Collaborators Meeting

Date: December 3, 2019

Time: 3:00 PM US/Eastern

Means: Skype

Attendees:

  1. Jason Halloran (WSU)
  2. Will Zaylor (CSU)
  3. Ammar Hafez (CSU)
  4. Carl Imhauser (HSS)
  5. Shady Elmasry (HSS)
  6. Kevin Shelburne (DU)
  7. Peter Laz (DU)
  8. Donald Hume (DU)
  9. Thor Andreassen (DU)
  10. Thor Besier (ABI)
  11. Nynke Rooks (ABI)
  12. Ahmet Erdemir (CC)
  13. Neda Abdollahi (CC)

Agenda:

  1. Model Calibration progress.
  2. Dissemination of phase specifications and outcomes.
  3. Upcoming conferences.
    1. Preparations for RIG meeting at ORS.
  4. Manuscripts.
    1. ORS Workshop manuscript.
    2. Manuscript on reproducibility of knee joint models.
    3. Manuscript on Model Development phase.
  5. Review of previous action items.
  6. Other business.
    1. Budgets.
    2. New logo.

Immediate Action Items:

Notes:

  1. Ahmet asked about model calibration progress and wondered if the participants had any questions. The anticipated delivery period is June 2020. He finished writing of the specifications but the Cleveland Clinic team did not start execution. Carl’s group has not started yet. Jason's team has started working with the Open Knee(s). They anticipate that Natural Knee Data will take more processing time. Thor's group has not started as well.
  2. Ahmet noted that at this moment, modeling outcomes are in large compressed files in source code repository. Model development specification/outcomes are hard to find now. He anticipates using the "Data Share" feature of SimTK to disseminate specifications and outcomes in a more easily accessible manner. He noted that when using source code repository, number of downloads for model and outcomes can not be counted. If the files are linked to the download section, download statistics can be monitored. With data share it is also a little bit difficult to monitor the number of downloads.
  3. The participants discussed activities in the upcoming ORS meeting.
    • Representatives from each team, except Auckland group, will be in the ORS. Will will talk about calibration progress at the ORS. The other groups have abstracts that are in computational biomechanics but not necessarily directly linked to the project.
    • Ahmet and Carl submitted a proposal for 2 hours research interest group (RIG) meeting, which was accepted. Topic is on fairness and credibility of modeling, where concepts such as data and model sharing, validation and credibility, findability, accessibility and reproducibility can be discussed. As our experience has shown, how we share our data and model is important. Ahmet will have a brief presentation about terminology and relationship between terms. He will pass the presentation to everyone to structure the session. The presentation will follow with a panel discussion between postdocs, students, engineers, etc. and the audience.
    • Overall the emphasis and structure of the RIG will be open, ideas under the umbrella of computational biomechanics can be discussed. To start, the organizers will likely ask about 5 important questions at audience, e.g., What is credibility? What are the interaction between credibility, validity, and reproducibility? What are the constraints? Do we have data and tools? How fair do model and data need to be? A count of abstracts in the conference that use computational modeling will be noted to emphasize the size of the community. Some of them, if related to credibility, can be highlighted.
    • Following the discussion on RIG meeting structure, Shady was asked to reach out to groups to organize a brainstorming session and make notes about important points. Ahmet will send relevant articles that can help to prepare.
    • The take home messages from the session will likely be: 1. Computational biomechanics is at the heart of orthopaedic research such that it can not be ignored. 2. Computational biomechanics can be more effective and unified if done as a community rather than in our corners. 3. A critical mass of modelers in orthopaedics will strengthen the need for a devoted session in computational biomechanics. 4. A community can be mobilized to brainstorm the next grant challenge.
    • It is recommended to include a clinician in the session, i.e., how computational modeling has influenced any medical field. Jason will follow up with a surgeon after session structure is decided.
  4. The participants moved on to discussions on manuscripts.
    • During the ORS 2019 workshop, Shady recorded the guest speakers, Carl has summarized their insight; he is finalizing the manuscript by making plots from survey results. He will provide a draft to everyone before ORS.
    • Manuscript on the review of knee joint models' reproducibility potential is still in progress. Jason has gone through what he received but couldn't find Auckland team's reviews. Auckland group has completed the reviews and will resend the files to Jason. Jason has a detailed outline and he suspects it will take a fair amount of time to compile. He will prepare a draft before ORS.
    • Auckland group has started a manuscript using the outcomes of the model development phase. Nynke submitted an early draft to the participants. They are mostly focusing on comparing the bone and cartilage meshes, i.e., thickness differences and articular geometry differences. They prepared some figures on cartilage boundaries and visualization. They may focus on ligament insertions in the next paper since the document is already bulky with all the mesh plots and comparison. Nynke thinks the detailed mesh comparison is ready to be out before Christmas. Ahmet wondered that the audience may be interested in how much these differences matter. Comparing the mesh without using them for simulations may be problematic. He also noted that we should have a more broad manusript on the model development phase paper, e.g., a master paper discussing: (1) Did all the groups model it the way they said? (2) Overall comparison on how the model differ between groups rather than going in depth and comparing meshes point by point. (3) Highlights of big deviations. (4) Overlaps to indicate potential for process standardization/unification. Such a manuscript may include examples of differences that have clinical importance like thickness of cartilage in contact points, ACL. It can highlight when modeler's decisions change; why some decisions are made just to get the model running, why everybody uses different cartilage models. This article can emphasize important model development topics and let the community follow up by providing links to model development specifications, deviations, and outcomes as raw data of the paper. Jason also mentioned that three teams had presentations in CMBBE 2019 focusing on deviations. These can be used as material for this manuscript. He also noted that his team performed some tissue insertions comparisons among all groups. Ahmet reminded that each of the modeling and simulation phase will have a strategy and summary paper.
    • Take home messages of an article on our experience during the model development phase may focus on what we actually did and what we said we would do: (1) Deciphering art of modeling in the model development stage. (2) Demonstrate the importance of sharing you workflow and your data, if someone wants to understand how the model works and what it is. (3) Flexibility of the decision making process, i.e., when people are asked to develop model from same data, their art and minds may change during the process. The output may differ and its impact on the reproducibility potential of the model is unknown. (4) List of areas of modeling art that requires further studies to assess the effect of these decisions on model development. (5) Bring awareness to the art that is not reported well in scholarly publishing and here is what can be captured by workflow documentation.
  5. Ahmet mentioned a few other topics.
    • Ahmet asked all the teams to send him their budget details: What has been awarded? How much of the funds have they spent? How much of the funds do they expect to spend until the end of year 3 (June 2020)?
    • Ahmet also mentioned the new logo. It is available in the source code repository for use in posters, presentations, etc.

2019-12-03 (last edited 2020-01-24 16:10:36 by hallorj)