安井 啓祐

This study aimed to evaluate the performance for answering the Japanese medical physicist examination and providing the benchmark of knowledge about medical physics in language-generative AI with large language model. We used questions from Japan's 2018, 2019, 2020, 2021 and 2022 medical physicist board examinations, which covered various question types, including multiple-choice questions, and mainly focused on general medicine and medical physics. ChatGPT-3.5 and ChatGPT-4.0 (OpenAI) were used. We compared the AI-based answers with the correct ones. The average accuracy rates were 42.2 ± 2.5% (ChatGPT-3.5) and 72.7 ± 2.6% (ChatGPT-4), showing that ChatGPT-4 was more accurate than ChatGPT-3.5 [all categories (except for radiation-related laws and recommendations/medical ethics): p value < 0.05]. Even with the ChatGPT model with higher accuracy, the accuracy rates were less than 60% in two categories; radiation metrology (55.6%), and radiation-related laws and recommendations/medical ethics (40.0%). These data provide the benchmark for knowledge about medical physics in ChatGPT and can be utilized as basic data for the development of various medical physics tools using ChatGPT (e.g., radiation therapy support tools with Japanese input).

Background This study evaluates dose perturbations caused by nonradioactive seeds in clinical cases by employing treatment planning system-based Monte Carlo (TPS-MC) simulation. Methodology We investigated dose perturbation using a water-equivalent phantom and 20 clinical cases of prostate cancer (10 cases with seeds and 10 cases without seeds) treated at Fujita Health University Hospital, Japan. First, dose calculations for a simple geometry were performed using the RayStation MC algorithm for a water-equivalent phantom with and without a seed. TPS-independent Monte Carlo (full-MC) simulations and film measurements were conducted to verify the accuracy of TPS-MC simulation. Subsequently, dose calculations using TPS-MC were performed on CT images of clinical cases of prostate cancer with and without seeds, and the dose distributions were compared. Results In clinical cases, dose calculations using MC simulations revealed hotspots around the seeds. However, the size of the hotspot was not correlated with the number of seeds. The maximum difference in dose perturbation between TPS-MC simulations and film measurements was 3.9%, whereas that between TPS-MC simulations and full-MC simulations was 3.7%. The dose error of TPS-MC was negligible for multiple beams or rotational irradiation. Conclusions Hotspots were observed in dose calculations using TPS-MC performed on CT images of clinical cases with seeds. The dose calculation accuracy around the seeds using TPS-MC simulations was comparable to that of film measurements and full-MC simulations, with differences within 3.9%. Although the clinical impact of hotspots occurring around the seeds is minimal, utilizing MC simulations on TPSs can be beneficial to verify their presence.

Deep learning-based CT image reconstruction (DLR) is a state-of-the-art method for obtaining CT images. This study aimed to evaluate the usefulness of DLR in radiotherapy. Data were acquired using a large-bore CT system and an electron density phantom for radiotherapy. We compared the CT values, image noise, and CT value-to-electron density conversion table of DLR and hybrid iterative reconstruction (H-IR) for various doses. Further, we evaluated three DLR reconstruction strength patterns (Mild, Standard, and Strong). The variations of CT values of DLR and H-IR were large at low doses, and the difference in average CT values was insignificant with less than 10 HU at doses of 100 mAs and above. DLR showed less change in CT values and smaller image noise relative to H-IR. The noise-reduction effect was particularly large in the low-dose region. The difference in image noise between DLR Mild and Standard/Strong was large, suggesting the usefulness of reconstruction intensities higher than Mild. DLR showed stable CT values and low image noise for various materials, even at low doses; particularly for Standard or Strong, the reduction in image noise was significant. These findings indicate the usefulness of DLR in treatment planning using large-bore CT systems.