From Slicer Wiki
Revision as of 16:21, 21 November 2012 by Gpernelle (talk | contribs)
Jump to: navigation, search
Home < Documentation < 4.2 < Extensions < iGyne

Introduction and Acknowledgements

Authors: Guillaume Pernelle, Xiaojun Chen
Contributors: iGyne Team
Contact: Guillaume Pernelle, <email></email>
License: Slicer license

Extension Description

iGyne is an open source software for MR-Guided Interstitial Gynecologic Brachytherapy. It enables on-time processing of the intra-operative MRI data via a DICOM connection to the scanner followed by a multi-stage registration of CAD models of the template and the obturator to the patient images. This allows the virtual placement of interstitial needles during the intervention. Software features include:

  • Loading of Applicator CAD models and MR images
  • Initial registration of CAD model of the Template to MR images
  • Registration Refinement
  • Visualization of Registration
  • Needle Selection and Display of Virtual Needles relative to Applicator CAD models and Tumor Segmentation


Use Cases

Phantom Study and Sample data


Igyne Tutorial (a .PPT file with 18 pages) will guide you step-by-step through the process of loading of applicator CAD models, performing registration, selecting virtual needle and viewing the results within Slicer4. It will take about 10 minutes to finish this tutorial.

Similar Extensions



  • Xiaojun Chen, Jan Egger, Akila Viswanathan, Neha Agrawal, William Wells, Ron Kikinis, Clare Tempany, Ferenc Jolesz and Tina Kapur. "Image Processing for MR-guided Gynecologic Interstitial Brachytherapy in AMIGO", The 9th Interventional MRI Symposium. September 22-23, 2012.

Information for Developers

iGyne is currently articulated in seven steps: 1) procedure selection 2) applicator selection 3) data importation 4) initial applicator registration 5) refined applicator registration 6) needle position planning 7) needle detection It offers also a way to go directly from 1) to 7), skipping the registration/planning steps

- (1) and (2) let you choose to use or not an applicator, and to select the most convenient one. There are different configurations for the fiducial markers:

  • 4 landmarks, ordered from the top left corner and counting counter-clockwise (new cases)
  • 3 landmarks, positioned at the corners of the template, ordered as above
  • 3 landmarks, starting from the one in the middle of the template and counting counter-clockwise

This step offers an automatic registration option if the CLI module Hough Transformed has been previously enabled

- (3) let you load the data while the scene is loaded depending on the made in step 2.

- (4) is the initial registration step. Depending on the choice you made on the previous step, you can click on the bright markers in the image or let the automatic registration find them for you (requires the CLI module called Hough Transform).

- (5) is the refined applicator registration step. It offers fully automated computation to complete manual parameterization if needed. Thus, several segmentation methods are available to segment the obturator, and the registration parameters can be tweaked. By default, the most efficient parameters are chosen and all steps of the refined registration are computed successively. Evaluation functions have also been integrated to allow developers to measure time and accuracy of the chosen methods and parameters. Beside the control interface, 2D views display cross sections of the applicator in axial, sagital and coronal planes giving the user a visual characterization of the registration state.

- (6) let you insert "virtual needles" and visualize their trajectory in the 2D/3D viewer. You can tweak the color and the length of the needles.

- (7) is the "keystone" of iGyne. Start by clicking on "Start Giving Needle Tips" and at each click in the 2D axial viewer, a needle will be segmented. You can display, delete the needles individually or all together. You can also differentiate the different insertion steps and separate each group of needles. A table shows some parameters of the caculated needles. If a registration has been done previously, it will assign to each needle the most probable label. It is also possible to tweak the needle segmentation parameters.