(Md. Rafiqul Islam, Tohoku University, Japan)

Radiotherapy is one of the main treatment modalities for cancer. It uses ionizing radiation to eradicate malignant tumor cells, either as a standalone treatment or in conjunction with surgery or chemotherapy. Proton therapy which has gained lots of attention in recent years, as a cancer treatment mode. Protons with different energies allow the high-dose region to conform to the tumor region. Considering the proton therapy, a major obstacle would be the uncertainties associated with the range of the proton beam, at which largest dose gradient is located. The uncertainties could be due to many factors such as; error in the estimated proton range, unexpected anatomical changes and issues with the patient or accelerator setup. A fully developed non-invasive range verification method could lead to reduced beam range uncertainties and providing a safe volume of dosage of proton therapy. Therefore, a reliable way to verify the range predicted either directly during the treatment or after the treatment, would be highly desirable for achieving the actual benefit of the proton therapy.

The goal of this thesis was to fully exploit the advantages of proton beams and enhance accuracy and precision of proton therapy by reducing range uncertainty. In order to verify the range of proton, a method for the quantitative determination of proton induced radioisotopes using a spectral analysis approach (SA) was proposed. The generated positron emitters as a result of proton interaction with major nuclei found in human tissues were quantified using the SA approach; this was investigated from both theoretical and experimental aspects. The PHITS (Particle and Heavy Ion Transport code System) Monte Carlo method was employed in this study to simulate the proton irradiation of a homogeneous, inhomogeneousslab and MIRD anthropomorphic phantoms. Positron emitting radioisotopes such as ¹⁵O,¹¹C and ¹³N were scored for different energies. It was found, among other radioisotopes the ¹³N produce a considerable peak near to the actual Bragg peak, whereas the ¹¹C and ¹⁵O peaks are far away of that. The simulation results showed the offset distances between the generated ¹³N peak and the actual Bragg peak with 1 to 2 mm for the homogeneous, inhomogeneous slab and MIRD anthropomorphicphantom studies, respectively. In the experimental study, a small scale prototype PEM (PEMGRAPH) system was used for detecting positron annihilated photons with 3D acquisition system. Experimental irradiation for a water-gel phantom wasperformed using an azimuthally varying field (AVF) cyclotron with an 80 MeV monoenergetic pencil beam. The detected gamma rays produced there are two peaks. In addition to these two peaks, a small tail peak was observed at the end of proton beam path and vanished rather quickly.

The SA was performed based on the “analysis scheme” to the dynamic time-course activity data findings from simulations and experimental studies. For both the simulation and experimental studies the SA technique was applied to the 40 frames (from 15 to 55 minutes; considering 15 minutes after proton irradiation) to quantify the positron emitting radioisotopes and also to predict the half-life of them in the different regions of interest (ROIs). The proposed scheme successfully extracted the 3D spatial distributions of positron emitting radioisotopes ¹⁵O, ¹¹C and ¹³N, respectively. The half-life of the SA extracted radioisotopes were confirmed by the ROIs analysis. In case of experimental data, the SA analysis confirmed the activity in the distal falloff region of proton induced ¹³N radioisotopes. The ROIs study also confirmed that the ¹³N radioisotope makes the highest contribution in the Bragg region. A quantitative comparison was performed between SA extracted and MC simulated radioisotopes. The results show there is no offset distance at thedistal depth region between the SA extracted and the MC simulated radioisotopes along the beam direction. It is observed that the offset distances between the SA extracted ¹³N peak and the actual Bragg peak with 1 to 2 mm, whereas the ¹¹C and¹⁵O peaks are very far away from the real Bragg peaks for the homogeneous, inhomogeneous slab and MIRD anthropomorphic phantom studies which is a good agreement with MC radioisotopes. On the other hand, when compared to the actual Bragg peak for the simulated homogeneous water-gel, the offset distance betweenthe SA extracted ¹³N peak from the experimental data is 3 mm. The simulated Bragg peak is very sharp, while the experimental ¹³N peak is rather wide, indicating reasonable agreement with the simulation results. This distinct ¹³N as well as the offset values could be used as an index for non-invasive PEM-based proton range verification using SA approach.

The SA extraction with 3D visualization showed promising results for proton induced radioisotope distribution. The proposed scheme and developed tools would be useful for the extraction and 3D visualization of positron emitting radioisotopes and in turn for proton range monitoring and verification. Future investigations into proton range monitoring for therapeutic purposes would benefit from the results obtained. It is concluded that using the developed tools, the obtained offset distance of the absolute proton range appears feasible for clinical proton fields delivered using pencil like protons beam. Further developments of the PEM system (PEMGRAPH) to facilitate an in-beam clinical trial would be recommended.

Go to AFOMP Pulse Contents/Index Page