Summary:
This paper describes a software framework (ESQUI) for surgical simulation that uses a number of open source software packages and development tools. The authors introduce a new Surgical Reality Modeling Language (SRML).

Hypothesis:
N/A

Evidence:
N/A

Open Science:
The paper is accompanied by all the source code.

Reproducibility:
I downloaded the source code and reviewed it. I did not run the precompiled Windows binaries or try to compile it.

Use of Open Source Software:
The paper mentions the use of the following open source packages: Blender, Python, VTK, and TCL/TK. The following open source development tools are used: CMake, Subversion, and Doxygen. There are plans to use Dart in the near future.

In addition to these packages, the ESQUI software includes a number of files taken from other open source projects. It seems that many source files were obtained from INRIA Cybermove [17]. In addition, the authors extended the Python module (Blender to VTK link) from the University of Alberta [8]. Thus, it seems incorrect to state (Section 1.1) that “In our environment, except Blender, the software has been programmed in our group.”

Open Source Contributions:
The authors state that the ESQUI software is available as open source, currently under the Creative Commons license. The mechanism of distribution, other than as an attachment to this paper, is not noted (e.g., is there a web page from which different versions can be downloaded and/or researchers can access the subversion repository?).

Code Quality:
Some files (such as the ones from INRIA) are written in English, but many files are written in Spanish so it is difficult for me to review them. All of the doxygen documentation is written in Spanish. I did not see any serious problems with the code quality. One minor point I would mention is that the implementation of the vector cross product (file vtkVector3f.h, obtained from INRIA) will not work if the output is the same as one of the inputs (e.g., x.crossProduct(x,y)).

Applicability to other problems:
This simulation framework and SRML format were initially developed for laparoscopic surgery, but as the authors note, they could be used for many other types of surgical simulations.

Suggestions for future work:
The authors state that future work will include the inclusion of more algorithms for collisions and deformations and that they are studying other systems such as Spring and OpenTissue. As far as I know, these two systems are similar to ESQUI and the authors should note what distinguishes ESQUI from them. The ESQUI software requires some additional work to be more usable by other researchers. Currently, some of the code and most of the documentation is written in Spanish and should be translated to English. An automated testing framework, using Dart, would also be helpful (the authors already mention the need for this). The current documentation consists of doxygen, a short (2-page) quickstart guide, and a few text files such as INSTALL.TXT. A “programmers guide” and/or “users guide” is needed to give an overview of the software architecture and illustrate how to use the software.

Requests for additional information from authors:
The authors should explain why it was necessary to introduce a new modeling language (SRML). What are the deficiencies/limitations of the existing modeling languages? Also, how does ESQUI compare to similar systems such as Spring and OpenTissue?

Additional Comments:
Although there are a few places where the English is not perfect, the paper is generally well written and is easy to understand. One typo to note: in Section 1.5 (Haptics), “keyword” should be “keyboard”.

Summary:
This paper proposes a surgical simulation framework including surgical scene visualization, haptic device support, and biomechanic modeling.

Hypothesis:
Non Applicable

Evidence:
The author claimed that the proposed SRML is more suitable for describing surgical scene, however the advantage of using this new XML format verses other existing virtual reality modeling languages is unclear

Open Science:
This work is based on existing open source software toolkits and also complies with the open source license terms as clearly stated in the “ESQUI Overview” section (http://creativecommons.org/licenses/by/2.5).

Reproducibility:
I wasn’t able to run the precompiled version of the source code under windows. The installation instruction didn’t give enough/accurate information on setting extra library paths.

Use of Open Source Software:
Has taken full advantage of available open source software toolkits

Open Source Contributions:
The source code is available for download.
More comprehensive user-guide might be helpful.
The source code documentation is not available for non-Spanish speaking developers.

Code Quality:
Well organized source code directories, and support cross-platform development

Applicability to other problems:
This work is originally devoted to the laparoscopic surgery simulation. The framework seems to be general enough to extend to other application areas.

Suggestions for future work:
Is there any speed and accuracy trade off issue in this kind of simulation? Are you considering doing any validation or verification study for the simulation?

Requests for additional information from authors:
None

Additional Comments:
None

Summary:
This paper describes the implementation of a surgical simulation framework supporting collisions and deformations of geometric models.

Hypothesis:
Non Applicable

Open Science:
The paper adheres to the principles of Open Science by providing all source code, and being based on a number of Open Source toolkits/utilities.
One example Surgical Reality Modeling Language (SRML) file was provided for testing.
The 'install.txt' meantioned a number of videos demonstrating the system, however these were not included with the article.

Reproducibility:
I found the installation instructions in 'install.txt' file a little too sparse. Perhaps there is a need for more detailed instructions regarding installation.
For example, it seems that VTK requires USE_HYBRID = ON, which was not listed anywhere in the documentation (as far as I could tell).

Unfortunately, after trying for some time, I could not get the project to compile. I encountered an error while compiling the vtkesquiColision project:

2>------ Build started: Project: vtkesquiColision, Configuration: Release Win32 ------

2>Compiling...

2>vtkModeloEstatico.cpp

2>c:\Program Files\Microsoft Visual Studio 8\VC\PlatformSDK\include\windef.h(152) : error C2632: 'int' followed by 'int' is illegal

2>c:\Program Files\Microsoft Visual Studio 8\VC\PlatformSDK\include\windef.h(152) : warning C4091: 'typedef ' : ignored on left of 'int' when no variable is declared

2>Build log was saved at "file://c:\Temp\IJ-MICCAI\ESQUI\Build\Colisions\vtkesquiColision.dir\Release\BuildLog.htm"

2>vtkesquiColision - 1 error(s), 1 warning(s)

I think this error has something to do with Comun\Macros.h. Editing out the line '#include Macros.h' in Colisions\vtkModeloEstatico.h allowed me to continue. The vtkesquiTMesh project also exhibited the same error for the following files: vtkTiposT2.cpp, vtkMallaT2.cpp, vtkCara.cpp, vtkContact.cpp, vtkOrgan.cpp. I was not able to successfully compile this project due to these errors.

2>------ Build started: Project: vtkesquiTMesh, Configuration: Release Win32 ------

2>Compiling...

...

2>vtkTiposT2.cpp

2>c:\Program Files\Microsoft Visual Studio 8\VC\PlatformSDK\include\windef.h(152) : error C2632: 'int' followed by 'int' is illegal

2>c:\Program Files\Microsoft Visual Studio 8\VC\PlatformSDK\include\windef.h(152) : warning C4091: 'typedef ' : ignored on left of 'int' when no variable is declared

2>vtkMallaT2.cpp

2>c:\Program Files\Microsoft Visual Studio 8\VC\PlatformSDK\include\windef.h(152) : error C2632: 'int' followed by 'int' is illegal

2>c:\Program Files\Microsoft Visual Studio 8\VC\PlatformSDK\include\windef.h(152) : warning C4091: 'typedef ' : ignored on left of 'int' when no variable is declared

2>vtkCara.cpp

2>c:\Program Files\Microsoft Visual Studio 8\VC\PlatformSDK\include\windef.h(152) : error C2632: 'int' followed by 'int' is illegal

2>c:\Program Files\Microsoft Visual Studio 8\VC\PlatformSDK\include\windef.h(152) : warning C4091: 'typedef ' : ignored on left of 'int' when no variable is declared

2>vtkLink.cpp

2>c:\temp\ij-miccai\esqui\t2mesh\vtkLink.h(17) : warning C4275: non dll-interface class 'vtkObject' used as base for dll-interface class 'vtkLink'

2>        C:\Dan\Tools\VTK\src\vtk-4.4.2\Common\vtkObject.h(48) : see declaration of 'vtkObject'

2>        c:\temp\ij-miccai\esqui\t2mesh\vtkLink.h(17) : see declaration of 'vtkLink'

2>c:\temp\ij-miccai\esqui\t2mesh\vtkContact.h(21) : warning C4275: non dll-interface class 'vtkObject' used as base for dll-interface class 'vtkContact'

2>        C:\Dan\Tools\VTK\src\vtk-4.4.2\Common\vtkObject.h(48) : see declaration of 'vtkObject'

2>        c:\temp\ij-miccai\esqui\t2mesh\vtkContact.h(20) : see declaration of 'vtkContact'

2>c:\temp\ij-miccai\esqui\t2mesh\vtkEsquiPolyData.h(12) : warning C4275: non dll-interface class 'vtkPolyData' used as base for dll-interface class 'vtkEsquiPolyData'

2>        C:\Dan\Tools\VTK\src\vtk-4.4.2\Common\vtkPolyData.h(52) : see declaration of 'vtkPolyData'

2>        c:\temp\ij-miccai\esqui\t2mesh\vtkEsquiPolyData.h(11) : see declaration of 'vtkEsquiPolyData'

...

I was using WinXP SP2, Visual Studio 8, CMake 2.4.2, VTK 4.4.2 and TCL/TK 8.4. Does the framework support the VS8 compiler?

Use of Open Source Software:
The proposed framework is based on a number of Open Source toolkits/utilities including:

• CMake: http://www.cmake.org


• VTK: http://www.vtk.org


• Python: http://www.python.org


• TCL/TK: http://tcl.tk


• Blender: http://blender.org

Code Quality:
All source code is provided and suitably modularised (ie. in folders Collisions, Haptic, Tools, etc.). A majority of the source code is in Spanish and is therefore difficult for English speaking developers (such as myself) to understand/use/maintain. The framework seems suitable for multi-platform deployment (resultant from being derived from other multi-platform toolkits such as VTK and TCL/TK).

Suggestions for future work:
Perhaps one area for future work is the discussed Surgical Reality Modeling Language (SRML). I believe rather than proposing a totally new XML format, the required mechanisms could be incorporated into an existing 3D modelling language such as VRML or X3D (see http://www.web3d.org). Specifically, X3D provides extensions for defining application specific scene-graph nodes. Once completed, MedX3D would be a good standard for defining 3D medical scenes for visualisation and simulation (see http://www.web3d.org/x3d/workgroups/medical). Have you considered using this (or a similar existing) format to define the scene?

Requests for additional information from authors:
Has the project been sucessfully compiled using Microsoft Visual Studio 8? If not, what compilers has the framework been successfully used with? Any advice for me to get the framework up-and-running?