PACS: Picture Archiving and Communication System

1
PACS Picture Archiving and Communication System

• Feipei Lai
• National Taiwan University

2
PACS

• A comprehensive computer system that is
  responsible for the electronic storage and
  distribution of medical images in the medical
  enterprise.
• Reduce costs
• Improve patient care

3

• A PACS stores, retrieves, distributes, and
  presents digital medical images.
• In a typical PACS, DICOM is used for image
  storage, retrieval, and data exchange.
• DICOM also provides a framework for image
  information management such as Modality Worklist
  (MWL), Modality Performed Procedure Step (MPPS),
  and Storage Commitment (STC).
• Furthermore, DICOM standardizes hard- and
  soft-copy display consistency and image
  import/export on rewritable media such as CDs or
  DVDs.

4
DICOM Digital Imaging and COmmunications in
Medicine

• The goals of DICOM are to achieve
  interoperability and to improve workflow
  efficiency between imaging systems and other
  information systems in healthcare environments
  worldwide.

5

• 1 Introduction and Overview
• 2 Conformance
• 3 Information Object Definitions
• 4 Service Class Specifications
• 5 Data Structures and Encoding
• 6 Data Dictionary
• 7 Message Exchange
• 8 Network Communication Support for Message
  Exchange
• 9 Retired
• 10 Media Storage and File Format for Media
  Interchange
• 11 Media Storage Application Profiles
• 12 Formats and Physical Media
• 13 Retired
• 14 Grayscale Standard Display Function
• 15 Security and System Management Profiles
• 16 Content Mapping Resource
• 17 Explanatory Information
• 18 Web Access to DICOM Persistent Objects (WADO)

6
1 Introduction and Overview

• provides an overview of the entire Digital
  Imaging and Communications in Medicine (DICOM)
  Standard.
• It describes the history, scope, goals, and
  structure of the Standard.
• In particular, it contains a brief description of
  the contents of each part of the Standard.

7
(No Transcript)
8
2 Conformance

• defines principles that implementations claiming
  conformance to the Standard shall follow
• Conformance requirements. PS 3.2 specifies the
  general requirements which must be met by any
  implementation claiming conformance. It
  references the conformance sections of other
  parts of the Standard.
• Conformance Statement. PS 3.2 defines the
  structure of a Conformance Statement. It
  specifies the information which must be present
  in a Conformance Statement. It references the
  Conformance Statement sections of other parts of
  the Standard.

9
CONSTRUCTION PROCESS FOR A NETWORK CONFORMANCE
CLAIM
10
CONSTRUCTION PROCESS FOR A MEDIA CONFORMANCE CLAIM
11
3 INFORMATION OBJECT DEFINITIONS

• PS 3.3 of the DICOM Standard specifies a number
  of Information Object Classes which provide an
  abstract definition of real-world entities
  applicable to communication of digital medical
  images and related information (e.g., waveforms,
  structured reports, radiation therapy dose,
  etc.).
• Each Information Object Class definition consists
  of a description of its purpose and the
  Attributes which define it.

12
4 SERVICE CLASS SPECIFICATIONS

• PS 3.4 of the DICOM Standard defines a number of
  Service Classes.
• A Service Class associates one or more
  Information Objects with one or more Commands to
  be performed upon these objects.
• Service Class Specifications state requirements
  for Command Elements and how resulting Commands
  are applied to Information Objects.
• Service Class Specifications state requirements
  for both providers and users of communications
  services.

13
5 DATA STRUCTURE AND SEMANTICS

• PS 3.5 of the DICOM Standard specifies how DICOM
  applications construct and encode the Data Set
  information resulting from the use of the
  Information Objects and Services Classes defined
  in PS 3.3 and PS 3.4 of the DICOM Standard.
• The support of a number of standard image
  compression techniques (e.g., JPEG lossless and
  lossy) is specified.

14
6 DATA DICTIONARY

• PS 3.6 of the DICOM Standard is the centralized
  registry which defines the collection of all
  DICOM Data Elements available to represent
  information, along with elements utilized for
  interchangeable media encoding and a list of
  uniquely identified items that are assigned by
  DICOM.

15
7 MESSAGE EXCHANGE

• PS 3.7 of the DICOM Standard specifies both the
  service and protocol used by an application in a
  medical imaging environment to exchange Messages
  over the communications support services defined
  in PS 3.8.
• A Message is composed of a Command Stream defined
  in PS 3.7 followed by an optional Data Stream as
  defined in PS 3.5.

16
8 NETWORK COMMUNICATION SUPPORT FOR MESSAGE
EXCHANGE

• PS 3.8 of the DICOM Standard specifies the
  communication services and the upper layer
  protocols necessary to support, in a networked
  environment, communication between DICOM
  applications as specified in PS 3.3, PS 3.4, PS
  3.5, PS 3.6, and PS 3.7.
• These communication services and protocols ensure
  that communication between DICOM applications is
  performed in an efficient and coordinated manner
  across the network.

17
10 MEDIA STORAGE AND FILE FORMAT

• PS 3.10 of the DICOM Standard specifies a general
  model for the storage of medical imaging
  information on removable media (see Figure
  6.10-1).
• The purpose of this Part is to provide a
  framework allowing the interchange of various
  types of medical images and related information
  on a broad range of physical storage media.

18
DICOM Media Communication Model
19
11 MEDIA STORAGE APPLICATION PROFILES

• PS 3.11 of the DICOM Standard specifies
  application specific subsets of the DICOM
  Standard to which an implementation may claim
  conformance.
• These application specific subsets will be
  referred to as Application Profiles in this
  section.
• Such a conformance statement applies to the
  interoperable interchange of medical images and
  related information on storage media for specific
  clinical uses.
• It follows the framework, defined in PS 3.10, for
  the interhcange of various types of information
  on storage media.

20
12 STORAGE FUNCTIONS AND MEDIA FORMATS FOR DATA
INTERCHANGE

• This part of the DICOM Standard facilitates the
  interchange of information between applications
  in medical environments by specifying
• a) A structure for describing the relationship
  between the media storage model and a specific
  physical media and media format.
• b) Specific physical media characteristics and
  associated media formats.

21
14 GRAYSCALE STANDARD DISPLAY FUNCTION

• PS 3.14 specifies a standardized display function
  for consistent display of grayscale images.
• This function provides methods for calibrating a
  particular display system for the purpose of
  presenting images consistently on different
  display media (e.g. monitors and printers).

22
15 SECURITY AND SYSTEM MANAGEMENT PROFILES

• PS 3.15 of the DICOM Standard specifies security
  and system management profiles to which
  implementations may claim conformance.
• Security and system management profiles are
  defined by referencing externally developed
  standard protocols, such as DHCP, LDAP, TLS and
  ISCL.
• Security protocols may use security techniques
  like public keys and smart cards.
• Data encryption can use various standardized data
  encryption schemes.

23
16 CONTENT MAPPING RESOURCE

• PS 3.16 of the DICOM Standard specifies
• templates for structuring documents as DICOM
  Information Objects
• sets of coded terms for use in Information
  Objects
• a lexicon of terms defined and maintained by
  DICOM
• country specific translations of coded terms

24
17 explanatory information

• This part of the DICOM Standard contains
  explanatory information in the form of Normative
  and Informative Annexes.

25
(No Transcript)
26
18 Web Access to DICOM Persistent Objects (WADO)

• This standard specifies a web-based service for
  accessing and presenting DICOM (Digital Imaging
  and Communications in Medicine) persistent
  objects (e.g. images, medical imaging reports).
• This is intended for distribution of results and
  images to healthcare professionals.
• It provides a simple mechanism for accessing a
  DICOM persistent object from HTML pages or XML
  documents, through HTTP/HTTPs protocol, using
  DICOM UIDs (Unique Identifiers).
• Data may be retrieved either in a
  presentation-ready form as specified by the
  requester (e.g. JPEG or GIF) or in a native DICOM
  format.
• It does not support facilities for web searching
  of DICOM images.
• This standard relates only to DICOM persistent
  objects (not to other DICOM objects or to
  non-DICOM objects).

27
1. Select patient information

• When a patient is registered and a procedure is
  ordered, this is typically communicated to a
  Radiology Information System (RIS), which does
  the scheduling and creates information about the
  procedure, scheduling, and patient demographics
  available (using the DICOM Modality Worklist) to
  the technologist at the modality.
• This info is then used to perform the
  acquisition, because the scheduled procedure
  information (such as CT-Abdomen) is available for
  the technologist.
• In addition, the patient demographic information
  is copied automatically and used to identify the
  images that are sent out.

28
2. (Optional) retrieve prior images

• In certain cases, prior images could be helpful
  to ensure that the acquisition uses the same
  techniques, and that those techniques can be
  easily compared with the previous study.
• In such cases, a technologist might pull them
  over by using DICOM Query/Retrieve.

29
3. Start procedure

• As soon as the acquisition starts, the PACS and
  Health Information System are notified using the
  DICOM Modality Performed Procedure Step.
• The Image Manager in the PACS knows that images
  will be forthcoming and it could remove the
  procedure from the scheduling list so that other
  devices will not attempt to perform the same
  procedure.

30
4. Send images

• Images are acquired and sent or pushed, using
  the DICOM Storage Service, to a PACS archive or
  QA station.

31
5. Update with exam complete status

• When completed, the DICOM Modality Performed
  Procedure Step will communicate what the actual
  performed procedure was (should it have been
  changed from the originally scheduled procedure),
  and how many images were created.
• The number of images is important for the Image
  Manager, which inventories them to make sure it
  has them all available in the PACS archive.
• Knowing that the exam was completed is important
  for the PACS scheduler so that it can alert a
  radiologist to read the study. (It does this by
  adding it to the interpretation Worklist.)
• The actual performed procedure information will
  be used by the RIS to ensure that the billing is
  done correctly.
• Radiation dose information might be exchanged as
  well, in the case of an X-Ray exam.

32
6. Read study

• A radiologist opens the to-be-read folder on a
  workstation, which was created using the DICOM
  General Purpose Worklist services.
• Prior exams might be pulled over using the DICOM
  Query/Retrieve.
• Depending on the system architecture, these prior
  images could have been pushed to the workstation
  prior to the reading in order to increase
  performance (by using the DICOM Storage).

33
7. Make images available for physician

• The images could be made available for the
  Referring Physician as well.
• They could access the images using a workstation
  in their office.
• The DICOM grayscale standard display function is
  used to make sure that the images look almost
  identical on the radiologist workstation and the
  physician workstation.
• In addition, any additional annotation and
  changes in the image appearance, such as window
  width and level, are exchanged between the
  radiologist workstation and the physician
  workstation.

34
8. Create hardcopy/softcopy

• The patient might need an additional hardcopy,
  such as film or a CD to take to the physician or
  specialist.
• A file room clerk can retrieve the images using
  DICOM Query/Retrieve and either print the
  significant images using DICOM Print or burn a
  DICOM-compatible disc (either CD or DVD).

35
9. Make images available for Primary Care
Physician

• Using the Internet, the images could also be made
  available to the Primary Care Physician.
• DICOM security will make sure that the
  information delivered over the Internet to the
  physician is encrypted and that the proper
  authorization has taken place prior to the image
  retrieval.
• The DICOM Key Object Note identifies which images
  are significant so that the complete study will
  not have to be reviewed.

36

• Wherever possible, DICOM utilizes relevant parts
  of other mature standards such as LOINC, SNOMED,
  JPEG, MPEG, BIRADS, TCP/IP and other Internet
  Standards.

37

• BI-RADS is an acronym for Breast
  Imaging-Reporting and Data System, a quality
  assurance tool originally designed for use with
  mammography.

38
Modality Work List

• DICOM Modality Work List (MWL), which allows for
  the scheduling, ordering and patient demographic
  information to be retrieved at a modality.

39
Modality Performed Procedure Step

• Modality Performed Procedure Step (MPPS) allows a
  device to communicate what the exam performed
  actually was (vs. the exam that was scheduled) to
  allow changes in the procedure to be
  communicated.
• In addition, it tells when a procedure has been
  started, indicating this on the scheduling list,
  and provides information about the number of
  images generated.

40
Storage Commitment

• Storage Commitment (STC), which transfers the
  responsibility for the images to the receiver, so
  they can be safely removed from the local disc.

41
LUT

• In order to achieve this consistency, devices
  have to support the DICOM grayscale or color
  standard, potentially implementing a so-called
  presentation Look Up Table to map the values.

42
Presentation State

• There is one other component of the presentation
  consistency, and that occurs when a physician
  does things like zoom an image, add annotation,
  or change the Window width and level. This
  information should be preserved in a standard
  manner.
• This issue is addressed by the DICOM Presentation
  State.

43
imaging integration

• to indicate information about the images, such as
  measurements, computer aided diagnostic data for
  significant images.
• DICOM identifies the key images of a study so
  that a physician does not have to review every
  image in the study (which is especially important
  when the initial study contains thousands of
  images such as for CT and MR).
• The measurements and findings, either by a human
  or computer, are generally encoded as a
  Structured Report (SR).

44
(No Transcript)
45
(No Transcript)
46
PACS ?????

• 1. Core Switch 6509 ????
• ??????????,?????????????????
• 2. Main PACS client ????
• Client ??? Edge Switch ??? Core Switch?,??
  Edge Switch????????????? Core Switch??????
• 3. ?? RAD PACS ???????????
• ????????????????,??????????????????,??????????
  ????????????? RAD PACS?????????? PACS Core
  Switch?,?????????????????,?????????????? PACS???

47

• 4. ?????????????
• ?????????? 30?,?????? 3???????????????48?10/100
  /1000????,?????????24?10/100/1000
  ?????????????????? Core Switch?
• 5. ???????? RAD PACS????
• ????????????,???????
• a.???????????,?? ACL (Access Control List)
  ???,?????????,??????? IP ???? RAD PACS ????
• b.?????????

48
Various Network Technologies, Bandwidths and
Typical Transfer Times
49

• 5MP?????????
• ? 64MB(?)??RAM???,
• ??? ? 25602048, Dual 8 /10 bit DVI 20(?)??TFT
  LCD?????
• ?? ? 700 cd/M2,
• ?????170?
• Contrast Ratio ? 6001

50

• 3MP?????????
• ? 64MB(?)??RAM???,
• ??? ? 15362048, Dual 8 /10 bit DVI 20(?)??TFT
  LCD?????,
• ?? ? 700 cd/M2,
• ???? ? 170?,
• Contrast Ratio ? 9001

51
Medical images

• Topographic imaging
• Represents the surface of the body
• Projection imaging
• The interactions of radiation penetrating along a
  known path of the radiation through the body
• Tomographic imaging
• The spatial distribution of the local interaction
  of the radiation with tissue in a thin slice
  through the body.

52
Medical images

• The quality of the images is represented by
  contrast and resolution.
• The contrast is determined by the nature of the
  interaction of the radiation with the tissue
  material (e.g., via partial absorption) or its
  structure (e.g., via reflection) or the
  preferential accumulation of indicator materials
  (e.g., iodine for X-ray, gadolinium ? for
  Magnetic Resonance Imaging, microbubles in the
  Ultrasound or radionuclides for Scintigraphy
  ????? ).

53
Medical images

• The resolution is expressed as spatial, temporal
  or contrast.
• Temporal resolution involves the exposure time
  required to complete the scan of a single image
  and the frame rate of the sequential individual
  images.

54
Medical images

• Plays a major role in medical research activities
  such as detection and quantitation of
  pathophysiological structure-to-function
  relationships, drug discovery and phenotyping.

55
phenotype

• A phenotype describes any observed quality of an
  organism, such as its morphology, development, or
  behaviour, as opposed to its genotype - the
  inherited instructions it carries, which may or
  may not be expressed.

56
Imaging instrument

• Components that generate the probing energy (such
  as electromagnetic radiation, ultrasound or
  electrical current)
• The detector system
• The tomographic image reconstructor (generally
  involves a mechanical or electronic scanning
  process and a variant of solving an inverse
  problem)
• The image display (generally involves a computer
  terminal).

57

• Novel ideas only survive if the environment for
  implementing the idea is present and/or the need
  is perceived.

58
PACS

• Ultrasound
• Magnetic Resonance Imaging
• Computed Tomography

59
NMR

• Lauterbur realized that the slight variation in
  magnet uniformity (the bane of spectroscopists)
  could be used to spatially localize the signal of
  interest and, hence, the controlled variation in
  magnetic field could form the basis of an imaging
  approach.

60
Ultrasound microbubbles

• Intravascular microbubbles were developed as an
  ultrasound contrast agent, but the harmonic
  frequencies generated which contaminated the
  Doppler signal used to measure their velocity
  became the basis of a great increase in
  specificity and sensitivity of the microbubble
  use in ultrasound.

61

• Specificity (true negative rate)
• TN / (TN FP)
• Sensitivity (true positive rate)
• TP / (TP FN)
• TP True Positive
• TN True Negative
• FP False Positive
• FN False Negative

62
Chemical shift

• The paramagnetic effect of oxygen in the blood
  could be used to generate highly specific images
  of cerebral oxygen use and spectroscopic
  evaluation of metabolic events in tissues such as
  in the brain and heart.

63
Positron Emission Tomography (PET)

• Measures radioactive traces injected into the body

64
reference

• Proceedings of the IEEE, Vol. 91, No. 10, October
  2003, pp. 1483-1491.

65
Basic concepts in Image Generation

• Spatial resolution
• The number of pixels per image area
• Contrast resolution
• The number of bit per pixel determines the
  contrast resolution
• Temporal resolution
• A measure of the time needed to create an image

66
Ultrasound
67
Principle of Echo Scanners

• In echo scanners, sound pulses are generated with
  frequencies of about a few MHz.
• These pulses are absorbed, scattered, or
  reflected in the patient.
• The reflections give rise to relatively strong
  echoes.

68

• Reflections occur at interfaces between media
  that are different with respect to density and/or
  the velocity of sound.
• Sound is reflected at interfaces with different
  acoustic impedances the so-called acoustic
  impedance is equal to the product of sound
  velocity and density.

69
Principle of Echo Scanners

• At an interface between soft tissue on one side
  and bone or air on the other side, a strong
  reflection is observed.
• Scattering takes place if the dimension of the
  object is small (i.e., about the wavelength of
  the incident radiation).
• The beam is then scattered in all directions, and
  therefore, the amplitude of the signal detected
  by the transducer is relatively small.

70
Principle of Echo Scanners

• The resolution of an echo scan, that is, the
  degree with which details located close together
  can still be distinguished, is determined by both
  the wavelength of the sound waves and the
  duration of the emitted pulse.
• The pulse is usually several wavelengths long. In
  practice, therefore, reflections from two points
  separated by a few wavelengths can be
  discriminated.
• The smaller the wavelength the better the
  resolution. Since the wavelength is inversely
  proportional to the frequency, the resolution is
  proportional to the frequency.

71
Principle of Echo Scanners

• The attenuation coefficient (which expresses how
  much the beam is attenuated per centimeter of
  tissue because of scatter and absorption) is
  proportional to the sound frequency for soft
  tissue and is even proportional to the square of
  the frequency for other types of tissues.
• The depth of penetration of the sound waves is
  inversely proportional to the frequency. The more
  the beam is attenuated, the more difficult it is
  to measure the reflections of deeper structures,
  since the signal-to-noise ratio gradually becomes
  smaller.

72
Principle of Echo Scanners

• Since resolution and penetration depth pose
  contradictory requirements
• Deeper structures can only be visualized with
  relatively low frequencies, with a concomitant
  lower resolution.
• The type of tissue influences the amount of
  absorption of the beam. Air and bone, for
  example, are strong absorbers, whereas muscle
  tissue and water hardly attenuate the beam.

73
Principle of Echo Scanners

• At a frequency of 3 MHz (wavelength of 0.5 mm)
  depths of up to 10 cm are well visualized, with
  an axial resolution on the order of 1 mm.
• For eye examinations a higher resolution is
  needed. In this case frequencies of between 5 and
  13 MHz (wavelengths of between 0.25 and 0.075 mm,
  respectively) are used.
• For brain examinations the sound beam must first
  pass bone structures (e.g., the temporal).
  Because of the high absorption of bone,
  especially for high frequencies, only low
  frequencies can be used, implying a lower
  resolution.

74
Temporal ????,??

• The space, on either side of the head, back of
  the eye and forehead, above the zygomatic arch
  and in front of the ear.

75
The 2003 Nobel Prize in Physiology or Medicine

• The Nobel Assembly at Karolinska Institutet
  awarded The Nobel Prize in Physiology or Medicine
  for 2003jointly to
• Paul C Lauterbur and Peter Mansfield for their
  discoveries concerning "magnetic resonance
  imaging"

76
Summary

• Imaging of human internal organs with exact and
  non-invasive methods is very important for
  medical diagnosis, treatment and follow-up.
• Seminal discoveries concerning the use of
  magnetic resonance to visualize different
  structures.
• These discoveries have led to the development of
  modern magnetic resonance imaging, MRI, which
  represents a breakthrough in medical diagnostics
  and research.

77

• Atomic nuclei in a strong magnetic field rotate
  with a frequency that is dependent on the
  strength of the magnetic field.
• Their energy can be increased if they absorb
  radio waves with the same frequency (resonance).
• When the atomic nuclei return to their previous
  energy level, radio waves are emitted.
• These discoveries were awarded the Nobel Prize in
  Physics in 1952.

78

• When the atom is placed in a magnetic field, the
  interaction energy -?B of the spin magnetic
  dipole moment with the field causes further
  splittings in energy levels and in the
  corresponding spectrum lines.

79

• During the following decades, magnetic resonance
  was used mainly for studies of the chemical
  structure of substances.
• In the beginning of the 1970s, this years Nobel
  Laureates made pioneering contributions, which
  later led to the applications of magnetic
  resonance in medical imaging.

80

• Paul Lauterbur (born 1929), Urbana, Illinois,
  USA, discovered the possibility to create a
  two-dimensional picture by introducing gradients
  in the magnetic field.
• By analysis of the characteristics of the emitted
  radio waves, he could determine their origin.
• This made it possible to build up two-dimensional
  pictures of structures that could not be
  visualized with other methods.

81

• Peter Mansfield (born 1933), Nottingham, England,
  further developed the utilization of gradients in
  the magnetic field.
• He showed how the signals could be mathematically
  analysed, which made it possible to develop a
  useful imaging technique.
• Mansfield also showed how extremely fast imaging
  could be achievable.
• This became technically possible within medicine
  a decade later.

82
(No Transcript)
83
Rapid development within medicine

• A great advantage with MRI is that it is harmless
  according to all present knowledge.
• The method does not use ionizing radiation, in
  contrast to ordinary X-ray (Nobel Prize in
  Physics in 1901) or computer tomography (Nobel
  Prize in Physiology or Medicine in 1979)
  examinations.
• However, patients with magnetic metal in the body
  or a pacemaker cannot be examined with MRI due to
  the strong magnetic field, and patients with
  claustrophobia may have difficulties undergoing
  MRI.

84
Especially valuable for examination of the brain
and the spinal cord

• Today, MRI is used to examine almost all organs
  of the body.
• The technique is especially valuable for detailed
  imaging of the brain and the spinal cord.
• Nearly all brain disorders lead to alterations in
  water content, which is reflected in the MRI
  picture.
• A difference in water content of less than a
  percent is enough to detect a pathological change.

85

• In multiple sclerosis ??? , examination with MRI
  is superior for diagnosis and follow-up of the
  disease.
• The symptoms associated with multiple sclerosis
  are caused by local inflammation in the brain and
  the spinal cord.
• With MRI, it is possible to see where in the
  nervous system the inflammation is localized, how
  intense it is, and also how it is influenced by
  treatment.

86
(No Transcript)
87

• Another example is prolonged lower back pain,
  leading to great suffering for the patient and to
  high costs for the society.
• It is important to be able to differentiate
  between muscle pain and pain caused by pressure
  on a nerve or the spinal cord.
• With MRI, it is possible to see if a disc
  herniation is pressing on a nerve and to
  determine if an operation is necessary.

88
Improved diagnostics in cancer

• MRI examinations are very important in diagnosis,
  treatment and follow-up of cancer.
• The images can exactly reveal the limits of a
  tumour, which contributes to more precise surgery
  and radiation therapy.
• Before surgery, it is important to know whether
  the tumour has infiltrated the surrounding
  tissue.
• MRI can more exactly than other methods
  differentiate between tissues and thereby
  contribute to improved surgery.

89

• MRI has also improved the possibilities to
  ascertain the stage of a tumour, and this is
  important for the choice of treatment.
• For example, MRI can determine how deep in the
  tissue a colon cancer has infiltrated and whether
  regional lymph nodes have been affected.

90
Magnetic Resonance Imaging

• The aim of MRI is to provide an image of the
  tissue distribution in a plane through the body,
  for example, by measuring the hydrogen density in
  that plane.
• The idea is, once again, to obtain a
  two-dimensional image of a two-dimensional slice
  through the body.

91
Principle of Magnetic Resonance Imaging

• How can the density of hydrogen nuclei at each
  location of interest in the body be obtained?
• The hydrogen atoms at each location have their
  own specific Larmor resonance frequency,
  depending on the local strength of the external
  magnetic field.

92
Principle of Magnetic Resonance Imaging

• By irradiating the body with EM radiation at a
  certain frequency in a direction perpendicular to
  the external magnetic field, only those hydrogen
  nuclei that have a Larmor frequency equal to the
  frequency of the RF excitation pulse will
  resonate.
• The Larmor frequency depends on the strength of
  the magnetic field, so these nuclei are located
  in a small volume.

93
Principle of Magnetic Resonance Imaging

• The RF excitation pulse has such a duration that
  after the pulse the magnetization vector will
  process perpendicularly to the external magnetic
  field (the 90o RF pulse).
• A current is then induced in a coil perpendicular
  to the external magnetic field. This current has
  an amplitude proportional to the number of the
  resonating nuclei in that volume and a frequency
  equal to the Larmor frequency.

94
Principle of Magnetic Resonance Imaging

• This frequency determines the position of the
  sampled volume. This procedure is repeated with
  other frequencies for all volumes with a specific
  Larmor frequency.
• By this procedure the density of hydrogen nuclei
  can be obtained at all locations of interest.

95
Principle of Magnetic Resonance Imaging

• The external magnetic field consists of two
  parts
• a strong homogeneous field
• a smaller magnetic field, of which the strength
  changes linearly in a certain direction. The
  linearly changing field can be applied in three
  directions by three orthogonally placed gradient
  coils. This changing magnetic field is also
  called the magnetic field gradient.

96

• If the field gradient is directed, for instance,
  from head to toe, every transverse slice in the
  patient resonates at a different Larmor
  frequency.
• RF coils may be used to detect the resultant
  magnetization changes, also called receiving
  coils.

97
Principle of Magnetic Resonance Imaging

• All nuclei in a slice orthogonal to the gradient
  direction will experience the same external
  magnetic field strength and therefore will have
  the same Larmor frequency (this gradient is
  called the slice selection gradient).
• The amplitude of the current in the receiving
  coil after the application of a 90o RF pulse will
  be proportional to the total number of hydrogen
  nuclei in this slice.

98
The phenomenon of spins aligning themselves to an
external magnetic field. At 0 K all spins are
aligned when an external magnetic field is
present (a) when no external magnetic field is
present the spins will point in all directions
(b) at room temperature only a small part of the
nuclei will align themselves 1 per million
(c) at a field strength of 0.1 tesla and 5 per
million (d) at a magnetic field strength of 0.5
tesla (after Philips). Figure 9.11
99
Tesla

• The units of B is the same as the units of F/qv.
• The SI unit of B is equivalent to 1 N.s/C.m
• 1 tesla 1 T 1 N/A.m
• Another unit of B, the gauss (1G 10-4 T)

100
Principle of Magnetic Resonance Imaging

• more position selectivity is attained as follows.
• after the 90o RF pulse, the slice selection
  gradient is switched off and another magnetic
  field gradient is applied orthogonally to the
  direction of the slice selection gradient (this
  additional gradient is called readout or
  measurement gradient, because it is applied after
  the 90o pulse and just before the induced current
  is measured in the receiving coil) the frequency
  of the resonating nuclei will change

101

• nuclei in the slice located along different rows
  orthogonally to the readout gradient direction
  will experience a different external magnetic
  field strength (the sum of the first homogeneous
  field and the readout gradient field) and
  therefore will have different Larmor frequencies
  in the different rows.

102
Principle of Magnetic Resonance Imaging

• This means that each row of hydrogen nuclei in
  the slice induces a current with a different
  frequency in the receiving coil. The amplitude of
  the component with a certain frequency, again, is
  proportional to the number of hydrogen atoms
  along the corresponding row, and the frequency
  now determines the position of the corresponding
  row.

103
Precession of magnetization under the influence
of an external magnetic field with strength Bo
and an oscillating field B1 (due to
electromagnetic radiation) during a 90? RF pulse
as seen from the observer (A) and as seen from
the standpoint of the rotating field (B)
(Philips). Figure 9.12
104
Principle of Magnetic Resonance Imaging

• Since the sum of signals with all of these
  different resonance frequencies is detected
  simultaneously, a Fourier transformation needs to
  be conducted to obtain amplitude and position
  information in the specific frequency spectrum.
• Each frequency corresponds to a row in the slice.
  The amplitude of each frequency component is
  proportional to the number of resonating nuclei
  in a certain row.

105
Principle of Magnetic Resonance Imaging

• The frequency spectrum corresponds to a profile
  representing the number of hydrogen nuclei in
  each row as a function of the location along the
  direction of the readout gradient.
• By measuring the profile for different directions
  of the readout gradient and using a
  back-projection technique, the distribution of
  the hydrogen nuclei over the slice can be
  obtained.
• The procedure can be sped up by applying a second
  field gradient orthogonally to the readout
  gradient before the latter is applied and by
  using the resulting phase differences.

106
Principle of Magnetic Resonance Imaging

• After termination of the 90o RF pulse, the
  magnetization gradually returns back to its
  equilibrium position, which is parallel to the
  external field. This phenomenon is called
  relaxation.

107
Computed Tomography

• Consider a cross section through the body of a
  patient with a thickness of about 1 mm. Now,
  divide the cross-section into a large number of
  small squares, each with an area of about 1 mm2.
• When a narrow X-ray beam, a so-called pencil
  beam, passes through the slice, each square
  through which the beam passes attenuates it to a
  certain extent.

108
Computed Tomography

• The amount of attenuation is determined by the
  molecular composition and the density of the
  tissue present in the square.
• The intensity of the eventually transmitted beam
  will be smaller than the intensity of the
  incident beam.
• The intensity reduction is caused by all the
  squares through which the beam passes.
• Each square may attenuate the beam differently
  because of the presence of different tissue.

109
Computed Tomography

• To measure the attenuation of a beam, it is
  necessary to use a combination of an X-ray tube
  and a detector.
• The absorbing tissue slice is located between
  them. The X-ray tube produces a pencil beam of
  known intensity and the detector measures the
  intensity of the transmitted beam.
• the combination of X-ray tube and detector
  (mounted into the so-called gantry) can both be
  shifted along a line ("translated") and rotated
  (see Fig. 9.6).

110
Principle of computed tomography. The combination
of X-ray tube and detector is translated across
the patient, producing a density profile p(k,
f). By rotating the X-ray tube-detector
combination, a number of profiles will be
obtained. From these profiles the attenuation
coefficients of each pixel can be
determined. Figure 9.6
111
Computed Tomography

• by displaying the attenuation coefficients (the
  values indicating the amount of absorption per
  millimeter) we can obtain an anatomical image, we
  must be able to determine the attenuation
  coefficients of each square separately.
• From one measurement it is not possible to deduce
  how much each separate square attenuated the
  beam.
• Yet, determination of the attenuation of each
  square in the cross-section is the purpose of the
  procedure.

112
Computed Tomography

• Since one pencil beam covers only one row of
  squares, we must translate the beam over a
  distance equal to its width.
• In this way we can take into account the
  attenuation coefficients of the squares located
  on a neighboring row.
• After measuring the transmitted intensity, the
  procedure is repeated by translating the beam and
  measuring the transmitted intensity until we have
  covered the total cross section.

113

• It is still not possible, however, to determine
  the individual attenuation coefficient of each
  square from these data for each position of the
  beam we obtain the total attenuation due to the
  attenuation caused by all squares that are passed
  by the pencil beam.

114
Computed Tomography

• What we have obtained is an intensity profile
  (and, therefore, a measure of the total
  attenuation) of the transmitted beam as a
  function of the position of the beam.
• Each point in the profile indicates how strongly
  the incident beam was attenuated by the row of
  squares that was passed by the beam in that
  position (Fig. 9.6).
• we repeat the procedure outlined above for
  various angles of the beam.
• It is possible to compute the attenuation per
  square from these data.

115

• an X-ray beam is attenuated in an exponential
  way, depending on the length of the path and the
  attenuation coefficients of the squares, denoted
  here as pixels encountered along its path.
• If we assume that the attenuation coefficient is
  constant over the whole pixel and if we represent
  the attenuation of pixel i by the attenuation
  coefficient mi, then the intensity of the
  transmitted beam (I) can be related to the
  intensity of the incident beam (I0) in the way
  represented in Fig. 9.7.

116

• If we take the natural logarithm of the ratio of
  I and I0, we obtain the following relation for
  each pixel
• In(I/I0) ?i1 N dimi

117
The intensity of the transmitted beam as a
function of the attenuation coefficient of the
pixels traversed. Upper part, the intensity after
crossing one volume element middle part, after
traversing n volume elements lower part, the
analog case. Figure 9.7
118

• Here, di is the length of the path that the beam
  traversed through each pixel i, and mi is its
  attenuation coefficient.
• We have obtained an equation with, to the left,
  the measured intensity ratio and, to the right, N
  unknown quantities the coefficients mi. Since
  the geometry of the beam and the cross-section
  are known, the length of the path di traversed
  through each pixel is also known.

119

• The number of unknown quantities therefore only
  increases.
• After the beam has covered the whole cross
  section, the equations contain as many unknowns
  as there are pixels in the cross section.
• By turning the gantry over a small angle and
  repeating the earlier procedure we obtain another
  intensity profile.

120

• The attenuation coefficients of the pixels remain
  the same, but the length of the paths traversed
  through each pixel has changed.
• In this way we obtain new equations with the same
  number of unknowns.
• By measuring the intensity profiles at enough
  angles we can, in principle, obtain as many
  equations as there are unknowns.

121

• As can be seen from Fig. 9.8, a visualization of
  the values of the attenuation coefficients by way
  of grey values indeed produces an anatomical
  image.
• The procedure of CT as explained here is not used
  in practice, because it would be too
  time-consuming, but it provides a good insight
  into the principles of CT.
• In practice, back-projection algorithms are used,
  since these are more efficient.

122
Example of cross-sections through several parts
of the body skull, thorax, and abdomen, obtained
by computed tomography. Figure 9.8
123

• Back projection is one of the techniques that is
  used in practice to obtain the attenuation
  coefficients mi.
• This technique can be used when intensity
  profiles that cover the total cross section under
  various angles are available.
• In an individual profile, each point represents
  the amount of attenuation by the pixels
  transmitted by the beam.

124

• Figure 9.9 shows the intensity profiles that
  result from a single attenuating pixel. Each
  profile shows a dip at the location where the
  beam passed through this pixel.

125
(No Transcript)
126

• Upper left, density distribution of a point
  absorber along a line through the object lower
  left, the resulting intensity profiles lower
  right, the back-projection upper right,
  reconstructed density distribution on a line
  through the object (after Philips).

127

• If we have only one intensity profile we cannot
  determine where on the path the pixel was
  located. We cannot even decide whether the
  absorption was due to a single pixel or was due
  to an attenuating medium that was present over
  the whole path.
• The only inference that we can make is that the
  attenuating medium was present only along one
  line in the cross section, since the intensity
  profile showed a dip at only one point.

128

• The back-projection method starts with the
  assumption that the absorbing medium is uniformly
  distributed over the line. Of course, this may be
  incorrect, but it will be demonstrated that the
  errors resulting from this assumption can be
  corrected.
• If we have several intensity profiles obtained at
  different angles, we get a reconstructed image,
  as shown in Fig. 9.9.
• The reconstruction has a star-like distribution.
  By increasing the number of angles, the intensity
  in the center will increase much faster than the
  intensity at the periphery.

129

• With the use of more angles, the back-projected
  image becomes more similar to the actual one,
  only it is less sharp instead of an image
  showing one attenuating pixel, the neighboring
  pixels are visible in the reconstructed image as
  well.
• This blurring effect can be corrected to a
  certain extent by using appropriate filtering
  techniques, resulting in a sharper image.

130

• Since a real cross section can be considered a
  union of cross sections, with each one containing
  only one attenuating pixel, the back-projection
  technique can also be applied to real patient
  cross sections.
• The back-projection technique can also be used in
  MRI (magnetic resonance imaging) and SPECT
  (single photon emission computed tomography).

131

• It appears that the attenuation coefficient is
  characteristic for the type of tissue (or more
  correctly, the chemical composition of the
  tissue), as is apparent from Fig. 9.10.
• When the values of the attenuation coefficients
  of the pixels are displayed on a monitor in the
  form of grey values, the result consists of
  anatomical images that can be directly
  interpreted.

132
Attenuation coefficients of several tissues
expressed in Hounsfield units.
133
(No Transcript)
134
Medical image modeling tools and applications
135
Hepatic surgery simulation

• Creating a simulator for training physicians to
  perform minimally invasive surgical procedures

136
Figure 1a. The different generations of surgical
simulators.
137
Figure 1b. The different technological components
of a second-generation simulator.
138
Figure 2. Extraction of the hepatic parenchyma ??
(parts a and b), the vascular trees and hepatic
lesions (parts c and d) from a CT scan image.
139
Figure 3a. Liver deformation using a linear
combination of precomputed elementary
deformations.
140
Figure 3b. Sequence of a simulated liver
resection that includes the clipping and cutting
of the portal vein.
141
Virtual colonoscopy
142
Figure 1. A user interface for the virtual
colonoscopy system (courtesy of Viatronix, Inc.).
143
Figure 2. Endoscopic view of painted information
after an antegrade flythrough (left) and an
example of a missed patch after both antegrade
and retrograde flythroughs (right). The green
areas were visualized, while the reddish areas
were missed.
144
Figure 3. A volume-rendered surface view (left)
and an electronic biopsy (right) of a polyp.
145
(No Transcript)
146
Volumetric Heart Modeling and Analysis
147
SPAMM SPAtial Modulation of Magnetization

• A magnetically tagged MRI technique
• Advantage a number of material points within the
  myocardium walls can be marked noninvasively and
  tracked, providing the true 3D motion of the
  heart muscle over time.

148
Figure 1. The essential data needed (boundaries
and tag lines) from each MRI-SPAMM short-axis
image (left) End-diastole short-axis view
(right) Mid-contraction short-axis view.
149
Figure 2. Sample results from the automated
boundary detection algorithm for the LV and RV
and inflow and outflow tracts of the RV.
150
(No Transcript)
151

• Figure 3. The top row shows the estimation of the
  LV-RV endocardium ??? and the epicardium ??? on
  the MRI-tagged slices for a normal heart.
• The second row shows a normal and an abnormal
  heart at end-diastole ???? .
• The RV endocardium of the RV hypertrophy ??
  patient is significantly larger than that of
  normal heart
• (a) Normal heart
• (b) RV hypertrophy heart.
• The third row shows the heart model's motion
  during systole ????.
• Finally, the fourth row shows a finite element
  model of the ventricles?? derived from MRI, with
  local fiber angles (blue) derived from in vitro
  ???? data superimposed at corresponding locations
  in (left) subendocardium and (right)
  subepicardium.

152
Incorporating 3D Virtual Anatomy into the Medical
Curriculum
153
(No Transcript)
154

• Figure 1. a) Automated segmentation of temporalis
  muscle (1) color VH Male slice, (2) a fuzzy
  connected component, (35) iterations of the
  VD-based algorithm, (6) an outline of the
  boundary
• b) 3D segmentation of the left kidney (1) input
  data, (2) fuzzy connectedness, (3, 4) VD
  classification (5) deformable model, (6) hand
  segmentation.
• c) 3D segmented and visualized left kidney
  derived from the Visible Human Male data set3D
  models of (1) fuzzy connectedness, (2) Voronoi
  Diagram classification, (3) deformable model, (4)
  hand segmentation.

155
Figure 2. Foot anatomy a) flexor muscles
(oblique ?? view), b) all the structures (oblique
view), c) a "reference" 3D model (plantar view),
d) corresponding medical illustration based on
the model in c).
156
Figure 3. Issues in biomedical imaging
informatics.
157
Open source software for Medical Image Processing
and Visualization
158
(No Transcript)
159

• Figure. Examples of medical image segmentation
  and registration algorithms available in ITK
• a) Functional MRI fused with MR angiography using
  landmark initialized mutual information
  registration (courtesy of UNC)
• b) 2D to 3D registration of angiography data
  (courtesy of the Imperial College of London)
• c) Inner-ear segmentation of the cochlea ?? and
  vestibular (??)??? system using fast-marching
  level-set methods (courtesy of Kitware)
• d) Eyes, muscles, and optic nerves of the Visible
  Human Project data using interactive color
  watershed segmentation (courtesy of the
  University of Utah).

160
(No Transcript)
161
The Visible Humans Project

• The goal of the Visible Humans project, sponsored
  by the U.S. National Library of Medicine, is to
  provide image data sets of the human body for use
  in the study of anatomy, for use in conducting
  imaging research, and for use in a wide range of
  educational, diagnostic and treatment planning
  and simulation applications.

162
The Visible Humans Project

• The first phase of the project has resulted in
  CT, MRI, and cryosection image sets for a human
  male and a human female.
• The complete male data set consisting of scans
  taken at 1 mm resolution is 15 gigabytes in size.
• The complete female data set consists of scans
  taken at 0.33 mm intervals and is 40 gigabytes in
  size.
• Both datasets may be downloaded over the Internet
  by interested individuals for research and
  experimentation.

163
(No Transcript)
164

• Reconstruction of a sagittal cross section and a
  few horizontal cross sections through one of the
  visible humans

165
(No Transcript)
166

• Example of 3-D presentation of the chest after
  boundary detection, labeling, shadowing and
  coloring of organs

167
Volume rendering

• The process of creating a 2D image directly from
  the 3D volumetric dataset of voxels.
• Voxel Short for volume pixel, the smallest
  distinguishable box-shaped part of a
  three-dimensional image.

168
Reference

• Communications of the ACM, February 2005, Vol.
  48, No. 2.

169
DICOM
170
Application data flow diagram
171

• The Storage Application Entity sends images and
  Presentation States to a remote AE.
• It is associated with the local real-world
  activity Send Images GSPS.
• Send Images GSPS is performed upon user
  request for each study completed or for specific
  images selected.
• When activated by users settings (auto-send),
  each marked set of images and associated
  Presentation States can be immediately stored to
  a preferred destination whenever a Patient/Study
  is closed by the user.

172

• If the remote AE is configured as an archive
  device the Storage AE will request Storage
  Commitment and if a commitment is successfully
  obtained will record this information in the
  local database.

173
SEQUENCING CONSTRAINTS
174
(No Transcript)
175
SEQUENCING OF ACTIVITY SEND IMAGES
176
Service-Object Pair (SOP) Class
177
SEQUENCING OF ACTIVITY ACQUIRE IMAGES
178

• 1. The Worklist AE opens an association with the
  Departmental Scheduler
• 2. The Worklist AE sends an N-CREATE request to
  the Departmental Scheduler to create an MPPS
  instance with status of IN PROGRESS and create
  all necessary attributes. The Departmental
  Scheduler acknowledges the MPPS creation with an
  N-CREATE response (status success).
• 3. The Worklist AE closes the association with
  the Departmental Scheduler.
• 4. All images are acquired and stored in the
  local database.

179

• 5. The Worklist AE opens an association with the
  Departmental Scheduler.
• 6. The Worklist AE sends an N-SET request to the
  Departmental Scheduler to update the MPPS
  instance with status of COMPLETED and set all
  necessary attributes. The Departmental Scheduler
  acknowledges the MPPS update with an N-SET
  response (status success).
• 7. The Worklist AE closes the association with
  the Departmental Scheduler.

180
SEQUENCING OF ACTIVITY FILM IMAGES
181

• 1. Hardcopy AE opens an association with the
  Printer
• 2. N-GET on the Printer SOP Class is used to
  obtain current printer status information. If
  the Printer reports a status of FAILURE, the
  print-job is switched to a failed state and the
  user informed.
• 3. N-CREATE on the Film Session SOP Class creates
  a Film Session.
• 4. N-CREATE on the Presentation LUT SOP Class
  creates a Presentation LUT (if supported by the
  printer).

182

• 5. N-CREATE on the Film Box SOP Class creates a
  Film Box linked to the Film Session. A single
  Image Box will be created as the result of this
  operation (Hardcopy AE only uses the format
  STANDARD\1,1)
• 6. N-SET on the Image Box SOP Class transfers the
  contents of the film sheet to the printer. If
  the printer does not support the Presentation LUT
  SOP Class, the image data will be passed through
  a printer-specific correction LUT before being
  sent.
• 7. N-ACTION on the Film Box SOP Class instructs
  the printer to print the Film Box
• 8. The printer prints the requested number of
  film sheets

183

• 9. The Printer asynchronously reports its status
  via N-EVENT-REPORT notification (Printer SOP
  Class). The printer can send this message at any
  time. Hardcopy AE does not require the
  N-EVENT-REPORT to be sent. Hardcopy AE is
  capable of receiving an N-EVENT-REPORT
  notification at any time during an association.
  If the Printer reports a status of FAILURE, the
  print-job is switched to a failed state and the
  user informed.
• 10. N-DELETE on the Film Session SOP Class
  deletes the complete Film Session SOP Instance
  hierarchy.
• 11. Hardcopy AE closes the association with the
  Printer

184
Network interfaces

• Physical network interface
• Ethernet 100baseT
• Ethernet 10baseT

185

• DICOMSRV DICOM MWL and MPPS application
• MWL Modality Worklist
• MPPS Modality Performed Procedure Step

186
(No Transcript)
187

• DICOMSRV accepts associations for Verification
  from Verification SCUs and responds automatically
  with Success status
• DICOMSRV accepts Association Requests for
  Modality Worklist from MWL SCUs and responds to
  queries from these SCUs. When a query is received
  DICOMSRV engages in local real-world activity
  Scheduled Procedure Queries. This results in a
  set of matching responses that DICOMSRV returns
  to the MWL SCU.

188

• DICOMSRV accepts Association Requests for
  Modality Performed Procedure Step from MPPS SCUs
  and responds to N-CREATE and N-SET Requests from
  these SCUs. When an N-CREATE or N-SET is received
  DICOMSRV engages in local real-world activity
  Update Procedure. This results in updates to the
  DICOMRis Database per the contents of the
  received message. DICOMSRV then returns N-SET or
  N-CREATE status to the MPPS SCU.

189

• SCU Service Class User
• SCP Service Class Provider
• SOP DICOM Service-Object Pair
• GSDF Grayscale Standard Display Function
• IOD Information Object Definition
• LUT Look-up Table
• UID Unique Identifier

190
(No Transcript)
191

• 1. The Modality queries for a worklist of
  Scheduled Procedure Steps
• 2. DICOMSRV searches its database and returns
  matches to the query
• 3. The Modality begins performance of a Procedure
  Step and sends the MPPS N-CREATE
• 4. The Modality completes or discontinues the
  procedure and sends the MPPS N-SET with status of
  COMPLETED or DISCONTINUED

192
SEQUENCING DIAGRAM FOR ACTIVITY CONFIGURED AE
REQUESTS MWL QUERY
193

• 1. The Modality opens an Association with
  DICOMSRV for the purpose of querying for a
  Modality Worklist
• 2. The Modality sends an MWL C-FIND query to
  DICOMSRV
• 3. DICOMSRV queries its database using the
  attributes from the C-FIND Request and returns 0
  to N C-FIND responses depending on matches
  returned from the database. DICOMSRV checks for a
  C-FIND Cancel Request after a configured number
  of responses are sent. If a Cancel is received
  then no further Pending responses are sent.
• 4. DICOMSRV sends the final C-FIND response