 Original research
 Open Access
 Published:
Phantom validation of quantitative Y90 PET/CTbased dosimetry in liver radioembolization
EJNMMI Research volume 7, Article number: 94 (2017)
Abstract
Background
PET/CT has recently been shown to be a viable alternative to traditional postinfusion imaging methods providing good quality images of ^{90}Yladen microspheres after selective internal radiation therapy (SIRT). In the present paper, first we assessed the quantitative accuracy of ^{90}YPET using an anthropomorphic phantom provided with lungs, liver, spine, and a cylindrical homemade lesion located into the hepatic compartment. Then, we explored the accuracy of different computational approaches on dose calculation, including (I) direct Monte Carlo radiation transport using Raydose, (II) Kernel convolution using Philips Stratos, (III) local deposition algorithm, (IV) Monte Carlo technique (MCNP) considering a uniform activity distribution, and (V) MIRD (Medical Internal Radiation Dose) analytical approach. Finally, calculated absorbed doses were compared with those obtained performing measurements with LiF:Mg,Cu,P TLD chips in a liquid environment.
Results
Our results indicate that despite ^{90}YPET being likely to provide highresolution images, the ^{90}Y low branch ratio, along with other imagedegrading factors, may produce nonuniform activity maps, even in the presence of uniform activity. A systematic underestimation of the recovered activity, both for the tumor insert and for the liver background, was found. This is particularly true if no partial volume correction is applied through recovery coefficients. All dose algorithms performed well, the worst case scenario providing an agreement between absorbed dose evaluations within 20%. Average absorbed doses determined with the local deposition method are in excellent agreement with those obtained using the MIRD and the kernelconvolution dose calculation approach.
Finally, absorbed dose assessed with MC codes are in good agreement with those obtained using TLD in liquid solution, thus confirming the soundness of both calculation approaches. This is especially true for Raydose, which provided an absorbed dose value within 3% of the measured dose, well within the stated uncertainties.
Conclusions
Patientspecific dosimetry is possible even in a scenario with low true coincidences and high random fraction, as in ^{90}Y–PET imaging, granted that accurate absolute PET calibration is performed and acquisition times are sufficiently long. Despite Monte Carlo calculations seeming to outperform all dose estimation algorithms, our data provide a strong argument for encouraging the use of the local deposition algorithm for routine ^{90}Y dosimetry based on PET/CT imaging, due to its simplicity of implementation.
Background
Radioembolization with ^{90}Y, or selective internal radiation therapy (SIRT), is a catheterbased therapy with the potential of delivering a highradiation dose directly to liver tumors or metastases meanwhile minimizing the effects on healthy liver parenchyma [1,2,3,4,5,6]. The radiopharmaceutical consists of nonbiodegradable ^{90}Yimprinted resin or glassbased microspheres with a diameter of approximately 20–40 μm. The tumoricidal effect is produced by beta particles emitted from ^{90}Y incorporated on the surface of the resin matrix or integrally bound within the microsphere matrix, in the case of glass beads. The selectivity of SIRT is based on anatomic and physiological aspects of liver tumors and metastases that can be exploited for the selective delivery of the microspheres [7,8,9]. Microspheres are injected into the hepatic artery and delivered directly into the smaller blood vessels that feed the tumor, therefore being trapped within the tumor microvasculature.
It is worth noticing that in recent years, radioactive ^{166}Ho poly(llactic acid) microspheres (166HoPLLAMS) have been developed as a viable alternative to liver radioembolization with ^{90}Y microspheres. In addition to highenergy betaradiation, ^{166}Ho emits gamma radiation that allows gamma camera quantitative imaging and dosimetry [10, 11]. Furthermore, holmium is highly paramagnetic and can be visualized using traditional magnetic resonance imaging (MRI), thereby allowing for MRIbased absorbed dose calculations [11,12,13].
^{90}Y is a pure βemitting radionuclide with maximum and average energies of 2.28 MeV and 933.7 keV, respectively. The corresponding maximum and average path lengths of the emitted β particles in soft tissue (1 g/cm^{3}) are 11 and 2.5 mm. Although ^{90}Y has been traditionally considered as a pure β^{−} emitter, the decay of this radionuclide has a minor branch to the 0^{+} first excited state of ^{90}Zr at 1.76 MeV that is followed by a β^{+}/β^{−} emission. This internal pair production has been largely studied in the past because it is generated by a rare electric monopole transition (E0) between the states 0^{+}/0^{+} of ^{90}Zr. A thorough explanation of the emission of β^{+} particles via internal pair production in the 0^{+}–0^{+} transition of ^{90}Zr is provided elsewhere [14].
In the last years, the small positronic emission has been exploited for ^{90}Y–PET imaging studies after liver radioembolization, with the aim to provide quantitative posttreatment imaging and to prospectively improve the therapy [15,16,17,18,19,20,21,22,23,24,25,26,27]. Furthermore, the possibility of detecting β^{+} emissions from ^{90}Y by PET scanners is paving the way for an accurate patientspecific dosimetry. Because microspheres can be assumed to be a permanent implant, the use of appropriate dose calculation algorithms allows quantitative ^{90}Y–PET images to be converted into absorbed dose maps. The accuracy of dose calculation is of paramount importance in liver radioembolization as side effects and treatment outcome are related to absorbed doses. Over the last few years, there has been an intensifying debate revolving around the most accurate dosimetric approach for the assessment of the absorbed dose in molecular radiotherapy. Many authors obtained absorbed dose maps from ^{90}Y–PET data performing threedimensional convolution of the kernel function with the cumulated activity. A recent analysis by Pasciak et al. [26] proposed the use of local deposition technique, while previous studies used the analytical MIRD (Medical Internal Radiation Dose) approach.
The aim of this study is to investigate the impact of different computational approaches on absorbed dose calculation, starting from the quantitative accuracy of ^{90}Y–PET. The workflow followed in the present work was structured into three logically sequential steps, which are detailed below.
(1) Firstly, the quantitative accuracy of reconstructed ^{90}Y–PET data was evaluated using a cylindrical phantom and an IEC (International Electrotechnical Commission) body phantom for the assessment of recovery coefficients. Then, the PET camera’s ability to produce ^{90}Y quantitative data was validated using an anthropomorphic phantom with a liver cavity and a background compartment. A hollow cylindrical insert was fixed into the liver cavity of the anthropomorphic phantom with the intent to simulate a hepatic lesion. Acquisitions were performed using a General Electric (GE) Discovery ST PET/CT scanner with a 6:1 tumor to background activity concentration ratio. The anthropomorphic phantom was acquired at days 1, 4, 5, 6, and 12 postphantom preparation down to an activity concentration of 0.31 ± 0.02 MBq mL^{−1} for the lesion insert (initial activity concentration 5.5 ± 0.3 MBq mL^{−1}) and 50 ± 3 kBq mL^{−1} for the liver background (initial activity concentration 0.89 ± 0.04 MBq mL^{−1}). In the present phantom study, activity measurements were performed using a welltype radionuclide calibrator available on site and traceable to the Italian National Institute of Ionizing Radiation Metrology for the geometry being measured (accuracy within ± 5% at k = 2 level, as recommended by AAPM report 181 [28]). PET images were reconstructed using the standard GE algorithm (full 3DOSEM VUEPoint HD two iterations, 15 subsets) and corrected for attenuation, scatter, random, and dead time using manufacturer software. Random and scatter corrections are incorporated into the iterative algorithm, with scatter correction performed using a fully 3D approach that considers both the axial and transaxial scatter components.
(2) The second part of the study was devoted to absorbed dose calculations. Absorbed doses to the cylindrical insert were calculated from ^{90}Y–PET images obtained in the previous stage, benchmarking different computational algorithms including (I) full Monte Carlo (MC) using Raydose [29]; (II) Kernel convolution using Philips Stratos [30]; (III) local deposition [26]; (IV) MC NParticle code [31] (MCNP4c); and (V) MIRD analytical approach [32].
(3) Ultimately, absorbed doses calculated using MC codes were compared with those obtained performing experimental measurements with highsensitivity LiF:Mg,Cu,P thermoluminescent dosimeters (TLDs) inside the same cylindrical insert filled with a homogenous ^{90}YCl_{3} solution. LiF:Mg,Cu,P chips were fully characterized in terms of absorbed dose to water at the Italian National Institute of Ionizing Radiation Metrology (ENEAINMRI) using the available ^{60}Co reference gamma beam [33]. In the present study, the average absorbed dose to water obtained from TLD measurements was considered the gold standard.
The observation reported here provides key insight into dosimetry calculations related with ^{90}Y–PET data. To the best of the authors’ knowledge, there are no published literature studies aimed to compare measured absorbed dose with dose values obtained from ^{90}Y–PET quantitative data.
The results reported in the present paper are organized into three separate sections, reflecting the abovementioned workflow.
Results
Quantitative analysis
A cylindrical phantom, an IEC body phantom, and an anthropomorphic phantom were used to assess and validate the PET camera’s ability to image ^{90}Y and produce quantitative data. ^{90}Ychloride in an aqueous solution of 0.1 mol dm^{−3} hydrochloric acid was used for phantom experiments as microspheres tend to settle over time during imaging, resulting in a nonuniform distribution of ^{90}Y.
Adsorption of ^{90}YCl_{3} on the inner walls of plastic phantoms may negatively affect PET quantitative imaging studies [34]. Therefore, in the present study, phantom preparation was performed using ultrapure water (18 MΩ cm) and adding diethylenetriaminepentaacetic acid (DTPA) at a concentration of 100 μg g^{−1} to the stock solution to prevent binding of ^{90}Y to phantom polymethylmethacrylate (PMMA) walls. Activity was dispensed into the phantoms by accurate volume measurements performed using a calibrated four significant digit balance. Double weighing of the volume to be dispensed was performed for each phantom.
The ^{90}Y calibration factor on our GE Discovery DST PET/CT scanner was evaluated rescaling the ^{18}F calibration factor by the ^{90}Y beta plus branch ratio (0.00318%). This procedure was validated performing an overnight acquisition of a cylindrical phantom (5640 mL volume, Fig. 1a) uniformly filled with (1.53 ± 0.08) GBq of ^{90}YCl_{3} mixed with pure water (273 ± 14 kBq mL^{−1}). The phantom was positioned at the center of the field of view (FOV) and a 16h single bed acquisition was launched. After the acquisition, data were reconstructed using two iterations, 15 subsets. Past research on the same Discovery ST PET/CT scanner [16] showed that two iterations represent a good tradeoff between signal recovery and image noise on object with volume larger than 10 mL, approximately. Images were elaborated with the builtin PET VCAR GE software. The recovered concentration in the uniform phantom was 257 ± 16 kBq mL^{−1} (6.2% standard deviation) with an underestimation of − 5.9% with respect to the true concentration (Table 1). This finding confirms the validity of the proposed calibration procedure as the true and the measured activity concentrations agree well within the stated uncertainties.
Partial volume effect (PVE) was studied through the assessment of absolute recovery coefficients (RCs) on spherical objects, i.e., the ratio of the measured activity concentration to true activity concentration. Regions of interest (ROI) as large as the physical diameter of the sphere were drawn. A NEMA IEC/2001 phantom with six fillable spheres of varying volume (0.52, 1.15, 2.57, 5.57, 11.49, and 26.52 mL) was used for the assessment of RCs (Fig. 1b–d). Each sphere was filled with a uniform ^{90}YCl_{3} liquid solution having an activity concentration of 2.28 ± 0.11 MBq mL^{−1} at the time of acquisition. An 8:1 spheretobackground ratio was used, as recommended by the NEMA Standards Publication NU 22007 [35]. PVE led to significant underestimation of the activity concentration, especially in spheres with a volume below 10 mL, approximately. RCs range from 0.2 for the smallest sphere to 0.86, approximately, for the largest sphere (Fig. 1e).
The PET camera’s ability to produce ^{90}Y quantitative data was validated using an anthropomorphic phantom with a liver cavity and a background compartment. A homemade cylindrical PMMA insert was manufactured and fixed into the liver cavity to simulate a hepatic lesion. The nominal volume of the insert is 19.13 mL (diameter 28.5 mm, height 30 mm). Measurements in anthropomorphic geometry were performed using a 6:1 tumor to background activity concentration ratio, as this uptake ratio is often encountered in the clinical practice. The first scan was performed with an activity concentration of 5.5 ± 0.3 MBq mL^{−1} for the lesion insert and 0.89 ± 0.04 MBq mL^{−1} for the liver. The anthropomorphic phantom was then acquired at days 4, 5, 6, and 12 (according to the scanner availability) down to an activity concentration of 0.31 ± 0.02 MBq mL^{−1} for the lesion insert and 50 ± 3 kBq mL^{−1} for the liver. The background compartment of the phantom was filled with nonradioactive water. Activity distributions of ^{90}Y–PET images were analyzed in terms of differential activity concentration volume histograms (dAVHs), generated by binning the activity concentration values from each voxel in the volume and calculating the volume of structure containing an activity concentration given by bin:
where Δc is the binning resolution in terms of activity concentration, v _{ i } is the ith volume element containing an activity concentration of at least c _{ i }, and V is the total volume. dAVHs of the tumor insert are reported in Fig. 2, while Fig. 3 shows the anthropomorphic phantom used to perform ^{90}Y–PET quantitative acquisitions. Despite the cylindrical insert being uniformly filled with ^{90}YCl_{3}, heterogeneous activity distributions were obtained from ^{90}Y–PET data. As a general rule, visual inspection of the images showed the presence of hot and cold spots in regions of the phantom uniformly filled with ^{90}Y (Fig. 2k–o). From a quantitative point of view, positiveskewed dAVHs were obtained for all acquisitions, i.e., activity concentration values clustered more toward lower activity concentration values with few higher values (Fig. 2a–j). This is in keeping with previous research on noise in PET imaging, modeling positiveskewed noise probability density function with Poisson, negative binomial, lognormal, and gamma distributions [36]. Consistently, we found that the lower the number of detected β^{+} events (i.e., the noise level), the higher the skewness value (Fig. 2a–j). A comprehensive study of noise in ^{90}YPET imaging lies outside the scope of the present study. The reader is referred to Carlier et al. [37] for details.
A systematic underestimation of the recovered activity, both for the tumor insert and for the liver background, was found. This is particularly true if no partial volume correction is applied through RCs. After the application of RCs, differences with measured activity concentration values reduce significantly with underestimations of the order of 4% down to an activity concentration of about 2 MBq mL^{−1} (Table 2 and Fig. 4a). For lower activity concentration values, deviations in the range 10–15% were obtained down to 0.24 MBq mL^{−1} (Tables 2 and 3). PVE typically occurs whenever the object size is less than three times the full width at half maximum of the reconstructed image resolution [38]. Therefore, PVE in the liver were assumed negligible.
A homogeneity index, HI _{ dAVH }, was derived from dAVHs to quantify the dispersion of the average activity within the volume of interest (VOI) [39]:
where C _{ mean } is the average activity concentration over the selected VOI and c _{ i } the activity concentration at the voxel level. For each acquisition, HI _{ dAVH } were plotted as a function of the number of positrons emitted from the VOI (Fig. 4b). Calculated HI _{ dAVH } are reported in Tables 1, 2, and 3. A power function of the form y = ax ^{b}was used to fit data points, where x is the number of emitted positrons from the VOI and a and b coefficients to be determined from the fit. From the regression analysis, it resulted b = − 0.54 ± 0.07 (with r ^{2} = 0.86), indicating that the homogeneity index is representative of image noise (thereby associated with low counting statistics) and it is likely to decrease in proportion to the square root of the number of emitted positrons.
Absorbed dose calculations
The absorbed dose to the tumor cylindrical insert (Fig. 2k–o) was calculated using different computational approaches, namely: I) Full MC using Raydose II) Kernelconvolution using Philips Stratos III) Local deposition (LD) algorithm IV) MC NParticle code (MCNP4c), assuming uniform activity distribution V) MIRD analytical approach. Methods I–III use ^{90}YPET images generated from the PET scanner for the assessment of the absorbed dose at the voxel level. Therefore differential dose volume histograms (dDVH) of the VOI can be produced and compared. Methods IV–V allow the average absorbed dose to the tumor insert to be determined assuming that the cylindrical insert is uniformly filled with a known ^{90}Y activity concentration. Therefore, possible activity concentration nonuniformities are not considered.
From a clinical point of view, microspheres become a permanent implant. Therefore, dose calculations were performed considering a single image (or activity concentration value) and assuming that the entire dose is delivered over the ^{90}Y physical halflife of 2.67 days. A comparison of the calculated average absorbed dose is shown in Fig. 5, where the different dose calculation algorithms were applied to each of the clinical condition reported in Fig. 2k–o. It is worth noticing that the possible influence of the wall thickness is considered only in Raydose and MCNP simulations. Raydose uses CT units to determine attenuation coefficients for radiation transport; therefore, perturbation effects of local heterogeneities (e.g., PMMA phantom walls) are correctly accounted for. Equivalently, a cylinder provided with a 1mm thick PMMA lateral wall (ρ = 1.19 g cm^{−3}) was implemented in MCNP simulations (elemental composition by weight: H = 0.080538, C = 0.599848, and O = 0.319614). Therefore, MCbased dose evaluations are reported with a different pattern to differentiate MC calculations with other algorithms that do not directly accounts for the presence of the acrylic wall.
As a general rule, all dose algorithms provided comparable dose estimates with deviations within 9%, except for dose calculations performed on day 1, where the standard deviation over the average absorbed dose values is 20%. When compared to each other, LD, MIRD approach, and Kernelconvolution methods provided comparable absorbed doses with an average deviation within 6% (maximum deviation 12% on day 1). Similarly, MC codes provided comparable absorbed dose values to the tumor insert, with an average deviation below 4%. Calculations performed using kernel convolution algorithm provided dose estimates in excellent agreement with the wellknown MIRD calculation method, with differences below 5% for all acquisitions, except for day 12 where the difference in the average absorbed dose is 17.7%.
Owing to the nonuniformity in the activity distribution, the standard deviation of each dose point is generally large (of the order of 30–40% of the dose value) and becomes unacceptable when the activity concentration decreases down to 0.22 MBq/mL (approximately 100% the dose value). Figure 6 shows a comparison between dDVHs generated by full MC calculations with Raydose, LD algorithm, and convolution approach with Philips Stratos. dDVHs were produced after a voxelbased dose estimation of the cylindrical insert. In the present study, the DICOM Structure Set format (RTSTRUCT) module was used to transfer phantom structures and related data between the workstations. Insert segmentation was performed so as to exclude the PMMA wall from the contour. Results in Fig. 6 are presented in terms of dv_{i}/dD, i.e., absolute volume element (dv_{i}) receiving dose in the corresponding dose bin (dD). The integral under each differential DVH yields the insert volume (nominal volume 19.13 mL). VOI contouring performed using methods I and II provided a volume of 19.60 mL (+ 2.45%), while insert contouring performed with Philips Stratos resulted in a volume of 19.21 mL (+ 0.42%).
Absorbed dose measurement using LiF:Mg,Cu,P TLDs
Absorbed dose calculations presented in the previous section were validated by absorbed dose to water measurements using LiF:Mg,Cu,P chips inside a PMMA phantom filled with a homogenous ^{90}YCl_{3} aqueous solution. Therefore, the average absorbed dose to water obtained from TLD measurements was considered the gold standard against which other dose algorithms can be benchmarked. Each TLD had 4.5 mm diameter and was 0.8 mm height. To this purpose, measurements were performed manufacturing six cylindrical phantoms the size of that used for quantitative analysis (Fig. 7). Each cylindrical phantom hosted a waterproof PMMA stick containing 3 TLD chips encapsulated by a polystyrene envelope (for a total of 18 chips). The PMMA phantoms were manufactured so that the radioactive liquid environment surrounds the whole stick. The ^{60}Co reference gamma beam available at ENEAINMRI was used to evaluate the calibration coefficient of each TLD chip in terms of absorbed dose to water. Relative standard uncertainty on the TLD calibration coefficient was found to be in the range 0.7–2.2%.
The measurement of the absorbed dose from betaemitting radionuclides in liquid solution is not straightforward. The exclusion of radioactivity from the volume occupied by the TLD is a major effect that needs to be accounted for. Furthermore, the waterproof polystyrene envelope unavoidably contributes to beta particle attenuation during the measurement. Finally, as TLD calibration was performed using a source and a geometry different from the real experimental condition, an energy and geometry correction factor needs to be considered. An accurate description of the measurements equation is reported in the “Methods” section, along with a detailed description of correction and conversion factors.
PMMA samples were filled with ^{90}YCl_{3} aqueous solution with activity concentration of 6.7 MBq/mL, approximately. Activity concentration was determined by measuring an aliquot of the stock solution in terms of activity per unit mass. Then, this was used to determine the activity of all subsequent sources produced from this stock solution. Accurate measurements of ^{90}Y activity concentration were performed using the ENEAINMRI portable tripletodouble coincidence ratio (TDCR) counter [40], with a relative standard uncertainty below 1%. Measurement time was 30 min approximately for each sample.
After measurement, each stick was removed from the sample and TLD chips were manually removed from the PMMA holder using dedicated tongs. One out of six sticks holding the chips was wasted because of leaks in the PMMA envelope that caused radioactive water to get in contact with the chips. Furthermore, 8 TLD chips were accidentally slightly contaminated during the removal procedures. These data points were removed from the final data as significantly higher.
The average absorbed dose to water per emitted beta particle was finally assessed as the weighted average of TLD measurements and resulted to be (1.09 ± 0.07) × 10^{−11} Gy/particle, with an associated relative weighted standard uncertainty of 6.3%. The measured value can be compared with the absorbed dose to water per emitted beta particle obtained using two MC codes, namely MCNP4c and Raydose.
With MCNP4c, the absorbed dose to water was calculated implementing the full cylindrical phantom geometry into the code. A homogenous ^{90}Y source was considered into the phantom, and the delivered energy to water was scored into geometrical cells the size of a TLD chip. Simulations provided absorbed dose values per unit emitted beta particle of (1.010 ± 0.005) × 10^{−11} Gy/particle.
On the other hand, imagebased calculations with Raydose were performed using the ^{90}YPET acquisition reported in Fig. 2k. Voxelbyvoxel absorbed dose calculations were performed on the cylindrical insert. Finally, VOIs the size of a TLD chip (4.5 mm diameter, 0.8 mm height) were drawn on the dose profile (Fig. 8). The average absorbed dose per unit emitted beta particle was determined to be (1.06 ± 0.04) × 10^{−11} Gy/particle.
Overall, absorbed dose assessed with MC codes is in excellent agreement with those obtained using TLD in liquid solution, thus confirming the soundness of both calculation approaches (Table 4). This is especially true for Raydose, which provided an absorbed dose value within 3% of the measured dose, well within the stated uncertainties.
Discussion
This study offers key insight into how different approaches to dosimetry calculations can accurately assess the absorbed dose after liver radioembolization using ^{90}Y–PET data. At a time of increasing evidence for an absorbed radiation doseeffect relationships in radioembolization with ^{90}Yladen microspheres [41, 42], it is the authors’ belief that there is an urgent need for accurate dosimetry in patients undergoing SIRT therapy. In fact, it is generally accepted that the absorbed dose both to tumor areas and to healthy liver is likely to have a significant impact on treatment’s clinical effectiveness. Furthermore, since microspheres administered during therapy remain trapped within the microvasculature, imaging requirements reduce to a singletime acquisition, thus simplifying the whole dosimetry process. As a consequence, there is great potential for SIRT dosimetry to become routine in the clinical practice.
Patientspecific dosimetry in SIRT suffers from a number of imagedegrading effects. Besides PVE, that is likely to play a major role in small lesions, low true coincidence statistic due to the combined interplay of low beta plus branching ratio and low achievable acquisition time are likely to impact the image quality. Furthermore, the high scatter component intrinsic of ^{90}Y PET imaging may negatively affect image quantification. Ultimately, patient respiratory motion is the primary cause of image blurring possibly leading to systematic dose underestimations [43].
In this study, we concerned ourselves with quantitative accuracy of ^{90}Y PET and the impact that different dose calculation approaches have on imagebased dosimetry after liver radioembolization. Yet another question of interest for us was to compare absorbed doses obtained from MC calculations with those obtained through absorbed dose measurements performed with TLD chips into a liquid radioactive environment. To the best of our knowledge, this line of research is the first to investigate the degree of agreement between calculated and measured dose using ^{90}Y–PET quantitative data.
Our data confirm and expand previous observations. Since most scanner do not support ^{90}Y as a viable radionuclide option, scanner calibration should be performed using surrogate radionuclides and applying proper decay correction factors (e.g., ^{86}Y [44, 45] or ^{22}Na [46]). Quantitative analysis showed that partial volume effects dominate spheres of volume below 11.5 mL (diameter < 28 mm), approximately. Recovery of activity concentration measured in the largest sphere on day 0 of imaging underestimated the true activity concentration of − 15%, approximately. Our results are in keeping with previous studies [18, 47]. However, the recent QUEST study [48] suggests that ToFPET scanners are likely to improve contrast of hot spheres and increase RCs.
Another important consideration is that the low branching ratio for ^{90}Y positron emission entails long scan times in order to reduce image noise. This is especially true for nonToF–PET scanner, where the acquisition time plays a key role. Quantitative analysis in anthropomorphic geometry showed that despite the tumor insert being uniformly filled with a homogenous ^{90}YCl_{3} solution, ^{90}Y–PET images presented significant spatial nonuniformity. Notwithstanding this, for activity concentrations exceeding 1.5 MBq/mL, the recovered activity concentration underestimated the known value by 4% approximately, demonstrating that accurate quantitation of ^{90}Y is possible if long scan time is performed and if partial volume effects are accurately compensated for. Quantitative accuracy decreases for decreasing activity concentration levels. Interestingly, our results are in close agreement with those obtained by Mille et al. [49] which recovered 95.4% of the known activity in a large cylindrical phantom despite obtaining noisy images with a voxel variability of 21%.
In the present study, the degree of uniformity was quantified in terms of homogeneity index. Quantitative data demonstrated a noiselike behavior of the homogeneity index, with the latter decreasing in proportion to the square root of the number of emitted positrons. This finding suggests that for the considered geometry and settings, image nonuniformities are mainly imputable to the lowcounting statistics. It is interesting to note that activity underestimation both in the tumor insert and in the liver compartment is associated with increasing homogeneity index.
Of additional concern has been the impact of different dose calculation algorithms in anthropomorphic geometry. As a matter of fact, to date only a few studies attempted to compare different dosimetry methods in liver radioembolization. Mikell and coworkers [50] compared four voxelbased dosimetry dose algorithms based on ^{90}Y bremsstrahlung SPECT images (namely, MC, softtissue kernel with density correction (SKD), softtissue kernel (SK), and LD method). Interestingly, the authors found that for tumor, nontumoral liver mean absorbed doses calculated with SKD, SK, and LD are equivalent to MC (within 5%). Deviations in the mean absorbed dose values increase when dosimetry is performed in the lungs, with right lung dosimetry being strongly influenced by the liver–lung interface [50]. In another work, Grassi et al. [51] compared STRATOS with a homemade software package (VoxelMed), both in phantom and in patients undergoing radiopeptide therapy with ^{177}Lu. Both software performed well, with an agreement within 5% in phantom. Larger deviations were observed in patients.
In the present study, the large voxel variability of absorbed dose values is related to the abovementioned intrinsically noisy nature of ^{90}Y PET images. However, on average, absorbed doses to the tumor insert have been found to be consistent to within 9%, except for dose calculations performed on day 1, where the standard deviation over the average absorbed dose values is 20%.
The most striking observation to emerge from the data comparison is the difference between dose values obtained with MC codes (namely, MCNP and Raydose) and other algorithms. The foremost cause of the slightly lower dose values obtained with Raydose is most probably due to the resampling of the CT grid from 512 × 512 to 128 × 128, executed for efficiency purposes. The larger voxel size, together with the inhomogeneity in the underlying PET images, produced a smoothing of the activity data. This is evident in the fine distribution of the dose (Fig. 6a–e): DVHs obtained with Raydose do not present any highdose tail as any hot spots in a single voxel are shared with adjoining voxels. This is likely to produce lower doses especially at the liquid–PMMA interface. On the other hand, the lower doses obtained using MCNP could be interpreted as being a result of: (i) the input activity concentration, supposed to be uniformly distributed over the entire cylindrical object and (ii) the absence of spillin activity from regions around the cylindrical insert in MCNP simulations. This is likely to result in a lower dose to the cylindrical insert.
Furthermore, when interpreting the results of the current study, it is also important to note that the presence of the PMMA wall of the cylindrical insert may had an impact on dose calculations, possibly leading to biased estimates between methods that directly account for the acrylic edges (Raydose, MCNP) and those that do not (kernel convolution, LD, and MIRD). The impact of the PMMA wall on dose calculations is somehow expected and inevitable, given the nonnegligible presence of acrylic material (4.98 g) if compared to the total mass of the insert when filled with ^{90}YCl_{3} (24.1 g). From a computational perspective, there is no substantial material difference between tumor and background liver. As a consequence, this effect is not expected in a clinical scenario, unless dose calculations are performed at the liver–lung interface, where difference in material density may play a key role [50].
Of note, average absorbed doses determined with the LD method are in excellent agreement with those obtained using the MIRD and the kernel convolution dose calculation approach. This result provides a strong argument for encouraging LD algorithm for the evaluation of absorbed doses in the clinical practice, where resource–intensive software packages are not always available. Similar conclusions have been drawn in recent research on the matter [26]. However, the large dose variability at the voxel level—most likely due to a scenario with low true coincidences and high random fraction—raises questions about the possibility of using dDVH for radiobiological modeling if acquisitions are performed with a nonTOF scanner.
During the last 20 years, there has been increasing interest in the measurement of absorbed doses from internal beta emitters used in molecular radiotherapy. In particular, TLDs have been extensively used in the past to measure the absorbed dose both in phantoms [52] and in animals [53,54,55]. However, to the best of the authors’ knowledge, this is the first work that compares absorbed doses obtained from ^{90}Y–PET quantitative imaging procedures with those obtained from experimental measurements using TL dosimeters. Direct Monte Carlo radiation transport is presently considered to be the most accurate of all currently available dose estimation algorithms [56]. As a consequence, its validation using dosimeters calibrated against absorbed dose primary standards has become even more pressing. Our research has confirmed the accuracy of MC calculations in ^{90}Y–PET dosimetry. \( {\overline{D}}_w^{meas} \) and \( {D}_w^{MCNP} \) are in the ratio 1.088. This difference needs to be further investigated, especially in the light of the wellknown uncertainty of MCNP cross sections at extremely low energies [57]. On the other hand, measured doses were in excellent agreement with absorbed doses calculated using a full MC approach, being \( {\overline{D}}_w^{\mathrm{meas}} \) and \( {D}_w^{\mathrm{Raydose}} \) in the ratio 1.03, well within the stated uncertainties.
The current research acknowledges a few limitations that should be noted to aid interpretation of the result. It is worth noting that the ability of a PET scanner to perform accurate ^{90}Y quantification relies on the knowledge of the internal pair production branching ratio. Therefore, precise knowledge of the branch ratio of the 0^{+}–0^{+} transition of ^{90}Zr is important for an accurate quantification of ^{90}Y accumulated inside the target region and detected via PET acquisition. Most recent literature findings report an internal pair production branch ratio as large as (3.186 ± 0.047) × 10^{−5} measured by Selwyn [58] and colleagues using a HPGe detector. This branch ratio value was used in the present study for the absolute calibration of the PET system. It is desirable that in the near future, the current uncertainty on the internal pair production branch ratio (about 1.5%) will be reduced by more accurate experimental measurements [59].
Furthermore, we focussed on absolute quantification and dosimetry on a single cylindrical insert simulating a liver lesion. The extent to which both our methods and results can be generalized to other geometries and different PET scanners certainly requires further investigation. Finally, it is worth noticing that PET calibration measurements should be repeated over a reasonable long period of time to assess the system stability and to evaluate the final uncertainty in the calibration factor. Unfortunately, it was not possible in the present experiment to investigate impact of repeated calibration measurements due to the limited resources available.
Contrary to external beam radiotherapy, in which individual patient dosimetry is mandatory and there are legal requirements for accuracy (within 5% to a reference point), there is a lack of standardization in molecular radiotherapy (MRT), and it suffers from isolated efforts to harmonize quantification and dosimetry approaches. The current main source of uncertainty in internal dosimetry is in taking the step from dose measurements on simple reference geometries to quantitative imaging measurements of the complex and varying geometries of the activity localized in real patients. This passage is essential to comply with EC Directive 2013/59/EURATOM, Article 56, which states that individual dose planning for radiotherapy patients (including MRT) must be enforced in legislation by EU member states by 6 February 2018. It is desirable that in the near future, the final standard uncertainty in absolute quantification in complex phantoms be well below well 5% to ensure the uncertainty in clinical absorbed dose estimations comply with the requirement of 5% to a reference point. Comprehensive guidance has yet to be presented in this field, and there is no doubt that an internationally endorsed protocol on ^{90}Y PET/CT quantitative imaging would lead to further advances in this area.
Conclusions
The results of this research are sufficient to conclude, with some caveats, that patientspecific dosimetry is possible even in a scenario with low true coincidences and high random fraction, as in ^{90}Y PET imaging, granted that accurate absolute PET calibration is performed and acquisition times are sufficiently long. In the final analysis, despite Monte Carlo calculations outperforming all dose estimation algorithms, we believe to have gathered ample evidence to recommend the use of the LD algorithm for routine ^{90}Y dosimetry based on PET/CT imaging, due to its simplicity of implementation.
Methods
Uncertainty in activity measurements
Activity measurements required for the phantom study were performed using the welltype radionuclide calibrator available on site and traceable to the Italian National Institute of Ionizing Radiation Metrology for the geometry being measured (accuracy within ± 5% at k = 2 level, as recommended by AAPM report 181). Weightings of liquid solutions were performed with highprecision balances providing accuracies below 0.01%. Therefore, uncertainties associated with volume measurements were considered negligible.
Activity measurements needed for absorbed dose measurements using LiF:Mg,Cu,P TL detectors in a liquid environment were performed using the ENEAINMRI portable tripletodouble coincidence ratio (TDCR) counter [35], providing a relative standard uncertainty below 1%.
PET scanner and reconstruction data
^{90}Y PET images were acquired with a GE Discovery ST PET/CT scanner (GE Healthcare, Milwaukee, WI, USA) with lower energy threshold at 425 keV. Acquisition time was 16 h for the reference phantom and 30 min per bed for the anthropomorphic phantom. PET images were reconstructed using the standard GE algorithm (full 3DOSEM VUEPoint HD two iterations, 15 subsets). PET coincidence window is 12 ns, Peak NECR (3D) 38 kcps, energy resolution 20%, and scatter fraction about 35%. PET scanner sensitivity and spatial resolution were assessed in a previous study [16]. PET data were corrected for attenuation, scatter, random, and dead time using manufacturer software. Random and scatter corrections are incorporated into the iterative algorithm [60]. Scatter correction was performed using a fully 3D approach that considers both the axial and transaxial scatter components, accurately modeling scatter from hot regions in neighboring super slices and outside the scan fieldofview resulting in greater quantitative accuracy (3D modelbased scatter correction, MBSC [61]).
Convolution algorithm
The convolution algorithm allows the absorbed dose following ^{90}Y microspheres administration to be calculated by the convolution of the 3D activity concentration matrix A(x, y, z)with the 3D dose kernel for ^{90}Y, K(x, y, z), according to the following equation:
with \( \lambda =\frac{\ln (2)}{T_{1/2}} \) decay constant of ^{90}Y. Dose calculations were performed using STRATOS, a voxelbased dose engine on the IMALYTICS Research Workstation.
Local deposition algorithm
Local deposition algorithm has been recently proposed for dose calculations whenever the energy released by charged particles can be assumed to be locally absorbed with in the same voxel as the decay. The other assumption of local deposition algorithm is that the full width at half maximum (FWHM) of the point spread function (PSF) should be larger than the dose point kernel. This is because the output images can be regarded as a convolution of the real object with the PSF of the system. In the ideal case where PSF = Kernel, the imaging system simulates the energy transport among voxels and the output image can be regarded as the absorbed dose map.
The dose at a single voxel, D _{ voxel }, can be calculated as the ratio of the energy delivered to the voxel, E _{ voxel }, and the mass of the voxel, m _{ voxel }. Assuming a voxel density ρ = 1 g cm^{3}, m _{ voxel }[g] = v _{ voxel }[cm^{3}], with v _{ voxel } voxel volume. In the present study, images were reconstructed using a voxel size as large as 0.273 cm × 0.273 cm × 0.327 cm. Therefore, m _{ voxel } = 0.024371 cm^{3}. The energy delivered to the voxel can be calculated as the number of particles totally emitted into the voxel (N _{ tot }) times the average energy delivered to the voxel by each particle (\( \overline{E} \)):
Since microspheres are permanently implanted in the patient with no removal from the region, the effective halflife is simply the radioactive halflife and the number of ^{90}Y particles totally emitted into the voxel can be calculated as follows:
With \( {A}_0^{voxel} \) initial activity into the voxel at t = 0, and T _{1/2} halflife of the radionuclide. Regarding the average energy delivered by each particle, \( \overline{E} \), the literature value 0.932 MeV can be assumed. Thus, the average energy delivered to the voxel is:
\( {A}_0^{\mathrm{voxel}} \) is obtained from ^{90}Y PET images. In fact, as ^{90}Y–PET images are calibrated in terms of activity concentration, C _{voxel}, (i.e., each voxel—i is expressed in term of Bq mL^{−1}) the activity into the voxel is given by:
Thus, the dose at a single voxel, D _{ voxel }, can be calculated as follows:
Absorbed dose measurement equation for TLD dosimetry
In the considered experimental geometry the absorbed dose to water was determined from the TLD reading through the following equation:
where

\( \mathrm{TL}\left({E}_{{}^{90}Y}\right) \) is the TLD reading after irradiation with ^{90}Y beta particles during measurement.

\( {N}_{D_w}^{{}^{60} Co}=\frac{D_w\left({}^{60} Co\right)}{TL\left({}^{60} Co\right)} \) is the TLD calibration coefficient at ^{60}Co quality, with D _{ w }(^{60} Co) absorbed dose to water delivered to the TLD and TL(^{60} Co) the TLD reading after irradiation.

k _{ fv }, finite volume effect, is the loss of absorbed dose due to exclusion of radioactivity from the volume occupied by the TLD.

\( {H}_{{}^{90}Y{,}^{60} Co}^{\exp, \mathrm{ref}}=\frac{{\left({D}_w^{90Y}/{D}_{\mathrm{TLD}}^{90Y}\right)}_{\exp \_\mathrm{geom}}}{{\left({D}_w^{60 Co}/{D}_{\mathrm{TLD}}^{60 Co}\right)}_{\mathrm{ref}\_\mathrm{geom}}} \) is the energy and geometry correction factor, given by the ratio of absorbed doses to water and to TLD material at different radiation qualities and different irradiation geometries. This is calculated by the ratios of D_{w}/D_{TLD} (TLD to water conversion factor), taking into account differences between measurement conditions (exp_geom) and reference calibration conditions (ref_geom).

k _{mat}, correction factor due to the presence of materials other than TLDs, i.e., PMMA stick hosting the three TLDs and the waterproof PMMA envelope surrounding the stick.
All correction factors reported in Eq. 3 were determined using Monte Carlo simulations with MCNP4c code.
Abbreviations
 AVH:

Activity volume histograms
 dDVH:

Differential dose volume histograms
 HI:

Homogeneity index
 LD:

Local deposition
 MC:

Monte Carlo
 PET:

Positron emission tomography
 PMMA:

Polymethylmethacrylate
 PVE:

Partial volume effect
 RC:

Recovery coefficient
 ROI:

Region of interest
 SIRT:

Selective internal radiation therapy
 TLD:

Thermoluminescent dosimeter
 TOF:

Time of flight
 VOI:

Volume of interest
References
 1.
Herba M, Thirlwell M. Radioembolization for hepatic metastases. Semin Oncol. 2002;29(2):152–9.
 2.
Houle S, Yip T, Shepherd F, Rotstein L, Sniderman K, Theis E, et al. Hepatocellular carcinoma: pilot trial of treatment with Y90 microspheres. Radiology. 1989;172(3):857–60.
 3.
Lau W, Ho S, Leung T, Chan M, Ho R, Johnson P, et al. Selective internal radiation therapy for nonresectable hepatocellular carcinoma with intraarterial infusion of ^{90}yttrium microspheres. Int. J. Radiat. Oncol. Biol. Phys. 1998;40(3):583–92.
 4.
Blanchard R, Morrow I, Sutherland J. Treatment of liver tumors with yttrium90 microspheres alone. Can Assoc Radiol J. 1989;40(4):206–10.
 5.
Gray B, Burton M, Kelleher D, Anderson J, Klemp P. Selective internal radiation (SIR) therapy for treatment of liver metastases: measurement of response rate. J Surg Oncol. 1989;42(3):192–6.
 6.
Yan Z, Lin G, Zhao H, Dong Y. An experimental study and clinical pilot trials on Yttrium90 glass microspheres through the hepatic artery for treatment of primary liver cancer. Cancer. 1993;72(11):3210–5.
 7.
Breedis C, Young G. The blood supply of neoplasms in the liver. Am J Pathol. 1954;30(5):969–85.
 8.
Lin G, Lunderquist A, Hägerstrand I, Boijsen E. Postmortem examination of the blood supply and vascular pattern of small liver metastases in man. Surgery. 1984;96(3):517–26.
 9.
Tygstrup N, Winkler K, Mellemgaard K, Andreassen M. Determination of the hepatic arterial blood flow and oxygen supply in man by clamping the hepatic artery during surgery. J Clin Investig. 1962;41(3):447–54.
 10.
Smits ML, Nijsen JF, van den Bosch MA, Lam MG, Vente MA, Huijbregts JE, van het Schip AD, Elschot M, Bult W, de Jong HW, Meulenhoff PC, Zonnenberg BA. Holmium166 radioembolization for the treatment of patients with liver metastases: design of the phase I HEPAR trial. J Exp Clin Cancer Res. 2010;15(29):70.
 11.
Smits ML, Elschot M, van den Bosch MA, van de Maat GH, van het Schip AD, Zonnenberg BA, Seevinck PR, Verkooijen HM, Bakker CJ, de Jong HW, Lam MG, Nijsen JF. In vivo dosimetry based on SPECT and MR imaging of 166Homicrospheres for treatment of liver malignancies. J Nucl Med. 2013;54(12):2093–100.
 12.
Seevinck PR, van de Maat GH, de Wit TC, Vente MA, Nijsen JF, Bakker CJ. Magnetic resonance imagingbased radiationabsorbed dose estimation of 166Ho microspheres in liver radioembolization. Int J Radiat Oncol Biol Phys. 2012 83(3):e437–e444.
 13.
van de Maat GH, Seevinck PR, Elschot M, Smits ML, de Leeuw H, van Het Schip AD, Vente MA, Zonnenberg BA, de Jong HW, Lam MG, Viergever MA, van den Bosch MA, Nijsen JF, Bakker CJ. MRIbased biodistribution assessment of holmium166 poly(Llactic acid) microspheres after radioembolisation. Eur Radiol. 2013;23(3):827–35.
 14.
D'Arienzo M. Emission of β+ particles via internal pair production in the 0^{+} – 0^{+} transition of ^{90}Zr: historical background and current applications in nuclear medicine imaging. Atoms. 2013;1(1):2–12.
 15.
Lhommel R, van Elmbt L, Goffette P, Van den Eynde M, Jamar F, Pauwels S, et al. Feasibility of ^{90}Y TOF PETbased dosimetry in liver metastasis therapy using SIRspheres. Eur J Nucl Med Mol Imaging. 2010;37(9):1654–62.
 16.
D’Arienzo M, Chiaramida P, Chiacchiararelli L, Coniglio A, Cianni R, Salvatori R, et al. ^{90}Y PETbased dosimetry after selective internal radiotherapy treatments. Nucl Med Commun. 2012;33(6):633–40.
 17.
hang T, Bourgeois A, Balius A, Pasciak A. Treatment modification of Yttrium90 radioembolization based on quantitative positron emission tomography/CT imaging. J Vasc Interv Radiol. 2013;24(3):333–7.
 18.
Elschot M, Vermolen B, Lam M, de Keizer B, van den Bosch M, de Jong H. Quantitative comparison of PET and bremsstrahlung SPECT for imaging the in vivo yttrium90 microsphere distribution after liver radioembolization. PLoS One. 2013;8(2):e55742.
 19.
D’Arienzo M, Filippi L, Chiaramida P, Chiacchiararelli L, Cianni R, Salvatori R, et al. Absorbed dose to lesion and clinical outcome after liver radioembolization with ^{90}Y microspheres: a case report of PETbased dosimetry. Ann Nucl Med. 2013;27(7):676–80.
 20.
Kao Y, Steinberg J, Tay Y, Lim G, Yan J, Townsend D, et al. Postradioembolization yttrium90 PET/CT  part 2: doseresponse and tumor predictive dosimetry for resin microspheres. EJNMMI Res. 2013;3(1):57.
 21.
Bourgeois A, Chang T, Bradley Y, Acuff S, Pasciak A. Intraprocedural yttrium90 positron emission tomography/CT for treatment optimization of Yttrium90 Radioembolization. J Vasc Interv Radiol. 2014;25(2):271–5.
 22.
Willowson K, Forwood N, Jakoby B, Smith A, Bailey D. Quantitative 90Y image reconstruction in PET. Med Phys. 2012;39(11):7153–9.
 23.
Carlier T, Eugène T, BodetMilin C, Garin E, Ansquer C, Rousseau C, et al. Assessment of acquisition protocols for routine imaging of Y90 using PET/CT. EJNMMI Res. 2013;3(1):11.
 24.
van Elmbt L, Vandenberghe S, Walrand S, Pauwels S, Jamar F. Comparison of yttrium90 quantitative imaging by TOF and nonTOF PET in a phantom of liver selective internal radiotherapy. Phys Med Biol. 2011;56(21):6759–77.
 25.
Bagni O, D’Arienzo M, Chiaramida P, Chiacchiararelli L, Cannas P, D’Agostini A, et al. 90YPET for the assessment of microsphere biodistribution after selective internal radiotherapy. Nucl Med Commun. 2012;33(2):198–204.
 26.
Pasciak A, Bourgeois A, Bradley YA. Comparison of techniques for 90Y PET/CT imagebased dosimetry following radioembolization with resin microspheres. Front Oncol. 2014;4
 27.
Pasciak A, Bourgeois A, McKinney J, Chang T, Osborne D, Acuff S, et al. Radioembolization and the dynamic role of ^{90}Y PET/CT. Front Oncol. 2014;4(38):1–12.
 28.
The Selection, Use, Calibration, and Quality Assurance of Radionuclide Calibrators Used in Nuclear Medicine. Report of AAPM Task Group 181 (2012), American Association of Physicists in Medicine, One Physics Ellipse College Park, MD 20740–3846.
 29.
Marcatili S, Pettinato C, Daniels S, Lewis G, Edwards P, Fanti S, et al. Development and validation of RAYDOSE: a Geant4based application for molecular radiotherapy. Phys Med Biol. 2013;58(8):2491–508.
 30.
http://www.philips.com/aw/imalytics/specializedmodules/dosimetry.html, Accessed 19 Oct 2016.
 31.
Briesmeister JF, MCNP—a general Monte Carlo Nparticle transport code—version 4C, Los Alamos National Laboratory report LA 13709M 2000.
 32.
Gulec SA, Mesoloras G, Stabin M. Dosimetric techniques in ^{90}Ymicrosphere therapy of liver cancer: the MIRD equations for dose calculations. J Nucl Med. 2006;47(7):1209–11.
 33.
Guerra A, Laitano R, Pimpinella M. Characteristics of the absorbed dose to water standard at ENEA. Phys Med Biol. 1996;41(4):657–74.
 34.
Park M, Mahmood A, Zimmerman R, LimpaAmara N, Makrigiorgos G, Moore S. Adsorption of metallic radionuclides on plastic phantom walls. Med Phys. 2008;35(4):1606–10.
 35.
NEMA Standards publication NU 2–2007, Performance measurements of positron emission tomographs, National Electrical Manufacturers Association, 2007.
 36.
Teymurazyan A, Riauka T, Jans H, Robinson D. Properties of noise in positron emission tomography images reconstructed with filteredbackprojection and rowaction maximum likelihood algorithm. J Digit Imaging. 2012;26(3):447–56.
 37.
Carlier T, Willowson K, Fourkal E, Bailey D, Doss M, Conti M. 90Y PET imaging: exploring limitations and accuracy under conditions of low counts and high random fraction. Med Phys. 2015;42(7):4295–309.
 38.
Soret M, Bacharach S, Buvat I. Partialvolume effect in PET tumor imaging. J Nucl Med. 2007;48(6):932–45.
 39.
Yoon M, Park S, Shin D, Lee S, Pyo H, Kim D, et al. A new homogeneity index based on statistical analysis of the dosevolume histogram. J Appl Clin Med Phys. 2007;8(2):9–17.
 40.
Capogni M, De Felice PA. Prototype of a portable TDCR system at ENEA. Appl Radiat Isot. 2014;93:45–51.
 41.
Fowler K, Maughan N, Laforest R, Saad N, Sharma A, Olsen J, et al. PET/MRI of hepatic ^{90}Y microsphere deposition determines individual tumor response. Cardiovasc Intervent Radiol. 2015;39(6):855–64.
 42.
van den Hoven AF, Rosenbaum CE, Elias SG, de Jong HW, Koopman M, Verkooijen HM. Insights into the doseresponse relationship of Radioembolization with resin 90Ymicrospheres: a prospective cohort study in patients with colorectal cancer liver metastases. J Nucl Med. 2016 Jul;57(7):1014–9.
 43.
Iaccarino G, Cacciatore A, Ungania S, Cazzato M, D'Andrea M, Strigari L. Breathing motion correction in dosimetry of liver lesion based on 90YPET/CT imaging. Phys Med. 2016;32:99.
 44.
Fourkal E, Veltchev I, Lin M, Koren S, Meyer J, Doss M, Yu JQ. 3D inpatient dose reconstruction from the PETCT imaging of ^{90}Y microspheres for metastatic cancer to the liver: feasibility study. Med Phys. 2013;40(8):081702.
 45.
Goedicke A, Berker Y, Verburg FA, Behrendt FF, Winz O, Mottaghy FM. Studyparameter impact in quantitative 90yttrium PET imaging for radioembolization treatment monitoring and dosimetry. IEEE Trans Med Imaging. 2013;32(3):485–92.
 46.
Attarwala AA, MolinaDuran F, Büsing KA, Schönberg SO, Bailey DL, Willowson K, Glatting G. Quantitative and qualitative assessment of yttrium90 PET/CT imaging. PLoS One. 2014 9(11):e110401.
 47.
Werner M, Brechtel K, Beyer T, Dittmann H, Pfannenberg C, Kupferschläger J. PET/CT for the assessment and quantification of ^{90}Y biodistribution after selective internal radiotherapy (SIRT) of liver metastases. Eur J Nucl Med Mol Imaging. 2009;37(2):407–8.
 48.
Willowson K, Tapner M, Bailey DA. Multicentre comparison of quantitative 90Y PET/CT for dosimetric purposes after radioembolization with resin microspheres. Eur J Nucl Med Mol Imaging. 2015;42(8):1202–22.
 49.
Mille M, Zimmerman B, Bergeron D, Fenwick A. An investigation of the quantitative accuracy of ^{90}Y PET and the impact on imagebased dosimetry. J Nucl Med. 2015;56(suppl.3):1845.
 50.
Mikell JK, Mahvash A, Siman W, Mourtada F, Kappadath SC. Comparing voxelbased absorbed dosimetry methods in tumors, liver, lung, and at the liverlung interface for ^{90}Y microsphere selective internal radiation therapy. EJNMMI Phys. 2015;2(16).
 51.
Grassi E, Fioroni F, Ferri V, Mezzenga E, Sarti MA, Paulus T, Lanconelli N, Filice A, Versari A, Iori M. Quantitative comparison between the commercial software STRATOS(®) by Philips and a homemade software for voxeldosimetry in radiopeptide therapy. Phys Med. 2015 31(1):7279.
 52.
Carita Aschan A, Matti J. Toivonen. The use of TL detectors in Dosimetry of systemic radiation therapy. Acta Oncol. 1999;38(2):189–96.
 53.
Yorke E, Williams L, Demidecki A, Heidorn D, Roberson P, Wessels B. Multicellular dosimetry for betaemitting radionuclides: autoradiography, thermoluminescent dosimetry and threedimensional dose calculations. Med Phys. 1993;20(2):543–50.
 54.
Strandh M, Strand S, Ljungberg M, Tagesson M. MiniTLdosimeters for in vivo measurements. Radiother Oncol. 1995;37:S52.
 55.
Griffith MH, Yorke ED, Wessels BW, DeNardo GL, Neacy WP. Direct dose confirmation of quantitative autoradiography with microTLD measurements for radioimmunotherapy. J Nucl Med. 1988;29:1795–809.
 56.
Dewaraja Y, Frey E, Sgouros G, Brill A, Roberson P, Zanzonico P, et al. MIRD pamphlet no. 23: quantitative SPECT for patientspecific 3dimensional dosimetry in internal radionuclide therapy. J Nucl Med. 2012;53(8):1310–25.
 57.
Hughes G. Recent developments in lowenergy electron/photon transport for MCNP6. Progress in Nuclear Sci Technol. 2014;4:454–8.
 58.
Selwyn R, Nickles R, Thomadsen B, DeWerd L, Micka JA. New internal pair production branching ratio of ^{90}Y: the development of a nondestructive assay for ^{90}Y and ^{90}Sr. Appl Radiat Isot. 2007;65(3):318–27.
 59.
D’Arienzo M, Capogni M, Smyth V, Cox M, Johansson L, Solc J, et al. Metrological issues in molecular radiotherapy. EPJ Web of Conferences. 2014;77:00022.
 60.
Iatrou M, Ross S, Manjeshwar R, Stearns C. A fully 3D iterative image reconstruction algorithm incorporating data corrections. IEEE Nucl. Sci. Symp. Conf. Rec. 2004;4:2493–7.
 61.
Iatrou M, Manjeshwar R, Ross S, Thielemans K, Stearns C. 3D implementation of scatter estimation in 3D PET. IEEE Nucl Sci Symp Conf Rec. 2006;4:2142–5.
Acknowledgements
The publication of this article was supported by funds of the European Association of Nuclear Medicine (EANM).
Funding
This research was funded by the EURAMET MetroMRT project (European Commission, Belgium Seventh Framework Programme, Belgium, Grant agreement no.217257. Funded partner: ENEA). Santa Maria Goretti Hospital and ENEA were supported in part by Sirtex Medical Limited, who provided ^{90}Y solution for the experiment. No financial support was received from SIRTeX and Philips.
Author information
Affiliations
Contributions
MD’Ar conceived and designed the study, oversaw all phases of the research, analyzed data, and wrote the paper. LF and OB jointly performed ^{90}Y–PET acquisitions. ES and NP performed dose calculations using Raydose software. FM and PF performed TLD reading at ENEAIRP. PC assisted with PET acquisitions. MT assisted with evaluation of results and manuscript preparation. MC performed ^{90}Y activity measurements by assuring traceability to the ENEAINMRI primary standards of activity. VDC performed TLD calibration at ENEAINMRI. MP assisted with TLD calibration, interpretation of the results, and revision of the manuscript. MD’An, GI, and LS helped with TLD irradiation at IFO Hospital. AF and TP performed dose calculation using STRATOS, a voxelbased dose engine on the IMALYTICS Research Workstation. RP assisted with analysis and interpretation of results. All authors read and approved the final manuscript.
Corresponding author
Correspondence to Marco D’Arienzo.
Ethics declarations
Consent for publication
Not applicable
Competing interests
M.T. is an employee of Sirtex (research and development). P.C. is an employee of GE HEALTHCARE. A.F. and T. P. are employees for Philips Research. The other authors declare that they have no competing interests.
Rights and permissions
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.
About this article
Cite this article
D’Arienzo, M., Pimpinella, M., Capogni, M. et al. Phantom validation of quantitative Y90 PET/CTbased dosimetry in liver radioembolization. EJNMMI Res 7, 94 (2017). https://doi.org/10.1186/s1355001703419
Received:
Accepted:
Published:
Keywords
 ^{90}Y–PET
 Liver radioembolization
 Molecular radiotherapy
 Dosimetry
 Quantitative imaging