Japanese Journal of Radiology最新文献

{"title":"Postmortem computed tomography assessment of intracardiac and intravascular blood clots and gravitational sedimentation: clinical and laboratory associations in 114 in-hospital deaths.","authors":"Masanori Ishida, Akira Katayama, Go Shirota, Naomasa Okimoto, Hiroyuki Abe, Keisuke Nyunoya, Kotaro Fujimoto, Mariko Kurokawa, Masumi Takahashi-Mizuki, Shohei Inui, Shunichiro Orihara, Kazuhiro Saito, Tetsuo Ushiku, Osamu Abe, Wataru Gonoi","doi":"10.1007/s11604-025-01797-3","DOIUrl":"10.1007/s11604-025-01797-3","url":null,"abstract":"<p><strong>Purpose: </strong>Postmortem computed tomography (PMCT) typically reveals blood clots and sedimentation in cardiac and vascular structures. We examined the associations between these postmortem findings and antemortem clinical and laboratory parameters in in-hospital death.</p><p><strong>Material and methods: </strong>This prospective study included 114 non-traumatic in-hospital deaths where PMCT was performed within 24 h postmortem. Two radiologists evaluated high-density areas in the right and left atria, pulmonary artery, and thoracic aorta, and classified them as blood clots or gravitational sedimentation. The clinical and laboratory data from the week before death were analyzed using univariate and multivariate logistic regression.</p><p><strong>Results: </strong>Interobserver agreement was excellent for all anatomical sites (κ = 0.87-0.91). Blood clot or blood sedimentation were observed in 34-53% of cases across different locations. Per univariate analysis, non-pneumonic infections, positive blood cultures, and elevated coagulation parameters (prothrombin time-international normalized ratio, activated partial thromboplastin time) were associated with gravitational sedimentation. In contrast, solid malignancies and higher values of hematologic parameters (platelet count, red blood cells, hemoglobin, neutrophil percentage) were associated with blood clot formation (all p < .05). Per multivariate analysis, non-pneumonic infections maintained strong associations with gravitational sedimentation across all sites (p < .05), while higher platelet counts independently predicted blood clot formation in the right atrium, left atrium, and thoracic aorta (p < .05).</p><p><strong>Conclusion: </strong>Postmortem gravitational sedimentation was associated with non-pneumonic infections, whereas clot formation correlated with higher platelet counts. These findings provide objective criteria for interpreting PMCT findings and may aid in evaluating patients' antemortem clinical status, particularly when clinical information is limited.</p>","PeriodicalId":14691,"journal":{"name":"Japanese Journal of Radiology","volume":" ","pages":"1465-1478"},"PeriodicalIF":2.1,"publicationDate":"2025-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12397116/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144142604","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Developing a CT radiomics-based model for assessing split renal function using machine learning.","authors":"Yihua Zhan, Junjiong Zheng, Xutao Chen, Yushu Chen, Chao Fang, Cong Lai, Mingzhou Dai, Zhikai Wu, Han Wu, Taihui Yu, Jian Huang, Hao Yu","doi":"10.1007/s11604-025-01786-6","DOIUrl":"10.1007/s11604-025-01786-6","url":null,"abstract":"<p><strong>Purpose: </strong>This study aims to investigate whether non-contrast computed tomography radiomics can effectively reflect split renal function and to develop a radiomics model for its assessment.</p><p><strong>Materials and methods: </strong>This retrospective study included kidneys from the study center and split them into training (70%) and testing (30%) sets. Renal dynamic imaging was used as the reference standard for measuring split renal function. Based on chronic kidney disease staging, kidneys were categorized into three groups according to glomerular filtration rate: > 45 ml/min/1.73 m², 30-45 ml/min/1.73 m², and < 30 ml/min/1.73 m².Features were selected based on feature importance ranking from a tree model, and a random forest radiomics model was built.</p><p><strong>Results: </strong>A total of 543 kidneys were included, with 381 in the training set and 162 in the testing set. In the training set, 16 features identified as most important for distinguishing between the groups were ultimately included to develop the random forest model. The model demonstrated good discriminatory ability in the testing set. The AUC for the > 45 ml/min/1.73 m², 30-45 ml/min/1.73 m², and < 30 ml/min/1.73 m² categories were 0.859 (95% CI 0.804-0.910), 0.679 (95% CI 0.589-0.760), and 0.901 (95% CI 0.848-0.946), respectively. The calibration curves for the kidneys in each group closely align with the diagonal, with Hosmer-Lemeshow test P-values of 0.124, 0.241, and 0.199 for the three groups, respectively (all P > 0.05). The decision curve analysis confirmed the radiomics model's clinical utility, demonstrating significantly higher net benefit than both treat-all and treat-none strategies at clinically relevant probability thresholds: 1-69% and 71-75% for the > 45 ml/min/1.73 m² group, 15-d50% for the 30-45 ml/min/1.73 m² group, and 0-99% for the < 30 ml/min/1.73 m² group.</p><p><strong>Conclusion: </strong>Non-contrast computed tomography radiomics can effectively reflect split renal function information, and the model developed based on it can accurately assess split renal function, holding great potential for clinical application.</p>","PeriodicalId":14691,"journal":{"name":"Japanese Journal of Radiology","volume":" ","pages":"1520-1530"},"PeriodicalIF":2.1,"publicationDate":"2025-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12397145/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144208567","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"The effect of a whole-heart motion correction algorithm on image quality, calcium score, and risk category classification of non-contrast-enhanced cardiac CT.","authors":"Lihong Chen, Gongting Chen, Pan Zhao, Huan Wang, Bing Liu, Jiahao Chun, Yanhui Hao, Ning Pan, Zhijie Jian, Jianxin Guo","doi":"10.1007/s11604-025-01801-w","DOIUrl":"10.1007/s11604-025-01801-w","url":null,"abstract":"<p><strong>Purpose: </strong>To evaluate the effect of a whole-heart motion correction algorithm (Snapshot freeze 2, SSF2) on image quality, calcium score, and risk category classification of non-contrast-enhanced cardiac CT.</p><p><strong>Materials and methods: </strong>Sixty-one coronary calcium scoring scans performed on a 256-row detector CT were included. Images of standard reconstruction without motion-corrected (STD) and with SSF2 were compared. The image quality was graded by consensus of two experienced radiologists using a four-point Likert scale. The diameter of the major and minor axes at the heaviest point of a dominant calcification, as well as its volume, were measured. The Agatston score, volume score, and corresponding risk category for each patient were recorded. Paired t test, Wilcoxon test, and Bland-Altman plot were used for statistical analysis.</p><p><strong>Results: </strong>Image quality was significantly improved with SSF2 (P < 0.001). The mean diameters of the major and minor axes were 10.5 ± 5.3 mm and 6.4 ± 6.8 mm for STD images, and 7.2 ± 4.6 mm and 4.2 ± 5.5 mm for SSF2 images, resulting in a mean volume of 149.6 ± 206.1mm³ and 118.7 ± 184.6mm³, respectively (all P < 0.001). The mean Agatston score and volume score for STD images were 493.9 ± 585.9 and 415.7 ± 478.4, respectively, and decreased to 409.8 ± 505.4 and 337.7 ± 399.7, respectively, after motion correction (P < 0.001). The risk category changed significantly as 8 (13.1%) and 10 (16.4%) patients, were downgraded after motion correction based on either Agatston score (P = 0.008) or volume score (P = 0.002), respectively.</p><p><strong>Conclusion: </strong>Motion correction with SSF2 in non-contrast-enhanced cardiac CT improves the image quality at both the patient and lesion level, resulting in a significant reduction in calcium score and risk category downgrade.</p>","PeriodicalId":14691,"journal":{"name":"Japanese Journal of Radiology","volume":" ","pages":"1479-1488"},"PeriodicalIF":2.1,"publicationDate":"2025-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144078196","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Artificial intelligence software to detect small hepatic lesions on hepatobiliary-phase images using multiscale sampling.","authors":"Shogo Maeda, Yuko Nakamura, Toru Higaki, Ayu Karasudani, Tatsuya Yamaguchi, Masaki Ishihara, Takayuki Baba, Shota Kondo, Dara Fonseca, Kazuo Awai","doi":"10.1007/s11604-025-01859-6","DOIUrl":"https://doi.org/10.1007/s11604-025-01859-6","url":null,"abstract":"<p><strong>Purpose: </strong>To investigate the effect of multiscale sampling artificial intelligence (msAI) software adapted to small hepatic lesions on the diagnostic performance of readers interpreting gadoxetic acid-enhanced hepatobiliary-phase (HBP) images.</p><p><strong>Methods: </strong>HBP images of 30 patients harboring 186 hepatic lesions were included. Three board-certified radiologists, 9 radiology residents, and 2 general physicians interpreted HBP image data sets twice, once with and once without the msAI software at 2-week intervals. Jackknife free-response receiver-operating characteristic analysis was performed to calculate the figure of merit (FOM) for detecting hepatic lesions. The negative consultation ratio (NCR), percentage of correct diagnoses turning into incorrect by the AI software, was calculated. We defined readers whose NCR was lower than 10% as those correctly diagnosed the false findings presented by the software.</p><p><strong>Results: </strong>The msAI software significantly improved the lesion localization fraction (LLF) for all readers (0.74 vs 0.82, p < 0.01); the FOM did not (0.76 vs 0.78, p = 0.45). In lesion-size-based subgroup analysis, the LLF (0.40 vs 0.53, p < 0.01) improved significantly with the AI software even for lesions smaller than 6 mm, whereas the FOM (0.63 vs 0.66, p = 0.51) showed no significant difference. Among 10 readers with an NCR lower than 10%, not only the LLF but also the FOM were significantly better with the software (LLF 0.77 vs 0.82, FOM 0.79 vs 0.84, both p < 0.01).</p><p><strong>Conclusion: </strong>The detectability of small hepatic lesions on HBP images was improved with msAI software especially when its results were properly evaluated.</p>","PeriodicalId":14691,"journal":{"name":"Japanese Journal of Radiology","volume":" ","pages":""},"PeriodicalIF":2.1,"publicationDate":"2025-08-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144954742","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Canadian radiology: 2025 update.","authors":"Jason Yao, Waqas Ahmad, Sarah Cheng, Andreu F Costa, Birgit B Ertl-Wagner, Savvas Nicolaou, Carolina Souza, Michael N Patlas","doi":"10.1007/s11604-025-01862-x","DOIUrl":"https://doi.org/10.1007/s11604-025-01862-x","url":null,"abstract":"<p><p>Radiology in Canada is evolving through a combination of clinical innovation, collaborative research and the adoption of advanced imaging technologies. This overview highlights contributions from selected academic centres across the country that are shaping diagnostic and interventional practice. At Dalhousie University, researchers have led efforts to improve contrast media safety, refine imaging techniques for hepatopancreatobiliary diseases, and develop peer learning programs that support continuous quality improvement. The University of Ottawa has made advances in radiomics, magnetic resonance imaging protocols, and virtual reality applications for surgical planning, while contributing to global research networks focused on evaluating LI-RADS performance. At the University of British Columbia, the implementation of photon-counting CT, dual-energy CT, and artificial intelligence tools is enhancing diagnostic precision in oncology, trauma, and stroke imaging. The Hospital for Sick Children is a leader in paediatric radiology, with work ranging from artificial intelligence (AI) brain tumour classification to innovations in foetal MRI and congenital heart disease imaging. Together, these initiatives reflect the strength and diversity of Canadian radiology, demonstrating a shared commitment to advancing patient care through innovation, data-driven practice and collaboration.</p>","PeriodicalId":14691,"journal":{"name":"Japanese Journal of Radiology","volume":" ","pages":""},"PeriodicalIF":2.1,"publicationDate":"2025-08-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144954779","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Exploring cerebral small vessel disease, choroid plexus, demographics and vascular risk factors that affect glymphatic function.","authors":"Yiwen Chen, Linfeng Yang, Meng Li, Pengcheng Liang, Zhenyu Cheng, Na Wang, Xinyue Zhang, Yu Wang, Changhu Liang, Jing Li, Lingfei Guo","doi":"10.1007/s11604-025-01843-0","DOIUrl":"https://doi.org/10.1007/s11604-025-01843-0","url":null,"abstract":"<p><strong>Background: </strong>Cerebral small vessel disease (CSVD) is a major contributor to cognitive decline and dementia. The glymphatic system, a recently discovered brain waste clearance system, plays a critical role in maintaining brain health. This study investigates glymphatic dysfunction in CSVD using diffusion tensor imaging along perivascular spaces (DTI-ALPS) and enlarged perivascular spaces (EPVS), to identify key risk factors and modifiable influences.</p><p><strong>Methods: </strong>373 participants were recruited and underwent magnetic resonance imaging. Multivariable regression and restricted cubic splines (RCS) were performed to investigate the associations between demographics, vascular risk factors, choroid plexus volume (CPV), imaging markers of CSVD, and time of day (TOD) with the glymphatic system estimated by DTI-ALPS and PVS.</p><p><strong>Results: </strong>L-ALPS correlated with age (β = - 0.233, p < 0.001), CPV (β = - 0.100, p = 0.047), and sex (β = 0.103, p = 0.027). R-ALPS correlated with age (β = - 0.108, p = 0.040), Fazekas score (β = - 0.151, p = 0.006), male sex (β = 0.209, p < 0.001), and cerebral microbleeds (β = - 0.146, p = 0.004). The risk factors for severe BG-PVS were age (OR: 1.090, P < 0.001) and Fazekas score (OR: 2.243, P < 0.001). L-ALPS correlated with TOD (β = - 0.091, p = 0.039), particularly in females (β = - 0.147, p = 0.016).</p><p><strong>Conclusion: </strong>Age, sex (male), white matter hyperintensity, CPV enlargement, and cerebral microbleeds are risk factors for glymphatic dysfunction. Circadian rhythms may modulate glymphatic activity.</p>","PeriodicalId":14691,"journal":{"name":"Japanese Journal of Radiology","volume":" ","pages":""},"PeriodicalIF":2.1,"publicationDate":"2025-08-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144954716","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Quantitative CT analysis for predicting the PD-L1 expression in lung adenocarcinoma.","authors":"Masaya Tanabe, Yoshie Kunihiro, Masahiro Tanabe, Fumi Kameda, Masatoshi Nakashima, Taiga Kobayashi, Toshiki Tanaka, Yoshinobu Hoshii, Katsuyoshi Ito","doi":"10.1007/s11604-025-01857-8","DOIUrl":"https://doi.org/10.1007/s11604-025-01857-8","url":null,"abstract":"<p><strong>Purpose: </strong>The objective of this study was to explore the relationship between a quantitative CT analysis and the expression of programmed death-ligand 1 (PD-L1) in lung adenocarcinoma.</p><p><strong>Materials and methods: </strong>This study included 116 patients diagnosed with lung adenocarcinoma who were assessed for the expression of PD-L1. Tumors were classified as pure ground-glass nodules (GGNs), part-solid nodules, and solid nodules. The quantitative CT analysis included the tumor diameter and volume, solid component diameter and volume, and rate of the solid components. The CT criteria, and PD-L1 expression rates were compared based on the tumor proportion score (TPS). Optimal cutoff values were obtained utilizing the maximized Youden index method based on the receiver operating characteristic (ROC) analysis. Univariate and multiple linear regression analyses were also performed to examine the influencing factors of 50% and 1% PD-L1 expression.</p><p><strong>Results: </strong>Solid nodules were significantly more frequent in the TPS ≥ 50% group (TPS ≥ 50% = 81.8% vs. TPS < 1% = 10.0%). The rate of solid component diameter and rate of solid component volume were significantly smaller in TPS < 1% than in TPS < 50% and 1-49% (p < 0.001, respectively). Multiple linear regression analysis identified the rate of solid component volume as a significant factor influencing 50% and 1% PD-L1 expression (p < 0.001 and p = 0.048, respectively).</p><p><strong>Conclusion: </strong>High PD-L1 expression rates may be associated with higher rates of solid components in lung adenocarcinoma.</p>","PeriodicalId":14691,"journal":{"name":"Japanese Journal of Radiology","volume":" ","pages":""},"PeriodicalIF":2.1,"publicationDate":"2025-08-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144954732","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Enhancing generalizability and clinical translation of ViT-CNA fusion for ASD diagnosis.","authors":"Yiming Wang","doi":"10.1007/s11604-025-01860-z","DOIUrl":"https://doi.org/10.1007/s11604-025-01860-z","url":null,"abstract":"","PeriodicalId":14691,"journal":{"name":"Japanese Journal of Radiology","volume":" ","pages":""},"PeriodicalIF":2.1,"publicationDate":"2025-08-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144954703","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Novel silicon-based material decomposition images in diagnosis of pancreatic ductal adenocarcinoma: comparison with iodine-based and 50-keV virtual monoenergetic images.","authors":"Yoshifumi Noda, Mayu Hattori, Nobuyuki Kawai, Tetsuro Kaga, Akio Ito, Takuma Ishihara, Toshiharu Miyoshi, Yukiko Takai, Masashi Asano, Hiroki Kato, Fuminori Hyodo, Avinash R Kambadakone, Masayuki Matsuo","doi":"10.1007/s11604-025-01856-9","DOIUrl":"https://doi.org/10.1007/s11604-025-01856-9","url":null,"abstract":"<p><strong>Objectives: </strong>To identify the optimal material decomposition (MD) images for diagnosis of pancreatic ductal adenocarcinoma (PDAC) and evaluate the added value of the MD image to 50-keV virtual monoenergetic images (VMIs) by comparing with iodine-based images and 50-keV VMIs.</p><p><strong>Methods: </strong>This retrospective study included patients who underwent pancreatic protocol dual-energy CT (DECT) between February 2019 and May 2023. First, a radiologist evaluated 702 image datasets generated using 27 different materials to identify the top three MD images which provided maximum contrast difference between normal pancreas and PDAC, and subsequently, the best MD image was selected based on z value and image quality by four radiologists. Then, another four radiologists independently interpreted the conventional image dataset (iodine-based images and 50-keV VMIs) and new optimal image dataset (optimal MD images and 50-keV VMIs), and graded the presence or absence of PDAC. The sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy were compared between the two image datasets using generalized estimating equations.</p><p><strong>Results: </strong>Overall, 110 patients (median age, 73 years; 63 men) were included. Among them, 67 patients (61%) had pathologically proven PDAC, and the optimal MD image selected was Silicon/Struvite. The optimal image dataset had higher specificity (88% vs. 81%; P = 0.006), PPV (93% vs. 89%; P < 0.001), and accuracy (94% vs. 92%; P = 0.01) than the conventional image dataset. No difference was found in the sensitivity (P = 0.34) and NPV (P = 0.33) between the two image datasets.</p><p><strong>Conclusion: </strong>Silicon/Struvite images provided high contrast difference between normal pancreas and PDAC and higher diagnostic performance for diagnosis of PDAC in combination of 50-keV VMIs compared to iodine-based images and 50-keV VMIs.</p>","PeriodicalId":14691,"journal":{"name":"Japanese Journal of Radiology","volume":" ","pages":""},"PeriodicalIF":2.1,"publicationDate":"2025-08-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144954749","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A systematic review of comparisons of AI and radiologists in the diagnosis of HCC in multiphase CT: implications for practice.","authors":"Jarrod Younger, Emily Morris, Nicholas Arnold, Chanchala Athulathmudali, Janani Pinidiyapathirage, William MacAskill","doi":"10.1007/s11604-025-01853-y","DOIUrl":"https://doi.org/10.1007/s11604-025-01853-y","url":null,"abstract":"<p><strong>Purpose: </strong>This systematic review aims to examine the literature of artificial intelligence (AI) algorithms in the diagnosis of hepatocellular carcinoma (HCC) among focal liver lesions compared to radiologists on multiphase CT images, focusing on performance metrics that include sensitivity and specificity as a minimum.</p><p><strong>Methods: </strong>We searched Embase, PubMed and Web of Science for studies published from January 2018 to May 2024. Eligible studies evaluated AI algorithms for diagnosing HCC using multiphase CT, with radiologist interpretation as a comparator. The performance of AI models and radiologists was recorded using sensitivity and specificity from each study. TRIPOD + AI was used for quality appraisal and PROBAST was used to assess the risk of bias.</p><p><strong>Results: </strong>Seven studies out of the 3532 reviewed were included in the review. All seven studies analysed the performance of AI models and radiologists. Two studies additionally assessed performance with and without supplementary clinical information to assist the AI model in diagnosis. Three studies additionally evaluated the performance of radiologists with assistance of the AI algorithm in diagnosis. The AI algorithms demonstrated a sensitivity ranging from 63.0 to 98.6% and a specificity of 82.0-98.6%. In comparison, junior radiologists (with less than 10 years of experience) exhibited a sensitivity of 41.2-92.0% and a specificity of 72.2-100%, while senior radiologists (with more than 10 years of experience) achieved a sensitivity between 63.9% and 93.7% and a specificity ranging from 71.9 to 99.9%.</p><p><strong>Conclusion: </strong>AI algorithms demonstrate adequate performance in the diagnosis of HCC from focal liver lesions on multiphase CT images. Across geographic settings, AI could help streamline workflows and improve access to timely diagnosis. However, thoughtful implementation strategies are still needed to mitigate bias and overreliance.</p>","PeriodicalId":14691,"journal":{"name":"Japanese Journal of Radiology","volume":" ","pages":""},"PeriodicalIF":2.1,"publicationDate":"2025-08-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144873211","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}