BME 311 Lecture 6

practicalities of clinical environment - cost

• purchase

• operation - trained and licensed technologist

• consumables - contrast or tracer

• maintenance

practicalities of clinical environment - time

• acquisition

• processing

• only enough time to get “sufficient” accuracy

noise

• fundamental randomness in data/images

• different data even if exact same experiment were repeated

• raw data (“counts”) often “poisson” whereby stdev = sqrt(mean

background

• signal that is not of interest - real or artifact

• reduces contrast of feature you are interested in

• can also have noise in it

contrast

• ability to distinguish different intensities

• often ration of signal to background

“count”

detection of a specific (integer) number of x-rays

projection imaging

acquisition of 2D (i.e. 2D) image of patients 3D anatomy

i.e. total beam attenuation along the line Z

what are we measuring in radiography

• counts at an image point relate to attenuation along line connecting to focal spot

• counts proportional to exp(-mu*x)

• what is mu varies along line? break into line segments dx:

  • counts proportional to exp(-mu1*dx)*exp(-mu2*dx)*exp(-mu3) = exp(-[sum of mu values]*dx)

basic geometric principles

• two triangles similar: a:A = b:B = c:C = h:H

• magnification (M) = I/O = SID/SOD

  • I = image size

  • O = object size

  • SID = source to image distance

  • SOD = source to object distance

what about imperfect focal spot

• extended source - think of it as a number of different points

• geometric blurring of objects occurs

• reduce OID to reduce blurring, but does it really help since this also reduces magnification? what really matters is to minimize f/I

f/F = g/G = OID/SOD = (SID-SOD)/SOD
= SID/SOD-1 = M - 1
so f = F (M-1)
Remember I/O = M, so I = M O
Ratio f/I = F/O (M-1) / M
Or f/I = F/O * (1 - 1/M)

closer detector produces sharper image

assuming detector has sufficient resolution

which condition is for compressed breast imaging

• very small focal spot is possible = small F, so small blurring

• need high resolution detector to acquire image

many optical technologies exist for imaging

• easy to detect in small pixels

• mature - commercial applications for >100 years

few detectors for x- or gamma-rays

• much more penetrating

• not many interactions in optical detectors

scintillator (phosphor)

• special material in which photoelectric/compton effect from x-ray interaction produces optical photons

• convert an x-ray photon into many optical photons

• “intensifying screen” contains. scintillator crystals

• want high absorption and conversion efficiencies

absorption efficiency aka QDE (Quantum Detection Efficiency (%))

• = fraction of incident x-rays that interact with detector and thus are deleted

  • efficiency of detector is proportional to 1 - exp(-mu*x)

  • determines radiographic “speed” - how long an exposure us needed

    • i.e., high speed = high absorption efficiency = short exposure time

total conversion efficiency

• fraction of energy deposited by x-ray absorption that is converted into optical photons

  • depends on the intrinsic properties of phosphor

conversion efficiency

• fraction of absorbed energy emitted as light

• % converted to photons

• e.g. Gd₂O₂S – conversion efficiency 15% - emits green light – 2.7 eV/photon

• What is the number of green light photons produced by absorption of 50keV x-ray photon?

  • 50000 x 15% = 7500 eV
    7500 / 2.7 ~ 2800 green photons

radiography detector types

• analog: screen-film

• digital: more modern, digitize earlier

  • computed radiography

  • CCD and CMOS

  • thin film transistor (TFT)

screen film cassette

• cost

  • film can’t be reused

• time

  • chemical developing takes time

intensifying screen function and geometry

• two functions:

  • absorb incident x-ray

  • emit visible light - isotropically = in all directions

• thicker → absorb more photons (i.e., more sensitive) → but reduced spatial resolution (due to more light spread)

• use 2 layers:

  • preserve spatial resolution by keeping scintillator thin

  • compensate for lost detection efficiency by using 2 layers

digital imaging: multiple technologies

conversion from analog to digital at different points in the imaging process

digital imaging: computed radiography (CR)

• partially digital

• instead of screen+film, uses special phosphor that converts x-ray to long-lived excited electron states

computed radiography: principle of readout

• A: x-rays are absorbed in the storage phosphor, electrons reach the conduction band

• B: electrons are trapped in this high-energy state

• C: red laser provides the trapped electrons to be excited, some of them will drop down to the valence band, emitting blue light during the transition

computed radiography

• raster scan

• once read out, ready for use again

first digital method: computed radiography

x-ray exposure, imaging plate, CR

computed radiography vs screen film

• CR provides greater dynamic range

• exposure is proportional to duration of x-ray pulse

  • which must be decided in advance by technologist

  • depends on size of patient, kVp, etc

  • can be incorrectly estimated

• more tolerant of over/under exposure

• retake due to overexposure or underexposure

  • cost

  • inconvenience

  • more dose

number of voxels, pixel exposure

resultant images

modern light sensors

silicon devices accumulate charge deposited by light

CCD = charge coupled device

• ~2.5 × 2.5 cm, 2048 × 2048 pixels

• great resolution (microns)

• but small and slower to read

CMOS

• microchip technology

• random access

• expensive, can be noisier

modern light sensors challenges

• pure crystalline silicon wafers

• need large areas = expensive

TFT arrays

• thin film transistors

• “hybrid” of CCD and CMOS

• some simple circuitry

• amorphous silicon = cheaper

• large areas relatively cheaply

TFT arrays: indirect vs direct

• indirect

  • scintillator to produce light

• direct

  • no scintillator, no optical photons at all

  • x-rays interact directly in detector: e-hole pairs

  • amorphous selenium semiconductor (Se Z=34 vs. Si Z=14)

  • charge travels straight: thick does not hurt resolution

  • lower absorption efficiency than scintillator (mu of Si is not that high)

dual energy radiography

soft tissue, bone enhanced

the anti-scatter grid

• why is scatter bad?

• tail, narrow lead walls = septa

• aligned with focal spot

• absorbs scattered radiation

• H/W in diagnostic radiology can be >10

• but work only at fixed distance between focal spot and detector