Calling Dr Watson to Radiology
Watson in Healthcare from Nuance
Nick van Terheyden, M.D.
Chief Medical Information Officer – Clinical Language Understanding
Nuance
Medicine used to be simple, ineffective
         and relatively safe.
  Now it is complex, effective and
       potentially dangerous
      Sir Cyril Chantler, Kings Fund Chantler C. The role and education of doctors in the delivery of health care.
                                                                                       Lancet 1999;353:1178-81u
Information Overload – Big Data

 Medical information is doubling approximately every 5 years
       -   Brent James, MD, MStat, Chief Quality Officer, Intermountain Health Care

 1.8 zetabytes of information created this year – majority of it
   unstructured – 57 Billion 32Gb iPods (Source: IDC)
   -   That’s enough information to fill 57 billion 32GB Apple iPads (which
       could build a mountain of iPads 25 times higher than Mt Fuji

 Watson can sift through 200 million pages in 3 secs
Watson DeepQA
HOW DOES IT WORK
DeepQA: The Technology Behind Watson

                                                                                                          Learned Models
                                                                                                         help combine and
                                                                                                        weigh the Evidence
                                                                  Evidence                           Balance
                                                                  Sources                           & Combine
                   Answer                                                                                   Models   Models
                   Sources                                                          Deep
Question                                         Answer           Evidence                                  Models   Models
                                                                                  Evidence
                             Candidate           Scoring          Retrieval 100,000’s Scores from
              Primary                                   1000’s of
                                                                                   Scoring
                                                                             many Deep Analysis
                              Answer                                                                        Models   Models
              Search                                Pieces of Evidence           Algorithms
                             Generation
                                100’s Possible
                                   Answers
         Multiple       100’s
     Interpretations   sources
Question &                                                                                                  Final Confidence
                    Question           Hypothesis          Hypothesis and Evidence
  Topic                                                                                  Synthesis             Merging &
                  Decomposition        Generation                 Scoring
 Analysis                                                                                                       Ranking


                                 Hypothesis         Hypothesis and Evidence                                     Answer &
                                 Generation                Scoring                                             Confidence
                                              ...
Architecture
User Experience
By Nuance and Partners…..

                            …..community of consumers
                            – large and small




               CLU……                                    Cloud to Cloud     DeepQA
                                                                         Solutions for
                                                ….community of            Healthcare
EMRs                                            Content
                                                Publishers




    Large
Institutiona            …..community of
 l Providers            CASE Content Partners
How Does Watson Perform in
Healthcare
Question and Answer Sets
Success

 Question: This hormone deficiency is associated with Kallmann's
  syndrome.
  -   Passage: Isolated deficiency of GnRH or its receptor causes failure of
      normal pubertal development and amenorrhea in women. This disorder
      is termed Kallmann syndrome when it is accompanied by anosmia and
      has also been termed idiopathic hypogonadotropic hypogonadism
      (IHH).”

 Answer: GnRH

 Notes: We know that “GnRH” is a hormone (from the ontology) so
  that lets us choose it as the most likely answer.
Question and Answer Sets
Miss

 Question: Eponym from Victorian literature for obesity
  hypoventilation syndrome.
  -   Correct passage: Obesity-hypoventilation syndrome is also known as
      pickwickian syndrome, in reference to Charles Dickens’…
  -   Correct answer: Pickiwickian Syndrome
  -   Wrong passage: Other clinical features associated with obesity-
      hypoventilation syndrome are daytime hypersomnolence and cor
      pulmonale.
  -   Wrong answer: cor pulmonale
Radiology Use Case
 Radiologists identifies radiological findings in a report
  - Hepatic Lesion in Woman with History of Breast CA
  - Too Small to Classify (TSTC)
  - Offer follow up and alternative assessment criteria helping classify more
      accurately for interval follow up or imaging work up

 Ask Watson
  - discrimination points, appropriate follow up, critical results etc

 Watson Consumes the report and patient context information
  - Offers next steps, differential list, new/recent information & links

 What possible next steps/studies would be appropriate to rule in/out
   diagnosis and/or determine best course of treatment
Generic Use Cases

 Medical Diagnosis
  -   Consumption of medical records, results etc offering differential
      diagnosis and probability analysis with links to underlying literature
      sources
  -   True personalization of medicine based on large cohort historical data
      analysis

 Interactive Model
  -   Requires real time access to patient data from multiple sources
      including provider/patient interaction
  -   Ongoing refinement based on dynamic interaction and learning
Challenges
 Ambiguous human language

 Integration with existing systems – extract of complete data set for
   history, results etc
   -   Often in disparate systems
   -   Non standard interfaces
   -   Non standard format
   -   Unstructured narrative

 Medical Data Sources and Ongoing Access and Updating

 Patient interaction with technology vs humans
  - Telemedicine and consumer trend towards home based care
Replacing the Doctor?
 Study done by the Mayo Clinic in 2006 identified the most important
   characteristics patients feel a good doctor must possess
 The Ideal clinician is
  - confident,
  - empathetic,
  - humane,
  - personal,
  - forthright,
  - respectful, and
  - thorough

 These facets are entirely human and will be hard for technology to replace

                                 Mayo Clin Proc. 2006;81(3):338-344
Calling Dr Watson to Radiology
Watson in Healthcare from Nuance
Nick van Terheyden, M.D.
Chief Medical Information Officer – Clinical Language Understanding
Nuance

Calling Dr Watson To Radiology - RSNA Presentation

  • 1.
    Calling Dr Watsonto Radiology Watson in Healthcare from Nuance Nick van Terheyden, M.D. Chief Medical Information Officer – Clinical Language Understanding Nuance
  • 2.
    Medicine used tobe simple, ineffective and relatively safe. Now it is complex, effective and potentially dangerous Sir Cyril Chantler, Kings Fund Chantler C. The role and education of doctors in the delivery of health care. Lancet 1999;353:1178-81u
  • 3.
    Information Overload –Big Data  Medical information is doubling approximately every 5 years - Brent James, MD, MStat, Chief Quality Officer, Intermountain Health Care  1.8 zetabytes of information created this year – majority of it unstructured – 57 Billion 32Gb iPods (Source: IDC) - That’s enough information to fill 57 billion 32GB Apple iPads (which could build a mountain of iPads 25 times higher than Mt Fuji  Watson can sift through 200 million pages in 3 secs
  • 6.
  • 7.
    DeepQA: The TechnologyBehind Watson Learned Models help combine and weigh the Evidence Evidence Balance Sources & Combine Answer Models Models Sources Deep Question Answer Evidence Models Models Evidence Candidate Scoring Retrieval 100,000’s Scores from Primary 1000’s of Scoring many Deep Analysis Answer Models Models Search Pieces of Evidence Algorithms Generation 100’s Possible Answers Multiple 100’s Interpretations sources Question & Final Confidence Question Hypothesis Hypothesis and Evidence Topic Synthesis Merging & Decomposition Generation Scoring Analysis Ranking Hypothesis Hypothesis and Evidence Answer & Generation Scoring Confidence ...
  • 8.
    Architecture User Experience By Nuanceand Partners….. …..community of consumers – large and small CLU…… Cloud to Cloud DeepQA Solutions for ….community of Healthcare EMRs Content Publishers Large Institutiona …..community of l Providers CASE Content Partners
  • 9.
    How Does WatsonPerform in Healthcare
  • 10.
    Question and AnswerSets Success  Question: This hormone deficiency is associated with Kallmann's syndrome. - Passage: Isolated deficiency of GnRH or its receptor causes failure of normal pubertal development and amenorrhea in women. This disorder is termed Kallmann syndrome when it is accompanied by anosmia and has also been termed idiopathic hypogonadotropic hypogonadism (IHH).”  Answer: GnRH  Notes: We know that “GnRH” is a hormone (from the ontology) so that lets us choose it as the most likely answer.
  • 11.
    Question and AnswerSets Miss  Question: Eponym from Victorian literature for obesity hypoventilation syndrome. - Correct passage: Obesity-hypoventilation syndrome is also known as pickwickian syndrome, in reference to Charles Dickens’… - Correct answer: Pickiwickian Syndrome - Wrong passage: Other clinical features associated with obesity- hypoventilation syndrome are daytime hypersomnolence and cor pulmonale. - Wrong answer: cor pulmonale
  • 12.
    Radiology Use Case Radiologists identifies radiological findings in a report - Hepatic Lesion in Woman with History of Breast CA - Too Small to Classify (TSTC) - Offer follow up and alternative assessment criteria helping classify more accurately for interval follow up or imaging work up  Ask Watson - discrimination points, appropriate follow up, critical results etc  Watson Consumes the report and patient context information - Offers next steps, differential list, new/recent information & links  What possible next steps/studies would be appropriate to rule in/out diagnosis and/or determine best course of treatment
  • 13.
    Generic Use Cases Medical Diagnosis - Consumption of medical records, results etc offering differential diagnosis and probability analysis with links to underlying literature sources - True personalization of medicine based on large cohort historical data analysis  Interactive Model - Requires real time access to patient data from multiple sources including provider/patient interaction - Ongoing refinement based on dynamic interaction and learning
  • 14.
    Challenges  Ambiguous humanlanguage  Integration with existing systems – extract of complete data set for history, results etc - Often in disparate systems - Non standard interfaces - Non standard format - Unstructured narrative  Medical Data Sources and Ongoing Access and Updating  Patient interaction with technology vs humans - Telemedicine and consumer trend towards home based care
  • 15.
    Replacing the Doctor? Study done by the Mayo Clinic in 2006 identified the most important characteristics patients feel a good doctor must possess  The Ideal clinician is - confident, - empathetic, - humane, - personal, - forthright, - respectful, and - thorough  These facets are entirely human and will be hard for technology to replace Mayo Clin Proc. 2006;81(3):338-344
  • 16.
    Calling Dr Watsonto Radiology Watson in Healthcare from Nuance Nick van Terheyden, M.D. Chief Medical Information Officer – Clinical Language Understanding Nuance

Editor's Notes

  • #3 Background on technology and Watson/Jeopardy and the data Tsunami we face in h/cHow Deep QA WorksDeep QA applied to HealthcareCurrent Example of Medical Intelligence (CTRM)Future Use Cases
  • #4 Brent James, MD, MStat, Chief Quality Officer, Intermountain Health Care; subject of The New York Times article “If Health Care is Going to Change, Dr. Brent James Will Lead the Way”http://www.nytimes.com/2009/11/08/magazine/08Healthcare-t.html?pagewanted=all
  • #8 Massively Parallel Probabilistic Evidence-Based Architecture  This is like looking inside the brain of the DeepQA system from about 30,000 feet high.
  • #12 Notes: This is easy if you know that Charles Dickens wrote Victorian literature. This is not part of medical inference, though, so we do not cover that, and an incorrect answer is preferred because its passage matched the query better. Without knowing about Victorian literature, there is not enough other information in the question to reliably find the correct answer.
  • #13 Incidentalomas
  • #16 doi: 10.4065/ 81.3.338Mayo ClinicProceedings March 2006 vol. 81 no. 3 338-344http://www.mayoclinicproceedings.com/content/81/3/338.fullhttp://www.mayoclinicproceedings.com/content/81/3/338/T2.expansion.htmlMayo ClinProc. 2006;81(3):338-344