Disruptive Technologies as a Change Agent in Pathology Informatics: How this ongoing transformation

1
Disruptive Technologies as a Change Agent in
Pathology InformaticsHow this ongoing
transformation will enrich our practice model

• Ulysses J. Balis, M.D.
• Director, Division of Pathology Informatics
• Department of Pathology
• University of Michigan
• ulysses_at_umich.edu

2
Disclosure

• Although no commercial products or services will
  be discussed by name, the following disclosures
  are provided as required by CME guidelines
• Aperio Technologies, Inc.
• Member, Technical Advisory Board
• Major Shareholder
• Living Microsystems, Inc./ Artemis Health, Inc.
• Founder
• Major Shareholder
• CellPoint Diagnostics, Inc.
• Founder
• Major Shareholder

3
Change Models for Innovation in Technical Fields

• Evolutionary
• Small incremental technological / process
  changes which improve quality, efficiency or
  economy
• Often expected
• Do not change the fundamentals of a domain in
  terms of workflow or practice
• Usually enabled by a plurality of contributors
  (multi-source solutions)
• Usually viewed by the field as a positive
  contribution

• Revolutionary
• Represents an entirely different way of solving
  or addressing a need or problem
• Usually unexpected
• Quality at first is problematic
• Can quickly change the fundamentals of a domain
• Can themselves be interrupted by subsequent
  disruptive changes
• Often provided by sole-source contributors
  (availability mired/ encumbered by intellectual
  property issues)
• Not always viewed by the field as positive
  contribution

4
The stepwise increase in circuit integration can
be seen as a global enabling condition for the
emergence of disruptive technologies in Pathology
Informatics.
5
(No Transcript)
6
(No Transcript)
7
Overview

• General Discussion on Disruptive Technologies
• Human Factors Engineering, Automation in workflow
  and quantization of human performance for
  specific tasks
• Transformative opportunities created by
  fundamentally new optics technologies
• Federation of the enterprise-wide data model
• New reporting models
• Combating commoditization of Surgical Pathology
  Services and The race to the bottom
• The systematic conversion of the combined
  Anatomic Pathology specialties to quantitative
  fields
• Some Demonstrations

8
Some Caveats
9
Overview

• General Discussion of Disruptive Technologies
• Human Factors Engineering, Automation in workflow
  and quantization of human performance for
  specific tasks
• Transformative opportunities created by
  fundamentally new optics technologies
• Federation of the enterprise-wide data model
• New reporting models
• Combating commoditization of Surgical Pathology
  Services and The race to the bottom
• The systematic conversion of the combined
  Anatomic Pathology specialties to quantitative
  fields
• Some Demonstrations

10
Disruptive Technologies

• A Disruptive Technology is a new technological
  innovation, product, or service that eventually
  overturns the existing dominant technology in the
  market
• This occurs, despite the fact that the disruptive
  technology is both radically different than the
  leading technology and that it often initially
  performs worse than the leading technology,
  according to existing measures of performance.

http//en.wikipedia.org/wiki/Disruptive_technology

11
Differential Concept

• Incremental technologies and disruptive
  technologies are fundamentally different in their
  impact and mechanism, yet they both serve the
  overall progression of any given field.
• Disruptive technologies are not incremental.
• Disruptive technologies, by their very
  definition, are difficult to predict and have
  both unpredictable, unintentional and even
  undesirable consequences.

12
Disruptive Technologies

• Disruptive technologies result in worse product
  performance, at least with the onset of their
  deployment.
• Emerge occasionally
• Bring to market a different value proposition
  than available previously
• Underperform established products in mainstream
  markets
• Have features that fringe/new customers value

13
Incremental Technologies May Perform Beyond the
Requisite Metric of Performance(over-engineered)

• Conversely, disruptive technologies may
  eventually attain adequate performance
• Servers vs. mainframes
• Foreign vs. U.S. automotive Industry
• Stent vs. CABG

14
http//upload.wikimedia.org/wikipedia/en/8/8e/Di
sruptivetechnology.gif
15
Examples
http//en.wikipedia.org/wiki/Disruptive_technology
Examples_of_disruptive_technologies
16
Overview

• General Discussion on Disruptive Technologies
• Human Factors Engineering, Automation in workflow
  and quantization of human performance for
  specific tasks
• Transformative opportunities created by
  fundamentally new optics technologies
• Federation of the enterprise-wide data model
• New reporting models
• Combating commoditization of Surgical Pathology
  Services and The race to the bottom
• The systematic conversion of the combined
  Anatomic Pathology specialties to quantitative
  fields
• Some Demonstrations

17
Human Factors Engineering

• Design devices, technologies and workflow to
  optimize human performance.
• A key element of process design and redesign is
  the monitoring of metrics, made possible by
  automation.
• An intrinsic element of cogent Lean and Six Sigma
  Implementation approaches

18
Human Factors EngineeringPerformance Review in
AP

• How could this become disruptive?
• The Pathologist becomes a data point of study for
  continuous quality monitoring
• TAT / Automated Kappa Statistic generation
• Credentialing could be tied to automated metrics
  of performance (similar to the U.S. Navy Pilot
  program for Carrier Pilots)
• Formalizes the peer consultation model, which is
  largely based at present on personal judgment,
  without automation.
• Initially, such measures might be poorly received
  by the Pathology Community

19
Human Factors EngineeringTransition from
Primary Diagnosis to Directed Expertise Models

• In this possible scenario, the pathologist would
  spend most of their time reviewing microscopy
  fields that were prescreened by quantitative
  algorithms.
• How could this become disruptive?
• Activity / effort shifted from review of all
  material to selected re-review of pre-screened
  content
• Potential multiplier effect for the effective
  throughput of each pathologist
• Fundamentally changes the role of the
  pathologist.
• May be an effective staffing solution for the
  anticipated shortage in Surgical Pathologists

20
Overview

• General Discussion on Disruptive Technologies
• Human Factors Engineering, Automation in workflow
  and quantization of human performance for
  specific tasks
• Transformative opportunities created by
  fundamentally new optics technologies
• Federation of the enterprise-wide data model
• New reporting models
• Combating commoditization of Surgical Pathology
  Services and The race to the bottom
• The systematic conversion of the combined
  Anatomic Pathology specialties to quantitative
  fields
• Some Demonstrations

21
Super Resolution Microscopy

• Super-resolution (SR) are techniques that in some
  way enhance the resolution of an imaging system.

http//en.wikipedia.org/wiki/Super-resolution
22
Super Resolution Microscopy

• In the specific application of Microscopy,
  resolution is currently limited by the
  diffraction-limited resolution of light
• 0.2 microns for a quality 100x oil objective
• In theory, such methods can approach the
  resolution of electron microscopy

Numerical Aperture (NA) n(sin m)
R /(2NA)
23

• Implications
• 5-40 x increase in spatial resolution for light
  microscopy
• Near-EM capabilities for both light and
  fluorescence microscopy
• Five to ten years from widespread availability
• May required an entirely new cohort of
  pathologists to effectively interpret resultant
  optical information

24
(No Transcript)
25
Overview

• General Discussion on Disruptive Technologies
• Human Factors Engineering, Automation in workflow
  and quantization of human performance for
  specific tasks
• Transformative opportunities created by
  fundamentally new optics technologies
• Federation of the enterprise-wide data model
• New reporting models
• Combating commoditization of Surgical Pathology
  Services and The race to the bottom
• The systematic conversion of the combined
  Anatomic Pathology specialties to quantitative
  fields
• Some Demonstrations

26
Should Complex Data Integration Be at the Lab or
Hospital IT Level?

• Current Hospital IT departments are at a
  crossroads in terms of being able to support a
  rapidly expanding portfolio of data sources and
  complexity.
• Concurrently, central IT departments are being
  pressured are to consolidate service models to a
  rarified strata of standardized solutions (single
  vendor approaches)
• Loss of capture of local unique and effective
  workflow
• Loss of application enhancement pathway
• Loss of productivity
• Decreased effectiveness with data management and
  clinical communication
• Decreased net quality of patient care

27
HL7 Receiver
HL7 Receiver
Just-in-time Aggregation
Multiple unique instances of customized HL7
interfaces which are not necessarily centrally
managed or organized for consistency in
deployment style
Just-in-time delivery
HL7 Receiver
Repository
Future State
Present State
28
What is Federation?
Conventional Data Model
LIS
User
Centralized HIS
RIS
OR
other
E.D.
Conventional HL7 interface
29
What is Federation?
Potential Revised Data Model
LIS
RIS
OR
E.D.
other
Web-based Just-in-time aggregation
User
Single SQL (or ODBC) query Concurrently vectored
to each Participatory Single Source of Truth
SSOT shifted to the appropriate domain-specific
stewards of data from the HIS domain
Participatory SQL servers
30
What is Federation?
Potential Revised Data Model
LIS
RIS
OR
E.D.
other
Web-based Just-in-time aggregation
User
Single SQL (or ODBC) query Concurrently vectored
to each Participatory Single Source of Truth

• Consequences of shifting to a SQL-based SSOT
  model
• Data only represented once in overall enterprise
  model
• Reduction in number of interfaces requiring
  support
• Potential to transfer classes information other
  than text
• Reduction in support responsibilities of central
  hospital IT.

Participatory SQL servers
31
Overview

• General Discussion on Disruptive Technologies
• Human Factors Engineering, Automation in workflow
  and quantization of human performance for
  specific tasks
• Transformative opportunities created by
  fundamentally new optics technologies
• Federation of the enterprise-wide data model
• New reporting models
• Combating commoditization of Surgical Pathology
  Services and The race to the bottom
• The systematic conversion of the combined
  Anatomic Pathology specialties to quantitative
  fields
• Some Demonstrations

32
Hematology as a model test-bed

• Reports often rely upon disparate types of data
  from separate laboratories and device/ laboratory
  information systems.
• Primary data is often acquired asynchronously
• Integrative reporting is co-dependent on a
  plurality of both current data are archived data
  (for longitudinal view of disease evolution)
• Reports can be monolithic or multi-part
• Billing can be complex
• Clinical communication can be complex and staged

33
Use Case 1 Integrate Hematology results with
Bone Marrow Intrepretations and Flow Cytometric
Data
Central LIS
Hem
Flow
AP
other
User
Web-based Just-in-time Aggregation, analysis
and report generation tools
LIS
HIS
demonstration
34
Use Case 2 Generate an Intramural longitudinal
View of patient data with legacy architectures
AP
other
User
Web-based Just-in-time Aggregation, analysis
and report generation tools
LIS Federated Shadow
LIS
Central LIS
HIS
Legacy Architecture
Hem
Flow
35
Use Case 3 Develop a Staged Notification Model
for Clinical Results Reporting
Central LIS
Hem
Flow
AP
other
User
Web-based Just-in-time Aggregation, analysis
and report generation tools
Rules Engine / Trigger Logic Database
LIS
HIS
36
Deployment Realities

• Federation is not limited to large departments
  with significant I.T. expertise.
• Legacy architecture is not a barrier for
  deployment of federated solutions.
• Solutions can be small in scope initially and
  then expanded as peers are added to the
  federation.
• Use of standard approaches in the federation
  ensures that the resultant architecture will
  stand the test of time.

37
Overview

• General Discussion on Disruptive Technologies
• Human Factors Engineering, Automation in workflow
  and quantization of human performance for
  specific tasks
• Transformative opportunities created by
  fundamentally new optics technologies
• Federation of the enterprise-wide data model
• New reporting models
• Combating commoditization of Surgical Pathology
  Services and The race to the bottom
• The systematic conversion of the combined
  Anatomic Pathology specialties to quantitative
  fields
• Some Demonstrations

38
Commoditization

• As national surgical pathology labs and
  franchises continue their predatory trend of
  infiltrating local markets, while making
  significant use of Information Technology as a
  differentiating component, price and quality will
  be driven to absolute minima this will bring
  into peril the current model of hospital-based
  practices.
• Bruce Friedman has coined the term Race to the
  Bottom, highlighting the intrinsic danger in
  quality for our profession, stemming from this
  systematic threat that reduces the entire
  specialty to a financial proposition, above all
  else.
• An effective counterstrategy may be for local
  practices to similarly leverage equivalent I.T.
  solutions, thus providing an equal measure of
  value-added features as compared to the
  infiltrating national labs.

39
Overview

• General Discussion on Disruptive Technologies
• Human Factors Engineering, Automation in workflow
  and quantization of human performance for
  specific tasks
• Transformative opportunities created by
  fundamentally new optics technologies
• Federation of the enterprise-wide data model
• New reporting models
• Combating commoditization of Surgical Pathology
  Services and The race to the bottom
• The systematic conversion of the combined
  Anatomic Pathology specialties to quantitative
  fields
• Some Demonstrations

40
(No Transcript)
41
ChallengesSize of whole slide images
X ray image from Dah http//farm3.static.flickr.c
om/2114/2141439225_4b996372df.jpg?v0
From Ian Fore - CaBIG/ National Cancer Institute
42
ChallengesStandards

• Image standards
• DICOM is not used for pathology images
• Usually proprietary or TIFF/JPEG file formats
• Annotation standards
• Different metadata
• E.g. stain type HE
• Efforts to agree standards for pathology imaging
  do exist
• DICOM, LDIP, OME

From Ian Fore - CaBIG/ National Cancer Institute
43
Current World View of Pathology Imagery
Repositories

• Model 1 Relational Database
• Image Metadata associated with case-level data
• Entire Schema required to carry out discovery
• Text-based
• Image data is a passive component of the query

• Model 2 Metadata-tagged Images
• Image Metadata associated with each image
• Image becomes a self-contained dataset available
  for discovery
• Text-based
• Image data is a passive component of the query

Entry in master accession table
Associated case and image descriptors
44
Highly Desirable World View of Pathology Imagery
Repositories (Future State)

• Model 3 Metadata-tagged surface map
• Image Metadata exists at the image level and is
  spatially coupled to underlying digital imagery
• Discovery can be carried out on the image-space
  itself, with retrieved metadata classifiers
  available for generating search result sets (e.g.
  differential diagnosis generation)
• Image-based

• Model 4 Surface discovery
• Non-metadata-associated digital imagery is
  spatially probed for statistical convergence with
  an image-based query set
• Imagery becomes a self-contained dataset
  available for discovery
• Image-based

?
?
45
Digital Representation of Images as the Key
Transformative Element Enabling Digital Microscopy

• Without the image data in digital format, there
  is no cogent question that can be asked, as there
  is not dataset available to query.
• With the advent of increasingly comprehensive
  digital image repositories, we encounter an
  entirely different situation essentially an
  embarrassment of riches as we now have more data
  than is easily parsed by conventional linear
  programming.
• this is a transformative enabling step,
  nonetheless
• As a confirming reality check Radiology has
  already firmly entered the realm of investigation
  of computer aided diagnosis (CAD), although it is
  cogent to recognize that their current datasets
  are much smaller that those now possible with
  digital whole-slide imaging
• And as such, the question becomes one of
  algorithmic and heuristic development.
• Hint, we know this is possible, as the human
  brain carries out real-time CBIR with high
  sensitivity and specificity.
• Caveat recognizing the that human brain is
  massively parallel in construction,
  recapitulating this with current computational
  technology may be impractical

46
Compelling Use Cases for Image Query

• Diagnostic decision support
• Longitudinal evaluation
• Differential diagnosis generation
• Detection of rare events
• Teaching
• Discovery

47
Attributes of an ideal search system

• Self-training, domain independent image
  segmentation / classification tool.
• Allows for at least two novel image search
  modalities
• Region of interest Query by example (image space
  search not text based)
• Retrieve diagnostic information associated with
  prior classified fields, enabling the generation
  of dynamically generated differential diagnosis
• Useful as a bridge for exploration of stochastics
  of multi-dimensional image space data when
  queried in tandem with high-dimensionality data
  sets types (genomics, proteomics, etc.)
• i.e. Morphogenomics
• Ability to carry out real time assessment of
  regions of interest against Terascale / Petascale
  image repositories.

48
1.415461031044954789001553027745e9864
2 x 2 vector 2564 possible values in a
four-dimensional space
What is an Image Vector?
4,294,967,296 possible values
Typically, vectors have ordinality of 8 x 8 or
greater
49
An Issue of Dimensional Reduction

• Problem With the prospect of a typical 100x100
  kernel (10,000 dimensional spaces), computational
  approaches carried out on raw data sets can take
  millions of years to complete, even with our
  fastest current supercomputers. (bad for
  turn-around time)
• Fortunately, there are mathematical operations
  that can sidestep this computational annoyance.
• Support Vector Engines
• K-means approaches
• Bayesian Networks
• Vector Quantization
• Galois Field Manifold Projection / Tensor
  Integration

50
Vector Quantization
Original Image
Division of image into local domains
Extraction of Local Domain Composite Vectors
?
VKSLx0y0Order , LxnymOrder
Vectorization of each local kernel
Individual assessment of each vector dimension
51
Vector Quantization
VKSLx0y0Order , LxnymOrder
Established Vocabulary
Query Against library (Vocabulary) of established
Galois Vectors
Novel Vector
Previously Identified Vector
Assignment of a unique serial number and
inclusion into global vocabulary
Assembly of compressed dataset
38857448643
52
VQ-Based Image Compressiona fantastic
opportunity for automated search
Raw Data
Restored Data
Compressed data (preserved spatial organization
of original data)
Depending on the selected compression ratio,
restored loss-compression imagery may or may not
be of diagnostic quality.
53
Pythagorean Theorem
b
916 does indeed result in 25
5 x 5
3 x 3
a
4 x 4
b
54
Which, after integration by parts, yields
55
Typical Galois Field mapped to the even
Jacobian/Chebyshev tensor polynomials manifested
on the edge of the complexity transition

• On Galois Fields
• Not merely a clustering algorithm
• The resulting field is a non-linear N-space
  manifold selected for its distinctiveness from
  all other modular functions in the Galois set
  space
• Fields may have local minima and local extrema
• Any Galois manifold is exclusive of any other
  Galois set
• Non-trivial to calculate trivial to query

56
Local Islands in Galois Field Space of
statistical convergence and near-convergence to
high-probability feature matches using support
vector analysis
57
Regions of a typical Galois manifold with no
correlation to established vocabulary tensors are
easily recognized as exhibiting chaotic behavior
and are therefore excluded.
58
Overview

• General Discussion on Disruptive Technologies
• Human Factors Engineering, Automation in workflow
  and quantization of human performance for
  specific tasks
• Transformative opportunities created by
  fundamentally new optics technologies
• Federation of the enterprise-wide data model
• New reporting models
• Combating commoditization of Surgical Pathology
  Services and The race to the bottom
• The systematic conversion of the combined
  Anatomic Pathology specialties to quantitative
  fields
• Some Demonstrations