Transcript “Its all about the timing…”

“Its all about the timing…
Echo cardiogram enhancement through
ECG/EKG triggering”
A presentation by Prof. David A. Rivkin, PhD
President & Chief Scientist, GCPI
Adjunct Professor, Drexel University
IEEE
Senior Member, EMBS Distinguished Lecturer
Chairman, IEEE SCV Instrumentation and Measurement Society
Chairman, IEEE Instrumentation and Measurement Society Technical
Committee 34 “Nanotechnology In Instrumentation and Measurement”
IEEE Nanotech Council Executive Member
AdvaMed Medical Device Industry Council Advisor
ISO TC-229 “Nanotechnology Standards” IEEE Liaison
Chairman, STI Emerging Technologies SIG
Fellow, Society for Industrial & Applied Mathematics
Timing of the Presentation







The Problem – Blurry Pictures…
Previous Solutions – Big Computers…
New Solution – Embedded Monitor/Trigger
Unified Solution – Best Timely Results
Overview of Design – How it works
Overview of Management – Why its on time
Conclusion – Its about time
The Problem – Blurry Pictures






X-Ray
Echo Cardiogram
CT
3D Ultrasound
PET
Electrical
Impedance
Tomography (IET)
The Problem – Blurry Pictures



Issue
Patient Moves
Breathing
Heart Beating
Possible Solution
 Restrain Patient
 Hold Breath
 Nothing…till now
The Problem – Echo Cardiograms



Measurements of
cardiac muscle and
arteries need to be
taken
Heart Beating
Dimensions Change
over Heart Rhythm
Cycle
The Problem – 3D Imaging




3D Ultrasound
3D Echocardiogram
X-Ray Computed
Tomography
MRI
The Problem – Invasion



When Imaging Is
Inconclusive, Invasive
measures must be
taken
E.g. Open Heart
Surgery
Always a risk of
complications
Previous Solutions –
Big Computers…






NASA
Bartron Medical
Imaging (defunct)
GCPI
Redundant Shadowing
Advanced CT algorithms
Cluster Computing &
Supercomputers
Previous Solutions –
Big Computers…






NASA
Bartron Medical Imaging
GCPI
Redundant Shadowing –
Determine Regions of
Motion Causing Error
Advanced CT algorithms
Compensate for Motion
with estimation and
redundant data
New Solution –
Embedded Monitor/Trigger



Electro Cardiogram
(ECG/EKG)
Use ECG Data to Trigger
Imager
Signal Processing to
Determine Trigger Point
New Solution –
Embedded Monitor/Trigger

Electro Cardiogram
(ECG/EKG) Data:



Well known
Simple to collect
Up to 12 Point Data
Collection
New Solution –
Embedded Monitor/Trigger


Peaks can act as
simple triggers
E.g. Trigger on
Ventricular
Contraction
New Solution –
Embedded Monitor/Trigger


ECG Signals Indicate
more complex
activities of the heart
than just Ventricular
Contraction
Any Phase of the
Cardiac Rhythm can
be determined
New Solution –
Embedded Monitor/Trigger


ECG with Signal
Processing can Trigger at
any point, on any peak or
with delays
Some imagers (such as
ultrasound) have inherent
delay allowing for
triggering to be after the
image is already collected
New Solution –
Embedded Monitor/Trigger


ECG based software
control of triggering
Advanced algorithms
can be used to trigger
only upon
abnormalities in ECG
data
New Solution –
Embedded Monitor/Trigger


Proper triggering can
enhance location of
abnormalities in heart
dimensions or
function
E.g. Abnormal Blood
Flow indicating hole in
heart wall otherwise
not detected
Unified Solution –
Best Timely Results

Combining Advanced
Triggering with
Advanced CT
Algorithms results in
the best images
Overview of Design – How it works


Standard 3M RedDot adhesive electrodes
Standard connectors and cabling
Overview of Design – How it works

Virtex-4 LX25 FPGA






Mechanical Specification


24,192 Logic Cells
128 MB RAM
64 MB Flash ROM
JTAG Hardware Debugging
48 Billion Multiply and Accumulates
/ Second
Temperature Range: -40° C to
+85° C
Developed on Standard Virtex4
Prototype Board
Overview of Design – How it works

Virtex4 Versions
Considered




DX - PowerPC
Core/DualCore
SX - XtremeDSP Slice
Overkill for the
application
Would have taken
longer to develop
Overview of Design – How it works





Developed Entirely in
SystemC
Prototyped on Dual
Processor AMD64 desktop
SystemC to VHDL by
SystemCrafter SC
Xilinx IDE WebPack
GBit Ethernet, USB or
Serial Update of EEPROM
Overview of Design – How it works




Isolation/Instrumentation
Amplifiers and Isolation
Power Supply
TTL, CMOS, RS-232 and
USB Trigger Lines Provided
for different manufacturers
interfacing needs
Miniature Stainless Steel
Enclosure
EMI Shielded Cabling
Overview of Management –
Why its on time…and under budget





5 Months from conception to First Articles
generated by CEM, scheduled for 6 months
Under Budget by 4%
Entering Clinical Trials currently
Compliant with ISO13485:2005 and 21CFR820
“Avoid Verbal Orders, Say It In Writing”
Overview of Management –
Why its on time…and under budget

Program Management
Under:



GPES
Stochastic Methods
Quality Scientific &
Engineering Collaboration


NASA
Distributed Engineering
Team at GCPI
D   Te  Z

2
Overview of Management –
Why its on time…and under budget




GPES – Global
Performance
Evaluation System
Real-Time Program &
Project Status
Developed by NASA at
Johnson Space Center
Open Source ASP on
MS SQLServer
Overview of Management –
Why its on time…and under budget

GPES Source Code
Modified to Use Stochastic
Methods



(o) = Most optimistic
duration - everything will
go according to plan with
a minimal amount of
difficulties.
(p) = Most pessimistic
duration - everything will
not go according to plan
and maximum potential
difficulties will develop.
(m) = Most likely duration
- will most often occur
should this effort be
repeated over and over
again
T

e
2
Z
= the sum of all critical path expected
times.
= the sum of all critical path variances.
= the number of standard deviations of a
normal distribution (the standard normal
deviate)
TE 
o  4m  p
6
 p  o
Variance   2  

 6 
D   Te  Z
2
2


Values for Z can be taken from “Project Management, A Managerial Approach,”
Jack R. Meredith and Samuel J. Mantel, Jr., John Wiley & Sons, NY, 1995
Overview of Management –
Why its on time…and under budget

Quality Scientific & Engineering Collaboration

NASA – 5 Researchers



GCPI – 18 Engineers (EEs, SEs, ME)



3 Offices: San Jose, Baltimore, London
Telecommute Design Staff
Medical Imaging Partners


AMES Research Center
Johnson Space Center
GE, Philips
Contract Manufacturing Partners

Flextronics, Benchmark
Conclusion – Its about time


New and exciting technologies are
changing the way products are developed.
SystemC offers new ways to design,
prototype and develop unified
Software/Hardware products using nontraditional computing methods
Conclusion – Its about time!

Any Questions?

Thank you for your patience!


By Prof. David A. Rivkin, PhD
[email protected]