What is Modeling and Simulation? What is a model? · PDF file · 2015-12-04What is...

Tina M. Morrison, Ph.D.Advisor, Computational Modeling

Office of Device Evaluation

CDRH/U.S. FDA

Regulatory Science

2

The science of developing new tools, standards,and approaches to assess the safety, efficacy, quality, and performance of FDA-regulated productso Benefits patients by speeding the rate of

important technologies reaching marketo Reduces time and resources needed for device

development, assessment, and review. For example:

• Can lead to quicker, more efficient device approvals

• Can decrease the size and duration of pre-market clinical trials

Faster, Cheaper, SaferFDA Strategic Plan, August 2011Advancing Regulatory Science at FDA

What is Modeling and Simulation?

Modeling refers to the development of a mathematical representation of a physical/chemical/biological reality‐of‐interest

Simulation refers to the procedure of solving the equations that resulted from model development

3

What is a model?

Model: a description or representation of a system, entity, phenomena, or process. • Data used to generate a model are considered part of the model

Models may be mathematical, physical, or logical representations of a system, entity, phenomenon, or process. Models can be used by simulation to predict a future state, if so desired.

In vitro(cells)

In vivo(animals)

In silico(mathematics)

4

What is a model?

At worst, a model is a concise description of a body of data.

At best, it captures the essential physics/chemistry/biology of the problem, it illuminates the principles that underline the key observations, and it predicts behavior under conditions which have not yet been studied (or conditions which are not evaluable with any other model).

5

What is a model?

Computational model: the numerical implementation of the mathematical model performed by a means of a computer.

Mathematical model: the equations, boundary values, initial conditions, and modeling data needed to describe the conceptual model.

Models can occur on one scale (e.g., structural mechanics of a hip implant) or on multiple scales (e.g., transmission of molecular dynamic properties of red blood cells and their affect on blood damage).

Models can also be multi‐discipline or multi‐physics (e.g., molecular dynamics and continuum mechanics)

6

What is a simulation?

Simulation: the imitation of the characteristics of a system, entity, phenomena, or process using a computational model.

Simulation refers to the procedure of solving the equations that resulted from model development.

For example, numerically solve a set of differential equations with different initial/boundary conditions.

7

Science Priority Areas#1 Modernize Toxicology#2 Stimulate Innovation in Clinical Evaluations and Personalized Medicine to Improve Product Development and Patient Outcomes#4 Ensure FDA Readiness to Evaluate Innovative Emerging Technologies#5 Harness Diverse Data through Information Sciences to Improve Health Outcomes

(Q)SAR models to predict human risk Computer models of cells, organs, and systems to better predict product safety and efficacy

Virtual physiologic patients for testing medical products

Clinical trial simulations that reveal interactions between therapeutic effects, patient characteristics, and disease variables

Knowledge building tools Methods to verify, store, share

http://www.fda.gov/downloads/ScienceResearch/SpecialTopics/RegulatoryScience/UCM268225.pdf

FDA has identified an important role for CM&S in its strategic priorities.

Center for Devices and Radiological Health (CDRH)

Center for Devices and Radiological Health

Office of the DirectorJeffrey Shuren, MD, JD

Office of In Vitro Diagnostic Device Evaluation and Safety

Office of Surveillance and Biometrics

Office of Compliance

Office of Communication, Education and Radiation Programs

Office of Management Operations

Office of Science and Engineering Laboratories

Office of Device Evaluation

What we do …

CDRH is responsible for regulating firms who manufacture, repackage, re‐label, and/or import medical devices sold in the U.S.

10

CDRH Mission

“The mission of the Center for Devices and Radiological Health (CDRH) is to protect and promote the public health. …We facilitate medical device innovation by advancing regulatory science, providing industry with predictable, consistent, transparent, and efficient regulatory pathways, and assuring consumer confidence in devices marketed in the U.S.”

11

CDRH Mission

Medical Device Classifications

Class IGeneral ControlsMost exempt from premarket submission

Class II Special ControlsPremarket Notification [510(k)]

Class IIILife‐supporting, life‐sustainingPremarket Application [PMA]

12

CM&S as valid scientific evidence

e.g., fatigue safety factors

CM&S is the medical device

e.g., clinical decision support

The Role of CM&S at CDRH

13

CM&S embedded in a medical device e.g., physiological

model in a control system


medicalnewstoday.com

CM&S embedded in a medical

device e.g., logical algorithm in a control system

14


15

heartflow.com

CM&S is the medical device

e.g., clinical decision support


16Kelm, Int J of Med Sci 2009

Bluestein, PLOS ‐ 2012

CM&S as valid scientific evidence

e.g., fatigue safety factors

Safety and Effectiveness

There is reasonable assurance that a device is safe when it can be determined, based upon valid scientific evidence, that the probable benefits to health from use of the device for its intended uses and conditions of use, when accompanied by adequate directions and warnings against unsafe use,outweigh any probable risks.

There is reasonable assurance that a device is effective when it can be determined, based upon valid scientific evidence, that in a significant portion of the target population, the use of the device for its intended uses and conditions of use, when accompanied by adequate directions for use and warnings against unsafe use, will provide clinically significant results.

17

Medical Device Evaluation

18

Comprehensive evaluation of a premarket submission for a therapeutic medical device is typically supported by a combination of valid scientific evidence from four types of models: animal, bench, computational, and human.

Because each model has different strengths and limitations for predicting real‐world clinical outcomes, the data portfolio for different use‐conditions will vary.

Models and Their Advantages

* M&S in medical devices, as compared to other industries, is nascent and is the one model with the most potential for refinement/improvement because the other models are fairly mature.

Adapted from Victor Krauthamer

Current Uses of CM&S in Premarket Review

20

Modeling is mainly considered a development and design optimization tool, rather than a method by which physical performance of final devices can be demonstrated.

Mainly the modeling studies are supplemental information (to complement mechanical bench testing) for

• 510(k) – class II devices• Pre‐Market Approval – class III devices• Investigational Device Exemption – Clinical Studies


21

Computational Solid MechanicsStents / Heart Valve Frames / Occluders / Vena Cava Filters / AnnuloplastyRings / Dental Implants / Spine & Joint Implants / Bone Plates & Screws / Surgical Tools

Determine the implant size in a device family that is expected to perform the worst under simulated in vivo conditions

o Reduces the amount of physical testingo Calculate Safety Factors for static and cyclic loads

Evaluate the effect of manufacturing tolerances Predicate Comparison Demonstrate a modification (e.g., dimensional) is minor and has minimal affect on performance


22

Computational Fluid DynamicsVentricular Assist Devices / Total Artificial Heart / Blood pumps / Heart Valves / Endovascular Grafts / Drug Eluting Devices

Characterize the flow field by identifying regions of high shear stress, wall shear stress, or areas of low flow or flow stagnation

o especially in regions that cannot be visualized on the bench

Determine blood damage, thrombosis potential, and drug transport using fluid flow properties


23

Computational ElectromagnetismPassive and Active Cardiology Implants / Peripheral Implants / Joint and Spinal Implants / Deep Brain Stimulators / MR‐guided Interventional Devices

Simulate the radiofrequency energy absorbed by patients undergoing magnetic resonance imaging (MRI)

o Especially worst‐case conditions that cannot be replicated in an animal model and cannot be tested ethically in humans

Radiofrequency‐induced currents and heating of (external) devices for electrophysiological recordings

Simulate the electric/magnetic field generated by a device during use to provide evidence of effectiveness


24

Physiological Closed‐Loop Controllers & AlgorithmsAnesthesiology Devices / Respiratory Devices / Artificial Pancreas / Neurodiagnostic Tools

Use the simulation as an alternative validation method to demonstrate device performance and robustness

In silico simulation model (control algorithm) of diabetes replaces in vivo animal testing for evaluating artificial pancreas

Signal modeling (EEG source localizing software) for brain activity analysis


25

Computational Thermal MappingAblation Devices

Determine the thermal field distributions generated by tissue ablation devices (e.g., High Intensity Ultrasound, radiofrequency)

Assess potential damage to surrounding tissue, organs and bones

CM&S are typically Relative• Design Optimization• Virtual Prototyping• Rationale for Pre‐Clinical Evaluation• Root Cause Investigations + Redesigns• Next Generation Designs

Medical Device Development with CM&S

26

CMS REIMBURSEMENT

POSTMARKET

MONITORING

CMSEVALUATION

CMSREIMBURSEDECISION

REAL‐WORLDPRODUCT

PERFORMANCE

FDAREGULATORYDECISION

PRODUCTLAUNCHCLINICAL

PRE‐CLINICALINVENTION

+PROTOTYPING

DISCOVERY+

IDEATION

NEXT GENERATIONPRODUCT

DEVELOPMENT

The Total Product Life Cycle

CM&S OpportunityPredict Product Success/Failure to support:

Regulatory decisionsReimbursement decisions

Medical Device Development with CM&S

27

CMS REIMBURSEMENT

POSTMARKET

MONITORING

CMSEVALUATION

CMSREIMBURSEDECISION

REAL‐WORLDPRODUCT

PERFORMANCE

FDAREGULATORYDECISION

PRODUCTLAUNCHCLINICAL

PRE‐CLINICALINVENTION

+PROTOTYPING

DISCOVERY+

IDEATION

NEXT GENERATIONPRODUCT

DEVELOPMENT

The Total Product Life Cycle

Some of the challengeswith the current practice …

… have led us to address the need for:

guidance on documentation and reporting CM&S results in premarket submissions;• FDA DRAFT Guidance on Reporting of Computational Modeling

Studies in Medical Device Submissions – published Jan 20142

28

2, http://www.fda.gov/MedicalDevices/DeviceRegulationandGuidance/GuidanceDocuments/ucm371016.htm

FDA Guidance

Reporting1 Computational Modeling Studies in Medical Device Regulatory Submissions (DRAFT)2

• Main body discusses the purpose of computational modeling and simulation in regulatory submissions

• Main body presents recommendations for reporting different elements of the computational modeling study

• There are five subject matter appendiceso Fluid & Mass Transport, Solid Mechanics, Electromagnetism, Thermal Transport, and Ultrasound

• DRAFT guidance is available for public comment untilApril 17th!

291, Erdemir, Guess, Halloran, Tadepalli, Morrison, J Biomech. 2012 February 23; 45(4): 625–6332, http://www.fda.gov/MedicalDevices/DeviceRegulationandGuidance/GuidanceDocuments/ucm371016.htm



guidance on documentation and reporting CM&S results in premarket submissions;• FDA DRAFT Guidance on Reporting of Computational Modeling

Studies in Medical Device Submissions – published Jan 20142

systematic assessment and understanding of device‐use conditions;• Critical Path Innitiative3

• FDA Library of Models and Simulation4

• Medical Device Innovation Consortium5

30

2, http://www.fda.gov/MedicalDevices/DeviceRegulationandGuidance/GuidanceDocuments/ucm371016.htm3, http://www.fda.gov/ScienceResearch/SpecialTopics/CriticalPathInitiative/default.htm4, http://www.fda.gov/MedicalDevices/NewsEvents/WorkshopsConferences/ucm346375.htm5, http://mdic.org/projects/computer‐modeling/

FDA Digital Library of M&S

The vision is that the Library will be a mechanism for curating public open‐use models and simulations in a non‐competitive space to foster collaboration and advance research, development and evaluation of medical devices.

Serve as a reference for access to state‐of‐the‐art CM&S and data related to medical products

Mechanism for FDA to transparently communicate utility and expectations of CM&S in a regulatory setting

Being a space for companies to share their “smaller datasets”, e.g., pediatric population, to create a “larger datasets”

31

FDA Digital Library of M&S

Hosted public workshop in June 2013 to introduce concept and openly discuss the Library

Developed key aspects of the infrastructure and framework for use of the Library

Anticipate initially that the Library will be used for curating DATA for creating models and validating simulations, and reference problems

Prototype is underway with sights on pilot in 2014

[email protected]

32“Virtual Physiological Patient”

Medical Device Innovation Consortium

A Public‐Private Partnership collaborating on Regulatory Science to make patient access to new medical device technologies faster, safer and more cost‐effective

Formed in December 2012

33

www.deviceconsortium.org

MDIC Strategic Approach

34

MDIC Initial Program Areas

35



methodologies to experimentally validate CM&S;• FDA Critical Path Initiative to evaluate CFD in blood pump

sensitivity analyses and uncertainty quantification; and• CDRH is actively engaged with the ASME Verification &

Validation Standards Committeeo ASME V&V 10 – Subcommittee on Solid Mechanicso ASME V&V 20 – Subcommittee on Fluid Dynamics and Heat Transfer

ASME V&V 40 – Subcommittee on CM&S for Medical Devices

36

ASME V&V 40 Subcommittee

Subcommittee of ASME V&V Committee• Established in January 2010

Charter: Provide procedures to standardize verification and validation (V&V) for computational modeling of medical devices

Developing a general methodology for industry and academia to establish a V&V plan to assess credibility of CM&S used in medical device design and evaluation

Subgroups working on general methodology, solid mechanics, fluid mechanics and some device specialties (e.g., cardiovascular, orthopedics)

37 38

Credibility Strategy Overview

39

Risk Assessment

Establish Context of Use

CM&S Risk: combination of decision influence and consequence Decision Influence: contribution of

CM&S outcome to the decision being made

Consequence: impact if the CM&S outcomes prove incorrect

Risk assessment• Directs/guides V&V activities• Defines model credibility thresholds

40

LOW

MEDIUM

HIGH

CONSEQUENCE

INFLUEN

CE

Framework to Assess Credibility

Ensuring that the credibility of the CM&S is commensurate with the associated risk provides the confidence that the CM&S is appropriate for the intended COU

Multiple factors contribute to CM&S credibility:

41

Code Verification Sample Characterization

Solution Verification Control Over Test Conditions

System Configuration Measurement Uncertainty

System Properties Discrepancy of Inputs/Outputs

Boundary Conditions Rigor of Output Comparison

Governing Equations Applicability to Context of Use

Topic of Lecture on April 28, 2014

Applicability of V&V Activities

42



methodologies to experimentally validate CM&S; sensitivity analyses and uncertainty quantification; and

• CDRH is actively engaged with the ASME Verification & Validation Standards Committee

o ASME V&V 10 – Subcommittee on Solid Mechanicso ASME V&V 20 – Subcommittee on Fluid Dynamics and Heat Transfer

ASME V&V 40 – Subcommittee on CM&S for Medical Devices

better elicitation of the consequence of the CM&S being incorrect.• CDRH is getting ready to launch a pilot program to expand the

current uses of CM&S in premarket submissions

43

Pilot Program at CDRH

CDRH is launching a pilot program this Spring to implement and evaluate the Credibility Strategy in the review of new uses of CM&S in premarket submissions• Collaboration with OSEL and ODE• Focus on Cardiovascular and Orthopaedic Devices

The Credibility Strategy will also be evaluated as part of the qualification process for the Medical Device Development Tool6 program

It will also be used to help determine which CM&S can be a part of the FDA Library of Models and Simulation

44

6, http://www.fda.gov/MedicalDevices/DeviceRegulationandGuidance/GuidanceDocuments/ucm374427.htm

Medical Device Development Tools

45

MDDT Qualification

Clinical Outcome Assessments

Biomarker Tests

Nonclinical Assessment Models

NAMS are used in device development or evaluation that reflect device function or in vivo performance

• In vitro test methods• In silicomodels

Research – The VPP“Virtual Physiological Patient”A collection of models and data for developing and evaluating medical devices (cardiovascular, orthopaedic, …)

Assessment of plaque composition, dynamic biomechanics, and therapeutic outcomes in subjects implanted with endovascular devices (ASPECT-1)

Goals1. Prospectively collect data on superficial femoral artery (SFA)

deformations and plaque composition in symptomatic patients before and after stent implantation.

2. Create a model using computational simulation techniques to investigate sensitivity to various dimensions and deformations observed in the clinical study.

3. Disseminate and leverage the data collected during the study using open source and educational initiatives.

Tortuosity, Bending & Twisting

Smouse et al. Presented at Transcatheter Cardiovascular Therapeutics (TCT 2004) – published with permission

Bending

Stent

Axial Compression

Walking, climbing stairs, sitting in a chair can deform a stent in different ways.

https://simtk.org/home/opensim

ASPECT‐1• ASPECT is designed to provide

vessel morphology, plaque composition1, and dynamic deformation data that can be used to create realistic models of symptomatic peripheral vascular disease

• Goal is to yield more clinically predictive results from nonclinical evaluation– less reliance on human clinical

testing and unrepresentative animal testing; more durable and effective medical implants.

[1] Bishop PD et al. Arterial calcification increases in distal arteries in patients with peripheral arterial disease. Ann Vasc Surg. 2008

Axial atherosclerotic tissue maps

3D Atherosclerotic tissue maps

ASTM Stent Axial Fatigue Round Robin

Examine stent durability using fatigue to fracture methodology • mechanical testing & FEA of generic, stainless steel design

CDRH/OSEL is conducting mechanical testing, along with many industry laboratories

Interlaboratory test data may be included in a potential “Fatigue to Fracture” standard and be compared to computational modeling results in the future

Typical FractureAxial Fatigue Test Sample Data50

“Virtual Physiological Patient”A collection of models and data for developing and evaluating medical devices (cardiovascular, orthopaedic, …)

CFD Round Robin Study (2010)

GOALS: Improve the use of CFD to evaluate device performance and blood damage safety;

create a data repository of results for public use, and contribute to standards and guidance for using CFD in

regulatory submissions

Phase 1Simple nozzle model

Interlaboratory evaluations• CFD modelers: 28 groups from 6 countries• Validation of velocity fields using quantitative flow visualization

techniques: 3 labs• Blood damage testing of the device: 3 labs

4 mm12 mm 12 mm

52

Laminar Flow Turbulent Flow

Axial Centerline Velocity

Lines: CFD Symbols: Experiment

Phase 1 Results

Large variations between CFD simulations and the experimental data1

531, Stewart SFC, Paterson EG, Burgreen GW, Hariharan P, Giarra M, Reddy V, Day SW, Manning KB, Deutsch S, Berman MR, Myers MR, Malinauskas RA, Cardiovasc Eng Technol. 2012:3;139‐60.

Phase 2: Blood Pump Model (2014)

Open to all CFD modelers Project website: visit https://fdacfd.nci.nih.gov

Blood Testing

Rotor Design

54

“Virtual Physiological Patient”A collection of models and data for developing and evaluating medical devices (cardiovascular, orthopaedic, …)

Investigate Effect of Non-Circular Valve Configurations on Leaflet Dynamics –

Preclinical Test Methods

Non‐circular configurations after percutaneouslyimplanted heart valves (PHVs) depend on several factors

• Such as patient anatomy, number of calcified leaflets, and degree of calcification (Zegdi 2008, 2010; Schultz 2009)

Zegdi et al., 2008

FDA recognizes that there is limited public information on: • Valve preclinical testing characteristics• Leaflet kinematics on non‐circular configurations, and the

affects on leaflet failures CDRH is currently investigating:

• Acute measure of performance using hydrodynamic and accelerated wear tester, and

• Correlations between computational and validation‐specific studies on the tissue leaflets

Demonstrating differences in leaflet kinematics (testing) and stresses (modeling)

Investigate Effect of Non-Circular Valve Configurations on Leaflet Dynamics –

Preclinical Test Methods

Research – The VPPA collection of models and data for developing and evaluating medical devices (cardiovascular, orthopaedic, …)

Virtual Physiological Patient

Background

59

Retrievals and clinical reports

Adverse Reactions to Metal Debris

Bearing wear and taper wear / corrosion

1 Matthies AK, et al. CORR 2012; 470(7):1895‐906 2 Natu S, et al. J Clin Path 2012; 65(5) 409‐183 Bowsher JG, et al. J Biomed Mater Res B. 2009; 91(1):297–3084 Goldberg et al. CORR. 2002; 401:149

1

2

3 4

60

Results under gait loading

Pressure (M

Pa)

Eqv. stress (M

Pa)

Impact Unload Stance Swing

Stance Swing

A

A

Research – The VPP“Virtual Physiological Patient”A collection of models and data for developing and evaluating medical devices (cardiovascular, orthopaedic, …)

Virtual Family

9 different models available, more than 200 organs and 43 tissues,Direct import and automatic material assignment, voxel import of models

female adult1.60m, 58kg

obese adult, 1.78m, 120kg

8 year old boy1.40m, 26kg

8 y girl1.35m, 30kg

5 y girl1.09m, 16kg

male adult, 1.74m, 70kg

14 y boy1.65m, 50kg

11 y girl1.46m, 36kg

6 y boy1.17m, 20kg

63

Human Head ModelsA functional high resolution human head model for analysis of efficacy of medical devices with electrical stimulation

neuronal connectome models for functional modeling of neuro‐stimulation devices.

Human Head Models

64

MRI Radio Frequency Coil ModelsDevelopment and distribution of validated MRI RF Coil Models

MRI Radio Frequency Coil Models

65

Future potential of CM&S for medical device development and evaluation

Digital Patients: Designers can access anatomic and physiologic computer models of (dozens, hundreds, thousands, …) of patients with a given disease.

Virtual Clinical Trials: New device concepts are “deployed” in digital diseased patients and performance is simulated leading to more effective bench testing, animal studies and (actual) clinical trials. Discover “Soft Failures”

Personalized Medicine: Physicians use simulation to predict safety and effectiveness of a given medical product for an individual patient.

66

Summary

Computational modeling and simulation has revolutionized many industries.

CDRH is in the unique position to stimulate and advance device innovation through computational modeling and simulation efforts.

CDRH, through its partnerships, initiatives and projects, is helping to revolutionize and realize the future of CM&S for both design and evaluation.

67

Lead Investigators at CDRH

ASTM Stent Axial Fatigue Non‐Circular Valve ConfigurationsMaureen Dreher, Ph.D. Nandini Duraiswamy, Ph.D.Jason Weaver, Ph.D.

CFD & Blood Damage Round RobinSandy Stewart, Ph.D.Rich Malinauskas, Ph.D.Prasanna Hariharan, Ph.D.

FDA Digital LibraryDonna Lochner

Virtual Family and MRI StudiesASPECT‐1 Leonardo Angelone, Ph.D.Lisa Lim, Ph.D. Wolfgang Kainz, Ph.D.

CD R H

68

ASME V&V40: Verification and Validation in Computational Modeling of Medical Devices

Contact

FDANIH

INDUSTRY

ACADEMIA M&S

NSF

NIST

NASA

DARPA

MDIC

69

Contact Information

[email protected]

Office of Device Evaluation, CDRH

What is Modeling and Simulation? What is a model? · PDF file · 2015-12-04What is...

Documents

Transcript of What is Modeling and Simulation? What is a model? · PDF file · 2015-12-04What is...