The SCAR TRIP Initiative

yardbellΤεχνίτη Νοημοσύνη και Ρομποτική

14 Νοε 2013 (πριν από 3 χρόνια και 1 μήνα)

69 εμφανίσεις

UC

SF LRI

The SCAR TRIP
TM

Initiative
& DICOM

Katherine P. Andriole

S
ociety for
C
omputer
A
pplications in
R
adiology


PACS Clinical Coordinator

University of California at San Francisco

Department of Radiology

Laboratory for Radiological Informatics

and

Department of Bioengineering

University of California at Berkeley


UC

SF LRI

OUTLINE


The Problem


The SCAR TRIP
TM

Initiative


Historical Review


Imaging in Other Fields vs Medicine

»
Entertainment Industry, DoD & NASA


UC

SF LRI

OUTLINE


Concepts Involved


Human Perception, Image Processing,
Visualization, Navigation, Usability,
Standards, Databases, Integration,
Evaluation, Validation

UC

SF LRI

OUTLINE


Affected Processes


Interpretation, Communication, Workflow
& Efficiency, Diagnostic Accuracy,
Quality of Care


Role of / Impact on DICOM


Incorporated but not widely used concepts


Necessary new features & functionality



UC

SF LRI

The Problem


Information & Image Data Overload


Requires
medical image interpretation
paradigm shift

to evaluate, manage &
exploit the massive amounts of data
acquired for improved


Efficiency


Accuracy


Survival

UC

SF LRI

The SCAR TRIP
TM

Initiative


T
ransforming the
R
adiological
I
nterpretation
P
rocess


to
spearhead

research, education, &
discovery of
innovative solutions to
address the problem of information
& image data overload
.

UC

SF LRI

SCAR TRIP
TM

Initiative


Radiology must
shift its image
interpretation & management processes

to deal with the burgeoning medical
image data sets acquired by digital
imaging devices.

UC

SF LRI

SCAR TRIP
TM

Initiative


Will
foster interdisciplinary research
on technological, environmental &
human factors

to better manage &
exploit the massive amount of data.

UC

SF LRI

SCAR TRIP
TM

Initiative


Will

focus on:


Improving
efficiency of interpretation


Improving
timeliness & effectiveness



Decreasing medical errors


Goal is to improve the quality & safety of
patient care.

UC

SF LRI

Historical Review


Why Is
Medicine So Far Behind?

(DoD, NASA, Hollywood)


Special & Challenging Environment


Urgency of Results


Safety Limitations & Restrictions


Cost of Error


Tremendous Variability of Human
Data within & between Individuals.


UC

SF LRI

Why Is Medicine So Far Behind?


Special & Challenging Environment


Difficult to Validate Performance


Poor Understanding of Human
Perception & its Relationship to the
Art of Medicine.

UC

SF LRI

Why Is Medicine So Far Behind?


Slower Adoption of Technology in General


Cultural & Practicality Barriers


More Difficult to See Clinical Impact Initially


Interdisciplinary

Nature of the Solution


UC

SF LRI

Often there is a disconnect between

Scientist
-
Researchers & End
-
Users
in the Clinical Arena

UC

SF LRI

Enabling Technologies

(creating urgency for TRIP
TM
)


Computing & Networking Capabilities


“Real
-
Time” Processing


Increased Bandwidth & Ubiquitous Access


Visualization Technologies


3
-
D Rendering, Color, Motion

UC

SF LRI

Enabling Technologies


Digital Imaging Modalities


True 3
-
D Data Acquisition & Isotropic Voxels


More Intuitive Graphical User Interfaces


Although much more needs to be done

UC

SF LRI

Concepts Involved


Human Perception


Image Processing & CAD


Visualization


Navigation


Usability


Standards, Databases & Integration


Evaluation & Validation

UC

SF LRI

Human Perception


Develop a
Standard for Image Quality


Develop
Objective Methodologies & Criteria


From which to determine optimal
presentation parameters


Based on Diagnostic Performance


Develop
Display Standards

UC

SF LRI

Psychophysical Models for
Detection of Abnormalities


Define & Develop Optimal Presentation
Parameters by understanding


What is desired by the observer


What properties of radiological images are
most useful in their interpretation


How can these properties be enhanced to
improve accuracy of interpretation.

UC

SF LRI

DICOM Role


WG 11
: Display Function Standard


Gray Scale Std Display Function GSDF


Presentation
-
LUT


IHE
: Consistent presentation of images


AAPM TF18
: Image Quality, QA


Still must address

Clinical Correspondence


UC

SF LRI

Image Processing & CAD


Man
-
Machine Systems for Image
-
Based
Diagnosis

which take advantage of both
human & machine capabilities.


Relinquish more routine chores to the
computer.


Have human concentrate on judgment
& comprehension tasks.

UC

SF LRI

Image Processing & CAD


Develop Computer Aids for Feature Perception


Cuing, Overlay & Annotation


Develop Radiology Workstation of the Future


Implement computer aids into a broadly
supportive workstation.


Decision Support, Data Mining & Reference
Libraries

UC

SF LRI

Image Processing & CAD


Design a workstation that can grow to
accommodate future computer tools &
advances.


Support clinical, research & teaching
needs.

UC

SF LRI

DICOM Role


Image processing
capabilities

at the
PACS display are currently very
minimal
.


Processing typically done at the modality
and/or required
specialty workstations
.



How can DICOM pass image processing
parameters without disclosing proprietary
information?


Structured Reporting & CAD (
WG8

&
15
)

UC

SF LRI

Visualization


Static Film


Dynamic Soft Copy & Image Manipulation


Tile Mode


Stack or Cine Mode


Linked Stack Mode for 3
-
D Correspondence


Multimodality Image Fusion

UC

SF LRI

UC

SF LRI

Combining Functional &
Anatomical Information

UC

SF LRI

3D Spectra Anatomy Overlay

“Normal”

Tumor

Necrosis

Courtesy Cynthia Chin, M.D., UCSF

UC

SF LRI

Visualization


Maximum Intensity Projection


Multi
-
Planar Reconstruction


3
-
D Surface/Volume Rendering


Virtual Reality Representations


???


UC

SF LRI

CT Cholangiogram
-

Axial

Courtesy Richard S.Breiman, M.D., UCSF

UC

SF LRI

Sliding MIP

Courtesy Richard S.Breiman, M.D., UCSF

Bile Duct
Anomalies
missed by
MRCP in
potential
partial liver
donors.

UC

SF LRI

Courtesy Gary R. Caputo, M.D., UCSF

3
-
D Surface/Volume Rendering

UC

SF LRI

Courtesy Cynthia Chin, M.D.,
UCSF

UC

SF LRI

DICOM Role


Currently most
3
-
D

representations must be



processed on
specialty workstations


some must be saved as screen
-
capture


manually push to PACS workstations &
Enterprise
-
wide Web (if capable of displaying)


Raw data
often

not stored
.


UC

SF LRI

DICOM Role


How can DICOM pass 3D Model without
disclosing proprietary information?



How simplify interoperability?


Unify Architecture


UC

SF LRI

DICOM Role


DICOM conceived as a strategy for moving &
storing
collections of single images
.


Network utilization is suboptimal


PACS must accommodate
multiple images

which can be treated as a
single unit


Series
-
Awareness, 3D, 4D, Functional Sets,
Cross
-
Referencing of Objects & Fusion


Unified presentation of
Color
WG11

& others.

UC

SF LRI

DICOM Role


WG16, Supplement 49

defines multiframe
(MR) images; model for CT;
WG17
,
20
,
21
.


enhanced image storage SOP class


allows multiple images to be combined
into one instance


Raw Data


Dimensionality


Context Info

UC

SF LRI

Navigation & Usability


3
-
D & Motion


Virtual Reality


Fly
-
Throughs


Hand
-
Eye Cues


Hand
-
Helds for Point
-
of
-
Care Delivery


Context Matching


Voice Activation


???


UC

SF LRI

3
-
D Surface Rendering

CABG

Courtesy Gary R. Caputo, M.D., UCSF

UC

SF LRI

Virtual Reality Fly
-
Through
of Coronary Arteries

Courtesy Gary R. Caputo, M.D., UCSF

UC

SF LRI

Sliding VR

Courtesy Richard S.Breiman, M.D., UCSF

UC

SF LRI


Michael Teistler, Technical Institute of Braunschweig

UC

SF LRI

Hand
-
Helds for Point
-
of Care
Delivery

UC

SF LRI

DICOM Role


Navigation by radiologist/clinician

at the PACS display (or enterprise
-
wide web) in real
-
time


Raw Data & Processing Model


Color Encoding


Overlays


Waveforms


Audio or Other Sense?

UC

SF LRI

Standards, Databases &
Integration


Open Standards


Real
-
Time Processing at PACS Display


3
-
D Integrated into PACS Display & Web


Other Relevant Data


Integrated HIS
-
RIS
-
PACS
-
Speech & IHE (maintaining
user & patient focus)


UC

SF LRI

Evaluation & Validation


Objective Methodologies


Standard Datasets for Performance Testing


Collaborative & Comparison Research

UC

SF LRI

Affected Processes


Interpretation


Communication


Workflow & Efficiency


Diagnostic Accuracy


Reduction of Medical Errors


Quality of Care


UC

SF LRI

We Have Come a Long Way,
But…

UC

SF LRI

What SCAR Hopes To Do


Bring Forward the Problem


Facilitate Exchange of Ideas


Between Researchers, End
-
Users,
Industry, Other Fields


Via Workshops & Forums


By Lobbing NIH & Other Agencies


Sponsor Research


Communicate Issues & Results

UC

SF LRI

DICOM Role

(especially)

WG4

Compression


WG8

Structured Reporting

WG10

Strategic
WG11

Display Function Std

WG16

Magnetic Resonance,
Sup49

WG17

3D

WG20

Imaging & Information Systems Integration

WG21

Computed Tomography


UC

SF LRI

DICOM Role

Join in the TRIP!