Frontiers in radiology最新文献

{"title":"Predictors of acute adverse reactions to non-ionic iodinated contrast media in CT imaging: a systematic review and meta-analysis.","authors":"Ke Liu, Xin Cheng, Yongli Zhu, Jun Long, Changsheng Li, Lijun Cui, Kang Li, Changping Mu","doi":"10.3389/fradi.2025.1656949","DOIUrl":"10.3389/fradi.2025.1656949","url":null,"abstract":"<p><strong>Background: </strong>Iodinated contrast media-acute adverse reactions (ICM-AARs) are frequent and clinically significant complications associated with radiological imaging. Despite investigation of their risk factors, there is no consensus, and no comprehensive synthesis has been conducted. This systematic review and meta-analysis aimed to investigate the factors influencing ICM-AARs.</p><p><strong>Methods: </strong>A systematic search for studies published in Chinese or English up to 22 July 2024 in the PubMed, Web of Science, Cochrane Library, Embase, CNKI, WanFang, CQVIP, and SinoMed databases was conducted. Studies on patients undergolng contrast-enhanced CT examinations with nonionic ICM were selected. The primary outcome measures were risk factors associated with ICM-AARs. The studies were analyzed for heterogeneity using the <i>Q</i>-test and I² statistic, while publication bias was assessed using funnel plots, Egger's test, and Begg's test. Stata 17 software was used for the meta-analysis.</p><p><strong>Results: </strong>Seventeen studies were included, encompassing 2,576,446 CT-enhanced examinations. Of these, 11,621 acute adverse reactions were reported, with a mean incidence of 0.45% and a quality score of ≥7. The meta-analysis showed that female sex (OR = 1.27, 95% CI = 1.13, 1.41), age <35 years (OR = 1.77, 95% CI = 1.19, 2.64), high body mass index (OR = 1.06, 95% CI = 1.01, 1.10), type of medical visit (outpatient) (OR = 2.23, 95% CI = 1.01, 4.93), history of adverse ICM reactions (OR = 11.03, 95% CI = 2.25, 53.97), history of other allergies (OR = 3.16, 95% CI = 1.27, 7.84), history of asthma (OR = 1.75, 95% CI = 1.19, 2.57), hyperthyroldism (OR = 4.59, 95% CI = 1.65, 12.82), and type of ICM (OR = 2.27, 95% CI = 1.68, 3.06) were risk factors for ICM-AARs. Age >60 years (OR = 0.71, 95% CI = 0.53, 0.95), pre-injection medication (OR = 0.56, 95% CI = 0.39, 0.79), and hypertensive disorders (OR = 0.78, 95% CI = 0.65, 0.94) were identified as protective against ICM-AARs.</p><p><strong>Conclusions: </strong>The incidence of ICM-AARs is influenced by a variety of clinical and demographic factors. Healthcare professionals may benefit from dynamically assessing patient-specific risk factors and considering targeted preventive measures for high-risk groups, particularly in populations similar to those studied.</p><p><strong>Systematic review registration: </strong>https://www.crd.york.ac.uk/PROSPERO/, PROSPERO (CRD42024571470).</p>","PeriodicalId":73101,"journal":{"name":"Frontiers in radiology","volume":"5 ","pages":"1656949"},"PeriodicalIF":2.3,"publicationDate":"2025-09-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12491283/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145234459","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"The evolution of artificial intelligence technology in non-alcoholic fatty liver disease.","authors":"Jiawen He, Yao Wu, Zhiyong Lin, Ruohong He, Li Zhuo, Yingying Li","doi":"10.3389/fradi.2025.1634165","DOIUrl":"10.3389/fradi.2025.1634165","url":null,"abstract":"&lt;p&gt;&lt;strong&gt;Background: &lt;/strong&gt;The incidence of Non-alcoholic Fatty Liver Disease (NAFLD) continues to rise, becoming one of the major causes of chronic liver disease globally and posing significant challenges to healthcare systems worldwide. Artificial intelligence (AI) technology, as an emerging tool, is gradually being integrated into clinical practice for NAFLD, providing innovative approaches to improve diagnostic efficiency, personalized treatment plans, and disease prognosis assessment. However, current research remains fragmented, lacking systematic and comprehensive analysis.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Objective: &lt;/strong&gt;This study conducts a bibliometric analysis of artificial intelligence applications in Non-alcoholic Fatty Liver Disease (NAFLD), aiming to identify research trends, highlight key areas, and provide comprehensive and objective insights into the current state of research in this field. We expect that these research results will provide valuable references for guiding further research directions and promoting the effective application of AI in liver disease healthcare.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Methods: &lt;/strong&gt;This study used the Web of Science Core Collection database as the data source, searching the Science Citation Index Expanded (SCI-Expanded) and Current Chemical Reactions (CCR-Expanded) citation indexes. The search timeframe was set to include all relevant literature from 2010 to March 25, 2025. The research methodology adopted a multi-software joint analysis strategy: First, HistCite Pro 2.1 was used to analyze the historical evolution and citation relationships of literature in this field. The tables generated by the tool systematically recorded the development process of the literature, clearly depicting the evolution of the research field over time. Second, Scimago Graphica was used to create a country/region collaboration network view, intuitively showing academic collaboration among countries/regions (SCImago Lab, 2022). VOSviewer 1.6.20 was used to analyze collaboration networks and visualize keyword co-occurrences to identify main research themes and clusters. CiteSpace was used for deeper scientific literature analysis, precisely capturing the dynamic changes of research hotspots and the evolution of frontier trends through Burst Detection algorithms and Timezone View.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Results: &lt;/strong&gt;A total of 655 papers were retrieved from 60 countries, 1462 research institutions, and 4,744 authors published in 279 journals. The number of papers surged dramatically during 2019-2024, with papers from these six years accounting for approximately 83.8% (549/655) of the total. Country-level analysis showed that the United States and China are the major contributors to this field; journal analysis indicated that Scientific Reports and Diagnostics are the journals with the highest publication volumes. In-depth analysis of 655 publications revealed four major research clusters: non-invasive assessment methods for liver fibrosis, ima","PeriodicalId":73101,"journal":{"name":"Frontiers in radiology","volume":"5 ","pages":"1634165"},"PeriodicalIF":2.3,"publicationDate":"2025-09-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12480972/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145208388","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Evaluation of the effectiveness of contrast-enhanced ultrasound in the diagnosis of early hepatocellular carcinoma: a systematic review.","authors":"Abdulaziz AlTaweel, Faisal Joueidi, Ahmad Joueidi, Ahmed AlDhubaiki, Hamad Mohammed Qabha, Homoud Abdulaziz AlZaid","doi":"10.3389/fradi.2025.1661522","DOIUrl":"10.3389/fradi.2025.1661522","url":null,"abstract":"<p><strong>Objectives: </strong>To investigate the evaluation of the effectiveness of contrast-enhanced ultrasound (CEUS) in the diagnosis of small hepatocellular carcinoma (HCC).</p><p><strong>Methods: </strong>A thorough search was conducted for pertinent literature using PubMed, SCOPUS, Web of Science, Science Direct, and Wiley Library. Rayyan QRCI was used throughout this extensive procedure.</p><p><strong>Results: </strong>Our results included thirteen studies with a total of 2016 patients, and 1672 (82.9%) were males. The follow-up duration ranged from 3 months to 24 months. CEUS was useful in anticipating the early recurrence of HCC, predicting the early recurrence of solitary lesion HCC patients, and differentiating between HCC and intrahepatic cholangiocarcinoma <3 Cm, distinguishing HCC from dysplastic nodules from tiny liver nodules, CEUS in cirrhotic patients. When paired with CEUS, conventional ultrasonography can detect minor HCC and assist in patient monitoring for those who receive an early diagnosis of HCC. CEUS showed high concordance with CECT for diagnosing lesions 2.1-3.0 cm in size. Notable limitations included heterogeneity in protocols and predominance of Asian populations (12/13 studies).</p><p><strong>Conclusion: </strong>CEUS offers significant clinical value as a noninvasive diagnostic tool, particularly for 1-3 cm lesions in cirrhotic patients and cases where CT is contraindicated, though protocol standardization and Western population validation remain needed.</p>","PeriodicalId":73101,"journal":{"name":"Frontiers in radiology","volume":"5 ","pages":"1661522"},"PeriodicalIF":2.3,"publicationDate":"2025-09-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12454341/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145139729","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Intervertebral disc anomaly intelligent classification system based on deep learning, IDAICS.","authors":"Zhiheng Gao, Yuchen Qian, Rongkang Fan, Yuqing Yang, Yu Wang, Lei Xing, Yu Chen, Yonggang Li, Haifu Sun, Yusen Qiao","doi":"10.3389/fradi.2025.1646008","DOIUrl":"10.3389/fradi.2025.1646008","url":null,"abstract":"<p><strong>Background: </strong>Intervertebral disc anomalies, such as degeneration and herniation, are common causes of spinal disorders, often leading to chronic pain and disability. Accurate diagnosis and classification of these anomalies are critical for determining appropriate treatment strategies. Traditional methods, such as manual image analysis, are prone to subjectivity and time-consuming. With the advancements in deep learning, automated and precise classification of intervertebral disc anomalies has become a promising alternative.</p><p><strong>Objective: </strong>This study aims to propose a deep learning-based method for classifying intervertebral disc abnormalities, with the goal of improving diagnostic accuracy and clinical efficiency in spinal health management.</p><p><strong>Methods: </strong>From August 2021 to March 2024, a dataset consisting of 574 CT images of intervertebral discs was collected and labeled into four clinically relevant categories: normal intervertebral discs, Schmorl's nodes, disc bulges, and disc protrusions. The dataset was divided into 500 images for model training, and 74 images for validation. A YOLOv8-seg network was employed for classification, with multiple preprocessing techniques applied to ensure data consistency and enhance model performance.</p><p><strong>Results: </strong>The IDAICS demonstrated high accuracy in classifying various intervertebral disc anomalies, including disc degeneration, herniation, and bulging, with a classification accuracy of over 93.2%, with a kappa coefficient of 0.905 (<i>P</i> < 0.001).</p><p><strong>Conclusion: </strong>This deep learning-based classification approach provides an efficient and reliable alternative to manual assessment, enabling automated diagnosis of intervertebral disc abnormalities. It offers significant potential to enhance clinical decision-making and improve spinal health management outcomes.</p>","PeriodicalId":73101,"journal":{"name":"Frontiers in radiology","volume":"5 ","pages":"1646008"},"PeriodicalIF":2.3,"publicationDate":"2025-09-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12454447/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145139701","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Editorial: Towards precision oncology: assessing the role of radiomics and artificial intelligence.","authors":"Salvatore Claudio Fanni, Damiano Caruso, Lorenzo Faggioni, Emanuele Neri, Dania Cioni","doi":"10.3389/fradi.2025.1676229","DOIUrl":"10.3389/fradi.2025.1676229","url":null,"abstract":"","PeriodicalId":73101,"journal":{"name":"Frontiers in radiology","volume":"5 ","pages":"1676229"},"PeriodicalIF":2.3,"publicationDate":"2025-09-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12440918/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145088369","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"The value of CT texture analysis in predicting mitotic activity and morphological variants of adrenocortical carcinoma.","authors":"N V Tarbaeva, A V Manaev, K V Ivashchenko, N M Platonova, D G Beltsevich, N V Pachuashvili, L S Urusova, N G Mokrysheva","doi":"10.3389/fradi.2025.1635425","DOIUrl":"10.3389/fradi.2025.1635425","url":null,"abstract":"<p><strong>Introduction: </strong>Adrenocortical carcinoma presents significant diagnostic challenges due to its histological heterogeneity and variable clinical behavior. This study aimed to evaluate the diagnostic value of radiomic features in predicting mitotic activity (low/high-grade) and morphological variants (conventional, oncocytic, myxoid) of adrenocortical carcinoma.</p><p><strong>Materials and methods: </strong>A retrospective analysis of 32 patients with histologically confirmed ACC (18 conventional, 9 oncocytic and 5 myxoid cases) was performed, with mitotic data available for 25 cases (13 low-grade and 12 high-grade cases). Radiomic features including Gray-Level Co-occurrence Matrix (GLCM), Run-Length (GLRLM), Size-Zone (GLSZM), Dependence (GLDM), Neighboring-Tone (NGTDM) and first order features were extracted from four-phase CT using PyRadiomics after manual 3D segmentation. Statistical analysis included Mann-Whitney <i>U</i>, Kruskal-Wallis tests, ROC curve (AUC, sensitivity, specificity) and PPV, NPV assessment.</p><p><strong>Results: </strong>Our analysis demonstrated statistically significant differences between tumor grades with firstorder_Skewness (AUC = 0.924, 95% CI: 0.819-0.986; <i>p</i> = 0.005) showing high predictive performance in the venous phase. Radiomic features did not show statistically significant differences between morphological variants of ACC after adjustment for multiple comparisons.</p><p><strong>Conclusion: </strong>Our results confirm the value of CT radiomics for preoperative stratification of ACC grade, but the question of differentiation of morphological variants remains unresolved and requires further validation in larger cohorts.</p>","PeriodicalId":73101,"journal":{"name":"Frontiers in radiology","volume":"5 ","pages":"1635425"},"PeriodicalIF":2.3,"publicationDate":"2025-08-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12367671/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144981116","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Evaluating the effect of voxel size on the accuracy of 3D volumetric analysis measurements of brain tumors.","authors":"Rithvik S Ghankot, Manwi Singh, Shelby T Desroches, Noemi Jester, Amit Mahajan, Samantha Lorr, Frank D Buono, Daniel H Wiznia, Michele H Johnson, Steven M Tommasini","doi":"10.3389/fradi.2025.1618261","DOIUrl":"10.3389/fradi.2025.1618261","url":null,"abstract":"<p><strong>Introduction: </strong>Neurofibromatosis type 2 related Schwannomatosis (NF2-SWN) is a genetic disorder characterized by the growth of vestibular schwannomas (VS), which often leads to progressive hearing loss and vestibular dysfunction. Accurate volumetric assessment of VS tumors is crucial for effective monitoring and treatment planning. Since tumor growth dynamics are often subtle, the resolution of MRI scans plays a critical role in detecting small volumetric changes that inform clinical decisions. This study evaluates the impact of MRI voxel resolution on the accuracy of manual and AI-driven volumetric segmentation of VS in NF2-SWN patients.</p><p><strong>Methods: </strong>Ten patients with NF2-SWN, totaling 17 tumors, underwent high-resolution MRI scans with varying voxel sizes on different MRI machines at Yale New Haven Hospital. Tumors were segmented using both manual and AI-based methods, and the effect of voxel size on segmentation precision was quantified through volume measurements, Dice similarity coefficients, and Hausdorff distances.</p><p><strong>Results: </strong>Results indicate that larger voxel sizes (1.2 × 0.9 × 4.0 mm) significantly reduced segmentation accuracy when compared to smaller voxel sizes (0.5 × 0.5 × 0.8 mm). In addition, AI-based segmentation outperformed manual methods, particularly at larger voxel sizes.</p><p><strong>Discussion: </strong>These findings highlight the importance of optimizing voxel resolution for accurate tumor monitoring and suggest that AI-driven segmentation may improve consistency and precision in NF2-SWN tumor surveillance.</p>","PeriodicalId":73101,"journal":{"name":"Frontiers in radiology","volume":"5 ","pages":"1618261"},"PeriodicalIF":2.3,"publicationDate":"2025-08-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12364922/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144981122","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Complications of percutaneously placed uncovered metallic biliary stents for malignant obstruction: a systematic review.","authors":"Jonathan Bock, Christopher J Reisenauer, Michael C Jundt, Matthew R Augustine, Richard G Frimpong, Edwin A Takahashi","doi":"10.3389/fradi.2025.1639323","DOIUrl":"10.3389/fradi.2025.1639323","url":null,"abstract":"<p><strong>Background: </strong>The aim of this systematic review was to determine the patency and complications related to percutaneous metallic biliary stent placement for malignant biliary obstruction in the current literature.</p><p><strong>Methods: </strong>This review was performed using the Preferred Reporting Items of Systematic Reviews and Meta-Analyses guidelines. EMBASE and PubMed were queried yielding 891 articles, 18 of which were included in the final analysis. The Newcastle-Ottawa Quality Assessment Scale was used to appraise article quality. Patient demographics, technical success rate, and procedure outcomes were recorded. Complications were classified as \"major\" if they resulted in blood transfusion or additional invasive procedures or were reported as such in the literature. Complications that did not meet these criteria were classified as \"minor\".</p><p><strong>Results: </strong>A total of 1,453 patients (677 female; weighted age 66.8 years) underwent biliary stent placement. The weighted technical success rate was 97.7%. The incidence of stent occlusion was 13.5% with 6.6% of patients requiring further intervention to maintain patency. There were 277 (19.1%) complications, of which 87 were classified as major. The most common complications were pancreatitis (93, 6.4%), cholangitis (69, 4.8%), and bleeding (64, 4.4%). In cases of bleeding, 4.7% of patients needed a blood transfusion and 15.6% required a procedure to treat bleeding. There were 6 (0.4%) procedure-related deaths.</p><p><strong>Conclusion: </strong>In conclusion, percutaneous metallic stent placement for malignant biliary obstruction has a high technical success rate and relatively low rate of occlusion. Although nearly one in five procedures resulted in a complication, most cases were minor.</p>","PeriodicalId":73101,"journal":{"name":"Frontiers in radiology","volume":"5 ","pages":"1639323"},"PeriodicalIF":2.3,"publicationDate":"2025-08-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12361158/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144981104","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Explainable AI in medicine: challenges of integrating XAI into the future clinical routine.","authors":"Tim Räz, Aurélie Pahud De Mortanges, Mauricio Reyes","doi":"10.3389/fradi.2025.1627169","DOIUrl":"10.3389/fradi.2025.1627169","url":null,"abstract":"<p><p>Future AI systems may need to provide medical professionals with explanations of AI predictions and decisions. While current XAI methods match these requirements in principle, they are too inflexible and not sufficiently geared toward clinicians' needs to fulfill this role. This paper offers a conceptual roadmap for how XAI may be integrated into future medical practice. We identify three desiderata of increasing difficulty: First, explanations need to be provided in a context- and user-dependent manner. Second, explanations need to be created through a genuine dialogue between AI and human users. Third, AI systems need genuine social capabilities. We use an imaginary stroke treatment scenario as a foundation for our roadmap to explore how the three challenges emerge at different stages of clinical practice. We provide definitions of key concepts such as genuine dialogue and social capability, we discuss why these capabilities are desirable, and we identify major roadblocks. Our goal is to help practitioners and researchers in developing future XAI that is capable of operating as a participant in complex medical environments. We employ an interdisciplinary methodology that integrates medical XAI, medical practice, and philosophy.</p>","PeriodicalId":73101,"journal":{"name":"Frontiers in radiology","volume":"5 ","pages":"1627169"},"PeriodicalIF":2.3,"publicationDate":"2025-08-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12391920/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144981053","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"WaveAttention-ResNet: a deep learning-based intelligent diagnostic model for the auxiliary diagnosis of multiple retinal diseases.","authors":"Biao Guo, Daqing Wang, Ruiqi Zhang, Jia Hou, Wenchao Liu, YongFei Wu, Xudong Yang, Lijuan Zhang","doi":"10.3389/fradi.2025.1608052","DOIUrl":"10.3389/fradi.2025.1608052","url":null,"abstract":"<p><strong>Objective: </strong>This study constructs a deep learning-based combined algorithm named WaveAttention ResNet (WARN) to investigate the classification accuracy for seven common retinal diseases and the feasibility of AI-assisted diagnosis in this field.</p><p><strong>Methods: </strong>First, a deep learning-based classification network is constructed. The network is built upon ResNet18, integrated with the Convolutional Block Attention Module (CBAM) and wavelet convolution modules, forming the WARN method for retinal disease classification. Second, the public OCTDL dataset is used to train WARN, which contains classification data for seven retinal disease types: age-related macular degeneration (AMD), diabetic macular edema (DME), epiretinal membrane (ERM), normal (NO), retinal artery occlusion (RAO), retinal vein occlusion (RVO), and vitreomacular interface disease (VID). During this process, ablation experiments and significance tests are conducted on WARN, and comprehensive analyses of various indicators for WARN, ResNet-18, ResNet-50, Swin Transformer v2, EfficientNet, and Vision Transformer (ViT) are performed in retinal disease classification tasks. Finally, data provided by Shanxi Eye Hospital are used for testing, and classification results are analyzed.</p><p><strong>Results: </strong>WARN demonstrates excellent performance on the public OCTDL dataset. Ablation experiments and significance tests confirm the effectiveness of WARN, achieving an accuracy of 90.68%, F1-score of 91.29%, AUC of 97.50%, precision of 93.31%, and recall of 90.68% with relatively short training time. In the dataset from Shanxi Eye Hospital, WARN also performs well, with a recall of 90.85%, precision of 79.94%, and accuracy of 89.18%.</p><p><strong>Conclusion: </strong>This study fully confirms that the constructed WARN is efficient and feasible for classifying seven common retinal diseases. It further highlights the enormous potential and broad application prospects of AI technology in the field of auxiliary medical diagnosis.</p>","PeriodicalId":73101,"journal":{"name":"Frontiers in radiology","volume":"5 ","pages":"1608052"},"PeriodicalIF":2.3,"publicationDate":"2025-07-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12339474/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144838763","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}