TY - JOUR
T1 - Comparison of reconstructed prompt gamma emissions using maximum likelihood estimation and origin ensemble algorithms for a Compton camera system tailored to proton range monitoring
AU - Valencia Lozano, Ingrid
AU - Dedes, George
AU - Peterson, Steve
AU - Mackin, Dennis
AU - Zoglauer, Andreas
AU - Beddar, Sam
AU - Avery, Stephen
AU - Polf, Jerimy
AU - Parodi, Katia
N1 - Funding Information:
We would like to thank Prof. Dr. Peter Thirolf of the Department of Medical Physics at Ludwig-Maximilians-Universität for helpful suggestions and regular discussions.
Publisher Copyright:
© 2022
PY - 2023/5
Y1 - 2023/5
N2 - Compton-based prompt gamma (PG) imaging is being investigated by several groups as a potential solution for in vivo range monitoring in proton therapy. The performance of this technique depends on the detector system as well as the ability of the reconstruction method to obtain good spatial resolution to establish a quantitative correlation between the PG emission and the proton beam range in the patient. To evaluate the feasibility of PG imaging for range monitoring, we quantitatively evaluated the emission distributions reconstructed by a Maximum Likelihood Expectation Maximization (MLEM) and a Stochastic Origin Ensemble (SOE) algorithm. To this end, we exploit experimental and Monte Carlo (MC) simulation data acquired with the Polaris-J Compton Camera (CC) prototype. The differences between the proton beam range (RD) defined as the 80% distal dose fall-off and the PG range (RPG), obtained by fitting the distal end of the reconstructed profile with a sigmoid function, were quantified. A comparable performance of both reconstruction algorithms was found. For both experimental and simulated irradiation scenarios, the correlation between RD and RPG was within 5 mm. These values were consistent with the ground truth distance (RD-RPGg≈ 3 mm) calculated by using the expected PG emission available from MC simulation. Furthermore, shifts of 3 mm in the proton beam range were resolved with the MLEM algorithm by calculating the relative difference between the RPG for each reconstructed profile. In non-homogeneous targets, the spatial changes in the PG emission due to the different materials could not be fully resolved from the reconstructed profiles; however, the fall-off region still resembled the ground truth emission. For this scenario, the PG correlation (RD-RPG) varied from 0.1 mm to 4 mm, which is close to the ground truth correlation (3 mm). This work provides a framework for the evaluation of the range monitoring capabilities of a CC device for PG imaging. The two investigated image reconstruction algorithms showed a comparable and consistent performance for homogeneous and heterogeneous targets.
AB - Compton-based prompt gamma (PG) imaging is being investigated by several groups as a potential solution for in vivo range monitoring in proton therapy. The performance of this technique depends on the detector system as well as the ability of the reconstruction method to obtain good spatial resolution to establish a quantitative correlation between the PG emission and the proton beam range in the patient. To evaluate the feasibility of PG imaging for range monitoring, we quantitatively evaluated the emission distributions reconstructed by a Maximum Likelihood Expectation Maximization (MLEM) and a Stochastic Origin Ensemble (SOE) algorithm. To this end, we exploit experimental and Monte Carlo (MC) simulation data acquired with the Polaris-J Compton Camera (CC) prototype. The differences between the proton beam range (RD) defined as the 80% distal dose fall-off and the PG range (RPG), obtained by fitting the distal end of the reconstructed profile with a sigmoid function, were quantified. A comparable performance of both reconstruction algorithms was found. For both experimental and simulated irradiation scenarios, the correlation between RD and RPG was within 5 mm. These values were consistent with the ground truth distance (RD-RPGg≈ 3 mm) calculated by using the expected PG emission available from MC simulation. Furthermore, shifts of 3 mm in the proton beam range were resolved with the MLEM algorithm by calculating the relative difference between the RPG for each reconstructed profile. In non-homogeneous targets, the spatial changes in the PG emission due to the different materials could not be fully resolved from the reconstructed profiles; however, the fall-off region still resembled the ground truth emission. For this scenario, the PG correlation (RD-RPG) varied from 0.1 mm to 4 mm, which is close to the ground truth correlation (3 mm). This work provides a framework for the evaluation of the range monitoring capabilities of a CC device for PG imaging. The two investigated image reconstruction algorithms showed a comparable and consistent performance for homogeneous and heterogeneous targets.
KW - Compton camera
KW - MLEM
KW - Prompt gamma imaging
KW - Proton therapy
KW - Range verification
KW - SOE
UR - http://www.scopus.com/inward/record.url?scp=85132926705&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85132926705&partnerID=8YFLogxK
U2 - 10.1016/j.zemedi.2022.04.005
DO - 10.1016/j.zemedi.2022.04.005
M3 - Article
C2 - 35750591
AN - SCOPUS:85132926705
SN - 0939-3889
VL - 33
SP - 124
EP - 134
JO - Zeitschrift fur Medizinische Physik
JF - Zeitschrift fur Medizinische Physik
IS - 2
ER -