Search

Recent developments in magnetic resonance (MR) to synthetic computed tomography (sCT) conversion have shown that treatment planning is possible without an initial planning CT. Promising conversion results have been demonstrated recently using conditional generative adversarial networks (cGANs). However, the performance is generally only tested on images from one MR scanner, which neglects the potential of neural networks to find general high-level abstract features. In this study, we explored the generalizability of the generator models, trained on a single field strength scanner, to data acquired with higher field strengths. T2-weighted 0.35T MRIs and CTs from 51 patients treated for prostate (40) and cervical cancer (11) were included. 25 of them were used to train four different generators (SE-ResNet, DenseNet, U-Net, and Embedded Net). Further, an ensemble model was created from the four network outputs. The models were validated on 16 patients from a 0.35T MR scanner. Further, the trained models were tested on the Gold Atlas dataset, containing T2-weighted MR scans of different field strengths; 1.5T(7) and 3T(12), and 10 patients from the 0.35T scanner. The sCTs were dosimetrically compared using clinical VMAT plans for all test patients. For the same scanner (0.35T), the results from the different models were comparable on the test set, with only minor differences in the mean absolute error (MAE) (35-51HU body). Similar results were obtained for conversions of 3T GE Signa and the 3T GE Discovery images (40-62HU MAE) for three of the models. However, larger differences were observed for the 1.5T images (48-65HU MAE). The overall best model was found to be the ensemble model. All dose differences were below 1%. This study shows that it is possible to generalize models trained on images of one scanner to other scanners and different field strengths. The best metric results were achieved by the combination of all networks.

Purpose: The dose response of Gafchromic EBT3 films exposed to proton beams depends on the dose, and additionally on the beam quality, which is often quantified with the linear energy transfer (LET) and, hence, also referred to as LET quenching. Fundamentally different methods to determine correction factors for this LET quenching effect have been reported in literature and a new method using the local proton fluence distribution differential in LET is presented. This method was exploited to investigate whether a more practical correction based on the dose- or fluence-averaged LET is feasible in a variety of clinically possible beam arrangements. Methods: The relative effectiveness (RE) was characterized within a high LET spread-out Bragg peak (SOBP) in water made up by the six lowest available energies (62.4-67.5 MeV, configuration b(1)) resulting in one of the highest clinically feasible dose-averaged LET distributions. Additionally, two beams were measured where a low LET proton beam (252.7 MeV) was superimposed on b1, which contributed either 50% of the initial particle fluence or 50% of the dose in the SOBP, referred to as configuration b(2) and b(3), respectively. The proton LET spectrum was simulated with GATE/Geant4 at all measurement positions. The net optical density change differential in LET was integrated over the local proton spectrum to calculate the net optical density and therefrom the beam quality correction factor. The LET dependence of the film response was accounted for by an LET dependence of one of the three parameters in the calibration function and was determined from inverse optimization using measurement b(1). This method was then validated on the measurements of b(2) and b(3) and subsequently used to calculate the RE at 900 positions in nine clinically relevant beams. The extrapolated RE set was used to derive a simple linear correction function based on dose-averaged LET (L-d) and verify the validity in all points of the comprehensive RE

Medical physics has made considerable contributions to recent advances in radiation oncology. Medical physicists are key players in the clinical and scientific radiation oncology context due to their unique skill sets, flexibility, clinical involvement and intrinsic translational character. The continuing development and widespread adoption of "high-tech" radiotherapy has led to an increased need for medical physics involvement. More recently, our field is rapidly changing towards an era of "precision oncology". These changes have opened new challenges for the definition of the professional and scientific roles and responsibilities of medical physicists. In this paper, we have identified four grand challenges of medical physics in radiation oncology: (1) improving target volume definition, (2) adoption of artificial intelligence and automation, (3) development of predictive models of biological effects for precision medicine, and (4) need for leadership. New visions and suggestions to orientate medical physics to successfully face these new challenges are summarized. We foresee that the scientific and professional challenges of our times are pushing medical physicists to accelerate toward multidisciplinarity. Medical physicists are expected to innovatively drive interactions and collaborations with other specialists outside radiation oncology while the radiation physics core will remain central. Medical physicists will retain strong and pivotal roles in quality, safety and in managing ever more complex technologies. The new challenges will require medical physicists to continuously update skills and innovate education, adapt curricula to include new fields, reinforce multi-disciplinary attitude and spirit of innovation. These challenges require visionary and open leadership, which is able to merge established roles with the exciting new fields where medical physics should increasingly contribute. (C) 2020 Elsevier B.V. All rights reserved.