Co-moderators:

Norbert J. Pelc, Sc.D. Stanford University

Cynthia McCollough, Ph.D.

Mayo Clinic

Multi-Energy CT: Current status and recent innovations

Multi-Energy CT: Current status and recent innovations

Basic concepts & current implementations

Norbert Pelc

Data and image analysis methods

Lifeng Yu

Clinical applications

Cynthia McCullough

Future directions

Taly Gilat Schmidt

Panel Discussion

Norbert J. Pelc, Sc.D.

Departments of Bioengineering and Radiology

Stanford University

Multi-Energy CT: Basic concepts & current

implementations

Acknowledgements

Uri Shreter and Robert Senzig: GE Healthcare

Thomas Flohr, Bernhard Schmidt and Karl Krzymyk: Siemens Healthcare

Ami Altman, Alex Ganin, Andy Mack, Efrat Shefer: Philips Healthcare

Cynthia McCollough: Mayo Clinic

Research support:

GE Healthcare

Philips Healthcare,

Samsung Electronics

Motivation

www.uhrad.com/

ctarc/ct153b2.jpg

Lower density or lower atomic number?

Motivation

www.uhrad.com/

ctarc/ct153b2.jpg

Is the HU increase due to underlying tissue

property or contrast agent?

Motivation

TN Sheth, et al, CARJ 56, 15, 2006.

Can we separate iodine and calcium,

or reduce blooming?

Attenuation coefficients depend on photon energy

0.1

1

10

100

35 55 75 95 115 135

bone

water

iodine

atte

nuat

ion c

oef

fici

ent

photon energy

use CT measurements at multiple energies

to add material specificity

Motivation

“Two pictures are taken of the same slice, one

at 100 kV and the other at 140 kV...so that

areas of high atomic numbers can be

enhanced... Tests carried out to date have

shown that iodine (Z=53) can be readily

differentiated from calcium (Z=20)”.

G.N. Hounsfield, BJR 46, 1016-22, 1973.

OUTLINE

• Physical principles of multi-energy x-ray

measurements

• Implementations

achieving spectral selectivity

obtaining multi-energy measurements

• Summary

attenuation and material identification

T

I0

I = I0 e-T

T = ln(I0/ I)

measures product of & T

not very material specific

unknown thickness two known materials

energy E1

I1

tb

I1 = I01 e -(w1tw+ b1tb)

I01

water

bone

10 100 20 50 200

cortical bone

water

photon energy (keV)

lin

ear

atte

nu

atio

n c

oef

ficie

nt

(cm

-1)

0.1

1.0

10.0

100.0

I2

E2

I02

I2 = I02 e -(w2tw+ b2tb)

E2 E1

solve for tw and/or tb

dual energy x-ray absorptiometry (DEXA)

2 energies 2 materials

• 2 energies 2 materials

material analysis with absorptiometry

• can we generalize this? N energies for N

materials?

• limitation: two strong interaction mechanisms

Compton scattering and photoelectric absorption

Barring a K-edge in the spectrum, the energy

dependence of each is the same for all elements!!

• Attenuation of any material is ~ a weighted sum of photoelectric and Compton functions

• Any material can be modeled as a weighted sum of two other materials (basis materials)

basis material decomposition:

Basis material decomposition

0.1

1

10

100

1000

0 20 40 60 80 100 120 140

O

Ca

Cu

Ca'

Cu

.61*O + .04*Cu

Ca

O

Basis material decomposition

M grams of Ca .04 M grams of Cu

.61 M grams of O

I

I0 I0

I

=

Indistinguishable at any x-ray energy above their K-edge

Common “basis functions”:

iodine and water, aluminum and plastic

photoelectric and Compton

material analysis with absorptiometry

• 1 energy

constant thickness

2 known materials

(constant thickness is a constraint)

• 2 energies

unknown thickness

2 known materials

• N energies

unknown thickness

still only 2 known materials

(only 2 independent functions)

• 2 energies

constant thickness

3 known materials

(constant thickness (CT voxel) is a constraint)

K-edges

10 100 20 50 200

iodine cortical bone

water

photon energy (keV)

linea

r at

tenuat

ion c

oef

fici

ent

(cm

-1)

0.1

1.0

10.0

100.0

Very specific material information

iodine image

SNR=3.4

Noise

low energy high energy

water image

SNR=37

~ a2 low + b2 high

Iodine data ~

a • Datalow - b • Datahigh

depends on:

- specific energies

- allocation of dose to

the two measurements

iodine image

SNR=1.3

iodine image

SNR=3.4

Noise depends on dose allocation

water image

SNR=34

water image

SNR=37

with 80/140 kVp dose

allocation that maximizes

iodine SNR

80 kVp dose, 140 kVp dose

same total dose

Spectral separation and control

• very critical for SNR efficiency, separation

robustness, etc.

• implementations

different kVp and/or filtration

detectors with energy discrimination

Principle of Dual Energy CT

Tube 2

Tube 1

Mean Energy:

56 kV 76 kV

Data acquisition with different X-ray spectra: 80 kV / 140 kV

Different mean energies of the X-ray quanta Courtesy of B. Krauss, B. Schmidt, and

Th. Flohr, Siemens Medical Solutions

SOMATOM Definition Flash

SPS= Selective Photon Shield

syngo Dual Energy

- Principle of Dual Energy

S1: 80 kV

x 104 15

10

5

0 50 100

150

80 kV 140 kV

photon energy (keV)

nu

mb

er

of

qu

an

ta

S2: 140 kV S1: 80 kV

15

10

5

0 50 100

150

80 kV

photon energy (keV)

nu

mb

er

of

qu

an

ta

140 kV

S2: 140 kV

x 104

+ SPS

140 kV + SPS

Dual kVp, dual filtration

135 kVp

1.5 mm bronze

85 kVp

0.1 mm erbium

• switched filtration

improves separation

• different mA helps

apportion dose

Lehmann et al: Med Phys 8, 659-67, 1981.

Slide 6 – Learn More 1 Perfect alignment Simultaneous alignment in time and space

NanoPanel Prism

Top Scintillator

X-Rays

Side-looking photodiode array

Low Energy Raw data E1 image

Top scintillator

Full CT Image

Bottom Scintillator:

2-mm GOS+ High Energy Raw data

Combined Raw data

Effective atomic number small but does not sacrifice light output Thickness optimized for energy separation and low-energy image noise

Bottom scintillator absorbs 99.5% of high-energy spectrum

E2 image

CT image

The IQon Spectral CT is not yet CE Marked and not available for sale in all regions. The IQon Spectral CT is pending 510(k) and not available for sale in the USA.

Carmi R, Naveh G, and Altman A: IEEE

NSS M03-367, 1876-78, 2005

Layered detector

• relatively poorer spectral separation

• simultaneous dual energy sensing

Investigational device Confidential

Absorbed single X-ray photon

High

Voltage

Counter 3

Counter 2

Counter 1

Counter 4

Charge pulse

Pulse height proportional

to x-ray photon energy

Discriminating

thresholds

Stored counts of all energy windows,

for each reading time period

Direct conversion

material

Electronics

Pixelated electrodes

Photon energy

Counts

Photon Counting Spectral CT – Detector Principle

Direct- Conversion Detector efficiently translates X-ray photons into large electronic signals

These signals are binned according to their corresponding X-ray energies

(CZT crystal: Cadmium Zinc Telluride)

Spectral separation

Separation and control of the spectra is important

reduced noise

improved material characterization

improved dose efficiency

Achieved by

choice of kVp (and mAs allocation)

use of filtration

performance of energy discriminating detectors

• ideal photon counting with K-edge filter

• ideal photon counting with energy analysis

• different kVp and filtration

• different kVp

• layered detector

Spectral separation implementations

better

spectral

separation

and

dose

efficiency

Data acquisition implementations

• Sequential scans at different kVp

motion sensitivity > scan time

• Two sources at ~90º on the same gantry

SOMATOM Definition Flash

syngo Dual Energy

- Principle of Dual Energy

Two X-ray tubes, 1st: 80 or 100 kVp, 2nd: 140 kVp

Dual Source Challenge: Inconsistent scans

Ideal

Moving Phantom

Simulation

Dual Source system

Moving Objects

Does not see movement

Courtesy of R. Senzig, GE Healthcare

Data acquisition implementations

• Sequential scans at different kVp

motion sensitivity > scan time

• Two sources at ~90º on the same gantry

some motion sensitivity (~ 25% Trot)

• Switching kVp within a single scan1, 2

1. Lehmann et al: Med Phys 8, 659-67, 1981.

2. Kalender et al, Med Phys 13, 334, 1986.

Courtesy of Uri Shreter, GE Healthcare

Rapid kVp switching Dual energy CT

• Requires fast

generator and

detectors

• Dose allocation

controlled by dwell

time

• Difficult to switch

filters

Data acquisition implementations

• Sequential scans at different kVp

motion sensitivity > scan time

• Two sources at 90º on the same gantry

some motion sensitivity (~ 25% Trot)

• Switching kVp within a single scan

• Energy discriminating detectors

layered detector, photon counting

better

immunity

to

motion

Summary of commercial systems

• Siemens: two sources, different kVp (80 or 100

/140) and filtration, direct control of mA

• GE: single source with rapid switching, same

filter for both kVps, control mAs by dwell time

• Philips: dual-layer detector, usual kVp mAs

control

• Lots of R&D work, especially on photon

counting detectors