Transcript - AAPM Chapter

Monte Carlo Demonstration of Small
Animal in-vivo Functional And
Anatomical Imaging With Neutrons
David C. Medich, Ph.D., CHP
Goals of this Presentation
1. To provide a justification as to why we could use a new
clinical and biomedical imaging tool.
2. To discuss how neutrons can be used to obtain an in-vivo
biological image that contains functional and anatomical
information. NOTE: focus here is on a small animal model.
3. To present Monte Carlo simulated results for:
1. Basic neutron images comparing quality both with and without
scatter contributions,
2. S/N ratio of different tissues and materials relative to water,
3. S/N ratio of tumors relative to their normal tissue counterpart,
4. S/N ratio of B-10 and Gd-157, to determine the concentrations
needed to obtain a useful functional image in a mouse model.
Current Technology
Technologies such as f-MRI, PET, and SPECT have been very
successful in identifying changes in tissue metabolism and
function, but these technologies have limitations...
Limitations:
• Poor spatial resolution : fMRI (2-3mm), PET (3-5mm), SPECT (510mm),
• fMRI – poor temporal resolution,
• Little to no anatomical imaging info is obtained from the technology
• PET – requires on-site (or close proximity) cyclotron to produce
positron emitters with very short half-lives,
• SPECT – isotopes produced at very limited number of foreign
reactors. A temporary shutdown of one of these reactors leads to
clinical isotope shortages...
Possible Role of Neutrons in
Functional Imaging...
Neutron Radiography
General Advantages: – why neutrons?
• Neutrons interact directly with the atomic nucleus (unlike photons which
interact with atomic electrons).
• Neutrons therefore have different reaction rates for isotopes of the
same element.
• Neutrons also have a high dynamic range of interaction cross-sections.
• neutron radiography, which our
technique is based, has very
high spatial resolution and
efficiency.
• Kardjilov, for example, used a
10 mm Gd2O2S(Tb) screen and
reported a 90% detection
efficiency with a spatial
resolution of 25 mm.(1)
1. N. Kardjilov et al. Nucl. Instr. and Meth. A 651 95-99 (2011)
Neutron Radiography
Potential Clinical Advantages
1. Neutrons interact readily with hydrogen, and nitrogen (and to a
lesser extent carbon and oxygen) compared to many other
materials.
•
Therefore, Neutrons are very sensitive to local changes in a tissue’s H,
C, O, N content.
•
This is important because, as we know from MRI, changes in hydrogen
content enables us to anatomically differentiate and image soft
tissues.
2. And, certain stable and non-toxic isotopes are much more opaque
to neutrons than lead is to diagnostic x-rays!
•
Therefore, these isotopes could be used with neutrons as a very
effective contrast agent.
Advantages of Neutron Imaging
Advantages of Neutron Imaging
Here is a quick summary of the thermal neutron absorption
in B-10 and Gd-157 as compared typical x-ray photon
absorption in lead...
Transmission through 10 um of shielding material and shield TVL
Radiation
Shielding
Cross Section
Transmission (10 mm)
Source
0.025 eV Neutron
0.025 eV Neutron
50 keV Photon
100 keV Photon
Material
Gd-157
B-10
lead
lead
(1/cm)
7716.8
501.5
90.9
62.7
(percent)
4.5E-02
60.56
91.31
93.92
TVL
( mm)
3
46
253
367
Advantages of Neutron Imaging
Identifiable
Tumors
Neuron
Radiography
X-ray
Radiography
Comparison of raw tumor imaging capabilities in neutron radiography and x-ray
radiography: The top images demonstrate that native neutron radiography has basic
tumor detection capabilities.
Image From: Brown and Parks, “Neutron Radiography in Biologic Media”, Am. J.
Roentg 106, 472 (1969)
Advantages of Neutron Imaging
Comparison between (a) histology and (b) Neutron Capture Radiography of a
70µm thick liver tissue sample with liver metastases after infusion of BPA.
From: Wittig, A., J. Michel, et al. (2008). "Boron analysis and boron imaging in
biological materials for Boron Neutron Capture Therapy (BNCT)." Critical reviews in
oncology/hematology 68(1): 66-90.
Advantages of Neutron Imaging
X-ray
Neutron
ImageImage
Advantages of Neutron Imaging
From: Metzke et. Al. “Neutron computed tomography of rat lungs” 2011 Phys. Med. Biol.
56 (2011) N1–N10
Advantages of Neutron Imaging
Potential Clinical Applications for Anatomical Imaging:
1. Both native and functional tumor detection have been
demonstrated.
2. Possible identification of early-stage osteosarcoma in both
trabecular and compact cortical bone.
3. Ability to image near or through metallic implants or
pacemakers.
• Also could assess blood flow around an aneurism embolization coil
or vascular stent.
• And assess the status and function of mechanical and cosmetic
implants.
More importantly, when combined with a neutron-opaque imaging
agent, neutron imaging could be an excellent alternative to current
functional imaging modalities (i.e. Neutron Functional Imaging)...
Functional Neutron Imaging
We are investigating B-10 and Gd-157 labeled contrast
agents to provide high contrast between diseased and
normal tissue and with spatial resolutions orders of
magnitude better than PET, SPECT, or f-MRI.
• We call our technique Contrast Enhanced Neutron Imaging
(CENI)
• In addition to high spatial resolution in functional imaging, CENI
also could offer anatomical information
• CENI would not use any radioactive materials.
•
•
•
•
No need to inject radioactive materials into patient.
No reliance on foreign nuclear reactors for SPECT isotopes.
No need for on-site cyclotrons for PET isotopes.
Lower patient radiation exposures.
Functional Neutron Imaging
A Picture is worth a Thousand Words...
Depiction of how spatial resolution affects the ability to identify disease.
Here, an ischemia is located on the distal end of a mouse heart which is
represented at spatial resolutions of 30mm, 250mm, and 2mm.
NOTE: UML is in the process of upgrading its CCD radiography camera. The resolution of
the new camera system (so far) appears to be around 80 mm.
Problems with Biological Imaging
So, why haven’t we been using this !
Previous biological imaging research with neutrons found that the
image quality obtained when imaging biological organisms thicker than
~1 mm is extremely poor .
A Neutron also pose a greater biological risk to tissue relative to an x-ray
photon
• concern exists that biological neutron radiography could lead to
potentially high radiation exposures to the patient…
Ultimately, many of the interesting interaction mechanisms between
neutrons and tissue also cause difficulties when trying to obtain a good
biological image...
•
We will show that neutron energy selection is very important...
•
And, that the interaction mechanism with hydrogen causes problems
for biological imaging of large objects.
Neutron Energy Selection
Most neutron imaging systems use thermal
neutrons but, to image large biological objects,
we will require energetic neutrons.
Neutron Transmission (%)
100
0.025 eV Neutron Transmission
2 MeV Neutron Transmission
I am focusing my research
on small animal functional
imaging using thermal
neutrons.
Why?
1. Present tech. can’t
functionally image small
animals!
10
1
2. The mouse is the most
studied, characterized, and
cheapest bio-model to use
(also the quickest to obtain
results due to life-span).
0.1
0.01
0.001
0
5
10
15
20
25
Tissue Thickness (cm)
30
35
40
3. This also is the easiest way
to being implementing
CENI.
Problems with Biological Imaging
A) Neutron Radiograph of a Mouse cranium using a 16:1 cadmium coated bucky-grid.
B) X-ray radiography image of the same mouse cranium. The grid ratio used to form this image
was not specified but is assumed to be 16:1.
Let us examine closely the neutron interaction with hydrogen…
Image From: Brown and Parks, “Neutron Radiography in Biologic Media”, Am. J.
Roentg 106, 472 (1969)
Neutron Scatter with Hydrogen…
Comparison of X ray and Thermal Neutron
Interaction Probabilities...
Photon Cross Section
Neutron Cross Section
1e+5
1e+5
Photon Absorption Cross Section with Water
Photon Scatter Cross Section with Water
1e+4
1e+3
Cross Section (barns)
Cross Section (barns)
1e+3
1e+2
1e+1
1e+0
1e-1
1e-2
1e+2
1e+1
1e+0
1e-1
1e-2
1e-3
1e-3
1e-4
1e-4
1e-5
0.001
Neutron Absorption Cross Section with Hydrogen
Neutron Scatter Cross Section with Hydrogen
1e+4
1e-5
0.01
0.1
Photon Energy (MeV)
1
1e-8
1e-7
1e-6
1e-5
1e-4
1e-3
1e-2
1e-1
1e+0
1e+1
Neutron Energy (MeV)
While photoelectrons will not degrade image quality, scattered
neutrons will strongly degrade image quality…
Research Hypotheses
Hypothesis #1: The limitations observed with biological
neutron radiography predominately are due to the
inability to adequately remove hydrogen scattered
neutrons from image formation (anatomical imaging).
Hypothesis #2: That 10Boron and 157Gadolinium have
sufficiently high absorption cross-sections for use as a
non-radioactive contrast agent for functional neutron
imaging studies (functional imaging).
Scatter vs. Image Quality
We tested our first hypothesis through a series of basic
computer simulations using the MCNP5 radiation transport
program.
• MCNP5 is a software package developed by Los Alamos National
Laboratories.
• It is used to model coupled neutron / photon / electron transport
through material and their resulting interactions.
• Images were formed using the FIR Radiography fluence tally.
• We included S(,) transport corrections for the thermal motion
(vibration/rotation) of hydrogenous molecules.
Scatter vs. Image Quality
First: We performed a simple Monte Carlo contrast analysis
of four major tissues and materials common in medical
imaging to see if we could natively differentiate anatomical
tissues with neutron biological imaging (non-optimized).
muscle
bone
adipose
water
Scatter vs. Image Quality
• We analyzed muscle, cortical bone, adipose tissue, water,
and air contrast from thermal neutrons (0.025 eV) .
• Two images were obtained: one with neutron-scatter
contributions to image formation and one without neutron
scatter contributions.
• We used an image resolution of 1mm2
• Each material was modeled as being a rectangle with a
thickness of 1 cm.
• No special image processing was done on the resulting
image.
Scatter vs. Image Quality
RESULTS:
1
Scatter vs. Image Quality
Neutron Transmission
1000
Air
(~800%)
100
Bone
(104%)
10
Muscle
Tissue
(5.4%)
Adipose
Tissue
(2.1%)
1
-5
-4
-3
-2
-1
0
1
Water
2
3
4
5
position
Resulting contrast differences between muscle, bone,
adipose tissue, water, and air under conditions of complete
neutron scatter removal.
Scatter vs. Image Quality
Next: We simulated (again using MCNP5) a basic arm
phantom consisting of: muscle, bone, and an added bonefracture.
Representation of the Monte Carlo
simulated phantom.
Scatter vs. Image Quality
• We modeled muscle, bone, and a bone fracture:
• Muscle OD=5cm,
• bone OD=2cm,
• Fracture 1mm thick through bone center.
• Again, (non-optimized) images were simulated in MCNP
environment both with-and without image contributions
from scattered neutrons
• Again, used a receptor grid resolution of 1mm2 (not all
that important)
• Again, no special image processing was done on the
resulting image.
Scatter vs. Image Quality
A: Scatter Included
B: Scatter Removed
Comparison of images formed with scattered neutron contributions and without…
Scatter vs. Image Quality
Neutrons of various energies
Scatter removed…
Scatter vs. Image Quality
Neutrons of various energies
Scatter included…
Tissue Contrast: S/N
Most recently, we used MCNP to quantify the neutron fluence
needed to obtain a suitable signal to noise ratio in a small
animal model for functional and anatomical imaging.
Our goal is to quantify object tissue and tracer contrast (and,
to an extent the dose equivalence delivered per image).
We analyzed:
• Different anatomical tissues contrasts relative to water.
• Tumor cell contrast vs. their normal cell counterpart.
• The required boron-10 concentration in tissue needed to act as a
suitable contrast agent in functional imaging.
• The required gadolinium-157 concentration in tissue needed to act
as a suitable contrast agent in functional imaging.
Tissue Contrast: S/N
Note: the dose equivalence from neutrons of fluence 0 also is presented as a
benchmark. This value was obtained using the 10CFR20 flux to dose conversion
factors.
Tissue Contrast: S/N
For this experiment, we simulated a basic 3 cm thick phantom
in the MCNP code (similar to the thickness of a large mouse);
the Monte Carlo methodology used was similar to that used for
the other studies...
1. Various target tissues (and materials) were modeled in a
water phantom with tissue thicknesses varied between:
0.1mm & 10mm. These target tissues studied included:
a)
b)
c)
d)
e)
f)
Adipose
Muscle
Soft Tissue
Bone (cortical)
Water
air
Tissue Contrast: S/N
2. Also, tumor cell types were compared to their normal cell
counter-parts at various target thicknesses (0.1-10mm):
a)
b)
c)
d)
Carcinoma to soft tissue
Melanoma to skin / soft tissue
Sarcoma to muscle tissue
Squamous lung to normal lung tissue
3. Lastly, B-10 and Gd-157 concentrations were simulated
as contrast agents in a water phantom. Again, target
tissues were modeled between 0.1mm and 10 mm.
a) [B-10]: Concentrations of B-10 simulated were:
10, 50, 100, 300, 500, 700, 1000, 2000 mg / g.
b) [Gd-157]: Concentrations of Gd-157 simulated were:
1, 5, 10, 30, 50, 80, 100, 200 mg / g.
Side Note: What Can I “See”?
Background
(1 sigma)
S/N = 2,
S/N = 2
S/N = 5
S/N = 1
K=1 :
K=2:
K=3 :
K=4 :
K=5 :
Probabilities of exceeding K:
15.9% chance of background exceeding limit
2.3% chance of background exceeding limit
0.13% chance of background exceeding limit
3E-3% chance of background exceeding limit
3E-5% chance of background exceeding limit
1 * (background & signal)
2 * (background and signal)
3 * (background and signal)
Note: S/N = 5 required
for single pixel contrast
(Rose Criteria)
Tissue Contrast
RESULTS: Differentiation of Tissue vs. Water
Signal To Noise Ratio between Tissue and Water:
Flux = 5E6 n/cm2. HE = 0.05 mSv (5 mrem)
Target tissue thickness (mm)
tissue type
0.1
0.3
0.5
0.7
1
3
5
Adipose
Air
Bone
Muscle
Soft tissue
1.3
6.7
4.0
0.1
0.1
2.7
21.3
13.3
0.4
0.4
4.0
34
24
0.7
0.6
5.3
55
33
1.0
0.8
6.7
84
51
1.5
1.2
20
377
194
4.5
3.5
32
984
426
7.5
5.9
7
10
43
55
2248 7072
798 1808
10.6 15.3
8.2 11.6
Tissue Contrast
Signal To Noise Ratio between Tissue and Water:
Flux = 2E7 n/cm2. HE = 0.2 mSv (20 mrem)
Target tissue thickness (mm)
tissue type
0.1
0.3
0.5
0.7
1
3
Adipose
Air
Bone
Muscle
Soft tissue
2.7
13.3
8.0
0.3
0.2
5.3
42.6
26.7
0.9
0.7
8.0
74.6
48.0
1.5
1.2
11
109
67
2
2
13
168
101
3
2
40
754
389
9
7
5
7
10
64
85
117
1970 4496 14144
853 1596 3617
15
21
31
12
16
23
Tissue Contrast
Signal To Noise Ratio between Tissue and Water:
Flux = 1E8 n/cm2. HE = 1mSv (100 mrem)
Target tissue thickness (mm)
tissue type
0.1
0.3
0.5
0.7
1
Adipose
Air
Bone
Muscle
Soft tissue
6.0
29.8
17.9
0.7
0.5
11.9
95.4
59.6
2.0
1.6
17.9
166.9
107.3
3.3
2.7
24
244
149
5
4
30
375
226
7
5
3
5
7
10
89
143 191 261
1687 4404 10054 31627
870 1907 3570 8087
20
34
47
68
16
26
37
52
Tumor Contrast
RESULTS: Differentiation of Tumor vs. Normal Tissue
S/N - Tumor to Normal Cell Type: Flux = 5E6 n/cm2. HE = 0.05 mSv
Target tumor thickness (mm)
tumor type
0.1
0.3
0.5
0.7
1
3
5
7
10
Carcinoma/soft
Melanoma/soft
Sarcoma/muscle
Squamous/lung
0.8
0.8
0.1
0.3
2.3
2.3
0.3
0.9
3.9
3.9
0.6
1.5
5.5
5.5
0.8
2.2
8
8
1
3
25
25
3
9
43
43
6
16
63
63
8
23
95
95
11
33
Tumor Contrast
S/N - Tumor to Normal Cell Type: Flux = 1E8 n/cm2. HE = 1.0 mSv
Target tumor thickness (mm)
tumor type
0.1
0.3
0.5
0.7
1
3
5
7
10
Carcinoma/soft
Melanoma/soft
Sarcoma/muscle
Squamous /lung
3.5
3.5
0.5
1.4
10.4
10.4
1.6
4.1
17.5
17.5
2.6
6.9
24.6
24.6
3.6
9.7
35
35
5
14
110
110
15
42
192
192
26
72
280
280
36
102
427
427
50
149
Boron Contrast
RESULTS: Contrast from Various Boron-10 Concentrations
S/N for Boron-10 Concentration: Flux = 5E8 n/cm2, HE = 5.1mSv
[B-10]
(ug/g)
10
50
100
300
500
700
0.3
0.1
0.6
1.3
4.0
6.6
7.9
0.5
0.2
1.0
2.2
6.6
11.0
13.2
1000
2000
9.3
13.2
15.4
22.0
Borated water thicknesses (mm)
0.7
1
3
5
0
0
1
2
1
2
6
10
3
4
13
22
9
13
39
65
15
22
65
107
18
26
78
128
22
31
31
44
91
128
149
209
7
2
15
30
90
148
177
10
3
21
43
127
208
248
20
7
41
85
246
394
463
204
286
286
394
528
709
Yellow: B-10 concentration presently used in 10BPA BNCT
Red: B-10 boron-carbide nanoparticle concentration used by Mortensen for BNCT
Boron Contrast
S/N for Boron-10 Concentration: Flux = 1.5E9 n/cm2, HE = 15.3 mSv
[B-10]
(ug/g)
10
50
100
300
500
700
1000
2000
0.3
0.2
1.1
2.2
6.8
11.4
13.7
16.0
22.9
0.5
0.3
1.8
3.7
11.4
19.0
22.9
26.7
38.1
Borated water thicknesses (mm)
0.7
1
3
5
0
1
2
3
3
4
11
18
5
7
22
37
16
23
68
112
27
38
113
186
32
46
135
222
37
53
157
257
53
76
222
362
7
4
25
52
156
257
306
353
495
10
6
36
74
221
360
429
495
683
20
11
71
147
426
683
802
914
1228
Yellow: B-10 concentration presently used in 10BPA BNCT
Red: B-10 boron-carbide nanoparticle concentration used by Mortensen for BNCT
Gadolinium Contrast
RESULTS: Contrast from various Gd-157 concentrations
S/N for Gd-157 Concentration: Flux = 5E8 n/cm2, HE = 5 mSv
[Gd-157]
(ug/g)
1
5
10
30
50
80
100
200
0.3
0.0
0.3
0.5
1.7
2.8
4.5
5.7
11.4
0.5
0.0
0.4
0.9
2.8
4.7
7.5
9.5
18.9
Gd-water thicknesses (mm)
0.7
1
3
5
0.1
0
0
0
0.6
1
3
4
1.3
2
5
9
3.9
6
17
28
6.6
9
28
46
10.6
15
45
74
13.2
19
56
92
26.5
38
111
181
7
1
6
13
39
65
103
128
248
10
1
8
18
55
92
145
180
345
Yellow: Gadolinium concentration presently for Gd-DTPA MRI
20
7
41
85
246
394
463
528
709
Gadolinium Contrast
S/N for Gd-157 Concentration: Flux = 1.5E9 n/cm2, HE = 15.3 mSv
[Gd-157]
(ug/g)
1
5
10
30
50
80
100
200
0.3
0.0
0.4
0.9
2.9
4.9
7.9
9.8
19.7
0.5
0.1
0.7
1.5
4.9
8.2
13.1
16.4
32.8
Gd-water thicknesses (mm)
0.7
1
3
5
0.1
0
0
1
1.0
1
4
7
2.2
3
9
16
6.8
10
29
48
11.4
16
49
81
18.3
26
78
128
22.9
33
97
160
45.9
65
192
314
7
1
10
22
67
112
179
222
429
Yellow: Gadolinium concentration presently for Gd-DTPA MRI
10
1
15
31
96
159
252
312
597
20
11
71
147
426
683
802
914
1228
Dose Eqvt. for 0.3 mm contrast
Summary: Dose Required for 0.3mm slice
identification for various S/N Ratios (B-10)
B-10: Dose Equivalence (mSv) to obtain S/N Ratio for 0.3mm slice
concentration
(ug/g)
10
50
100
300
500
700
1000
2000
1
585
13
3
0.33
0.12
0.08
0.06
0.03
2
2339
52
12
1.3
0.47
0.32
0.24
0.12
S/N =
3
5262
117
27
2.9
1.05
0.73
0.54
0.26
4
9355
208
49
5.2
1.87
1.30
0.95
0.47
5
14617
324
76
8.2
2.92
2.03
1.49
0.73
Dose Eqvt. for 0.3 mm contrast
Summary: Dose Required for 0.3mm slice
identification for various S/N Ratios (Gd-157)
Gd-157: Dose Equivalence (mSv) to obtain S/N Ratio for 0.3mm slice
concentration
(ug/g)
1
5
10
30
50
80
100
200
1
7168
79
18
2
1
0.25
0.16
0.04
2
28673
318
72
7
3
0.99
0.63
0.16
S/N =
3
64515
715
161
16
6
2.23
1.42
0.35
4
114693
1271
287
29
10
3.97
2.53
0.63
5
179208
1986
448
45
16
6.20
3.96
0.99
Conclusions
1. With proper scatter-removal, neutrons can
obtain usable anatomical information in a
mouse model
a) This information would be complementary to that
obtained using x-ray, CT, and MRI.
b) All tissue types can be imaged with this technique and
with minimal radiation exposure to the mouse
c) CENI also has been shown to be able to natively
differentiate tumors from their normal cell counterpart!
Conclusions
2. More importantly, neutrons can be used with
B-10 or Gd-157 to obtain metabolic and
functional imaging information in a mouse
model:
a) Our technique would have spatial resolutions better than
10x that presently obtained with PET/SPECT and fMRI.
b) It would require quantities of B-10 and Gd-157 that are
biologically safe and in quantities used in present medical
procedures.
c) CENI can be used with any type of functional imaging
study presently performed.
d) From a safety perspective, CENI also is expected to
provide lower (or at least equivalent) radiation exposures
since no radioactive materials are used.
Acknowledgements
I wish to thank my collaborators:
Dr. Andrew Karellas
Director of Radiological Physics
Professor of Radiology
University of Massachusetts Medical School
Dr. Peter Gaines
Assistant Professor of Biology
University of Massachusetts Lowell
And my research assistants, who have or are working
on this project:
• Blake Currier
• Daniel Cutright, Ph.D.
Questions?