{"title":"[A Comprehensive Evaluation of the Performance of Large Language Models on the Japanese National Examination for Radiological Technologists].","authors":"Taku Kuramoto, Shinya Takarabe, Rikito Katsuda, Nao Ichikawa, Hiroyuki Tsushima","doi":"10.6009/jjrt.26-1639","DOIUrl":"10.6009/jjrt.26-1639","url":null,"abstract":"<p><strong>Purpose: </strong>This study aimed to evaluate the performance of several large language models (LLMs) on the Japanese National Examination for Radiological Technologists and to characterize their performance profiles.</p><p><strong>Methods: </strong>We utilized a dataset comprising questions from 12 consecutive years of the national examination (the 65th to the 76th iterations), excluding items that were officially retracted or deemed inappropriate. 5 distinct LLMs (ChatGPT-3.5, Gemini 2.5 Flash, Gemini 2.5 Pro, Copilot, and Claude Sonnet 4) were prompted to answer these questions. The accuracy of each LLM was calculated for the entire question set and for subsets categorized by question format.</p><p><strong>Results: </strong>Across the entire examination and within numerous subject areas, Gemini 2.5 Pro achieved the highest accuracy. An analysis by question format revealed a general trend: most LLMs demonstrated superior performance on text-based questions, followed by calculation-based and then image-based questions. However, some models exhibited notably strong performance specifically on calculation-based problems.</p><p><strong>Conclusion: </strong>While LLMs demonstrate considerable proficiency in answering questions from the National Examination for Radiological Technologists, our findings also reveal significant limitations, particularly in their capacity to interpret image-based problems. This study highlights both the potential utility and the current challenges of leveraging LLMs as supplementary learning tools for this professional certification examination.</p>","PeriodicalId":74309,"journal":{"name":"Nihon Hoshasen Gijutsu Gakkai zasshi","volume":"82 5","pages":""},"PeriodicalIF":0.0,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147679199","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"[Dose Management and Estimation Model Development Using a Dose Area Product Meter during kV Planar Imaging in Image-guided Radiotherapy].","authors":"Hideharu Miura, Toshiya Okazue, Kenji Kanemoto","doi":"10.6009/jjrt.26-1614","DOIUrl":"10.6009/jjrt.26-1614","url":null,"abstract":"<p><strong>Purpose: </strong>This study aimed to measure patient exposure dose during kilovoltage (kV) planar imaging in image-guided radiotherapy (IGRT) using a dose area product (DAP) meter and to estimate the DAP under unmeasured imaging conditions using two-dimensional cubic interpolation.</p><p><strong>Methods: </strong>Dose measurements during kV planar imaging were performed using a DAP meter (VacuDAP duo; VacuTec Meßtechnik, Dresden, Germany) attached to the kV X-ray tube of the On-Board Imager mounted on a TrueBeam linear accelerator (Varian Medical Systems, Palo Alto, CA, USA). DAP was measured under various imaging conditions, with tube voltages ranging from 60 to 140 kV and tube currents from 2.5 to 20 mAs. A two-dimensional cubic interpolation based on the measured data was implemented to estimate the DAP for unmeasured imaging conditions.</p><p><strong>Results: </strong>The deviation from linearity was relatively larger under low-dose conditions, with a maximum deviation of 1.8% at 100 kV and 2.5 mAs. Nevertheless, the overall DAP for each mAs value exhibited high linearity (R<sup>2</sup>=1.00). The relative error of the estimated DAP was within 2.0%.</p><p><strong>Conclusion: </strong>The DAP meter was demonstrated to be effective for dose management in kV planar imaging in IGRT, and use of the two-dimensional cubic interpolation enabled accurate dose estimation even under unmeasured imaging conditions.</p>","PeriodicalId":74309,"journal":{"name":"Nihon Hoshasen Gijutsu Gakkai zasshi","volume":"82 4","pages":""},"PeriodicalIF":0.0,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147357993","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"[Development and Accuracy Evaluation of the Beam Width Using a Tungsten Ring in a Wide-beam CT System].","authors":"Takuma Hayashi, Atsushi Fukuda, Nao Ichikawa, Kosuke Matsubara","doi":"10.6009/jjrt.26-0302","DOIUrl":"https://doi.org/10.6009/jjrt.26-0302","url":null,"abstract":"","PeriodicalId":74309,"journal":{"name":"Nihon Hoshasen Gijutsu Gakkai zasshi","volume":"82 3","pages":"93-95"},"PeriodicalIF":0.0,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147500934","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"[The Forefront of Image Information Sharing between Medical Facilities].","authors":"","doi":"10.6009/jjrt.26-0609","DOIUrl":"https://doi.org/10.6009/jjrt.26-0609","url":null,"abstract":"","PeriodicalId":74309,"journal":{"name":"Nihon Hoshasen Gijutsu Gakkai zasshi","volume":"82 6","pages":""},"PeriodicalIF":0.0,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148297694","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"[Quantitative Assessment of Artifacts Specific to Compressed Sensing-sensitivity Encoding (CS-SENSE) MRI].","authors":"Shun Ono, Kenichi Sakai, Masaki Ohkubo, Tatsuya Kondo","doi":"10.6009/jjrt.26-1641","DOIUrl":"10.6009/jjrt.26-1641","url":null,"abstract":"<p><strong>Purpose: </strong>Compressed sensing-sensitivity encoding (CS-SENSE), which combines sensitivity encoding (SENSE) with compressed sensing, is utilized to reduce magnetic resonance imaging (MRI) acquisition time. The aim of this study was to propose a novel quantitative index for evaluating the curved, stripe-like artifacts that specifically appear in the phase-encoding direction of images acquired using the CS-SENSE technique.</p><p><strong>Methods: </strong>MRI scans were performed using a 3.0-T MR system and a uniform phantom. The reduction factor (R) in the CS-SENSE protocol, which controls the degree of data undersampling, was varied from 2 to 5. To extract artifact components, band-pass filtering was applied to each signal-intensity profile in the phase-encoding direction. The standard deviation of the filtered image was normalized by the mean signal intensity of the original image, yielding an artifact quantitative index.</p><p><strong>Results: </strong>The artifact quantitative index increased progressively with higher R values, with a pronounced increase observed at R≥4. This trend was consistent with visual assessments, indicating that the index appropriately reflected the severity of artifact occurrence.</p><p><strong>Conclusion: </strong>The proposed index enables quantitative evaluation of artifacts specific to CS-SENSE images and may be useful for determining optimal R settings.</p>","PeriodicalId":74309,"journal":{"name":"Nihon Hoshasen Gijutsu Gakkai zasshi","volume":"82 6","pages":""},"PeriodicalIF":0.0,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147965935","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"[Anemia Assessment Using Aortic Attenuation on Dual-energy CT].","authors":"Ryoya Kato, Tadanori Takata, Keiri Kawara, Wataru Mitsui, Takashi Hamaguchi, Kota Ikarashi, Hiroki Kawashima","doi":"10.6009/jjrt.26-1611","DOIUrl":"https://doi.org/10.6009/jjrt.26-1611","url":null,"abstract":"<p><strong>Purpose: </strong>This study aimed to evaluate the relationship between computed tomography (CT) attenuation values obtained from virtual monochromatic images (VMI) and virtual non-contrast (VNC) images generated by dual-energy CT (DECT) and hemoglobin concentration (Hb), and to investigate the feasibility of anemia assessment using DECT.</p><p><strong>Methods: </strong>A total of 400 patients (200 males and 200 females) who underwent abdominal contrast-enhanced CT with dual-energy mode were included. Non-contrast VMIs at 70 keV (diaphragmatic dome and celiac artery levels) and equilibrium-phase VNC images (celiac artery level) were generated, and CT attenuation values were measured in the aortic lumen. Correlations between CT values and Hb levels, as well as receiver operating characteristic (ROC) analyses, were performed.</p><p><strong>Results: </strong>For non-contrast VMI, correlation coefficients at the diaphragmatic dome were 0.582 in males and 0.562 in females, and at the celiac artery level were 0.829 in males and 0.860 in females. For VNC images, the coefficients were 0.628 in males and 0.545 in females. ROC analysis of non-contrast VMI at the celiac artery level showed the highest diagnostic performance, with area under the curve values of 0.914 in males and 0.920 in females.</p><p><strong>Conclusion: </strong>CT attenuation values on non-contrast VMI at the celiac artery level demonstrated strong correlations with Hb and support the clinical utility of DECT for noninvasive anemia assessment.</p>","PeriodicalId":74309,"journal":{"name":"Nihon Hoshasen Gijutsu Gakkai zasshi","volume":"82 9","pages":""},"PeriodicalIF":0.0,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148665238","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"[Impact of Source Position Error in Tandem-ovoid Applicator on Dose Distribution in Cervical Cancer Brachytherapy].","authors":"Junichi Akagawa, Tatsuhiro Sera, Yuki Yuasa","doi":"10.6009/jjrt.26-1651","DOIUrl":"10.6009/jjrt.26-1651","url":null,"abstract":"<p><strong>Purpose: </strong>This study aimed to determine the discrepancy between planned and actual source positions in the Geneva applicator for cervical cancer brachy therapy and to evaluate its impact on dose distribution.</p><p><strong>Methods: </strong>The Geneva applicator was placed on a computed tomography (CT) couch top, and CT images were acquired at each source dwell position to determine the actual source locations. The discrepancy between the actual source positions and the planned source positions reconstructed using the applicator modeling (AM) method was calculated. These discrepancies were then incorporated into the treatment plans of 13 patients who underwent brachytherapy for cervical cancer at our institution. Dose metrics, including D<sub>90%</sub> for the high-risk clinical target volume (HR-CTV) and D<sub>2cc</sub> for the bladder, rectum, small bowel, and large bowel, were evaluated between the original and the modified treatment plans.</p><p><strong>Results: </strong>The overall three-dimensional displacement of the source was 0.86±0.15 mm. Regarding dose metrics, the D<sub>90%</sub> for the HR-CTV was significantly higher (mean increase, 1.38%) in the plans reflecting the actual source positions, while the D<sub>2cc</sub> for the bladder and rectum was significantly lower (mean decrease, 2.29% and 2.57%, respectively) compared to the original plans.</p><p><strong>Conclusion: </strong>The results suggest that incorporating actual source positions into treatment planning can improve the accuracy of applicator reconstruction in cervical cancer brachytherapy.</p>","PeriodicalId":74309,"journal":{"name":"Nihon Hoshasen Gijutsu Gakkai zasshi","volume":"82 7","pages":""},"PeriodicalIF":0.0,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148159274","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"[Usefulness of a Split-bolus Injection Protocol for Pediatric Cardiac CT in Congenital Heart Disease].","authors":"Yusuke Ikeno, Masaru Morishita","doi":"10.6009/jjrt.26-1700","DOIUrl":"https://doi.org/10.6009/jjrt.26-1700","url":null,"abstract":"<p><strong>Purpose: </strong>To propose a split-bolus contrast injection protocol for pediatric cardiac CT and to evaluate its usefulness by comparing it with the early phase protocol and systemic circulation protocol.</p><p><strong>Methods: </strong>Thirty patients were included in each of the following three protocols: the early phase protocol; the systemic circulation protocol, in which contrast medium was injected slowly over a prolonged period and CT was performed after systemic circulation; and the split-bolus protocol. CT attenuation values were measured in the ascending aorta, descending aorta, pulmonary artery, pulmonary veins, and superior vena cava and compared among the three protocols. In addition, two radiologic technologists independently assessed vascular visualization using a four-point visual scoring system.</p><p><strong>Results: </strong>CT attenuation values obtained with the split-bolus protocol showed no significant differences from those of the early phase protocol in the ascending aorta, descending aorta, pulmonary artery, or pulmonary veins (p=0.282, 0.174, 0.632, and 0.159, respectively), whereas significant differences were observed compared with the systemic circulation protocol (all p<0.001). In contrast, no significant difference in CT attenuation of the superior vena cava was observed between the split-bolus protocol and the systemic circulation protocol (p=0.923). In the visual evaluation, the split-bolus protocol achieved a median score of 4 for all evaluated vessels across all cases.</p><p><strong>Conclusion: </strong>The split-bolus protocol may enable visualization of the superior vena cava while maintaining contrast enhancement of the aorta, pulmonary artery, and pulmonary veins, suggesting that it may be a useful contrast injection protocol for pediatric cardiac CT.</p>","PeriodicalId":74309,"journal":{"name":"Nihon Hoshasen Gijutsu Gakkai zasshi","volume":"82 12","pages":""},"PeriodicalIF":0.0,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149240737","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}