Regarding the tissue elemental composition calculation from "Dynamic PET/CT measurements of induced positron activity in a prostate cancer patient after 50-MV photon radiation therapy"
osama Mawlawi, MD Anderson Cancer Center
21 November 2013
Jongmin Cho1,2, Geoffrey Ibbott2, and Osama Mawlawi3
1The University of Texas Graduate School of Biomedical Sciences at Houston, Houston, TX 77030, USA
2Department of Radiation Physics, The University of Texas MD Anderson Cancer Center, Houston, TX 77030, USA
3Department of Imaging Physics, The University of Texas MD Anderson Cancer Center, Houston, TX 77030, USA
We describe limitations of the approach used by Janek et al to determine tissue elemental composition from radioisotope count rate measurements following photon therapy. The approach used by Janek et al lacked three correction factors that are essential to convert the count rate to the elemental composition. We describe these factors here.
Correspondence
Dear Editor,
We read with interest the original research work of Janek et al [1] that was recently published in the European Journal of Nuclear Medicine and Molecular Imaging (2013, 3:6). Janek et al evaluated the feasibility of determining patient tissue elemental composition following photo-nuclear activation by positron emission tomography. The authors compared the initial count rate of each radioisotope from activated tissue with the International Commission on Radiation Units and Measurements (ICRU) elemental composition data to determine the accuracy of their elemental tissue composition calculation. The method they used was:
ρC12 ~ S1, ρO16~ S2, ρN14 ~ S3, (1)
where ρC12 , ρO16 , and ρN14 are the percentage composition (number of atoms/cm3 ) of 12 C, 16 O, and 14 N, respectively and S1 , S2 , and S3 are the initial count rates of 11 C, 15 O, and 13 N, respectively following activation. For the volume of interest (VOI) containing adipose tissue only, the authors calculated an S1 and an S2 of 32% and 65% while the published ICRU values are 67.7% and 31.5%, respectively. The difference in the elemental compositions between the approach used by the authors and the ICRU data is more than 100%, which can hardly be interpreted as ‘in moderate agreement,’ as indicated by Janek et al.
We believe the main reason for the discrepancy between the two measurements in the adipose tissue VOI is the omission of two fundamental factors. The first is the build-up factor 1– e-λItR , where λI and tR are the decay constant of radioisotope I and irradiation time, respectively. This factor accounts for the amount of radioactivity generated during the time of its creation. The second factor accounts for the activation rate and is expressed as ∫σTI ΦE dE, where σTI is the photo-nuclear cross-section of an atom T converting to a radioisotope I and ΦE is the photon flux. This factor accounts for the amount of radioisotope created due to the product of the cross-section and photon flux.
When these two factors are included in the model, the count rate for each element can be calculated by:
ST ~ ρT (1–e-λItR) ∫σTI ΦE dE (2)
and the percentage composition of an element T can be obtained by:
ρT ~ ST / ((1–e- λItR) ∫σTI ΦE dE) (3)
Using the suggested model, the elemental composition for the adipose VOI used by Janek et al. result in a percentage composition of 67.1% carbon, 31.5% oxygen, and 1.4% nitrogen which are in excellent agreement with the published ICRU data (67.7% carbon, 31.5% oxygen, and 0.58% nitrogen). For this calculation, a continuous irradiation time of tR= 6 min 52 s (the total irradiation time, including rotation of the gantry as described by Janek et al) and ∫σC12C11 ΦE dE= ∫σO16O15 ΦE dE= 0.4 × ∫σN14N13 ΦE dE were assumed because ∫σC12C11 dE and ∫σO16O15 dE are somewhat similar over the photon energy range of 0 ~ 50 MeV, whereas ∫σN14N13 dE are two to three times greater than ∫σC12C11 dE, according to the Experimental Nuclear Reaction Data Database [2] and Fuller [3].
In addition to the above two factors, there is a third factor that should also be taken into consideration –the perfusion-driven activity, wash-in/out factor. This factor accounts for the accumulation or removal of activity in a VOI due to wash in/out effects during scan time. For the adipose VOI, this factor is negligible because perfusion-driven wash in/out in the adipose tissue is slow [4]. However, for the bladder VOI used by Janek et al, this factor has a much bigger impact. For example, if the wash in/out factor is ignored when using our suggested model, the carbon and oxygen fractions in the bladder VOI will be 27.4% and 72.8%, respectively, which are much worse than the values reported by Janek et al (8% and 92%) compared to the ICRU data (4% and 96%). Modeling the exact amount of new radioisotopes added (wash in) to the bladder VOI is very difficult. However, when such a component is included in the model, the carbon and oxygen fractions are expected to be closer to the published ICRU values. This is primarily due to correcting the overestimation of the slowly decaying 11C (reducing the 12C fraction) from the continuous radioisotope influx to the bladder.
Recently, our group published the results of using PET/CT measurements to calculate patient tissue elemental composition after proton radiation therapy [5]. We calculated patients’ soft tissue elemental composition (C and O) for the entire activated region and compared seven regions of interest with the ICRU data. We used a previously published perfusion driven wash-out model [4,6,7], and the calculated C and O compositions in all of these regions were on average within 10% (range 3% ~ 15%) of the ICRU values. The omission of a wash-out model increased the differences between our calculations and ICRU data by as much as 30%. We think that the mathematical approach described by Cho et al [5], which includes all factors described above, should be used to improve the approach used by Janek et al to determine C, O, and N compositions after photon radiation therapy.
Authors’ contributions
Jongmin Cho first read the article by Janek et al and found the limitations of their approach. Jongmin Cho was then assisted by Geoffrey Ibbott and Osama Mawlawi in developing a methodology to improve the approach of Janek et al. All authors read and approved the final manuscript.
Acknowledgement
The authors thank Carlos Gonzalez-Lepera, Michael Gillin, and Frances Stingo at the University of Texas MD Anderson Cancer Center and Harald Paganetti at Massachusetts General Hospital and Harvard Medical School for valuable comments. This research was supported in part by the National Institutes of Health through MD Anderson’s Cancer Center Support Grant (CA016672).
References
1. Janek Straat S, Jacobsson H, Noz ME, Andreassen B, Naslund I, Jonsson C: Dynamic PET/CT measurements of induced positron activity in a prostate cancer patient after 50-MV photon radiation therapy. EJNMMI Res 2013, 3:6.
3. Fuller EG: Photonuclear reaction cross sections for 12C, 14N and 16O. Phys Rep 1985, 127:185–231.
4. Parodi K, Bortfeld T and Haberer T 2008 Comparison between in-beam and offline positron emission tomography imaging of proton and carbon ion therapeutic irradiation at synchrotron- and cyclotron-based facilities Int. J. Radiat. Oncol. Biol. Phys. 71 945–56
5. Cho J, Ibbott G, Gillin M, Gonzalez-Lepera C, Min CH, Zhu X, El Fakhri G, Paganetti H, Mawlawi O: Determination of elemental tissue composition following proton treatment using positron emission tomography. Phys Med Biol 2013, 58:3815–3835.
6. Mizuno H, Tomitani T, Kanazawa M, Kitagawa A, Pawelke J, Iseki Y, Urakabe E, Suda M, Kawano A, Iritani R, Matsushita S, Inaniwa T, Nishio T, Furukawa S, Ando K, Nakamura YK, Kanai T, Ishii K: Washout measurement of radioisotope implanted by radioactive beams in the rabbit. Phys Med Biol 2003, 48:2269–2281.
7. Tomitani T, Pawelke J, Kanazawa M, Yoshikawa K, Yoshida K, Sato M, Takami A, Koga M, Futami Y, Kitagawa A, Urakabe E, Suda M, Mizuno H, Kanai T, Matsuura H, Shinoda I, Takizawa S: Washout studies of 11C in rabbit thigh muscle implanted by secondary beams of HIMAC. Phys Med Biol 2003, 48:875–889.
Competing interests
The authors declare they have no competing interests.
Regarding the Letter to the Editor by Jongmin Cho, Geoffrey Ibbott, and Osama Mawlawi.
Sara Janek Strååt, Department of Medical Physics, Section of Imaging Physics, Karolinska University Hospital, Stockholm, SE-171 76, Sweden
21 November 2013
Thank Dr. Cho et al. for the information regarding how to correct our calculation of activated nuclides to bring them closer to the ICRU data.
Our work presented in EJNMMR was intended as a continuation of a previous published paper by Janek et al. 2006 [1] where measurements of activated tissue after 50MV photon irradiation of a piece of animal meat were performed using PET imaging. There, a method was developed where the separation of activated 11C, 15O and 13N distributions was possible using time frame measurements in PET. Comparison with tissue data from the animal meat was also performed. The aim of the report published in EJNMMR was to highlight this particular way of producing radionuclides by using high energy photon irradiation, this time performed in a patient for the first time.
As described in Janek et al 2006 (in Eqs 2 & 3), and correctly pointed out by Cho et al., the observed specific count rate in the PET scan (the number of disintegrations) of a certain radionuclide, after photon irradiation, will depend on a number of parameters. It should be noted however, that Cho et al. [2] are using proton activation in their modeling and that the nuclear processes as well as nuclear products involved here differ from that in photon activation.
Build-up factor. It is correct that the build-up is one of the fundamental factors in determining the true tissue composition. The reason that we did not do this in the paper (but this should have been discussed) is that the experiments were performed on the particular therapy machine (MM50) which at the time was running in research mode. This means that the irradiation is not as straight forward as in a clinical situation as the beam varys on and off during treatment. The irradiation time of 6 min and 52 sec in reality means irregular beam on and off times between gantry rotations. The activity and decay calculations during irradiation will hereby differ considerably depending on how and when the actual beams were delivered (and cannot be calculated assuming constant irradiation as suggested by Cho et al). However, based on earlier, very similar, phantom experiments performed at the MM50, it was estimated that this factor should affect the calculated nuclide distribution only moderately. Unfortunately, the MM50 was removed about a month after the experiments and further experiments have not been possible to perform.
Photoneutron cross section factor. As shown in Janek et al 2006, the cross sections for 12C(g,n)11C, 16O(g,n)15O and 14N(g,n)13N are similar in shape and magnitude. 12C(g,n)11C and 16O(g,n)15O are actually almost identical when integrated and multiplied for the given photon energy spectrum interval of the MM50. (This was also pointed out by Cho et al). This is why these factors do not need to be included in the calculation. The 14N(g,n)13N cross section on the other hand is about 3 times larger than the C and O cross sections but will despite this not affect the relative nuclide distribution due to the very low abundance of N in the examined VOIs.
Perfusion driven wash-out/-in factor. Studying the literature, the examination of washout components needs activity measurements to be performed directly after the irradiation as most of the biological decay processes are taking place during the very first minutes after activation. As stated in the paper, the therapy machine and the PET/CT unit were not co-located and after 7 minutes it was concluded that no washout processes were measurable from PET-data. Also, in the bladder VOI it was estimated that the urine excretion into the bladder during the period of imaging was small compared to the total urine volume, and hence that dilution of activity could be neglected. That our analysis most probably is correctly modeled without biological parameters taken into account is also indicated by the fact that
for both VOIs the decay constants of 11C and 15O were extracted with excellent goodness of fit (R2=0.9993). Nevertheless, we would like to do wash-in and wash-out studies if we get the chance to do another experiment.
Also, in the case of photon activation, the amount of activated tissue for each radionuclide will, besides the above given correction factors, also have to include the number of target atoms per volume as well as the photon fluence rate of the photon beam. Although these factors have less effect on the final nuclide distribution, they will be important to consider when performing further experiments.
References
1. Janek S, Svensson R, Jonsson C, Brahme A: Development of dose delivery verification by PET imaging of photonuclear reactions following high energy photon therapy. Phys Med Biol 2006, 51:5769-83.
2. Cho J, Ibbott G, Gillin M, Gonzalez-Lepera C, Min CH, Zhu X, El Fakhri G, Paganetti H, Mawlawi O: Determination of elemental tissue composition following proton treatment using positron emission tomogr aphy. Phys Med Biol 2013, 58:3815–3835.
Regarding the tissue elemental composition calculation from "Dynamic PET/CT measurements of induced positron activity in a prostate cancer patient after 50-MV photon radiation therapy"
21 November 2013
Jongmin Cho1,2, Geoffrey Ibbott2, and Osama Mawlawi3
1The University of Texas Graduate School of Biomedical Sciences at Houston, Houston, TX 77030, USA
2Department of Radiation Physics, The University of Texas MD Anderson Cancer Center, Houston, TX 77030, USA
3Department of Imaging Physics, The University of Texas MD Anderson Cancer Center, Houston, TX 77030, USA
Corresponding author email: omawlawi@mdanderson.org
Abstract
We describe limitations of the approach used by Janek et al to determine tissue elemental composition from radioisotope count rate measurements following photon therapy. The approach used by Janek et al lacked three correction factors that are essential to convert the count rate to the elemental composition. We describe these factors here.
Correspondence
Dear Editor,
We read with interest the original research work of Janek et al [1] that was recently published in the European Journal of Nuclear Medicine and Molecular Imaging (2013, 3:6). Janek et al evaluated the feasibility of determining patient tissue elemental composition following photo-nuclear activation by positron emission tomography. The authors compared the initial count rate of each radioisotope from activated tissue with the International Commission on Radiation Units and Measurements (ICRU) elemental composition data to determine the accuracy of their elemental tissue composition calculation. The method they used was:
ρ C12 ~ S1 , ρ O16 ~ S2 , ρ N14 ~ S3 , (1)
where ρ C12 , ρ O16 , and ρ N14 are the percentage composition (number of atoms/cm3 ) of 12 C, 16 O, and 14 N, respectively and S1 , S2 , and S3 are the initial count rates of 11 C, 15 O, and 13 N, respectively following activation. For the volume of interest (VOI) containing adipose tissue only, the authors calculated an S1 and an S2 of 32% and 65% while the published ICRU values are 67.7% and 31.5%, respectively. The difference in the elemental compositions between the approach used by the authors and the ICRU data is more than 100%, which can hardly be interpreted as ‘in moderate agreement,’ as indicated by Janek et al.
We believe the main reason for the discrepancy between the two measurements in the adipose tissue VOI is the omission of two fundamental factors. The first is the build-up factor 1– e-λI tR , where λ I and tR are the decay constant of radioisotope I and irradiation time, respectively. This factor accounts for the amount of radioactivity generated during the time of its creation. The second factor accounts for the activation rate and is expressed as ∫σ TI Φ E dE, where σ TI is the photo-nuclear cross-section of an atom T converting to a radioisotope I and Φ E is the photon flux. This factor accounts for the amount of radioisotope created due to the product of the cross-section and photon flux.
When these two factors are included in the model, the count rate for each element can be calculated by:
ST ~ ρ T (1– e-λ I tR ) ∫σ TI Φ E dE (2)
and the percentage composition of an element T can be obtained by:
ρ T ~ ST / ((1– e- λ I tR ) ∫σ TI Φ E dE) (3)
Using the suggested model, the elemental composition for the adipose VOI used by Janek et al. result in a percentage composition of 67.1% carbon, 31.5% oxygen, and 1.4% nitrogen which are in excellent agreement with the published ICRU data (67.7% carbon, 31.5% oxygen, and 0.58% nitrogen). For this calculation, a continuous irradiation time of tR = 6 min 52 s (the total irradiation time, including rotation of the gantry as described by Janek et al) and ∫σ C12C11 Φ E dE = ∫σ O16O15 Φ E dE = 0.4 × ∫σ N14N13 Φ E dE were assumed because ∫σ C12C11 dE and ∫σ O16O15 dE are somewhat similar over the photon energy range of 0 ~ 50 MeV, whereas ∫σ N14N13 dE are two to three times greater than ∫σ C12C11 dE, according to the Experimental Nuclear Reaction Data Database [2] and Fuller [3].
In addition to the above two factors, there is a third factor that should also be taken into consideration –the perfusion-driven activity, wash-in/out factor. This factor accounts for the accumulation or removal of activity in a VOI due to wash in/out effects during scan time. For the adipose VOI, this factor is negligible because perfusion-driven wash in/out in the adipose tissue is slow [4]. However, for the bladder VOI used by Janek et al, this factor has a much bigger impact. For example, if the wash in/out factor is ignored when using our suggested model, the carbon and oxygen fractions in the bladder VOI will be 27.4% and 72.8%, respectively, which are much worse than the values reported by Janek et al (8% and 92%) compared to the ICRU data (4% and 96%). Modeling the exact amount of new radioisotopes added (wash in) to the bladder VOI is very difficult. However, when such a component is included in the model, the carbon and oxygen fractions are expected to be closer to the published ICRU values. This is primarily due to correcting the overestimation of the slowly decaying 11C (reducing the 12C fraction) from the continuous radioisotope influx to the bladder.
Recently, our group published the results of using PET/CT measurements to calculate patient tissue elemental composition after proton radiation therapy [5]. We calculated patients’ soft tissue elemental composition (C and O) for the entire activated region and compared seven regions of interest with the ICRU data. We used a previously published perfusion driven wash-out model [4,6,7], and the calculated C and O compositions in all of these regions were on average within 10% (range 3% ~ 15%) of the ICRU values. The omission of a wash-out model increased the differences between our calculations and ICRU data by as much as 30%. We think that the mathematical approach described by Cho et al [5], which includes all factors described above, should be used to improve the approach used by Janek et al to determine C, O, and N compositions after photon radiation therapy.
Authors’ contributions
Jongmin Cho first read the article by Janek et al and found the limitations of their approach. Jongmin Cho was then assisted by Geoffrey Ibbott and Osama Mawlawi in developing a methodology to improve the approach of Janek et al. All authors read and approved the final manuscript.
Acknowledgement
The authors thank Carlos Gonzalez-Lepera, Michael Gillin, and Frances Stingo at the University of Texas MD Anderson Cancer Center and Harald Paganetti at Massachusetts General Hospital and Harvard Medical School for valuable comments. This research was supported in part by the National Institutes of Health through MD Anderson’s Cancer Center Support Grant (CA016672).
References
1. Janek Straat S, Jacobsson H, Noz ME, Andreassen B, Naslund I, Jonsson C: Dynamic PET/CT measurements of induced positron activity in a prostate cancer patient after 50-MV photon radiation therapy. EJNMMI Res 2013, 3:6.
2. Experimental Nuclear Reaction Data (EXFOR) Database. [http://www.nndc.bnl.gov/exfor/exfor00.htm]. Accessed May 2013.
3. Fuller EG: Photonuclear reaction cross sections for 12C, 14N and 16O. Phys Rep 1985, 127:185–231.
4. Parodi K, Bortfeld T and Haberer T 2008 Comparison between in-beam and offline positron emission tomography imaging of proton and carbon ion therapeutic irradiation at synchrotron- and cyclotron-based facilities Int. J. Radiat. Oncol. Biol. Phys. 71 945–56
5. Cho J, Ibbott G, Gillin M, Gonzalez-Lepera C, Min CH, Zhu X, El Fakhri G, Paganetti H, Mawlawi O: Determination of elemental tissue composition following proton treatment using positron emission tomography. Phys Med Biol 2013, 58:3815–3835.
6. Mizuno H, Tomitani T, Kanazawa M, Kitagawa A, Pawelke J, Iseki Y, Urakabe E, Suda M, Kawano A, Iritani R, Matsushita S, Inaniwa T, Nishio T, Furukawa S, Ando K, Nakamura YK, Kanai T, Ishii K: Washout measurement of radioisotope implanted by radioactive beams in the rabbit. Phys Med Biol 2003, 48:2269–2281.
7. Tomitani T, Pawelke J, Kanazawa M, Yoshikawa K, Yoshida K, Sato M, Takami A, Koga M, Futami Y, Kitagawa A, Urakabe E, Suda M, Mizuno H, Kanai T, Matsuura H, Shinoda I, Takizawa S: Washout studies of 11C in rabbit thigh muscle implanted by secondary beams of HIMAC. Phys Med Biol 2003, 48:875–889.
Competing interests
The authors declare they have no competing interests.Regarding the Letter to the Editor by Jongmin Cho, Geoffrey Ibbott, and Osama Mawlawi.
21 November 2013
Thank Dr. Cho et al. for the information regarding how to correct our calculation of activated nuclides to bring them closer to the ICRU data.
Our work presented in EJNMMR was intended as a continuation of a previous published paper by Janek et al. 2006 [1] where measurements of activated tissue after 50MV photon irradiation of a piece of animal meat were performed using PET imaging. There, a method was developed where the separation of activated 11C, 15O and 13N distributions was possible using time frame measurements in PET. Comparison with tissue data from the animal meat was also performed. The aim of the report published in EJNMMR was to highlight this particular way of producing radionuclides by using high energy photon irradiation, this time performed in a patient for the first time.
As described in Janek et al 2006 (in Eqs 2 & 3), and correctly pointed out by Cho et al., the observed specific count rate in the PET scan (the number of disintegrations) of a certain radionuclide, after photon irradiation, will depend on a number of parameters. It should be noted however, that Cho et al. [2] are using proton activation in their modeling and that the nuclear processes as well as nuclear products involved here differ from that in photon activation.
Build-up factor. It is correct that the build-up is one of the fundamental factors in determining the true tissue composition. The reason that we did not do this in the paper (but this should have been discussed) is that the experiments were performed on the particular therapy machine (MM50) which at the time was running in research mode. This means that the irradiation is not as straight forward as in a clinical situation as the beam varys on and off during treatment. The irradiation time of 6 min and 52 sec in reality means irregular beam on and off times between gantry rotations. The activity and decay calculations during irradiation will hereby differ considerably depending on how and when the actual beams were delivered (and cannot be calculated assuming constant irradiation as suggested by Cho et al). However, based on earlier, very similar, phantom experiments performed at the MM50, it was estimated that this factor should affect the calculated nuclide distribution only moderately. Unfortunately, the MM50 was removed about a month after the experiments and further experiments have not been possible to perform.
Photoneutron cross section factor. As shown in Janek et al 2006, the cross sections for 12C(g,n)11C, 16O(g,n)15O and 14N(g,n)13N are similar in shape and magnitude. 12C(g,n)11C and 16O(g,n)15O are actually almost identical when integrated and multiplied for the given photon energy spectrum interval of the MM50. (This was also pointed out by Cho et al). This is why these factors do not need to be included in the calculation. The 14N(g,n)13N cross section on the other hand is about 3 times larger than the C and O cross sections but will despite this not affect the relative nuclide distribution due to the very low abundance of N in the examined VOIs.
Perfusion driven wash-out/-in factor. Studying the literature, the examination of washout components needs activity measurements to be performed directly after the irradiation as most of the biological decay processes are taking place during the very first minutes after activation. As stated in the paper, the therapy machine and the PET/CT unit were not co-located and after 7 minutes it was concluded that no washout processes were measurable from PET-data. Also, in the bladder VOI it was estimated that the urine excretion into the bladder during the period of imaging was small compared to the total urine volume, and hence that dilution of activity could be neglected. That our analysis most probably is correctly modeled without biological parameters taken into account is also indicated by the fact that
for both VOIs the decay constants of 11C and 15O were extracted with excellent goodness of fit (R2=0.9993). Nevertheless, we would like to do wash-in and wash-out studies if we get the chance to do another experiment.
Also, in the case of photon activation, the amount of activated tissue for each radionuclide will, besides the above given correction factors, also have to include the number of target atoms per volume as well as the photon fluence rate of the photon beam. Although these factors have less effect on the final nuclide distribution, they will be important to consider when performing further experiments.
References
1. Janek S, Svensson R, Jonsson C, Brahme A: Development of dose delivery verification by PET imaging of photonuclear reactions following high energy photon therapy. Phys Med Biol 2006, 51:5769-83.
2. Cho J, Ibbott G, Gillin M, Gonzalez-Lepera C, Min CH, Zhu X, El Fakhri G, Paganetti H, Mawlawi O: Determination of elemental tissue composition following proton treatment using positron emission tomogr aphy. Phys Med Biol 2013, 58:3815–3835.
Competing interests
None declared