Intrinsic Radiation in Lutetium Based PET Detector: Advantages and

Disadvantages
Qingyang Wei ( 魏清阳) 1,2

1. Department of Electrical Engineering, Tsinghua University, Beijing 100084, China

2. Key Laboratory of Particle & Radiation Imaging, Ministry of Education, Beijing, 100084, China

Abstract ：Lutetium (Lu) based scintillators such as LSO and LYSO, are widely used in modern PET
detectors due to their high stopping power for 511 keV gamma rays, high light yield and short decay time.

However, 2.6% of naturally occurring Lu is 176Lu, a long-lived radioactive element including a beta decay

and three major simultaneous gamma decays. This phenomenon introduces random events to PET systems

that affects the system performance. On the other hand, the advantages of intrinsic radiation of 176Lu (IRL)

continues to be exploited. In this paper, research literatures about IRL in PET detectors are reviewed. Details

about the adverse effects of IRL to PET and their solutions, as well as the useful applications are presented

and discussed.

：
Key words PET; Lutetium based scintillator; intrinsic radiation; influence

PACS: 87.57.uk, 29.40.Mc

1 PET and PET detector

Positron emission tomography (PET) is a nuclear medicine and functional imaging technique [1]. The

system detects pairs of 511keV gamma rays emitted indirectly by positron-emitting radionuclides (11C, 13N,
15 18
O, F, etc.), which are labeled on biologically active molecules and introduced into the bodies. Three-

dimensional images of tracer concentration within the bodies are then mathematically reconstructed. In

clinical examinations, PET is a powerful tool for early diagnosing cancers, heart diseases and brain diseases,

et al [1]. And PET is also applied in pre-clinical small animal imaging for drug development, pathological

studies and gene expression studies, et al [3].

PET machines are mainly built from scintillator detectors consisting of inorganic scintillators and optical

photon detectors (PMT, APD or SiPM)[4]. The gamma rays deposit energy in the scintillators and generate

optical photons, which are detected by the optical photon detectors. The scintillator of the early PET detector

is high light yield NaI(Tl) [6]. Whereas, it has disadvantages for PET including low stopping power, long

The paper has been submitted to Chinese Physics C

Supported by China Postdoctoral Science Foundation (2014M550745), National Natural Science Foundation of China (No.
11375096), National Natural Science Foundation of China (11275105), and Tsinghua University Initiative Scientific Research
Program (20131089289).
Email: weiqingyang@tsinghua.edu.cn
decay time and hydroscopic property. In 1970s, BGO with high stopping power was discovered and

substituted NaI(Tl) for PET. However, BGO is still not ideal for PET due to its low light yield and long decay

time [5]. In the late 1980s, lutetium oxyorthosilicate (LSO) were discovered by Melcher et al [7] and then

other lutetium based scintillators (LBS) such as LYSO and LFS were also found. With properties of high

stopping power, high light yield and fast decay time, LBS is quite suitable for PET especially time of flight

PET (TOF-PET)[8], thus LBS becomes the best choice of modern PET detectors. The parameters of some

scintillators used in PET detectors are listed in table1.

Table 1 Properties of some scintillators used in PET detectors. Note that the parameters have slight variation due to the change
of the dopants in the scintillator growth [5].

NaI(Tl) BGO LSO LYSO LFS

Linear attenuation coeff.(cm-1) 0.34 0.92 0.87 0.86 0.82

Light yield (%NaI(Tl) ) 100 15 75 80 77

Decay constant (ns) 230 300 40 41 35

2 Intrinsic radiation of Lutetium-176 (IRL)

The naturally occurring Lutetium consists a fraction of 176Lu (abundance 2.6%), a long-lived radioactive

element with a half-life (T1/2) of ∼3.6×1010 year [9]. Figure 1 shows the decay scheme of 176Lu.The main

process of 176Lu 176Hf decay is a β-particle emission (maximum energy 596 keV) and three simultaneous

γ-ray emissions (energies: 88, 202, and 307 keV). The count rate (R) of a milliliter scintillator can be

calculated with equation 1:

nρ N A nρ N A 0.693
R= × 2.6% × (1 − e −0.693/T1/2 ) ≈ × 2.6% × , (1)
M M T1/2

Where, n is the subscript number of Lu in the scintillator molecular formula, ρ is the density, NA is the

Avogadro constant (6.02×1023), and M is the molar mass.

Take LSO (Lu2SiO5) for example, calculated as equation 2, RLSO equals to 307 Bq/ml.

2 × 7.4 g / ml × 6.02 × 10 23 0.693

RLSO = × 2.6% × 10
= 307 Bq / ml , (2)
460 3.6 × 10 × 365 × 24 × 3600
 Figure 1 Decay scheme of 176Lu[10]

In a PET detector, the β particles emitted from the LBS array, given the short range, mostly deposit

energies in the same LBS array. While, γ-rays can be detected not only in the same LBS array, but also in

other LBS arrays by emitting outside the first LBS array. Therefore, there are different types of events

including single photon events, double photon events and multiple photon events. The decay particles of a
176
Lu element only detected in a same LBS array is a single photon event (SPE) (Fig. 2(a)). An escaped γ-

particle detected by another LBS array is a true intrinsic coincidence event (TICE) (Fig. 2(b)). Two single

events occurring in a coincidence time window (CTW) is a random intrinsic coincidence event (RICE) (Fig.

2 (c)). More than one γ-particle escaping and detected by different LBS arrays is a multiple photon event. A true

intrinsic coincidence event and a single event generated in a CTW is also identified as a multiple photon

event (Fig2.d).

(a) (b) (c) (d)

Figure 2 Several typical IRL events in a PET system. (a) a single-photon event, (b) a double-photon event (true intrinsic

coincidence), (c) a double-photon event (random intrinsic coincidence), (d) a multiple-photon event.

The count rate of SPE is proportional to the amount of the scintillator being used. The count rate of TICE is

determined by the mount of the scintillator, the size of crystal and the geometry of the system. And the count rate

of RICE is proportional to the square of the mount of scintillator and the value of CTW.

Figure 3 shows the GATE-simulated [11] spectrums of IRL in a small animal PET (InliView 3000 [13]). As
shown in Fig. 3(a), the energy spectrum of the SPE in PET detector is a broad beta spectrum distorted by discrete

shifts in energy caused by the simultaneous detection of one, two or three of the gamma photons. The count rate

of TICE is ~5 kcps. Fig. 3(b) shows the energy spectrum of the TICE without TOF. The peaks of 202 and 307 keV

are more obvious than the peaks in SPE. Fig. 3(c) shows the energy spectrum of the TICE with TOF. The first

trigger and the second trigger can be distinguished where the second triggers are the escaping gamma photons.

Thus the photopeaks of 202 and 308 keV are easiest distinguished.

(a) (b) (c)

Figure 3 The simulated energy spectrums of Lu background radiation in a small animal PET (InliView 3000) , energy resolution

20%@511keV. (a) the spectrum of SPE. (b) the spectrum of the TICE without TOF information. (c) the spectrum of the TICE

with TOF.

3 The drawbacks of IRL in PET

The effects IRL in PET imaging have been well documented. It is usually not a big concern in most in-

vivo situations. However, IRL becomes problematic for some special studies such as low activity imaging,

long axial FOV PET, and PET/SPECT imaging.

3.1 common PET imaging

The IRL affects the system performance by generating additional prompt and random coincidences.

Because the SPE energy spectrum of the 176Lu is a wider distribution covering the photopeak of 511keV, the

intrinsic random coincidences could not be prevented completely. Although, the intrinsic random

coincidences can be corrected by measuring the delayed coincidence or using some other correction methods

[14], they alters the timing performance of the system prolonging the dead time of the detectors [15]. The

intrinsic true coincidences will introduce blank noise to the reconstructed image and the scatter fraction will

be overestimated with NEMA protocol without considering the Lu intrinsic true coincidences [16].

The count rates of intrinsic random coincidences and intrinsic true coincidences are both depended on

the energy window. Usually, the energy window is recommended as 350-650keV in a Lu-based clinical PET

system. Under this energy window, most intrinsic true coincidence events can be prevented because most of
the prompt gamma of Lu background radiation is 202 and 307 keV which below 350 keV. Moreover, the Lu

background radiation is usually very low relative to the injected activity in clinical applications [19], therefore,

IRL is not sufficient to affect routine clinical PET scanning.

Compared to the clinical systems, preclinical small animal PETs have much lower injected activities,

poorer energy resolution due to the small crystal and less scatter effect in small animals. To improve the

sensitivity, the energy window is usually recommended as 250keV-750 keV. Thus, the impact of IRL on the

performance of small animal PET systems must not be disregarded [19].

The effects of IRL on PET systems will be further reduced, but not eliminated entirely, as newer

generations of detectors and electronics permit the use of more narrow energy acceptance windows and

coincidence time windows.

3.2 low activity PET imaging

There are some applications of PET imaging done with a low activity level (the injection activity below

1 kBq for some situations), for instance, cell-trafficking studies [20], gene expression imaging [21] or in-

beam PET imaging [22]. In these cases, the IRL may seriously affect the imaging performance with the

conventional energy window [23]. The minimum detectable activity value is increased due to the Lu

background radiation. An available method to reduce the number of Lu intrinsic random coincidence events

is using a narrow energy window and the TOF information. Therefore, it requires the Lu based system with

a good energy resolution and a good TOF information. In 2014, Yoshida [23] proposed a novel method by

combing TOF and multiple coincidence information to further reduce the effects of IRL.

3.3 Long axial FOV PET

The axial field of view (AFOV) of current clinical whole-body PET scanners range from 15–22 cm

which have limited sensitivity. Recently, a long AFOV (2m) PET was proposed [25] which could image the

whole-body at one time and would significantly increase the sensitivity. It would open the door to new

applications, such as whole-body parametric imaging of pharmacological kinetics and systemic imaging of

radiolabeled stem cell progenitor cell populations [25]. The single event count rate of the IRL is proportional

to the amount of the scintillator being used that means it will 10 times of the common PET system in a 2m

long AFOV PET. Thus, there will be very high random events in the long AFOV PET. To reduce the impact

of random events, variable coincidence time windows for LORs should be employed [25].

3.4 PET/SPECT imaging

Animal PET scanners with inserted collimators were developed to do animal SPECT imaging [26]. Due
to the low efficiency of SPECT imaging and the energy range of SPECT imaging overlapping with IRL energy

spectrum, IRL may have serious impacts on SPECT imaging. The solution method is subtracting the

contribution of Lu background by doing a long time of background scan [27]. The pattern of Lu background

radiation can be reduced with the correction. However, the minimal detectable target activity will be higher

for the Lu-based SPECT system as compared to typical SPECT systems which do not have crystal intrinsic

activity [27].

4 The benefits of IRL in PET

Everything has two sides. The IRL is an interest spot for researchers and has been utilized in PET in

many directions.

4.1 For detector module developing

The IRL can be used as an approximate flood source to assist PET detector module design instead of

using additional sources. The background radiation is fast enough to generate a flood histogram of a block

detector in several minutes [28]. It can help to quickly test the block design of crystal surface processing and

reflect schemes. In additional, the Lu-based scintillator array can also be used as a multiple gamma-rays

source for other detectors to do energy calibration or energy nonlinear response measurement.

The IRL can also help the study of depth of interaction PET detector (DOI-PET). For example, the

scintillators' internal Lu-176 beta decay was applied to measure DOI functions for a dual end readout DOI -

PET [29].

4.2 For PET quality assurance (QA)

PET requires a common QA procedure to ensure the system working with a good condition. The IRL is

able to do a fast test for daily or weekly QA. Fig. 4 shows the 2Dmaps of the Lu coincidence event, where

Fig. 4(a) is a relative normal one, while Fig. 4(b) shows that a sector of the system has some problems which

need to be fixed.

(a)
 (b)

Figure 4 2Dmaps of the Lu background coincidence events of the InliView-3000 small PET system. (a) A relative normal

2Dmap, some blocks have bad performance and (b) is an un-normal 2Dmap indicated that a problem occurred in the eighth

sector.

4.3 For monitoring PET channel gain drift

It is necessary to do energy calibration for PET detectors in order to identify the energy of each detector

unit. Moreover, the calibration should be repeated over time due to the condition variation of the photon

detectors, such as the PMT aging [30] and the temperature changing of APD or SiPM. The IRL was applied

as a system self-calibration means to monitor the random or systematic changes or drift in the system [30].

The method consists of tracking the position of the energy peak 597 keV (sum of all three gamma rays), it

was proved that the Lu based method obtained a calibration functionally equivalent to the standard calibration

method with external 511-keV sources.

4.4 For Time Alignment of TOF-PET

Time alignment of a TOF-PET system is an important calibration procedure as the nature of the data

collected in PET is inherently with variable time shifts. There are many methods to obtain the time corrections

for the crystals in the PET scanner [31]. All of which require a radioactive source, except the method

proposed by Rothfuss which is based on IRL [32]. The time of flight is measured between the two crystals

involved in a Lu true intrinsic coincidence event. A time of flight can also be calculated between any two

crystals with positional information acquired from the fixed geometry of PET scanner. The time correction

for the crystal pair can then be equated by taking the measured time difference and subtracting the geometrical

calculated time of flight. It was found the timing alignment performed with IRL resulted in performance as

good as the method using a uniform gamma source phantom [32].

4.5 For generating transmission image

Lu intrinsic true coincidence events are feasibility to perform as a transmission source to generate
transmission image of the scan object in PET. Furthermore, in a TOF-PET system, with information of the

TOF and the chord length between two crystals in coincidence, a second time window can be set to observe

transmission events simultaneously to positron emission events. Thus, the transmission and emission image

can be reconstructed simultaneously. It is observed that the flux of the background activity is high enough to

create useful transmission images with acquisition time of 10 min [33]. This method can provide an

attenuation solution to the system that do not have or require a coupled CT scanner, such as PET-MR

machines or dedicated PET machines with specific applications such as brain imaging. The transmission

image can also be employed to perform scatter corrections or starting images for algorithms that estimate

emission and attenuation simultaneously [33].

4.6 For system geometrical calibration

The transmission image of Lu intrinsic coincidence events could also be applied for system geometrical

calibration. We developed a small animal PET/SPECT/CT system (InliView 3000) consisting of a LYSO

based PET scanner, a collimator SPECT by sharing the PET scanner and a cone-beam CT based on CMOS

detector. The IRL was utilized to image the collimator to do geometrical calibration. Fig. 5 shows the

transmission image of the collimator. Moreover, we designed a multi-tungsten-alloy-ball phantom to be

imaged by IRL and CT scanner. The two types of transmission images are used to generate the transformation

matrix between PET and CT system. Our proposed method is expected to replace the traditional multi-point

or multi-line gamma sources which make the geometrical calibration more convenient.

Figure 5 The inserted multi-pinhole collimator in the InliView-3000 (left), the trasaxial, views of the collimator reconstructed

image using IRL data (right)[34]

5 Conclusions
Lutetium based scintillators with good properties are the first choice of PET detectors. The IRL affects

the performance of the PET systems. With a narrow energy window and a narrow coincidence time window,

the IRL can be ignored for conventional clinical PET studies while it cannot be disregarded for small animal

PETs. For some special systems and applications, such as low activity imaging, long axial FOV PET and

PET detector based SPECT imaging, the impact of IRL will be significant.
 On the other hand, IRL becomes a powerful tool. The existing of IRL can be utilized to assist detector

design, to do system quality assurance, to monitor channel gain drift, to perform time alignment of TOF-PET,

to generate transmission image and to do geometrical calibration, et al.

References ：
[1] Muehllehner, Gerd, and Joel S. Karp. Positron emission tomography. Phy. Med. Biol., 2006, 51.13: R117.
[2] Townsend, D. W., Carney, J. P., Yap, J. T., et al. (2004). PET/CT today and tomorrow. J. Nucl. Med., 45(1 suppl), 4S-
14S.
[3] YAO Ru-Tao, Roger Lecomte, Elpida S. Crawford. Small-animal PET: what is it, and why do we need it?. J. Nucl.
Med. Tech., 2012, 40.3: 157-165
[4] WEI Qing-Yang, WANG Shi, MA Tian-Yu, et al. Influence factors of two dimensional position map on photomultiplier
detector block designed by quadrant sharing technique. Nucl. Sci. Tech.. 2011, 22, 224-229.
[5] Lewellen T K. Recent developments in PET detector technology[J]. Phy. Med. Biol., 2008, 53(17): R287-R317.
[6] Terpogossian M M, Phelps M E, Hoffman E J, et al. A positron-emission transaxial tomograph for nuclear imaging
(PETT) [J]. Radiology. 1975, 114(1): 89-98.
[7] Melcher, C. L., and J. S. Schweitzer. Cerium-doped lutetium oxyorthosilicate: a fast, efficient new scintillator. IEEE T.
Nucl. Sci., 1992, 39.4: 502-505.
[8] Moses W W. Time of flight in PET revisited[J]. IEEE T. Nucl. Sci., 2003, 50(52): 1325-1330.
[9] Kossert K, Joerg G, Gostomski C L V. Experimental half-life determination of Lu-176[J]. Applied Radiation and
Isotopes. 2013, 81: 140-145.
[10] WEI Qing-Yang. Performance analysis and decoding optimization of PET detector modules (Ph. D. Thesis). Beijing,
Tsinghua University, 2013. (in Chinese)
[11] Jan, S., G. Santin, D. Strul, et al. GATE: a simulation toolkit for PET and SPECT. Phy. Med. Biol., 2014 49(19): 4543.
[12] Mcintosh B, Stout D B, Goertzen A L. Validation of a GATE model of 176Lu intrinsic radioactivity in LSO PET
systems[J]. IEEE T. Nucl. Sci., 2011, 58(3): 682-686.
[13] WEI Qing-Yang, WANG Shi, MA Tian-Yu, et al. Development and performance evaluation of a compact
PET/SPECT/CT tri-modality system for small animal imaging applications. Submitted to Nuclear Inst. and Methods in
Physics Research, A.
[14] Yamamoto S, Horii H, Hurutani M, et al. Investigation of single, random, and true counts from natural radioactivity in
LSO-based clinical PET[J]. Ann. Nucl. Med., 2005, 19(2): 109-114.
[15] Efthimiou N, Loudos G, Karakatsanis N A, et al. Effect of Lu-176 intrinsic radioactivity on dual head PET system
imaging and data acquisition, simulation, and experimental measurements[J]. Med. Phy., 2013, 40(11250511).
[16] Huber J S, Moses W W, Jones W F, et al. Effect of Lu-176 background on singles transmission for LSO-based PET
cameras[J]. Phy. Med. Biol., 2002, 47(19): 3535-3541.
[17] Watson C C, Casey M E, Eriksson L, et al. NEMA NU 2 performance tests for scanners with intrinsic radioactivity[J].
J. Nucl. Med., 2004, 45(5): 822-826.
[18] Eriksson L, Watson C C, Wienhard K, et al. The ECAT HRRT: An example of NEMA scatter estimation issues for
LSO-based PET systems[J]. IEEE T. Nucl. Sci., 2005, 52(11): 90-94.
[19] Efthimiou N, Loudos G, Karakatsanis N A, et al. Effect of Lu-176 intrinsic radioactivity on dual head PET system
imaging and data acquisition, simulation, and experimental measurements[J]. Med. Phy.. 2013, 40(11250511).
[20] Adonai N, Nguyen K N, Walsh J, et al. Ex vivo cell labeling with 64Cu-pyruvaldehyde-bis (N4-
methylthiosemicarbazone) for imaging cell trafficking in mice with positron-emission tomography[J]. Proc Natl Acad
Sci U S A. 2002, 99(5): 3030-3035.
[21] Sundaresan G, Yazaki P J, Shively J E, et al. 124I-labeled engineered anti-CEA minibodies and diabodies allow high-
contrast, antigen-specific small-animal PET imaging of xenografts in athymic mice[J]. J. Nucl. Med., 2003, 44(12):
1962-1969.
[22] Parodi K, Enghardt W, Haberer T. In-beam PET measurements of beta+ radioactivity induced by proton beams[J]. Phy.
Med. Biol., 2002, 47(1): 21-36.
[23] Yoshida E, Tashima H, Nishikido F, et al. Reduction method for intrinsic random coincidence events from (176)Lu in
low activity PET imaging[J]. Rad. Phy. Tech., 2014, 7(2): 235-245.
[24] Goertzen A L, Suk J Y, Thompson C J. Imaging of weak-source distributions in LSO-based small-animal PET
scanners[J]. J. Nucl. Med., 2007, 48(10): 1692-1698.
[25] Poon J K, Dahlbom M L, Moses W W, et al. Optimal whole-body PET scanner configurations for different volumes of
LSO scintillator: a simulation study[J]. Phy. Med. Biol., 2012, 57(13): 4077-4094.
[26] YAO Ru-Tao, DENG Xiao, Beaudoin J. F., Ma, T., et al. Initial evaluation of LabPET/SPECT dual modality animal
imaging system [J]. IEEE T. Nucl. Sci., 2013, 60, 76-81.
[27] YAO Ru-Tao, Ma Tian-Yu, Shao Yi-Ping. Lutetium oxyorthosilicate (LSO) intrinsic activity correction and minimal
detectable target activity study for SPECT imaging with a LSO-based animal PET scanner[J]. Phy. Med. Biol. 2008,
53(16): 4399-4415.
[28] WEI Qing-Yang, MA Tian-Yu, WANG Shi, et al. Measure PET detector performance with the intrinsic radioactivity of
scintillator[M]. IEEE NSS/MIC Conf. Rec., Yu B, 2012, 3544-3547.
[29] Bircher C, SHAO Yi-ping. Use of internal scintillator radioactivity to calibrate DOI function of a PET detector with a
dual-ended-scintillator readout[J]. Med. Phy., 2012, 39(2): 777-787.
[30] Conti M, Eriksson L, Hayden C. Monitoring energy calibration drift using the scintillator background radiation[J].
IEEE T. Nucl. Sci., 2011, 58(31): 687-694.
[31] Moses, W.W. and C.J. Thompson, Timing calibration in PET using a time alignment probe. IEEE T. Nucl Sci, 2006.
53(5): p. 2660-2665.
[32] Rothfuss, H., Moor, A., Young, J., et al. Time alignment of time of flight positron emission tomography using the
background activity of LSO.2013 NSS/MIC Conf Rec, pp. 1-3.
[33] Rothfuss H, Panin V, Moor A, et al. LSO background radiation as a transmission source using time of flight[J]. Phy.
Med. Biol. 2014, 59(18SI): 5483-5500.
[34] WEI Qing-Yang, WANG Shi, DAI Tian-Tian, et al. Geometrical Calibration of a PET-scanner-based Multi-pinhole
SPECT Using LYSO Background Radiation [M]. IEEE NSS/MIC Conf. Rec., Yu B, 2014.