Date: December 14, 2017
Time: 10:00 AM US/Eastern
Means: In person at Cleveland Clinic
- Ahmet Erdemir (CC)
- Jason Halloran (CSU)
- Carl Imhauser (HSS)
- Peter Laz (DU)
- Kevin Shelburne (DU)
- Follow up on previous meeting on December 13, 2017.
- Identify data for use in the project.
- Identify prospective gatherings and outreach activities.
- Outreach to and meet with local scientific and clinical community.
Immediate Action Items:
Kevin & Pete (DU)
- Provide short ranked list (at least 2 knees - one male and one female) of possible data sets from and Natural Knee Data.
- Ahmet (CC)
- Provide short ranked list (at least 2 knees - one male and one female) of possible data sets from Open Knee(s).
The group followed up the discussions in previous meeting, see 2017-12-13. Main focus was on the identification data sets for project's use. Ahmet described the Open Knee(s) data in more detail using experimentation specifications of Open Knee(s). Kevin described the Natural Knee data browsing through the dissemination site and showing some slides from his previous work. More detailed notes are provided below.
- Notes on identification of data for use.
- For the project, one male and one female data set are needed. Ahmet summarized the available specimens from Open Knee(s). The male 40 years old specimen may be a decent candidate (oks008) though Ahmet noted that his "gut" feeling is oks001 (elderly male) and oks003 (young female) will likely serve the project best. Ahmet also pointed out that oks001 has already had some modeling and segmentation performed, which has helped to clarify some of the documentation for this specimen.
- Ahmet also summarized the imaging protocols for each specimen, along with the registration procedure. Three different types of MR sequences were used: one general purpose, one for cartilage, and three along with multiple planes for connective tissue (complete details are in Open Knee(s) wiki). The tibia, patella, and femur registration spheres were shown, along with the mounting screws for the Optotrak markers.
- Mechanical testing procedures (tibio- and patellofemoral) for Open Knee(s) were also summarized. Laxity tests were performed along with combined loading scenarios. For patellofemoral joint testing, tibiofemoral joint was flexed at a prescribed angle, quadriceps load was applied (along a measured line of action) and contact pressures were measured. Mechanical testing data provides kinematics and kinetics of bones in different coordinate systems. Anatomical landmarks on the whole leg were also collected.
Kevin summarized the data from the Denver group. Kevin commented that he would like to take a closer look at the available data to determine what set is the most complete. There are both CT and MR for most specimens (but not all). Imaging included just the knees but the leg may have been intact (Kevin will check). All loading cases were manually applied with visual feedback for the tester (to know what loads were being applied). Laxity testing was applied from 0 to 90 degrees knee flexion. Manual laxity testing was repeated after cutting cruciate ligament(s). After laxity testing, the knees were loaded into a "quadriceps loading rig", which leaves the tibiofemoral joint free while the the quadriceps was loaded to achieve knee extension. The loading rig components and the line of action of the quadriceps was digitized during the experiment. No additional mass was added to the tibia during extension, though there was a hamstrings load applied. The tissue cut (ACL and PCL) states me be good candidates for benchmarking.
- Kevin's initial feeling is specimen number 1 is a good candidate, but number 2 would also be acceptable even though it appeared quite different, especially in terms of anatomy.
- Natural Knee Data from Denver group does not have registration data but includes anatomical landmarks. Ligament insertion sites were also digitized.
- There were some discussions about the state of the data to be compiled for the group. In particular, whether raw or processed (e.g. into anatomical coordinate frames), or both, should be used for model development, calibration and benchmarking. Ahmet favored raw data for segmentation (image only) but both raw and processed for joint level mechanics.
- Ahmet also noted that the specifics of earmarked data and raw data presentation will be decided at each modeling phase. At this moment, it is important to make sure that selected knee specimens have as complete data as possible to accomplish all modeling phases and their anatomical and mechanical states have relevance to model re-use cases.