Development of patient-specific 3D models from histopathological samples for applications in radiation therapy

A library of proton lineal energy spectra spanning the full range of clinically relevant energies

PURPOSE

In locations where the proton energy spectrum is broad, lineal energy spectrum-based proton biological effects models may be more accurate than dose-averaged linear energy transfer (LETd) based models. However, the development of microdosimetric spectrum-based biological effects models is hampered by the extreme computational difficulty of calculating microdosimetric spectra. Given a precomputed library of lineal energy spectra for monoenergetic protons, a weighted summation can be performed which yields the lineal energy spectrum of an arbitrary polyenergetic beam. Using this approach, lineal energy spectra can be rapidly calculated on a voxel-by-voxel level.

ACQUISITION AND VALIDATION METHODS

Monoenergetic proton tracks generated using Geant4-DNA were imported into SuperTrack, a GPU-accelerated software for calculation of microdosimetric spectra. Libraries of proton lineal energy spectra which span the energy range of 0-300 MeV were computed. The libraries were validated by comparison to Monte Carlo calculations in the literature, as well as lineal energy spectra measured experimentally with a tissue equivalent proportional counter.

DATA FORMAT AND USAGE NOTES

The lineal energy libraries have been made available in three data formats, two are plain-text, .csv and .les, and one binary encoded, root. Library files include the lineal energy bin abscissa in keV per micron and the unnormalized number of counts occurring within that bin. A computational technique for summation of the library files to yield the lineal energy of a polyenergetic beam is described in this work.

POTENTIAL APPLICATIONS

The lineal energy libraries can be used to rapidly determine the lineal energy spectra at the location of cell cultures for in-vitro experiments and in each voxel of a treatment plan for in-vivo outcome modelling. These libraries have already been incorporated into RayStation 2023B-IonPG for lineal energy spectra calculation and we anticipate they will be incorporated into further dose calculation engines and Monte Carlo toolkits.

Relative biological effectiveness of clinically relevant photon energies for the survival of human colorectal, cervical, and prostate cancer cell lines

Objective.Relative biological effectiveness (RBE) differs between radiation qualities. However, an RBE of 1.0 has been established for photons regardless of the wide range of photon energies used clinically, the lack of reproducibility in radiobiological studies, and outdated reference energies used in the experimental literature. Moreover, due to intrinsic radiosensitivity, different cancer types have different responses to radiation. This study aimed to characterize the RBE of clinically relevant high and low photon energiesin vitrofor three human cancer cell lines: HCT116 (colon), HeLa (cervix), and PC3 (prostate).Approach.Experiments were conducted following dosimetry protocols provided by the American Association of Physicists in Medicine. Cells were irradiated with 6 MV x-rays, an192Ir brachytherapy source, 225 kVp and 50 kVp x-rays. Cell survival post-irradiation was assessed using the clonogenic assay. Survival fractions were fitted using the linear quadratic model, and survival curves were generated for RBE calculations.Main results.Cell killing was more efficient with decreasing photon energy. Using 225 kVp x-rays as the reference, the HCT116 RBESF0.1for 6 MV x-rays,192Ir, and 50 kVp x-rays were 0.89 ± 0.03, 0.95 ± 0.03, and 1.24 ± 0.04; the HeLa RBESF0.1were 0.95 ± 0.04, 0.97 ± 0.05, and 1.09 ± 0.03, and the PC3 RBESF0.1were 0.84 ± 0.01, 0.84 ± 0.01, and 1.13 ± 0.02, respectively. HeLa and PC3 cells had varying radiosensitivity when irradiated with 225 and 50 kVp x-rays.Significance.This difference supports the notion that RBE may not be 1.0 for all photons through experimental investigations that employed precise dosimetry. It highlights that different cancer types may not have identical responses to the same irradiation quality. Additionally, the RBE of clinically relevant photons was updated to the reference energy of 225 kVp x-rays.

GPU-accelerated calculation of proton microdosimetric spectra as a function of target size, proton energy, and bounding volume size

Objective.Shortcomings of dose-averaged linear energy transfer (LETD), the quantity which is most commonly used to quantify proton relative biological effectiveness, have long been recognized. Microdosimetric spectra may overcome the limitations of LETDbut are extremely computationally demanding to calculate. A systematic library of lineal energy spectra for monoenergetic protons could enable rapid determination of microdosimetric spectra in a clinical environment. The objective of this work was to calculate and validate such a library of lineal energy spectra.Approach. SuperTrack, a GPU-accelerated CUDA/C++ based application, was developed to superimpose tracks calculated using Geant4 onto targets of interest and to compute microdosimetric spectra. Lineal energy spectra of protons with energies from 0.1 to 100 MeV were determined in spherical targets of diameters from 1 nm to 10μm and in bounding voxels with side lengths of 5μm and 3 mm.Main results.Compared to an analogous Geant4-based application, SuperTrack is up to 3500 times more computationally efficient if each track is resampled 1000 times. Dose spectra of lineal energy and dose-mean lineal energy calculated with SuperTrack were consistent with values published in the literature and with comparison to a Geant4 simulation. Using SuperTrack, we developed the largest known library of proton microdosimetric spectra as a function of primary proton energy, target size, and bounding volume size.Significance. SuperTrack greatly increases the computational efficiency of the calculation of microdosimetric spectra. The elevated lineal energy observed in a 3 mm side length bounding volume suggests that lineal energy spectra determined experimentally or computed in small bounding volumes may not be representative of the lineal energy spectra in voxels of a dose calculation grid. The library of lineal energy spectra calculated in this work could be integrated with a treatment planning system for rapid determination of lineal energy spectra in patient geometries.

Patient-specific microdosimetry: a proof of concept

Microscopic energy deposition distributions from ionizing radiation are used to predict the biological effects of an irradiation and vary depending on biological target size. Ionizing radiation is thought to kill cells or inhibit cell cycling mainly by damaging DNA in the cell nucleus. The size of cells and nuclei depends on tissue type, cell cycle, and malignancy, all of which vary between patients. The aim of this study was to develop methods to perform patient-specific microdosimetry, that being, determining microdosimetric quantities in volumes that correspond to the sizes of cells and nuclei observed in a patient's tissue. A histopathological sample extracted from a stage I lung adenocarcinoma patient was analyzed. A pouring simulation was used to generate a three-dimensional tissue model from cell and nucleus size information determined from the histopathological sample. Microdosimetric distributions including f(y) and d(y) were determined for Co-60,Ir-192,Yb-169 and I-125 in a patient-specific model containing a distribution of cell and nucleus sizes. Fixed radius models and a summation method (where f(y) from many fixed radii models are summed) were compared to the full patient-specific model to evaluate their suitability for fast determination of patient-specific microdosimetric parameters. Fixed radius models do not provide a close approximation of the full patient-specific model y ̅_f or y ̅_d for the lower energy sources investigated, Yb-169 and I-125. The higher energy sources investigated, Co-60 and Ir-192 are less sensitive to target size variation than Yb-169 and I-125. A summation method yields the most accurate approximation of the full model d(y) for all radioisotopes investigated. A summation method allows for the computation of patient-specific microdosimetric distributions with the computing power of a personal computer. With appropriate biological inputs the microdosimetric distributions computed using these methods can yield a patient-specific relative biological effectiveness as part of a multiscale treatment planning approach.