Knowledge-based Information Management for Biomedical Applications

1
Knowledge-based Information Management for
Biomedical Applications

• Wesley Chu
• Computer Science Department
• University of California
• Los Angeles, CA
• wwc_at_cs.ucla.edu
• www.kmed.cs.ucla.edu

2
Outline

• Data types
• Uses of knowledge bases to enhance information
  management
• Sample systems
• Structured data
• Multi-media
• Free-text
• Conclusion

3
Information Formats used in Biomedical
Applications

• Structure Data
• Multi-media Images
• Semi-structure
• Free-text

4
Uses of Knowledge Bases to Enhance Information
Management

• Approximate matching
• Query conditions
• Image features
• Similar conceptual terms

5
Uses of Knowledge Bases to Enhance Information
Management

• KB query processing
• Similarity query answering
• Associative query answering
• Scenario-specific query answering
• Sentinel --Triggering and alerting

6
Examples of KB Information Systems

• CoBase (1990-1998), DARPA
• A database that cooperates with the user for
  structure data
• KMeD (1991-2000), NSF
• A Knowledge-based medical multi-media database
• Medical Digital Library (2001-2005), NIH
• A knowledge-based digital file room for patient
  care, education, and research.

7
CoBase www.cobase.cs.ucla.edu

• Project leader Wesley W. Chu

• Graduate studentsK. Chiang
• C. Larson
• R. Lee
• M. Merzbacher
• M. Minock
• Frank Meng
• Wenlei MaoMark Yang
• K. Zhang

• Staff
• Q. Chen
• Gladys ChowHua Yang

8
CoBase Cooperative Databases

• Conventional query answering
• Need to know the detailed data based schema
• Cannot get approximate answers
• Cannot answer conceptual queries
• Cooperative query answering
• Derive approximate answers
• Answer conceptual queries
• Provide additional relevant answers that user
  does not (or does not know how to) ask for

9
Cooperative Queries
CoBase Servers
Heterogeneous Information Sources
Find a nearby friendly airport that can land F-15
Find hospitals with facility similar to St.
Johns near LAX
CoBase provides Relaxation
Approximation Association Explanation
Domain Knowledge
Find a seaport with railway facility in Los
Angeles
10
Generalization and Specialization
11
Cooperative Querying for Medical Applications

• Query
• Find the treatment used for the tumor similar-to
  (loc, size) X1 on 12 year-old Korean males.
• Relaxed Query
• Find the treatment used for the tumor Class X on
  preteen Asians.
• Association
• The success rate, side effects, and cost of the
  treatment.

12
Type Abstraction Hierarchies forMedical Domain
13
KB Type Abstraction Hierarchy

• Using clustering technique to group similar
• Attribute values
• Image features
• Spatial relationships among objects
• Provides multi-level knowledge (conceptual)
  representation

14
Data mining for TAH for NumericalAttribute Values

• Clustering metrics relaxation error
• Difference between the exact value and the
  returned approximate value
• Relaxation error is weighted by the probability
  of occurrence of each value
• Can be extended to multiple attributes

15
Query Relaxation
16
Summary CoBase

• Derive Approximate Answers
• Answer Conceptual Queries
• Provide Associative Query Answers

17
KMeD www.kmed.cs.ucla.edu

• PI Wesley Chu, Ph.D, Computer Science Department
• Co-PIs
• A. Cardenas, Ph.D, Computer Science Department
• Ricky Taira , Ph.D, School of Medicine

• Graduate studentsAlex Bui
• Chrisitna Chu
• John Dionisio
• T. PlattnerD. Johnson
• C. Hsu
• T. Ieong

• ConsultantsDenies Aberle, M.D.
• C.M. Breant, Ph.D

18
KMeD Goal Retrieval of Images by Features
Content

• Features
• size, shape, texture, density, histology
• Spatial Relations
• angle of coverage, shortest distance, overlapping
  ratio, contact ratio, relative direction
• Evolution of Object Growth
• fusion, fission

19
(No Transcript)
20
(No Transcript)
21
(No Transcript)
22
Characteristics of Medical Queries

• Multimedia
• Temporal
• Evolutionary
• Spatial
• Imprecise

23
(No Transcript)
24
(No Transcript)
25
Knowledge-Based Image Model
Representation Level (features and content)
26
Queries
Query Analysis and Feature Selection
Knowledge- Based Query Processing
Knowledge-Based Content Matching Via TAHs
Query Relaxation
Query Answers
27
User Model

• To customize users
• interest and preference, needs, and goals.
• e.g. query conditions, relaxation control,
  etc.
• User type
• Default Parameter Values
• Feature and Content Matching Policies
• Complete Match
• Partial Match

28
User Model (cont.)

• Relaxation Control Policies
• Relaxation Order
• Unrelaxable Object
• Preference List
• Measure for Ranking
• Triggering conditions

29
Query Preprocessing

• Segment and label contours for objects of
  interest
• Determine relevant features and spatial
  relationships (e.g., location, containment,
  intersection) of the selected objects
• Organize the features and spatial relationships
  of objects into a feature database
• Classify the feature database into a Type
  Abstraction Hierarchy (TAH)

30
(No Transcript)
31
Similarity Query Answering

• Determine relevant features based on query input
• Select TAH based on these features
• Traverse through the TAH nodes to match all the
  images with similar features in the database
• Present the images and rank their similarity
  (e.g., by mean square error)

32
Visual Query Language and Interface

• Point-click-drag interface
• Objects may be represented by icons
• Spatial relationships among objects are
  represented graphically

33
(No Transcript)
34
Visual Query Example
Retrieve brain tumor cases where a tumor is
located in the region as indicated in the picture
35
(No Transcript)
36
(No Transcript)
37
(No Transcript)
38
Implementation

• Sun Sparc 20 workstations (128 MB RAM, 24-bit
  frame buffer)
• Oracle Database Management System
• C
• Mass Storage of Images (9 GB)

39
(No Transcript)
40
(No Transcript)
41
(No Transcript)
42
Summary KMeD

• Image retrieval by feature and content
• Matching images based on features
• Processing of queries based on spatial
  relationships among objects
• Answering of imprecise queries
• Expression of queries via visual query language
• Integrated view of temporal multimedia data in a
  timeline metaphor

43
(No Transcript)
44
Medical Digital Librarywww.kmed.cs.ucla.edu

• Project leader Wesley W. Chu

• Graduate studentsVictor Z. LiuWenlei
  MaoQinghua Zou

• ConsultantsHooshang Kangaloo, M.D.Denies
  Aberle, M.D.

45
Data Types Used in a Medical Digital Library

• Structured data (patient lab data, demographic
  data,)--CoBase
• Images (X rays, MRI, CT scans)--KMeD
• Free-text (Patient reports, Teaching files,
  Literature, News articles)--FTRS (Free-text
  retrieval system)

46
A Free-Text Retrieval System (FTRS)
Knowledge-based Free- Text Retrieval System
(FTRS)
Ad hoc query
Patient report for content correlation
Query results
News Articles
Patient reports
Medical literature
Teaching materials
47
A Sample Patient Report

• Tissue Source
• LUNG (FINE NEEDLE ASPIRATION) (LEFT LOWER LOBE)
• FINAL DIAGNOSIS
• - LUNG NODULE, LEFT LOWER LOBE (FINE NEEDLE
  ASPIRATION)
• - LUNG CANCER, SMALL CELL, STAGE II.

Tissue Source LUNG (FINE NEEDLE ASPIRATION)
(LEFT LOWER LOBE) FINAL DIAGNOSIS - LUNG
NODULE, LEFT LOWER LOBE (FINE NEEDLE
ASPIRATION) - LUNG CANCER, SMALL CELL, STAGE
II.
48
Scenario-Specific Retrieval
Tissue Source LUNG (FINE NEEDLE ASPIRATION)
(LEFT LOWER LOBE) FINAL DIAGNOSIS - LUNG
NODULE, LEFT LOWER LOBE (FINE NEEDLE
ASPIRATION) - LUNG CANCER, SMALL CELL, STAGE
II.
49
Challenge I Indexing for Free-Text

• Extracting key concepts in the free-text for
  indexing
• Free-text Lung cancer, small cell, stage II
• Concept terms in knowledge source stage II small
  cell lung cancer
• Conventional methods use NLP
• Not scalable

50
Challenge II Mismatch between terms used in
query and documents

• Example

Query lung cancer,
?
?
?
Document 3 anti-cancerdrug combinations
Document 1 lung carcinoma
Document 2 lung neoplasm
51
Challenge III Terms used in the query are too
general

• Expanding the general terms in the query to
  specific terms that are used in the document

Query lung cancer, diagnosis options
Query lung cancer, chest x-ray, bronchography,
Document the effectiveness of chest x-ray and
bronchography on patients with lung cancer
52
A Medical KBUnified Medical Language System
(UMLS)

• Meta-thesaurus - control vocabulary (1.6M
  biomedical phrases, representing 800K concepts)
• Semantic Network classify concepts into classes
  (e.g. disease and syndrome, treated by,
  therapeutic procedure, etc.)
• Specialized Lexicon

53
Using knowledge sources to resolve these
challenges

• Challenge I Automatic indexing of free text
• Challenge II Mismatch between terms in the
  query and the documents
• Challenge III Terms in the query are too general

54
IndexFinder Extracting domain-specific key
concepts

• Technique
• Permute words from text to generate concept
  candidates.
• Use knowledge base to select the valid
  candidates.
• Problem
• Valid candidates may be irrelevant to the
  document.
• Redundant concept

55
Filtering out Irrelevant Concepts

• Syntactic filter
• Limit permutation of words within a sentence.
• Semantic filter
• Use the semantic type (e.g. body part, disease,
  treatment, diagnosis) to filter out irrelevant
  concepts
• Use ISA relationship to filter out general
  concepts and yield specific concepts.

56
IndexFinder Performance

• Two orders of magnitude faster than conventional
  approaches
• No NLP
• Time complexity is linear with the number of
  distinct words in the text
• Preliminary Evaluation
• IndexFinder generates more valid terms than that
  of NLP (using a single noun phrase)
• Filtering is effective to eliminate irrelevant
  terms

57
Using knowledge sources to resolve these
challenges

• Challenge I Automatic indexing of free text
• Challenge II Mismatch between terms in the
  query and the documents
• Challenge III Terms in the query are too general

58
Phrase-based Vector Space Model (VSM)
Query lung cancer,
Query lung cancer,
lung cancer lung carcinoma
missing!!!
parent_of
anti-cancer drug combinations
Document lung carcinoma
Document lung neoplasm
Document anti-cancer drugcombinations
Document anti-cancer drugcombinations
lung neoplasm
Knowledge source
59
Phrase-based VSM Examples
(C0242379) lung cancer
(C0003393) anti cancer drug combin
60
Using knowledge sources to resolve these
challenges

• Challenge I Automatic indexing of free text
• Challenge II Mismatch between terms in the
  query and the documents
• Challenge III Terms in the query are too general

61
Query Expansion (QE)

• Queries in the following form benefit from
  expansionltkey conceptgt ltgeneral supporting
  concept(s)gte.g. lung cancer e.g. treatment
  options

ltkey conceptgt ltspecific supporting
concept(s)gte.g. lung cancer e.g.
chemotherapy, radiotherapy
62
Knowledge-based Scenario-specific Expansion
63
Retrieval Effectiveness Comparison (Corpus
OHSUMED, KB UMLS)
Overall improvement 33,100 queries vs. 5, 50
queries
64
FTRS Scenario-specificQuery Answering

• Sample templatesltdiseasegt, treatment,ltdiseas
  egt, diagnosis

65
FTRS Scenario-specific content correlation

• IndexFinder extracts key concepts from free-text
  for content correlation

66
Summary KB Free-text retrieval

• Technologies
• IndexFinder extracts key concepts from the
  free-text
• Phrase-based VSM a new document indexing
  paradigm (concept and its word stems) to improve
  retrieval effectiveness
• Knowledge-based query expansion match query
  with scenario-specific documents
• provides scenario-specific free-text retrieval

67
Conclusions

• Knowledge sources
• provides
• Approximate matching
• Query conditions
• Image features
• Query processing
• Similarity query answering
• User modeling
• Associative answering
• Triggering and alerting
• Document retrieval
• Convert ad hoc free-text into controlled
  vocabulary
• Phrase-based VSM
• Content correlation
• Scenario-specific retrieval
• Increase capabilities and effectiveness
  Information Management

68
Acknowledgement
This research is supported by DARPA, NSF Grant
9619345, and NIC/NIH Grant4442511-33780