Upgrade to Pro — share decks privately, control downloads, hide ads and more …

Improving the breast cancer diagnosis using digital repositories

CETA-Ciemat
March 22, 2013
170

Improving the breast cancer diagnosis using digital repositories

CETA-Ciemat

March 22, 2013
Tweet

Transcript

  1. breast cancer
    diagnosis
    using
    digital repositories
    improving the
    César Suárez Ortega
    [email protected]
    18-20 March 2013
    International Work-Conference on Bioinformatics and Biomedical Engineering

    View full-size slide

  2. The IMED project

    View full-size slide

  3. introduction
    » IMED is a collaboration among:
    » CETA-CIEMAT (Trujillo, Spain)
    » INEGI (Porto, Portugal)
    » FMUP-HSJ (Porto, Portugal)
    » Since 2008
    » Created to ease the cancer diagnosis.
    » Focused on breast cancer.

    View full-size slide

  4. Breast cancer
    is most common cancer
    for women.
    According to World Health Organization:
    http://www.who.int/cancer/events/breast_cancer_month/en/
    1
    1
    1

    View full-size slide

  5. Early detection
    is the cornerstone
    of its control.

    View full-size slide

  6. Double checking of cases
    reduces the ratio
    of missed cancer
    R. Ramos-Pollan, et al., Discovering Mammography-based Machine Learning Classifiers
    for Breast Cancer Diagnosis. Journal of Medical Systems (9 April 2011), pp. 1-11.
    1
    1

    View full-size slide

  7. But this is expensive!

    View full-size slide

  8. But this is expensive!
    Solutions?

    View full-size slide

  9. But this is expensive!
    Solutions?
    Computer-aided
    diagnosis

    View full-size slide

  10. But this is expensive!
    Solutions?
    Computer-aided
    diagnosis
    CAD Systems
    1
    1

    View full-size slide

  11. what are CAD
    systems?
    » Assistance in the diagnosis of medical images
    » Multiple disciplines are combined:
    » Artificial intelligence
    » Digital image processing
    » Radiology

    View full-size slide

  12. what are CAD
    systems?
    » Assistance in the diagnosis of medical images
    » Multiple disciplines are combined:
    » Artificial intelligence
    » Digital image processing
    » Radiology
    They don't replace doctors, they support them.

    View full-size slide

  13. MIWAD
    BCDR
    MLCs
    three pillars
    Breast
    Cancer
    Digital
    Repository
    Machine
    Learning
    Classifiers
    Mammography
    Image
    Workstation for
    Analysis and
    Diagnosis

    View full-size slide

  14. MIWAD
    BCDR
    MLCs
    three pillars
    BCDR
    – A repository of
    annotated breast
    cancer cases.
    MLCs
    Machine learning
    classifiers for breast
    cancer diagnosis
    MIWAD
    – A workstation
    prototype with
    a CAD system.

    View full-size slide

  15. MIWAD
    BCDR
    MLCs

    View full-size slide

  16. MLCs
    » Artificial intelligent systems.
    » They learn from data.

    View full-size slide

  17. benign
    malign
    Examples
    MLC
    info + classification
    info + classification

    View full-size slide

  18. benign
    malign
    Examples
    MLC
    – benign? malign?
    benign
    classification
    info + classification
    info + classification
    info

    View full-size slide

  19. MLCs
    » Artificial intelligent systems.
    » They learn from data.
    » Lot of them!

    View full-size slide

  20. MLCs
    » Artificial intelligent systems.
    » They learn from data.
    » Lot of them!
    Decision Tree

    View full-size slide

  21. MLCs
    » Artificial intelligent systems.
    » They learn from data.
    » Lot of them!
    Decision Tree
    Association Rule

    View full-size slide

  22. MLCs
    » Artificial intelligent systems.
    » They learn from data.
    » Lot of them!
    Decision Tree
    Association Rule
    Neural Networks

    View full-size slide

  23. MLCs
    » Artificial intelligent systems.
    » They learn from data.
    » Lot of them!
    Decision Tree
    Association Rule
    Neural Networks
    Genetic Programming

    View full-size slide

  24. MLCs
    » Artificial intelligent systems.
    » They learn from data.
    » Lot of them!
    Decision Tree
    Association Rule
    Neural Networks
    Genetic Programming
    Support Vector Machines

    View full-size slide

  25. MLCs
    » Artificial intelligent systems.
    » They learn from data.
    » Lot of them!
    Decision Tree
    Association Rule
    Neural Networks
    Genetic Programming
    Support Vector Machines
    Clustering

    View full-size slide

  26. MLCs
    » Artificial intelligent systems.
    » They learn from data.
    » Lot of them!
    Decision Tree
    Association Rule
    Neural Networks
    Genetic Programming
    Support Vector Machines
    Clustering
    Bayesian Networks

    View full-size slide

  27. MLCs
    » Artificial intelligent systems.
    » They learn from data.
    » Lot of them!
    Decision Tree
    Association Rule
    Neural Networks
    Genetic Programming
    Support Vector Machines
    Clustering
    Bayesian Networks
    Inductive Logic

    View full-size slide

  28. MLCs
    » Artificial intelligent systems.
    » They learn from data.
    » Lot of them!
    Decision Tree
    Association Rule
    Neural Networks
    Genetic Programming
    Support Vector Machines
    Clustering
    Bayesian Networks
    Inductive Logic
    Sparse Dictionary

    View full-size slide

  29. What is the best one for
    breast cancer?

    View full-size slide

  30. We are working on it :)

    View full-size slide

  31. current results
    » GRID and HPC to speed up the research.
    » Metrics
    » ROC → True positives VS. False positives
    » AUC → Probability of a true positive
    » Our classifiers?

    View full-size slide

  32. current results
    » GRID and HPC to speed up the research.
    » Metrics
    » ROC → True positives VS. False positives
    » AUC → Probability of a true positive
    » Our classifiers?
    AUC = 85%

    View full-size slide

  33. What data we use
    to train our MLCs?

    View full-size slide

  34. MIWAD
    BCDR
    MLCs

    View full-size slide

  35. BCDR is a wide-ranging annotated repository
    of anonymous breast cancer studies

    View full-size slide

  36. BCDR
    » Our reference dataset for MLCs.
    » It contains a custom subset of DICOM and:
    » Reviewed by specialists and biopsy proven
    » Cases classified using BI-RADS
    » Useful for researchers.
    23
    17
    per lesion region
    intesity, texture and shape features
    clinical features per case

    View full-size slide

  37. data model
    patients
    images
    lesions
    segmentations
    classifications
    studies
    *
    *
    *
    *
    *
    *

    View full-size slide

  38. data model
    patients
    images
    lesions
    segmentations
    classifications
    studies
    *
    *
    *
    *
    *
    *
    ?

    View full-size slide

  39. segmentation
    of a lesion

    View full-size slide

  40. two flavours
    BCDR-FMR
    1010 cases
    795 segmented lesions
    3073 mammographies
    720x1168 TIFF images (8bpp)

    View full-size slide

  41. two flavours
    BCDR-DMR
    600 cases
    828 segmented lesions
    2837 mammographies
    3328x4084 TIFF images (14bpp)

    View full-size slide

  42. two flavours
    BCDR-DMR
    600 cases
    828 segmented lesions
    2837 mammographies
    3328x4084 TIFF images (14bpp)
    + 2073 ultrasound images 800x600 (8bpp)

    View full-size slide

  43. http://bcdr.inegi.up.pt

    View full-size slide

  44. http://bcdr.inegi.up.pt

    View full-size slide

  45. http://bcdr.inegi.up.pt

    View full-size slide

  46. http://bcdr.inegi.up.pt

    View full-size slide

  47. MIWAD
    BCDR
    MLCs

    View full-size slide

  48. MIWAD
    » A workstation for analysing and diagnosing.
    » It's multiplatform (Windows, Linux & Mac)
    » Composed by two connected apps:
    MIWAD-DB MIWAD-CAD

    View full-size slide

  49. MIWAD-DB
    » It manages...
    » ...the clinical data
    » ...the mammographies
    » BCDR as data source.
    » It operates over BCDR data.
    » Executes MLCs with a 1-click process.

    View full-size slide

  50. patients
    studies
    lesions
    segmentations
    classifications

    View full-size slide

  51. mammographies

    View full-size slide

  52. MIWAD-CAD
    » Screening tool (based on TUDOR Viewer)
    » It includes image processing tools.
    » Tools to help in the segmentation process.
    » Three functions:
    1 It stores the segmentations in BCDR.
    2 It calculates image features.
    3 It helps doctors to find lessions.

    View full-size slide

  53. CAPTURAS CAD

    View full-size slide

  54. this is real

    View full-size slide

  55. CAD
    DB
    Bruno
    São João Hospital

    View full-size slide

  56. summarizing...

    View full-size slide

  57. MIWAD
    BCDR
    MLCs
    BCDR
    – A repository of
    annotated breast
    cancer cases.
    MLC
    Machine learning
    classifiers for breast
    cancer diagnosis
    MIWAD
    – A workstation
    prototype with
    a CAD system.

    View full-size slide

  58. MIWAD
    BCDR
    MLCs
    BCDR is used to train MLCs

    View full-size slide

  59. MIWAD
    BCDR
    MLCs
    BCDR is the data source of
    MIWAD...

    View full-size slide

  60. MIWAD
    BCDR
    MLCs
    ...and MIWAD feeds the BCDR

    View full-size slide

  61. MIWAD
    BCDR
    MLCs
    So, MIWAD helps to
    improve the MLCs

    View full-size slide

  62. MIWAD
    BCDR
    MLCs
    The MLCs are
    integrated in MIWAD

    View full-size slide

  63. future work
    » IMED goes educational!
    » MIWAD as a training app for radiologists:
    » Web architecture & responsive design.
    » Auto-evaluation using MLCs.
    » BCDR as a formative resource.
    » Continuation of the improvement of MLCs. (+90% AUC)
    » Certification for MLCs
    » Continuation of the BCDR growth (searching new sources)

    View full-size slide

  64. conclusions
    » CAD systems can achieve great results in:
    » Improving the breast cancer diagnosis
    » Training specialists
    » Specialists demand better educational apps (difficult to find)
    » All is about data!
    » It's difficult to get reliable data.
    » BCDR is the most complete open breast cancer repository.
    » MLCs data

    View full-size slide

  65. CETA-CIEMAT acknowledges the support received from the
    European Regional Development Fund through its
    Operational Programme Knowledge-based Economy.

    View full-size slide

  66. any questions?
    thanks for your attention! :)

    View full-size slide