Third Period Review Neurological & Neuromuscular Diseases Brussels – May 29th 2015
Aim ... to make people walk better By: • Better understanding • learning from others Brussels – 12 May 2016
Gait analysis by motion tracking Brussels – 12 May 2016
3D joint kinematics Brussels – 12 May 2016 5
Interpretation… Brussels – 12 May 2016 6
NND goals • To develop a repository of gait analysis data of CP children to enable similarity searches and other probabilistic modelling, to exploit retrospective evidence to the point of care • To standardize clinical and gait analysis protocols in paediatrics • To use those to produce prospective data for modelling in CP • To develop and validate subject specific modelling of the musculoskeletal system in relation to gait • To explore the potential of gait analysis and subject specific modelling to develop sensitive measures of disease progression in DMD and CMT1
Clinical expectation: Improve walking function • Patient-specific biophysical modeling – For detailed insight of different aspects of the disease – Inform treatment decision – Monitor treatment effects • Big-data infostructure – For outome prediction – For similarity searches – Baselineinfo for disease progression Brussels – 12 May 2016
Technical goals biophysical approach • Construct accurate personalized musculoskeletal models for NND children • Driven by the needs in clinical practice • Muscle lengths – typical pathological adaptation in CP • Muscle forces – typically increased in CP – spasticity – – Decreased in DMD, CMT1 Brussels – 12 May 2016
Technical workflow MRI images S TATISTICAL S HAPE MODEL Parameter extraction for + Gait analysis Physical exam musculo-skeletal modelling Brussels – 12 May 2016
DATAFLOW - EXTENDED Semantic infrastructure
NND - WP6 Tasks • T6.1 QA on data collection and protocols [M1-18] • T6.2 Gait analysis collection for CP [M1-36] • T6.3 Gait analysis collection for DMD & CMT1 [M12-36] • T6.4 Image acquisition [M 3-36] Brussels – May 29th 2015
Data acquisition: consensus and QA on data collection and protocols [M1-24] • A complete description of the protocols used in the clinical institutes (6.1-1) Technical Quality Assurance protocols ; Marker placement protocols & Operational protocols and workflow described and compared • Survey by the partners (6.1-2) Inventory performed along 11 gait labs inside + outside EU • A Consensus Proposal for EU CMA gait labs (6.1-2) Consensus protocol finalized amongst clinical partners (next slide) Final details of protocol + list of outcome parameters for database currently written down D6.1 : CGA clinical protocols @M18 D6.2 : Standard minimal dataset for data exchange and modeling (including TQA results) @M24 Brussels – 12 May 2016
Overview consensus protocol 1. Standardized patient history Clinical patient history and background 2. 3D Clinical gait analysis Kinematic data, kinetic data, EMG 3. Standard physical exam Joint range of motion, spasticity, bone deformities Basic strength, selective motor control 4. Walking oxygen consumption data Energy expenditure (oxygen uptake) during walking 5. Isometric muscle strength tests using hand-held dynamometry 6. Lower body MRI Muscle-tendon lengths, joints rotation centers/axes, muscle volumes, muscle attachment sites and anatomical landmarks Brussels – 12 May 2016
QA on data collection and protocols • Reliability measures of the protocols , to assure quantitative levels of reliability Technical quality assurance : low level validation of labs performed Repeatability and inter-rater reliability: currently being tested (one lab done) Brussels – 12 May 2016
TQA : repeatability analysis CMC w CMC w CMC w OPBG KUL VUA Joint Angle Subject #1 In the sagittal plane the repeatability within Right Left Right Left Right Left laboratory was excellent Hip flexion/extension 0.99 0.98 0.99 0.98 0.98 0.98 Hip abduction/adduction 0.89 0.80 0.94 0.96 0.75 0.72 In the frontal and transverse plane the repeatability Hip rotation 0.85 0.88 0.80 0.84 0.20 0.21 was lower than sagittal plane Knee flexion/extension 0.99 0.98 0.99 0.99 0.97 0.93 CMC for hip rotation was the lowest value, it could Ankle dorsiflexion/plantar 0.94 0.91 0.97 0.95 0.70 0.85 Ankle abduction/adduction 0.83 0.92 0.94 0.96 na na be due to a different marker placement between Ankle rotation 0.77 0.93 0.94 0.94 na na therapists. Comparable values of CMC for second Hip moment flexion/extension 0.83 0.89 0.96 0.94 0.80 0.90 Knee moment flexion/extension 0.90 0.90 0.97 0.95 0.60 0.77 subjects were found (0.78, 0.83) Ankle moment dorsiflexion/plantar 0.92 0.96 0.99 0.99 0.99 0.94 KUL OPBG VUA 60 60 60 40 40 40 Angle (°) Angle(°) Angle (°) 20 20 20 OP1 0 0 OP1 0 OP2 OP1 OP2 OP2 -20 -20 0 10 20 30 40 50 60 70 80 90 100 -20 0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100 %stride %stride %stride
WP6 Tasks • T6.1 QA on data collection and clinical protocols [M 1-18] • T6.2 Gait analysis collection for CP [M 1-36] * • T6.3 Gait analysis collection for DMD and CMT [M12 - M44] • T6.4 Image acquisition [M03 - M36] * * Deliverable 6.3 (first 130 CP patients) was due March 2016
Brussels – 12 May 2016
T6.2 – T6.3 Gait analysis collection NND OPBG – 3 May 2016 NND KUL – 6 May 2016 Complete Complete Patient Reference Acquired GOAL Patient Reference Acquired GOAL 222 290 TOTAL OVERALL 451 490 TOTAL OVERALL Total CP prospective extended 6 10 8 10 Total CP prospective extended Total CP prospective clinical 27 40 33 40 Total CP prospective clinical Total CP retrospective 400+ 400 150 200 Total CP retrospective Total DMD T0 9 10 9 10 Total DMD T0 Total DMD T1 2 10 7 10 Total DMD T1 Total CMT T0 7 10 10 10 Total CMT T0 Total CMT T1 0 10 5 10 Total CMT T1 24 22 (Healthy MRI NND VUMC – 6 May 2016 Complete Patient Reference Acquired GOAL TOTAL OVERALL 43 50 Total CP prospective extended 7 10 Total CP prospective clinical 36 40 Total TD reference data 14 20 Brussel 12 May 2016
Issues / Corrective actions Data acquisition is slightly behind (89%) does NOT affect the workflow of follow up partners (WP11, WP16) Actions: • OPBG: finalize CPretro (0616), complete CPprosp, CPext, DMD and CMT • KUL: complete CPprosp, CPext, DMD T0 and CMT T0, and T1 • VUmc : complete CPprosp, CPext (after sorting out MRI) • Perturbation experiments
Gait Perturbations
Application Scenario : Similarity search Patient information Patient measurements MD-Paedigree Conversion Anamnesis to standard database DB format Physical exam O2 MRI Gait Analysis information Similar case Gait Analysis trial data
exploratory phase similarity search (November 2015 – February 2016) Gait patterns from 357 patients (children with Cerebral Palsy ) from KULeuven, involving 1731 trials Brussels – May 29th 2015
WP11 – Modelling pipe-line Functional Calibration TUD Siemens USFD TU Delft Motek Segmented Complete Personalized Human OpenSim Model MRI 3D Models Anatomical Model Body Model MD-Paedigree/NND wp11 workflow
WP11 – Modelling pipe-line Functional Calibration TUD Siemens USFD TU Delft Motek Segmented Complete Personalized Human OpenSim Model MRI 3D Models Anatomical Model Body Model MD-Paedigree/NND wp11 workflow
MRI segmentation From MRI …to individual images… bone and muscle models Quantitative evaluation shows similar segmentation quality in Dr. Maria Costa/Siemens/2016 both healthy and ill cases.
WP11 – Modelling pipe-line Functional Calibration TUD Siemens USFD TU Delft Motek Segmented Complete Personalized Human OpenSim Model MRI 3D Models Anatomical Model Body Model MD-Paedigree/NND wp11 workflow
Patient-Specific Complete Anatomical Model: Morphing of Template to MRI Segmentation
Patient-Specific Complete Anatomical Model: Morphing of Template to MRI Segmentation
Patient-Specific Complete Anatomical Model: Geometric Parameters Extraction
WP11 – Modelling pipe-line Functional Calibration TUD Siemens USFD TU Delft Motek Segmented Complete Personalized Human OpenSim Model MRI 3D Models Anatomical Model Body Model MD-Paedigree/NND wp11 workflow
OpenSIM: personalize segment length + mass • Use data from either FC or MRI – Unified file format (.MPF) • Scale segments – Relative joint center distance – Either FC or MRI • Proportional mass distribution – Weight from FC • Patient specific OpenSIM model
Constructing a scaled model Functional joint center calibration Scaled musculoskeletal models Brussels – May 29th 2015
WP11 – Modelling pipe-line Functional Calibration TUD Siemens USFD TU Delft Motek Segmented Complete Personalized Human OpenSim Model MRI 3D Models Anatomical Model Body Model MD-Paedigree/NND wp11 workflow
• Functional calibration (video) Visualization done using MITK MITK .org Brussels – May 29th 2015
Visualization done using MITK MITK .org Brussels – May 29th 2015
Recommend
More recommend
