PET/CT Principles D.W. Townsend, PhD Departments of Medicine and Radiology, University of Tennessee, Knoxville, TN
Cancer Imaging Imaging and Tracer Development
Advances in PET-SPECT/CT and RTP Melbourne, Australia December 13th 2005 2005
Principles of PET/CT
principles of design
!
Hardware fusion: anatomy + function • to image different aspects of diseas
Why combine anatomy and function?
• to identify non-specific tracer uptake • to facilitate image interpretation • to provide unique added value to bo
Fused image accurately localizes uptake into a lymph node and thus demonstrates spread of disease. Fused images can improve staging of head and neck cancer
CT (anatomy)
PET/CT
PET (function)
Designing a PET/CT scanner
Gantry dimensions: 228 cm x 200 cm x 168 cm
168 cm 80 cm
CT rotation: rotation: 0.4 s; s; 16 slice 200 cm
CT CT
PET 190 cm
biograph 16
Dual-modality imaging range
PET/CT design choices PET parameters
CT parameters detectors:
ceramic; 1 – 24
sc scintillator:
BGO; GSO; LSO
slices:
2, 4, 6, 8, 16, 40, 64
det detec ecto torr size size::
4 x 4 mm; mm; 6 x 6 mm
trans. FOV:
45 – 50 cm
tr trans. FOV:
55 – 60 cm
rota rotati tion on spe speed ed:: 0.4 – 2.0 2.0 s
re resolution:
~ 4 – 6 mm
tube current:
80 – 280 mA
ax axial extent:
15 – 18 cm
heat heat capa capac city ity:
3.5 3.5 – 6.5 MHU MHU
se septa:
2D/3D; 3D only
topogram:
128 – 2000 cm
att attenuation:
CT-based (r (rod; po point)
time /100 cm:
13 – 90 s
patient port:
60 cm; 70 cm
slice width:
0.6 – 10 mm
pe peak NECR:
30 @ 9 kBq/ml
patient port:
70 cm
(3D)
– 93 @ 29 kBq/ml
PET/CT patient support designs
CT
PET
Fixed cantilever point; floor-mounted rails
1
2
Variable cantilever point; dual positions CT
PET
CT Variable cantilever point; support in tunnel
PET
Stationary bed; gantries travel on rails
Current PET/CT scanner designs BGO, LYSO
GSO (Zr)
mm 3
4 x 6 x 30 mm 3 3D only (no sep 6, 10, 16 slice 70 cm port 6 ns coincidenc bed supported
6 x 6 x 30 2D/3D (septa) 8, 16, 64 slice CT 70 cm port dual-position bed
Discovery ST, STE, RX
Gemini GXL
LSO
LSO mm 3
6 x 6 x 25 mm 3 3D only; rotatin 4 slice CT 70 / 60 cm port 4.5 ns coincide bed on rails
4 x 4 x 20 3D only (no septa) 8, 16, 64 slice CT 70 cm port 4.5 ns coincidence bed on rails
biograph 6, 16, 64
SceptreP3
Current PET/CT scanner designs (continued) NEMA - US Shipments ($M) $120.0 $100.0
$45
$80.0 $60.0 $42
Aquiduo
$80
$49
$40.0
$41
$51 $82
$48
LSO
$39 $3 9
4 x 4 x 20 3D only (no septa) 16 slice CT 70 cm port 4.5 ns coincidence gantry on rails
FY
Q3 Q4 02 02
$76
$89
$6
$3
$ 39 $13
$0.0
$105 $74
$74
$20.0 $37
mm 3
$92
$8
$4
$1
Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 03 03 03 03 04 04 04 04 05 05
PET/CT PET
PET/CT clinical impact
Innovations in PET technology -1• improved PET scintillator performance • increased axial field-of-view for volume sensitivity • Scintillators
Density Zeff Decay (ns)
• Sensitivity
BGO
LSO
GSO
7.13
7.4
6.7
74
66
61
300
35-45
30-60
Light
8,200
28,000
10,000
% NaI
15
75
25
3D (septa retracted)
Panel detectors and P-5H scanner 36 cm 52 cm
• 5 heads; 30 rpm rotation of assembly • 68Ge point sources (2 x 10 x 1 mCi)
P-5H schematic
• list mode Em and Tx acquisition • simultaneous Em and Tx acquisition
Innovations in PET technology -2• improvements in spatial resolution 17
13
22
10 28
37
6.4 4.0 mm mm xx 6.4 4.0 mm mm BGO LSO 8x8 crystals/detec tor r 13 13crystals/detector crystals/detector crystals/detecto 3.4 mmslice slicewidth width 2 mm
) 100 % ( y r e v o c e 50 R
22
28
37
17 13 10
0
Sphere diameter
Innovations in PET technology -3• fully 3D statistical reconstruction algorithms • incorporation of time-of-flig time-of-flight ht (TOF) information • Reconstructio Reconstruction n algorithms
• Time-of-flight (TOF)
t (ns)
3DRP
FORE+AWOSEM
x (cm)
SNR*
0.1
1.5
5.2
0.3
4.5
3.0
0.5
7.5
2.3
1.2
18.0
1.5
• accurate system model
* SNR gain for 40 cm phantom
Innovations in PET technology -4• continuous bed motion acquisition
23 23
35
23
35
23 23
23 35
35
23
End
23
Data Collection
Start
255 Planes
step-and-shoot
continuous movement
Why acquire PET data with continuous bed motion?
•
Unif Unifor orm m axia axiall sign signal al-t -too-no nois ise e
•
Pote Potent ntia iall lly y imp impro rove ve lesi lesion on dete detect ctio ion n
•
Elim Elimin inat ate e axia axiall under under-s -sam ampl plin ing g arti artifa fact cts s
•
Redu Reduce ce pati patien entt mov movem emen entt art artef efac acts ts
•
Redu Reduce ce the the noi noise se from from norm normal aliz izat atio ion n
•
Natu Natura rall way way to defi define ne the the axi axial al FOV FOV
Innovations in CT scanner design
T C l a r i p S
T rot
collimation
typical 30 cm scan 1
slices/s
1972
300 s !4
13 mm
---
0.007 /4
1980
2s
2 mm
20 mm, 30 s
0.5
1990
1s
1 mm
10 mm, 30 s
1
1995
0.75 s
1 mm
8 mm, 30 s
1.3
1998
0.5 s
4 ! 1 mm
4 ! 1 mm, 30 s
12
2002
0.4 s
16 ! 0.75 mm
16 ! 0.75 mm, 12 s
60
2004
0.3 s
64 ! 0.5 mm
64 ! 0.5 mm, 3 s
240
2010
0.2 s
512 ! 0.5 mm
512 ! 0.5 mm, 0.2 s
2500
1 assuming
a breath-hold limit of 30 s p = 1, otherwise S eff is increased 3 assuming p = 1.5 since image quality is independent of pitch for MSCT 2 assuming
2
2
3
3
3
Principles of PET/CT
principles of operation
!
CT-based attenuation correction 0 7
µ
/
1 1 5
0.5
µ
adipose tissue lung liver, blood, pancreas, skin spongiosa
0.4
femur
air-water mix
-1000 air
-500
water-bone mix
0 water
500
1000
cortica 1500
CT number (HU)
Mixture Model 1500 Break
point ~80 HU
1000 ) m c / ( ) V e k 1 1 5 ( !
air-water mix
water-bone mix
p 500 V k mix 1 0 4 1 t 0 a U H
µ
= c µwater
µmix = c1 µbone + (1-c1) µwater phantom patient, soft
-500
patient contrast patient, bone
-1000 -1000
-500
0
500
HU at 80 kVp
Hounsfield Units (HU)
1000
1500
2000
PET/CT imaging: some practical issues CT breathing protocols
Metal artifacts
Breath-hold CT in 420 lb patient. Note diaphragm is even resolved from liver. Few artefacts can be seen even with shallow breathing
10 s scan time Intravenous contrast
Bolus of iv contrast may cause artefacts in PET image. Can be identified on CT and uncorrected PET image. Use saline flush to reduce effect
biograph 16
Oral contrast
Oral contrast CT
Enhancement up to ~700 HU. Segment bone and oral contrast
B one
Barium contrast
Summary: CT-based attenuation correction • bi-linear scaling model validated in human tissue • no diagnostic issues with iv contrast (dual injector) • oral contrast effect can be corrected if wamted • negative contrast (water) can be used instead • CT and non-corrected PET used as reference • respiration protocol, particularly for lung cancer "
CT acquired with breath hold at partial expiration
"
PET gated for improved match of lung lesions
PET/CT radiation dosimetry: scan protocols Effective FDG dose: E int = Effective CT dose:
Scout
0.2-0.8 mSv
"FDG
20
. A , where "FDG = 19 µSv/MBq
E ext = "CT . CTDIvol , where
Spiral CT
D-CT: 5 - 30 mSv LD-CT: 0.5 - 3 mSv
"CT =
1.47 mSv/MGy
PET
5 - 7 mSv Reconstruction
PET/CT
D-CT: 10.2 – 37.8 mSv LD-CT:
5.7 – 10.8 mSv
LD + D: 10.7 – 40.8 mSv
Average total dose: dose: 25 mSv
PET/CT scan protocol topogram
spiral CT
• arms up (except neck) • acquired CT scan with breath hold Mixing model: CT-based attenuation correction • partial or full expiration • bi-linear scaling model • 10 – 15 s scan time Fusion • intravenous water-air +/- oral contrast • threshold at ~ 50 HU mix • kVp-dependent scaling • 120 kVp, 140 - 160 mAs water-bone 0.18 0.16 0.14 0.12
0.1
oral contrast • little error from iv contrast • 10 mCi of FDG • artifactsattenuation from metal implants • CT-based correction • use CT or non-corrected PET scan • 90PET min uptake period • model-based scatter correction Hounsfield units < 15% • 2 - 6 min per bed position• 336• xACF 336error reconstruction matrix • validated in human tissue CT
0.08
attenuation correction • correction for
PE T mix
0.06 0.04 0.02
0
-1000
-500
0
500
1000
1500
• 4 - 15 bed positions • Fourier rebinning Reconstruction • 10 - 40 min scan duration• 2D-OSEM, 4i/8s; fully 3D-OSEM fused imag • respiratory gating • 5 mm axial smoothing CT PET PET image
Principles of PET/CT
principles of application
!
Mandibular cancer
biograph 16 Cancer Imaging an Tracer Developm e
83 year-old female with mandibular cancer. PET/CT scan acquired pre-surgery identified 3 left-side positive nodes 5-12 mm in size with increased FDG uptake. Post surgery, pathology identified 35 nodes positive for cancer.
Primary (1.5 x 3.8 cm)
5 mm lytic spine lesion
Nodes (12 mm, 7 mm, 5 mm)
Bone lesions, 6-7 mm in diameter
Staging disease
biograph 16 Cancer Imaging an Tracer Developm e
Left side: side: 9/30 nodes positive. 6/16 level IV; 3/5 level V (<1.5 cm) III
Right side: side: 1/16 nodes positive. 1/12 level III (1.3 cm) Primary cancer found in the right palatine tonsil
49 year-old male with mass in left neck. Unknown primary.
Prostate cancer
biograph 16 Cancer Imaging an Tracer Developm e
82 year-old male, 189 lbs, with 17 year history of prostate cancer, referred for PET/CT for staging following recent rise in PSA. Elevated uptake of FDG in left prostate bed (SUV=8.7) suspicious for recurrent disease. Metastatic uptake in L3, sacral and right proximal femur noted. Scan Sc an pr prot otoc ocol ol::
CT: 16 CT: 168 8 mA mAs, s, 12 120 0 kV kV,, 5 mm sl slic ices es at 0. 0.75 75 mm PET: 11.1 mCi FDG, 120 min pi, 3 min/bed, 8 beds; 4i/8s; 5F
Benign pleomorphic adenoma Cancer Imaging and Tracer Developm ent
PET
• 10.1 mCi injected; 133 lbs; 143 min post-injection; • 7 bed positions; 2 min/bed position • SUV (mean) = 6 PET/CT
Summary: the impact of FDG-PET/CT • localize pathological FDG uptake • distinguish normal uptake from pathology • improve accuracy of interpretation interpretation • improve confidence of clinical reading • add value to both CT both CT and PET PET for staging • improve accuracy of therapy planning planning and and biopsy • accurately monitor therapeutic response • reduce scan duration and increase throughput • simplify scheduling for patients and physicians
The future for PET/CT s 800 t i n 700 U
PET
600
© 2004 Frost & Sullivan Sullivan
PET/CT
500
PET/CT will likely replace PET even without extensive clinical validation of PET/CT compared to PET only.
?
400 300 200 100 0 2000
2001
2 002
2003
2004
2005
2 006
2007
• greater clinical flexibility • yields better clinical results • increased confidence • CT can be used stand-alone
20 08
2009
2 010
• 99% PET/CT by 2010 • 10% growth rate (units) • PET/CT is 100% of growth
