Comparison of image quality of different iodine
isotopes (I-123, I-124 and I-131)
Erwann Rault, Stefaan Vandenberghe, Roel Van Holen, Jan De Beenhouwer,
Steven Staelens, and Ignace Lemahieu
MEDISIP, ELIS department, Ghent University, Sint-Pietersnieuwstraat 41, B-9000, Ghent, Belgium
Abstract
pretherapeutic patient-specific dosimetry. The latter is usually performed by directly imaging the uptake of the tracer
in the body. This can be done either by imaging the radioactive tracer used for therapy or by replacing it by one of its
isotopes, more suitable for imaging.
Due to its beta- emission (606 keV, 90%), I-131 is often
used for therapy. It also emits some photons that can be
used for SPECT imaging. Their main emission energy peak
is about 364 keV and requires the use of a HEAP collimator. However, its spectrum of emission is complex and some
emission peaks, above 364 keV, cause high energy contamination [1] and degrade the dosimetry.
I-123 is more suitable for imaging. The energy of its main
gamma emission peak is 159 keV which is close to the 140
keV from Tc-99m for which gamma-camera design has traditionally been optimized. I-123 can be imaged in a SPECT
system with a LEHR (Low Energy High Resolution) collimator, optimized for the Tc-99m (140 keV), or with a ME
(Medium Energy) collimator, optimized for energies up to
300 keV. However, this radionuclide also emits peaks of a
higher energy which again can be responsible for high energy contamination.
Due to its emission of positrons (23% of the decays), I-124
is used in PET imaging. PET systems are using an electronic collimation based on interaction time. The emission
spectrum of I-124 is also very complex [2]. In 50% of the
cases, the emission of a positron is followed by the emission
of a gamma of 602 keV, which is likely to create a false coincidence (similar to scattered coincidences) [3].
The aim of this study was to evaluate the image quality and
quantify the impact of high energy contamination for each
isotope. High energy contamination [4, 5] ends up in the
main energy window by one or by a combination of the following effects: scatter in the object, Compton effect in the
crystal, penetration of the collimator, and backscatter from
the camera end parts (light guides, PMTs, lead ending).
I-131 is a frequently used isotope for radionuclide therapy.
This technique for cancer treatment requires a pretherapeutic dosimetric study. The latter is usually performed (for this
radionuclide) by directly imaging the uptake of the therapeutic radionuclide in the body or by replacing it by one of
its isotopes, more suitable for imaging. This study proposes
to compare the image quality that can be achieved by three
iodine isotopes: I-131 and I-123 for Single Photon Emission Computed Tomography (SPECT) imaging, and I-124
for Positron Emission Tomography (PET) imaging.
The imaging characteristics of each isotope were investigated by simulated data. Their spectrums, point spread
functions and contrast recovery curves were drawn and
compared. I-131 was imaged with a HEAP (High Energy
All Purpose) collimator while two collimators were compared for I-123: LEHR (Low Energy High Resolution) and
ME (Medium Energy). No mechanical collimation was used
for I-124. The influence of small high energy peaks (>
0.1%) on the main energy window contamination were evaluated. Furthermore, the effect of a scattering medium was
investigated and the TEW (Triple Energy Window) correction was used for spectral-based scatter correction.
Results show that I-123 gives best results with a LEHR collimator when the scatter correction is applied. Without correction, the ME collimator reduces the effects of high energy
contamination. I-131 offers the worst results. This can be
explained by the large amount of septal penetration from
the photopeak and by the collimator which gives a low spatial resolution. I-124 gives the best imaging properties due
to its electronic collimation (high sensitivity) and a short
coincidence time window.
1. Introduction
Radionuclide Therapy is a promising technique for cancer
treatment. However, this technique requires an accurate
1
2. Materials and Methods
2.1. Simulations and Reconstruction
These studies were performed on two simulated systems,
the Philips Axis (SPECT) [6] for I-123 and I-131 and the
Philips Allegro (PET) [7] for I-124. Two types of collimators were compared for I-123, as it is mainly imaged
with a LEHR (Low Energy High Resolution) collimator or
a ME (Medium Energy) collimator. I-131 was studied with
a HEAP (High Energy All Purpose) collimator.
The GATE (GEANT4 Application for Tomographic Emission) [8] package was used for the simulations. It is a
Monte Carlo simulation tool well validated, and its use of
the GEANT4 libraries offers a high accuracy. Gate models
of the aforementioned scanners were extensively validated
[6, 9].
The simulated data were reconstructed with the MLEM algorithm [10, 11]. The number of iterations was chosen to
allow the better contrast with a reasonable level of noise for
each of the isotopes (around 100 iterations in each case).
The images were reconstructed in 2 dimensions and their
matrix size was chosen to 256*256 pixels.
(a) I-123
(b) I-124
2.2. Spectrum analysis
This study aims to point out the proportion of contamination
for each of the three isotopes. Those characteristics will
help explaining the imaging possibilities of each isotope.
The detected spectrum of each isotope (without collimator)
is shown on figure 1. All the peaks representing more than
0.1% of the emission and that can be detected in the main
energy window, have been investigated. To do so, a point
source has been placed at 15 cm from the collimator. In order to better differentiate the history of detected photons,
the sources have been simulated by their discrete energy
peaks. For I-123, six peaks have been simulated (159 keV:
83%, 346 keV: 0.13%, 440 keV: 0.42%, 505 keV: 0.32%,
529 keV: 1.39%, 539 keV: 0.38%). Six peaks have also been
simulated for I-131 (364 keV: 82%, 326 keV: 0.27%, 503
keV: 0.36%, 637 keV: 7.18%, 643 keV: 0.22%, 723 keV:
1.77%). For I-124, the source can not be decomposed in
discrete energy peaks without changing the results because
of the time coincidence between emission of positrons and
gammas. This relation is important as it can create false coincidences. For this isotope, proportion of scatter and high
energy contamination was quantified as a whole.
For each of those isotopes, the efficiency of detection was
calculated. Then the proportion of high energy contamination was determined and finally, the peaks responsible for
that contamination were pointed out for the SPECT studies. The effect of a scattering medium (cylinder of water
of 10cm radius and 18cm height centred in the source) was
(c) I-131
Figure 1: Various spectras of photons (singles) on the detector without
the use of a collimator. The energies shown are the energies of the main
emission peaks of the isotopes. The energy of the main peak used for the
imaging is in a red box.
additionally investigated.
2.3. Resolution study
The PSF (Point Spread Functions) was studied for each isotope. A point source of 0.1 mm diameter, at the centre of
the field of view, was acquired with the different systems
and the sinograms were constructed. The radius of rotation
of the camera was set to 15 cm. As the point source was
placed in the centre of the scanner, the PSF was defined by
the sum of all the angles of the sinograms.
In order to quantify the resolution, the FWHM (Full Width
at Half Maximum) and the FWTM (Full Width at Tenth
Maximum) have been computed on the PSF. Those two parameters have been compared with or without a scattering
2
medium (cfr. section 2.2).
The influence of the TEW correction [12] was also investigated. This is a scatter correction method based on the
estimation of the scatter in the main energy window by interpolating it from two adjacent energy windows. In this
study, the main energy window was centered on the main
energy peak of the isotope with a width of 20% of this energy (I-123: 143-175 keV, and I-131: 328-400 keV). The
two other windows were adjacent to the main window with
a width of 10% of the main energy (I-123: 135-143 keV and
175-183 keV, and I-131: 291- 328keV and 400-437 keV).
I-131 and 155 times than the one with I-123 on a ME collimator. Figure 3 points out the influence of each high energy peak on the photopeak contamination. I-123 imaged
with a LEHR collimator suffers from 49.1% of high energy
contamination, compared to 7.8% with a ME collimator if
no scattering medium is present. Those results are higher
(LEHR: 63.9% and ME: 19.9%) with a scattering medium.
The high energy contamination represents 35.9% for I-124
and 27.4% for I-131 with a scattering medium.
2.4. Contrast Recovery Curves
A 2D source, composed of a circular background of 17 cm
diameter and eight hotspots of diameters from 8 mm to 20
mm, was designed to study the contrast evolution with the
size of the objects. The contrast between the hotspots and
the background was set to 4:1. The effect of a scattering
medium (cfr: section 2.2) was investigated.
The contrast is defined as the ratio between the mean value
in the object and the mean value in the background. The
calculated contrast recovery factor is then the ratio between
the contrast that has been found on the images and the contrast of the source. This parameter gives an estimation of
the accuracy of quantification that can be achieved with the
studied isotope without applying any kind of partial volume
correction.
(a) I-123 LEHR
3. Results
(b) I-123 ME
3.1. Spectrum analysis
(c) I-131 HEAP
Figure 3: History of the photons reaching the detector for I-123 with
Figure 2: Efficiency of detection achieved with the isotopes. The grey
a LEHR collimator (a), I-123 with a ME collimator (b) and I-131 with
a HEAP collimator (c). Penetration photons are the photons which penetrate the collimator and get detected in the crystal (back scatter is not
considered as scatter). Scatter photons are the ones which scatter in the
collimator. Finally the geometric ones are the photons detected on the
detector without any scatter nor penetration.
part of the plots corresponds to the scatter and penetration photons.
The efficiency of detection for all of these isotopes is shown
in figure 2. When only the geometric photons are considered, the ME collimator offers a better efficiency (1.4 times
the efficiency of the LEHR collimator) for I-123. Due to its
electronic collimation, the efficiency that can be achieved
with the I-124 is 96 times higher than the one achieved with
Results for I-123 show that peaks of 0.4% of the emission
can be responsible for up to 7.9% (539 keV) of contamination with a LEHR and without a scattering medium. With
3
a scattering medium, this proportion of contamination goes
up to 10.4%. The main peak responsible for contamination
for I-123 is the 529 keV peak which represents 1.39% of
the emission (LEHR: 28% and ME:4.8% without a scattering medium). For I-131, the main peak responsible for high
energy contamination is the peak of 637 keV (12.5%).
Figure 3 also shows for each emission peak, the proportion
of penetration (photons which penetrates the collimator and
that did not scatter in the collimator nor in the object), scatter (photons that scatter in the collimator or in the object)
and geometric photons detected in the crystal. Although the
amount of detected photons from the main energy peak is
high for I-131, it is important to consider that a large amount
from this peak is detected after septal penetration. This fraction is very important compared to the two other isotopes:
52% of the photons from the main energy peak are penetration photons compared to 18% for I-123 with a LEHR
collimator and 3% for I-123 with a ME collimator.
Figure 5: Comparison of the FWHM and the FWTM for each of those
isotopes (with or without attenuation matter and with or without TEW correction). FWHM and FWTM are respectively written (in mm) under the
name of the isotopes
3.2. Resolution study
Figure 6: Images of the contrast study sources for all of the isotopes.
Figure 4: Point Spread Functions (in mm, without a scattering medium).
3.3. Contrast Recovery Curves
Reconstructed images used for the contrast study are displayed on figure 6. As shown on figure 7, the best contrast recovery has been found for I-124. It reaches 0.935
for objects of 20mm diameter. I-123 with a ME collimator
reaches better results than with a LEHR collimator when no
TEW correction is applied (ME: 0.634 and LEHR: 0.617
for 20mm). However, the results are the contrary when the
TEW correction is applied (ME: 0.653 and LEHR: 0.669 for
20mm). I-131 imaged with HEAP collimator shows very
low results (0.363 for 20mm).
The scattering medium lowers those results for I-123 (ME:
0.634 to 0.559) and I-124 (0.935 to 0.746). I-131 does not
suffer from the effects of the scattering medium (0.363 to
The PSF obtained for all of the isotopes, without a scattering medium, are shown in figure 4. I-124 offers a better
resolution than the two other isotopes. The curve of I-123
with the ME collimator shows the effects of the collimator
(septas) which lowers the resolution. The use of a LEHR
collimator is therefore better for good spatial resolution. For
I-131, the septas of the collimator are clearly visible on the
PSF and the resolution obtained is very poor.
Results for both FWHM and FWTM are shown on figure 5.
It shows that, for each isotope, the scattering medium is increasing the FWTM while the FWHM remains roughly the
same. The TEW correction helps lowering the first effect.
4
emission peak. This result and the one obtained for its spatial resolution explained the poor results for this isotope.
I-124 offers the best image quality due to its electronic
collimation which gives a better efficiency and resolution.
The short coincidence window also limits influence of high
energy peaks. Given those results, I-124 seems to be
the more promising isotope for a pretherapeutic dosimetry
study. Such a methodology has already been developed by
Sgouros for I-131 cancer therapy [13].
5. Conclusion
Figure 7: Contrast Recovery obtained for all isotopes without a scattering medium as a function of the hotspot size (in mm).
For imaging I-123, the use of a LEHR collimator is important when a good resolution is required. However, a ME
collimator reduces the influence of the high energy peaks.
For I-123 and for contrast, it is better to use a ME collimator when no scatter correction is applied. With TEW correction, it is then better to use a LEHR collimator because
of the better spatial resolution.
Due to its electronic collimation which gives a better efficiency and resolution than in SPECT, I-124 offers the best
image quality. The short coincidence window also limits
influence of high energy peaks.
Due to its relatively high energy imaging peak, the use of
a HEAP collimator is required to image I-131. The image
quality that can be achieved with this collimator is much
lower than the one that has been achieved with the I-123
or the I-124. This result can be explained by the important
amount of septal penetration from the main energy peak.
0.361).
4. Discussion
This study was performed with an accurate Monte Carlo
simulator allowing us to store history of the detected events
which is not possible with measured data. The number of
simulated photons varied from 50000 detected photons in
the photopeak window for the spectrum analysis to 1000000
for the imaging characteristics definitions. Those values
were sufficient to obtain good statistics over the distribution
of the energies and over the spatial distribution. However, to
compare the imaging characteristics of the isotopes (spatial
resolution and contrast recovery), the same source activity
and equal acquisition time have been used. This aims to
give comparable results between the isotopes.
This study shows that, for SPECT, even small peaks of
emission (> 0,1%) in the spectrum of those isotopes can
have a high influence and degrade significantly the image
quality. High energy contamination is mainly due to septal
penetration in the collimator followed by Compton detection in the crystal and backscatter from the endparts. For
those peaks, of a higher energy than the photopeak, scatter
in the collimator or in the scattering medium has a lower
influence.
By looking at the spectrums of the I-123, one can clearly
see that using a ME collimator helps lowering the influence
of high energy contamination. However, contrast is also
directly related to resolution. The results obtained for the
contrast recovery can then be explained by those found for
the resolution. I-123 is best imaged with a LEHR collimator when the TEW correction is applied while the use of
a ME collimator is required when no scatter correction is
made. Those results are a good approximation of what can
be achieve for quantification by these isotopes.
Even if I-131 does not suffer a lot from high energy contamination, it suffers from a lot of penetration from the main
References
[1] Y. K. Dewaraja, M. Ljungberg, and K. F. Koral.
Accuracy of 131I tumor quantification in radioimmunotherapy using SPECT imaging with an ultrahigh-energy collimator: Monte Carlo study. J Nucl
Med, 41(10):1760–1767, Oct 2000. non lu.
[2] Allan C. G. Mitchell, Jose O. Juliano, Charles B. Creager, and C. W. Kocher. Disintegration of I124 and
I123. Phys. Rev., 113:628–633, 1959.
[3] Stefaan Vandenberghe. Three-dimensional positron
emission tomography imaging with 124i and 86y.
Nucl Med Commun, 27(3):237–245, Mar 2006.
[4] Steven Staelens, Tim de Wit, Ignace Lemahieu, and
Freek Beekman. Significant septal penetration effects
caused by impurities in cyclotron-produced thallium201. J. Nucl. Med., 47:381, 2006.
[5] A. Cot, E. Jane, J. Sempau, C. Falcon, S. Bullich,
J. Pavia, F. Calvino, and D. Ros. Modeling of high5
energy contamination in SPECT imaging using Monte
Carlo simulation. IEEE TRANSACTIONS ON NUCLEAR SCIENCE, 53(11):198–203, FEB 2006.
[6] Steven Staelens, Daniel Strul, Giovanni Santin, Stefaan Vandenberghe, Michel Koole, Yves D’Asseler,
Ignace Lemahieu, and Rik Van de Walle. Monte carlo
simulations of a scintillation camera using gate: validation and application modelling. Phys Med Biol,
48(18):3021–3042, Sep 2003.
[7] Suleman Surti and Joel S Karp. Imaging characteristics of a 3-dimensional gso whole-body pet camera. J
Nucl Med, 45(6):1040–1049, Jun 2004.
[8] S. Jan, G. Santin, D. Strul, and S. Staelens et al.
GATE: a simulation toolkit for PET and SPECT. Phys
Med Biol, 49(19):4543–4561, Oct 2004.
[9] F. Lamare, A. Turzo, Y. Bizais, C. Cheze Le Rest, and
D. Visvikis. Validation of a monte carlo simulation
of the philips allegro/gemini pet systems using gate.
Phys Med Biol, 51(4):943–962, Feb 2006.
[10] L. A. Shepp and Y. Vardi. Maximum likelihood reconstruction for emission tomography. MI-1(2):113–122,
1982.
[11] K. Lange and R. Carson. EM reconstruction algorithms for emission and transmission tomography. J.
Comput. Assist. Tomogr., 8(2):306–316, 1984.
[12] K. Ogawa, Y. Harata, and T. Ichihara. A practical
method for position dependent scatter correction in
single photon emission ct. IEEE Trans. Med. Imag,
10(3):408–412, 1991.
[13] George Sgouros, Katherine S Kolbert, Arif Sheikh,
Keith S Pentlow, Edward F Mun, Axel Barth,
Richard J Robbins, and Steven M Larson. Patientspecific dosimetry for 131I thyroid cancer therapy using 124I PET and 3-dimensional-internal dosimetry
(3D-ID) software. J Nucl Med, 45(8):1366–1372, Aug
2004.
6