Monte Carlo simulations and phantom measurements towards more quantitative dosimetry and imaging in nuclear medicine

Ashok Tiwari

doi:10.17077/etd.006424

Targeted radiopharmaceutical therapy (TRT) is an explosively growing and evolving cancer treatment approach that uses therapeutic levels of biochemically targeting radioactive molecules to attach to cancer cells and subsequently irradiate them with toxic levels of radiation. This approach is currently being used clinically to treat neuroendocrine tumors, and a new drug to treat prostate cancer is likely to be approved in the coming months. In TRT, molecules that are chemically engineered to seek out and attach to proteins highly specific to a particular tumor type are labeled with a particle-emitting (beta or alpha) radionuclide and injected intravenously. Over a period of several hours, the radioactively-tagged molecules (radiopharmaceuticals) chemically bind with high specificity to cancer cells, where they typically remain for several days until the radionuclide decays and emits its high-energy particulate radiation. These particles, through Coulombic interaction with molecules in the cells, break critical molecular bonds, inflicting biological damage in its wake. The mass quantity of these therapeutic radiopharmaceuticals is small, typically 1-10 micrograms, but the number of potentially therapeutic molecules injected number on the order of 1015. Currently, clinical use of these therapeutic radiopharmaceuticals uses a “one dose fits all” approach, where regardless of patient size, sex, age, the same quantities of radioactivity are administered (the dose that was used in the clinical trial that resulted in the drug’s approval). Although this single-dose approach has been demonstrated to be clinically useful, it is by no means optimized; in fact, most patients are substantially and systematically under-dosed, and do not receive the full potential therapeutic benefit of TRT, as the clinical trials often emphasize safety over potential efficacy. There is a growing interest in personalizing TRT using quantitative imaging techniques like positron emission tomography (PET) and Single Photon Emission Computed Tomography (SPECT) to quantitatively measure the spatial and temporal biological distribution of the therapeutic radiopharmaceutical. From the in vivo imaging data, where we estimate the number of radioactive decays per second in each voxel of the image, we can calculate the amount of energy deposited both in the tumor, to determine whether therapeutic levels of radiation are being achieved, and in normal organs, to be assured that we are not inadvertently delivering toxic levels of radiation to the normal organs. Using this information, it should be possible to tailor a radiotherapeutic dosing approach that is tuned to a patient’s own tumor and normal tissue uptake, as measured by imaging that both optimizes dose to tumor, while assuring that healthy organs are not over-dosed. Several methods are under-study for absorbed dose calculations post radiopharmaceutical therapy. The current gold standard for 3D voxel-wise dosimetry is patient-specific Monte Carlo calculations using the quantitative nuclear imaging activity distributions over time (SPECT or PET) as the input data for the absorbed dose deposition map. Monte Carlo simulations are, however, highly computationally intensive if one wants to achieve low statistical noise at the voxel level. The dose point kernel (DPK) method is a more computationally efficient approach, which uses pre-calculated, radionuclide- and tissue-specific DPKs and image-based patient specific radionuclide distributions as input data to generate patient-specific absorbed dose maps. This method relies on convolving pre-calculated isotope-specific energy deposition kernels with the cumulative activity distribution, obtained from patient imaging, such that the absorbed dose map is obtained. The first project of this thesis generated Monte Carlo-based dose point kernels for clinically relevant radionuclide beta decays and monoenergetic electrons in various tissues to understand the impact of tissue type on dose point kernels. One fundamental limitation to using this DPK method in clinical dosimetry is that the dose kernels are based on analytic or numerical calculations, or Monte Carlo simulations of beta absorbed dose deposition, yet these probabilistic physics-based energy deposition calculations have not been experimentally validated. The lack of experimental validation work in the literature is primarily due to the challenge of accurately measuring absorbed dose deposition along the relatively short beta range of therapeutic radionuclides (1–10 mm) with sufficient spatial resolution to meaningfully compare with Monte Carlo simulations. As a second project of this work, physical measurements were performed using radiochromic film to measure the beta absorbed dose distributions of 90Y and 177Lu. Excellent agreement was observed between the experimental beta absorbed doses in the linear region of the radiochromic film and the GATE Monte Carlo simulations, demonstrating that radiochromic film dosimetry has sufficient sensitivity and spatial resolution to be used as a tool for measuring beta decay absorbed dose distributions. It also demonstrated, for the first time, that Monte Carlo simulations appear to be accurate to within several percent, as compared with careful physical measurements. There is increasing evidence that the use of alpha-emitters as radiolabels on some of these therapeutic radiopharmaceuticals results in even more effective treatment than beta-emitters. This is an emerging and promising cancer treatment procedure that is fundamentally different from all other cancer treatments due to the very highly localized nature of energy deposition. Monte Carlo simulations suggested that the alpha-emitters travel < 100 µm (only several cell diameters) in the tissue and they are highly potent because their ionization density is 100-1000x greater compared to therapeutic beta-emitters. This dense ionization track is highly damaging to DNA and extremely effective in tumor killing. Another project of this thesis work was to study clinically relevant therapeutic alpha-emitters energy deposition details in several tissues. This project provides a comprehensive study on α-emitting radionuclides for the purposes of its micro-dosimetric calculations for the DPKs generation and to study the impact of their Bragg peaks on overall dose distribution. Since the voxel sizes used in nuclear imaging modalities are > 1 mm, and the range of therapeutic α-emitters are substantially less than 1 mm, so the image-based dosimetry using alpha DPKs are not feasible at this point. However, these kernels may be useful to study the micro metastasized tumor dosimetry in the context of pathology slides demonstrating the microscopic distribution of cancer cells in a tumor. In nuclear imaging there has been a great excitement in the research and development of the PET scanners for monitoring radiopharmaceutical therapy especially for neuroendocrine tumors and prostate cancer. Total-body PET, where the PET detection ring is extended to multiple rings such that extend nearly fully from head to toe is a new scanner design. New prototype scanners have been designed and built by academia and now industry are demonstrating 20-30X increases in photon detection sensitivity. Major PET scanner manufacturers are beginning to offer new versions of PET systems with unprecedently large axial fields of views from 1-2 meters. This could be an ideal tool for both identifying patients eligible to TRT, as well as evaluating the results of these therapies in late-stage metastatic cancers. There has been recent significant interest in the development of a total-body PET scanners in academia for research such as uEXPLORER (United Imaging Healthcare), PennPET EXPLORER (University of Pennsylvania) and the Siemens Vision Quadra for a commercial production. However, GE Healthcare, a prominent PET system manufacturer, yet seems undecided whether to enter the TB-PET market. The latest generation of PET scanners manufactured by GE Healthcare is the Discovery MI (DMI) scanner with an AFOV up to 25cm and a crystal thickness of 25 mm. This crystal thickness is more than 20% longer than the crystals used in other commercial PET systems. As a next project, we assessed the DMI’s potential as a total-body scanner using Monte Carlo simulations. This work investigated the imaging properties of a large extended AFOV DMI scanner by looking at the performance gain with increasing AFOV through simulation. We have found that the AFOV of 2 meters with its 25 mm thick LYSO crystals resulted in ~(28-60)-fold performance gain relative to the current 4-ring DMI architecture, and interestingly, even a potential 2X sensitivity enhancement over other similarly configured TB-PET systems.

Monte Carlo simulations and phantom measurements towards more quantitative dosimetry and imaging in nuclear medicine

Files and links (1)

Abstract

Details

Metrics