IGT:Sept28 2005
Convenor: Randy Ellis
Attendees: Anton Deguet, Noby Hata, Michael Halle, Tina Kapur, Peter Kazanzides, Ron Kikinis, William Lorensen, Eigil Samset, W "Sandy" Wells
Regrets: Kevin Cleary, Gabor Fichtinger, Steve Pieper
Agenda: Brainstorming the development process for a new Slicer intraoperative navigation system
Highlights: The project goal is production of high-quality, open-source navigation software. Major research efforts, including the Brigham and Women's Hospital effort on the Advanced Multimodality Image-Guided Operating suite (the AMIGO suite), will need navigation software that merges multiple intraoperative imaging sources. This software is currently not available and will be needed for research approved by the Institutional Review Board (IRB) of organizations pursuing navigation research.
The software will need to be [reasonably] compatible with existing and projected Slicer modules. From the start, all were agreed that we should follow Good Manufacturing Practices (GMP) in the design, development, and maintenance of the software. All agreed that the current NAMIC/Kitware process is robust and suitable.
The project management will be driven by use cases, which for this purpose are descriptions of interventional medical procedures using navigation. It was generally felt that storyboards for the use cases would be the best way of capturing the user requirements, to which the technical requirements and constraints will be added. From these requirements, the common ones will be aggregated into "base software" that will go through rigorous review and testing. The specialized requirements, flow, and additional components will be kept in software "silos" specific to each use case, thus restricting the development and testing to that silo and avoiding the quadratic expansion of component interaction that has been found in other large systems.
All agreed that there is a great deal of software, existing and currently in development, that is likely to be useful. In addition to Slicer modules, the entire VTK and ITK suites are likely to be used. The new IGSTK effort is also producing components that will likely prove useful. Tina Kapur noted that this effort and the IGSTK effort are complementary, with this effort focused on the interventional procedures (captured in use cases) and the IGSTK effort focused on components likely to be needed for a wide range of procedures.
There was some discussion over the role of a Control Board, which would be responsible for adding and deleting major features of the software. It was generally agreed that this would be useful, but details were not settled in the brainstorming session.
The participants have all been part of previous efforts in navigation and noted special requirements that might not come out of the use cases directly. Some of these were:
- IRB needs and regulatory-approval needs (e.g., FDA) are best bulit in early
- the system (software plus hardware) must robustly recover from inadvertent or intentional power cycling that routinely happens in operating rooms as various pieces of equipment compete for space and power
- ancillary to this is a notion of atomicity in saving the software state
- logging was widely seen as necessary, possibly down to logging screen images as a procedure is performed
- other systems will need to "slave" to the navigator to support multiple sites collaborating during a procedure
- vendor-specific implementations are likely in the early stages, ideally progressing to more generic implementations as the project matures
Randy Ellis will proceed to collect potential interventional procedures and lead the work on elaborating these into use cases that can drive the process of extracting the critical requirements of the next Slicer-based navigation system.
