e 2D area ove on-rigid registratio 0.54%. ssion utlined a method fo ty information and Early experiments is needed in order ences zlowski et al., Ma ew et al., J. Magn. um et al., British Jo Model-Based M Guy N uter Engineering, U BC e-based characteriz that correlate histo asily registered to h and deformation o nical models which itrary and typicall deformation maps prostate and peripr nd, to the best of ou d by the institution 2-weighted (T2w) application of vibr state specimen wa te to be constructe egistration scheme view of the in vivo t translates, scales lly, we employ a n rming the ex vivo . In el of with ng's eter- n. In oach rive be volume, not only a mpute the inverse m n process on a tran performance, we are illustrated in Fi after 6.8% tion. the erlap on is or a model-based d the measured ela s on both syntheti to build cancer di gn. Reson. Imagin Reson. Imaging 3 ournal of Radiolog (a) I (a) Method for Re Nir 1 , Ramin S. Sahe University of Britis C, Canada, 3 Centr zation of prostate opathology with in histology using me of the prostate betw h are deformed in ly constant elastic s. We propose the rostatic tissue. The ur knowledge, has al ethics board an ) MRI scan follow ations to the prost s fixed in 10% bu ed easily. In contr e, in which the ex volume. Next, we and rotates the m novel iterative ela model, the in vivo around the model. mapping and warp nsverse slice of the compared the fina igure 2. Quantitati registration of an asticity to warp the c and clinical data stribution probabi ng 28 (2010) 621-6 32 (2010), 992-996 gy 77 (2004) S140 Initial ) 3D egistration of E ebjavaher 1 , Piotr K sh Columbia, Vanc re de Recherche Bi cancer to date is g n vivo MRI, one ca echanical constrai ween the scans. It a physical manne ity properties that novel use of mag e incorporation of not been reported d a signed consen wed by MRE in a 3 tate. We obtained uffered formalin an rast, the in vivo im vivo model is mat e align the 3D mod model over the volu astic registration a o volumetric imag Therefore, displa p the model onto t e volume. al registered ex viv ively, we found th ex vivo model to e volumetric imag a show promising lity maps. 628. 6. 0-S153. (b) Rigidly aligned (b) Slice 16 Ex Vivo to In Kozlowski 2 , Ralph couver, BC, Canad iomédicale Bichat- generated from mu an use ex vivo MR nts [2], it remains is the goal of this r have been used e t are optimized to gnetic resonance el such elastography d before. nt was obtained pri .0-Tesla system (P the elasticity ima nd scanned in a 7.0 mages contain surr tched to the in vivo del to fit the image ume in order to mi algorithm to comp ge is warped in ord acements of the vo the volume. We im vo model to a mod he relative 3D volu an in vivo T2w M e. To the best of o results. An ongoi [4] Muthupillai et [5] Tsai et al., IEE [6] Modersitzki, O d (c) Warpe (c) Sli Vivo Prostate Sinkus 3 , and Sept da, 2 MRI Research -Beaujon (CRB3), ulti-parametric mag RI scans of the fixe a challenge to reg s work to develop extensively for reg produce a low tar lastography (MRE y data into the regi ior to experiments Philips, The Nethe age using local fre 0-Tesla system (B rounding anatomy o volumetric imag e with respect to tr inimize the intensi pute the residual n der to match the m oxels are propagat mplement our met del constructed fr ume overlap betwe MRI volume with our knowledge, thi ing study will prov t al., Science 269 ( EE Trans. on Med Oxford University ed image ( ice 18 e MRI Using E timiu E. Salcudean h Centre, Universi Paris, France gnetic resonance i ed prostate specim gister ex vivo to in a method for accu gistration of medi arget registration e E) [4] that assigns istration framewor s. A patient schedu erlands). We used equency estimation Bruker, Germany). y and tissue, with ge. As a pre-proce translations, scalin ity variations on th non-rigid mapping model, by minimiz ted through the im thod using a nume rom manual segme een the two model a corresponding M is is the first mode vide further evalu (1995) 1854-1857 d. Imaging 22 (200 y Press (2004). (d) Warped model (d) Slice 20 Elastography n 1 ity of British Colum imaging (MRI) [1] men after radical pr n vivo MRI due to urate registration b cal imaging data [ error. This is not a actual in vivo me rk ensures a realist uled for radical pr a dynamic harmon n (LFE) of the dis High quality ex v which the prostat essing step, we inte ng and rotations. W he regions inside a g between the alig zing global intensi mage in a physical erical scheme bas entation of the pro ls (Dice’s similarit MRE data. The m el-based registratio uation of the meth 7. 03) 137-154. Figure 1. Re Cross-section model (red) o vivo T2w slic (c) the image the prostate i The inverse m model, produ Figure 2. Re (a) Registere (red) and ma in vivo mode (b-d) Selecte the registere on correspon and manual mbia, Vancouver, ]. In order to obtai rostatectomy. Whil unknown misalign between ex vivo an [3]. However, thes always realistic an asurements of elas tic regularization o rostatectomy under nic MRE techniqu splacement image vivo images allow te blends. Thus, w erpolate the ex viv We use a rigid regis and outside the sur gned model and th ity variations insid fashion. As a pos ed on [6]. Figure ostate in the in viv ty coefficient) to b method utilizes bot on method that use od on clinical data egistration process n of the ex vivo 3D overlaid on an in ce. Notice that in e is warped to fit inside the model. map, applied to the uces (d). egistration results. ed ex vivo model anually segmented el (cyan) in 3D. ed cross-sections o ed model overlaid nding T2w slices segmentations. in le n- nd se nd s- of r- ue s. a we vo s- r- he de t- 1 vo be th es a, . D e d of 2568 Proc. Intl. Soc. Mag. Reson. Med. 20 (2012)