Collaborators Meeting
Date: April 5, 2022
Time: 5:00 PM US/Eastern
Means: MS Teams
Attendees:
- Jason Halloran (WSU)
- Carl Imhauser (HSS)
- Shady Elmasry (Exponent)
- Kevin Shelburne (DU)
- Peter Laz (DU)
- Thor Andreassen (DU)
- Thor Besier (DU)
- Marco Schneider (ABI)
- Ahmet Erdemir (CC)
Agenda:
- Model Reuse specifications progress.
- Model Recalibration and Benchmarking status.
- Updates on manuscripts.
- Renewal grant resubmission.
- Other.
Immediate Action Items:
- Ahmet
- Draft renewal grant resubmission introduction.
- Jason
- Curate renewal grant resubmissions responses.
- Carl
- Finalize ORS Workshop manuscript and submit final version to co-authors.
- Submit Model Recalibration and Benchmarking outcomes.
- Shady
- Prepare Model Reuse specifications after submission of Model Recalibration and Benchmarking outcomes.
Thor & Kevin
- Confirm assembled results for Model Calibration manuscript.
- All
- Send their confirmation of the final ORS workshop manuscript to Andy.
- Submit specifications for Model Reuse phase.
- Fill in ideas for addressing reviewer comments.
Notes:
- Ahmet asked about the progress of Model Reuse specifications. DU is close, they set a deadline for April 15. ABI is waiting for Thor's review. Jason is working on it. CC has already prepared the documents. HSS will work on it after submitting outcomes of recalibration and benchmarking.
- Shady finalized Model Recalibration and Benchmarking package. It is awaiting Carl's review.
- Jason resubmitted the JOR manuscript on reporting and reproducibility assessment of knee models. Carl will finalize the ORS workshop manuscript and will send the final version to the group. Co-authors should respond to Andy's e-mail confirming their involvement and approval of the manuscript. This is needed for FDA internal assessment.
- Thor has been working on synchronizing his understanding of Model Calibration manuscript data with the other teams' descriptions. There are a few discrepancies, he will contact the HSS group.
- Thor showed the passive flexion kinematics predictions for all groups overlaid on experimental data. He tried to put all data in the same coordinate system (experimental) with the same reference state. For oks003, the kinematics was represented as rotations and translations of joint connectors with model's original state as zero reference. For DU02 joint rotations and translations matched Grood and Suntay (rotations around connectors, translations based on clinical description).
- The group also discussed the results of passive flexion, in particular internal-external rotation, in relation to plots of the Model Development manuscript. Ahmet noted that the Model Calibration manuscript may want to emphasize if the distribution of predictions get tighter (after calibration).
- Ahmet asked for help in responding summary statement of the grant application. He asked all to provide their ideas in a bullet point manner and maybe identify where in the grant application the responses may fit. Jason will curate these, he already created a shared folder in Google Drive. Ahmet aims to utilize these to write the introduction for the resubmission.
- The group discussed main concerns of review. We need to establish a strategy to quantify and evaluate accuracy and precision of a consensus workflow. This may be achieved by elaborating on the credibility and VVUQ frameworks; quantify and interpret in light of the context of use. Understanding what steps contribute to uncertainty will also be important to elaborate on. We can utilize outcome of the original project to quantify where variability is introduced and maybe perform sensitity on these to establish their impact. There is a concern raised about the rightness of consensus workflow. We may need to emphasize that consensus is aimed among experts in knee modeling. Using the outcomes of the original project, we may identify relevant accuracy of workflows (and introduce weighting). Or even, utilize the consensus workflow with the same goals of the original project modeling to quantify their relative predictive capacity against individualized workflows (threshold based on benchmarking results). We may also alleviate perceptions of innovation being incremental. Ahmet noted that from a metascience of modeling and simulation, the project is novel and we may want to emphasize this. Pete indicated the possibility to utilize new data sets (previously collected or to be collected), e.g. in vivo using biplanar radiography. Thor noted that this may add some innovation, particularly as a segue for modeling of in vivo systems. The group agreed that this can be driven by the contexts of use emerging from the community. In response to Thor, the participants noted to keep the Delphi approach as there were no significant issues raised by it.
- Thor asked if anyone would be attending to ISTA. He will be there. DU team will likely be there as well.