{"title":"Myocardial late enhancement using dual-source CT: intraindividual comparison of single-energy shuttle and dual-energy acquisition.","authors":"Takanori Kokawa, Kakuya Kitagawa, Satoshi Nakamura, Masafumi Takafuji, Takashi Oya, Hajime Sakuma","doi":"10.1186/s13244-025-01944-4","DOIUrl":"10.1186/s13244-025-01944-4","url":null,"abstract":"<p><strong>Objectives: </strong>Myocardial computed tomography late enhancement (CT-LE) is a valuable modality used for the assessment of myocardial infarction and fibrosis and is effective in detecting latent cardiac amyloidosis. However, the optimal acquisition mode for CT-LE remains unknown. Here, we compared single-energy shuttle mode and DE mode for improving the quality of CT-LE imaging using dual-source CT.</p><p><strong>Methods: </strong>Fifteen patients with suspected or known ischemic heart disease underwent CT-LE imaging 5 min after coronary CT in both shuttle and dual-energy (DE) modes. In DE mode, virtual monoenergetic images at various keVs were reconstructed, and extracellular volume (ECV) was quantified using iodine-specific images. For shuttle mode, ECV was assessed by subtracting the volume from pre-contrast images from CT-LE after non-rigid registration.</p><p><strong>Results: </strong>In DE mode, signal-noise-to-ratio was the highest at 70 keV, but it was still lower than that in shuttle mode (p < 0.001). Contrast-noise-to-ratio was the highest on DE mode at 40 keV and was comparable with that in shuttle mode (p = 0.51). Interobserver agreement for infarct detection was higher in shuttle mode (kappa = 0.981) compared to DE mode (kappa = 0.808). Global ECV was comparable between shuttle and DE modes (p = 0.96). However, the coefficient of variation of segmental ECV was significantly lower in shuttle mode (p < 0.001).</p><p><strong>Conclusion: </strong>Shuttle mode CT-LE demonstrates superior image quality, better agreement in infarct detection, and ECV consistency in comparison to DE mode, suggesting its potential as the preferred approach for CT-LE imaging using dual-source CT despite limited z-axis coverage of 10.5 cm.</p><p><strong>Clinical relevance statement: </strong>CT late enhancement imaging in shuttle mode provides superior image quality and consistent extracellular volume measurements compared to dual-energy mode, highlighting its potential as the preferred acquisition method for CT late enhancement imaging in dual-source CT.</p><p><strong>Key points: </strong>Shuttle mode and dual-energy acquisition are compared for optimal myocardial CT-late enhancement (CT-LE) imaging. Shuttle mode can provide better image quality and more consistent extracellular volume measurements. Despite limited coverage, shuttle mode may be preferred for myocardial CT-LE imaging.</p>","PeriodicalId":13639,"journal":{"name":"Insights into Imaging","volume":"16 1","pages":"64"},"PeriodicalIF":4.1,"publicationDate":"2025-03-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11929652/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143691828","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Insulinoma detection on low-dose pancreatic CT perfusion: comparing with conventional contrast-enhanced CT and MRI.","authors":"Shiwei Luo, Xilong Mei, Youlan Shang, Jiaqi Yao, Nuerbiya Keranmu, Shaqi He, Cheng Yu, Fei Tang, Cong Li, Wenhan Yang, Jun Liu","doi":"10.1186/s13244-025-01943-5","DOIUrl":"10.1186/s13244-025-01943-5","url":null,"abstract":"<p><strong>Objectives: </strong>To evaluate the efficacy of low-dose pancreatic CT perfusion (pCTP) in detecting insulinomas in patients with recurrent hypoglycemia, and to compare its diagnostic performance with conventional contrast-enhanced CT (CECT) and MRI.</p><p><strong>Methods: </strong>This study retrospectively collected 53 patients with recurrent hypoglycemia (28 with insulinomas; 25 without insulinomas). PCTP image analysis was conducted by two radiologists. Quantitative perfusion parameters of insulinomas vs. tumor-free pancreatic parenchyma were analyzed. For cases where both pCTP and CECT/MRI were performed, six radiologists blinded to the patients' diagnosis independently evaluated the pCTP and CECT/MRI to determine the presence and location of insulinoma. The diagnostic performance of insulinoma detection between pCTP and CECT/MRI was compared.</p><p><strong>Results: </strong>For patients who underwent both CECT and pCTP, the sensitivity (CECT 0.167-0.333 vs. pCTP 0.667-1.000) of tumor detection was higher for five of six radiologists on pCTP than on CECT. For patients who underwent both MRI and pCTP, four radiologists showed higher sensitivity (MRI 0.400-600 vs. pCTP 0.700-0.800) of tumor detection on pCTP than on MRI, while two radiologists showed slightly lower sensitivity (MRI 0.800, 1.000 vs. pCTP 0.700, 0.900) on pCTP. Among perfusion parameters, peak enhancement, blood flow, and mean transit time exhibited higher AUC than blood volume and time to peak.</p><p><strong>Conclusion: </strong>PCTP demonstrated superior diagnostic performance in insulinoma detection among less-experienced radiologists compared to CECT and MRI, while more-experienced radiologists achieved marginally better results with MRI. These findings suggest pCTP's potential as a complementary imaging modality, particularly beneficial for junior radiologists in insulinoma detection.</p><p><strong>Critical relevance statement: </strong>Pancreatic CT perfusion exhibited promising diagnostic performance in insulinoma detection, particularly among junior radiologists, demonstrating the potential to complement conventional imaging modalities and serve as a valuable clinical tool for the detection and localization of insulinoma.</p><p><strong>Key points: </strong>Accurate preoperative identification and localization of insulinomas is important for appropriate treatment. Peak enhancement, blood flow, and mean transit time outperformed blood volume and time to peak in insulinoma detection. Pancreatic CT perfusion has the potential to complement conventional imaging modalities for insulinoma detection.</p>","PeriodicalId":13639,"journal":{"name":"Insights into Imaging","volume":"16 1","pages":"63"},"PeriodicalIF":4.1,"publicationDate":"2025-03-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11929649/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143691792","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Heping Deng, Xiaolei Dong, Yu Zhang, Peng Zhou, Yakun He, Liu Yang
{"title":"Preliminary study on the feasibility of united compressed sensing with radial acquisition as a routine method for liver dynamic contrast-enhanced examination in elderly patients with malignancy.","authors":"Heping Deng, Xiaolei Dong, Yu Zhang, Peng Zhou, Yakun He, Liu Yang","doi":"10.1186/s13244-025-01936-4","DOIUrl":"10.1186/s13244-025-01936-4","url":null,"abstract":"<p><strong>Objective: </strong>To explore the value of the united imaging compressed sensing with radial acquisition (uCSR) in liver dynamic contrast-enhanced examinations for elderly patients with malignancy.</p><p><strong>Methods: </strong>Hundred patients aged 65 years or over were randomly divided into two groups: 50 patients underwent liver dynamic contrast-enhanced scanning using the uCSR sequence during free breathing, and 50 patients underwent scanning using the three-dimensional volume interpolated breath-hold examination (3D-VIBE) sequence while holding breath. Two radiologists independently and subjectively evaluated the overall image quality and image artifacts with a five-point scale. Concurrently, two technologists measured the signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) of the arterial, portal venous and delay phase images in both groups.</p><p><strong>Results: </strong>uCSR has superior overall image-quality and image-artifact scores (z = 2.342, p = 0.019; z = 2.105, p = 0.035). The 3D-VIBE images of the arterial phase have higher SNR than uCSR (t = 4.988, p = 0.000), with no significant difference in the CNR (z = 0.676, p = 0.499). In the portal venous phase, the SNR and CNR of the 3D-VIBE images are superior to those of uCSR (z = 5.674, p = 0.000; t = 3.638, p = 0.000). In the delay phase, the SNR of the 3D-VIBE is slightly better than the uCSR (t = 5.471, p = 0.000), and the CNR shows no significant difference (z = 1.258, p = 0.208).</p><p><strong>Conclusion: </strong>uCSR can be used as a method for liver dynamic contrast-enhanced scans in elderly patients with malignancy. It can improve patient comfort and reduce the failure rate of scans.</p><p><strong>Critical relevance statement: </strong>Our findings suggested that uCSR can be used for liver dynamic contrast-enhanced scans in elderly patients with malignancy, this preliminary study provided basis for it.</p><p><strong>Key points: </strong>The uCSR can suppress the impact of respiratory motion artifacts on images. The UCSR can perform dynamic enhanced scanning of the liver under free breathing dynamics. The uCSR is suitable for dynamic contrast-enhanced MR imaging of the liver in elderly patients with malignancy.</p>","PeriodicalId":13639,"journal":{"name":"Insights into Imaging","volume":"16 1","pages":"65"},"PeriodicalIF":4.1,"publicationDate":"2025-03-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11929642/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143691889","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Jakub Pristoupil, Laura Oleaga, Vanesa Junquero, Cristina Merino, Ozbek Suha Sureyya, Martin Kyncl, Andrea Burgetova, Lukas Lambert
{"title":"Generative pre-trained transformer 4o (GPT-4o) in solving text-based multiple response questions for European Diploma in Radiology (EDiR): a comparative study with radiologists.","authors":"Jakub Pristoupil, Laura Oleaga, Vanesa Junquero, Cristina Merino, Ozbek Suha Sureyya, Martin Kyncl, Andrea Burgetova, Lukas Lambert","doi":"10.1186/s13244-025-01941-7","DOIUrl":"10.1186/s13244-025-01941-7","url":null,"abstract":"<p><strong>Objectives: </strong>This study aims to assess the accuracy of generative pre-trained transformer 4o (GPT-4o) in answering multiple response questions from the European Diploma in Radiology (EDiR) examination, comparing its performance to that of human candidates.</p><p><strong>Materials and methods: </strong>Results from 42 EDiR candidates across Europe were compared to those from 26 fourth-year medical students who answered exclusively using the ChatGPT-4o in a prospective study (October 2024). The challenge consisted of 52 recall or understanding-based EDiR multiple-response questions, all without visual inputs.</p><p><strong>Results: </strong>The GPT-4o achieved a mean score of 82.1 ± 3.0%, significantly outperforming the EDiR candidates with 49.4 ± 10.5% (p < 0.0001). In particular, chatGPT-4o demonstrated higher true positive rates while maintaining lower false positive rates compared to EDiR candidates, with a higher accuracy rate in all radiology subspecialties (p < 0.0001) except informatics (p = 0.20). There was near-perfect agreement between GPT-4 responses (κ = 0.872) and moderate agreement among EDiR participants (κ = 0.334). Exit surveys revealed that all participants used the copy-and-paste feature, and 73% submitted additional questions to clarify responses.</p><p><strong>Conclusions: </strong>GPT-4o significantly outperformed human candidates in low-order, text-based EDiR multiple-response questions, demonstrating higher accuracy and reliability. These results highlight GPT-4o's potential in answering text-based radiology questions. Further research is necessary to investigate its performance across different question formats and candidate populations to ensure broader applicability and reliability.</p><p><strong>Critical relevance statement: </strong>GPT-4o significantly outperforms human candidates in factual radiology text-based questions in the EDiR, excelling especially in identifying correct responses, with a higher accuracy rate compared to radiologists.</p><p><strong>Key points: </strong>In EDiR text-based questions, ChatGPT-4o scored higher (82%) than EDiR participants (49%). Compared to radiologists, GPT-4o excelled in identifying correct responses. GPT-4o responses demonstrated higher agreement (κ = 0.87) compared to EDiR candidates (κ = 0.33).</p>","PeriodicalId":13639,"journal":{"name":"Insights into Imaging","volume":"16 1","pages":"66"},"PeriodicalIF":4.1,"publicationDate":"2025-03-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11929644/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143691712","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Zhengping Zhang, Kede Mi, Zhaojun Wang, Xiaoyan Yang, Shuping Meng, Xingcang Tian, Yanzhu Han, Yuling Qu, Li Zhu, Juan Chen
{"title":"Using optimized CT type to predict histological classifications of thymic epithelial tumors: a radiomics integrated analysis.","authors":"Zhengping Zhang, Kede Mi, Zhaojun Wang, Xiaoyan Yang, Shuping Meng, Xingcang Tian, Yanzhu Han, Yuling Qu, Li Zhu, Juan Chen","doi":"10.1186/s13244-025-01933-7","DOIUrl":"10.1186/s13244-025-01933-7","url":null,"abstract":"<p><strong>Objective: </strong>To develop and externally validate an integrated model that utilizes optimized radiomics features from non-contrast-enhanced CT (NE-CT) or contrast-enhanced CT (CE-CT), along with morphological features and clinical risk factors, to predict histological classifications of thymic epithelial tumors (TETs).</p><p><strong>Methods: </strong>A total of 182 patients with TET, classified as the low-risk group and the high-risk group based on histology, were divided into a training cohort (N = 122, center 1) and an external validation cohort (N = 60, center 2). Radiomics features were extracted from different CT types, followed by feature selection, including consistency, correlation, and importance tests, to generate Rad-scores for both NE-CT and CE-CT. The integrated model was developed by combining the optimal Rad-score, morphological features, and clinical risk factors using multivariate logistic regression. Model performance was assessed by the area under the receiver operating characteristic curve (AUC) and compared by Delong test. A nomogram was used to visually present the integrated model.</p><p><strong>Results: </strong>A total of 851 radiomics features were extracted, with NE-CT and CE-CT Rad-scores consisting of four and five features, respectively. The AUCs of the CE-CT Rad-score were higher than those of the NE-CT Rad-score in both the training cohort (0.783 vs 0.749) and the external validation cohort (0.775 vs 0.723, p = 0.361). The integrated model, combining five morphological features and the CE-CT Rad-score, achieved AUCs of 0.814 and 0.802 in the training and external validation cohorts, respectively.</p><p><strong>Conclusion: </strong>The integrated model, incorporating radiomics features from CE-CT and morphological features, can help to identify the histological classifications of TETs.</p><p><strong>Critical relevance statement: </strong>This study developed an integrated model based on radiomics features from contrast-enhanced CT and morphological features, demonstrating that the integrated model has impressive predictive capability in distinguishing histological classifications of thymic epithelial tumors through external validation.</p><p><strong>Key points: </strong>Radiomics features extracted from CT more effectively represented thymic epithelial tumor (TET) heterogeneity than morphological features. The radiomics model using contrast-enhanced CT outperformed that using non-contrast-enhanced CT in identifying histological classifications of TET. The integrated model, combining radiomics and morphological features, exhibited the highest performance in predicting TET histological classifications.</p>","PeriodicalId":13639,"journal":{"name":"Insights into Imaging","volume":"16 1","pages":"67"},"PeriodicalIF":4.1,"publicationDate":"2025-03-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11929666/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143691879","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Intestinal fibrosis assessment in Crohn's disease patient using unenhanced spectral CT combined with 3D-printing technique.","authors":"Qiapeng Huang, Zhihui Chen, Ruonan Zhang, Huasong Cai, Xufeng Yang, Xiaodi Shen, Lili Huang, Xinyue Wang, Qingzhu Zheng, Mingzhe Li, Ziyin Ye, Xubin Liu, Ren Mao, Yangdi Wang, Jinjiang Lin, Zhoulei Li","doi":"10.1186/s13244-025-01914-w","DOIUrl":"10.1186/s13244-025-01914-w","url":null,"abstract":"<p><strong>Objectives: </strong>To integrate multiple parameters derived from unenhanced spectral CT with 3D-printing technique to accurately evaluate intestinal lesions in patients with Crohn's disease (CD).</p><p><strong>Methods: </strong>Patients with proven CD who underwent preoperative spectral CT and surgery were included. The spectral CT images and histopathological specimens were achieved by employing 3D-printing technique. Diagnostic models were developed utilizing Z-Effective, Electron Density (ED), and Hounsfield unit (HU) values derived from spectral CT, along with spectral curve slopes λ<sub>1</sub> and λ<sub>2</sub>, as well as ΔHU<sub>MonoE</sub>. The area under the receiver operating characteristic curve (AUC) and the influence of inflammation on the efficacy of the models were analyzed.</p><p><strong>Results: </strong>The ED and HU at MonoE 50 keV of the spectral CT were determined to exhibit the highest correlation with the fibrosis degree of the diseased intestine. The training dataset yielded an AUC of 0.828 (95% CI: 0.705-0.951). The sensitivity and specificity were calculated to be 77.3% and 82.6%, respectively. The AUC of the validation set was 0.812 (95% CI: 0.676-0.948) with a sensitivity of 63.6% and specificity of 89.7%. Moreover, our model demonstrated enhanced diagnostic accuracy for detecting fibrosis with an AUC value of 0.933 (95% CI: 0.856-1.000), sensitivity of 90.9%, and specificity of 87.0%, after regulating the influence of inflammation.</p><p><strong>Conclusion: </strong>The integration of unenhanced multi-parametric spectral CT and 3D-printing technique seems to be able to assess the intestinal fibrosis. Our diagnostic model remains effective in assessing the severity of fibrosis under presence of inflammation.</p><p><strong>Critical relevance statement: </strong>Our diagnostic model accurately assessed the degree of intestinal wall fibrosis in Crohn's disease patients by using unenhanced spectral CT and 3D-printing technique, which could facilitate individualized treatment.</p><p><strong>Key points: </strong>Evaluating the extent of Crohn's disease-related fibrosis is important. The combination of 3D-printing technique and multi-parametric spectral CT enhances diagnostic accuracy. The developed model using spectral CT allows for the assessment of intestinal fibrosis using multi-parameters.</p>","PeriodicalId":13639,"journal":{"name":"Insights into Imaging","volume":"16 1","pages":"62"},"PeriodicalIF":4.1,"publicationDate":"2025-03-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11926292/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143669818","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Naomi S Sakai, Andrew A Plumb, Norin Ahmed, Kashfia Chowdhury, Yakup Kilic, Maira Hameed, Anisha Patel, Anisha Bhagwanani, Emma Helbren, Rachel Hyland, Gauraang Bhatnagar, Harbir Sidhu, Hannah Lambie, James M Franklin, Maryam Mohsin, Elen Thomson, Darren Boone, Damian Tolan, Safi Rahman, Nik Ding, Gordon W Moran, Stuart Bloom, Ailsa Hart, Alex Menys, Simon Travis, Steve Halligan, Stuart A Taylor
{"title":"MRI assessment of body composition for prediction of therapeutic response to biologic agents in patients with Crohn's disease.","authors":"Naomi S Sakai, Andrew A Plumb, Norin Ahmed, Kashfia Chowdhury, Yakup Kilic, Maira Hameed, Anisha Patel, Anisha Bhagwanani, Emma Helbren, Rachel Hyland, Gauraang Bhatnagar, Harbir Sidhu, Hannah Lambie, James M Franklin, Maryam Mohsin, Elen Thomson, Darren Boone, Damian Tolan, Safi Rahman, Nik Ding, Gordon W Moran, Stuart Bloom, Ailsa Hart, Alex Menys, Simon Travis, Steve Halligan, Stuart A Taylor","doi":"10.1186/s13244-025-01930-w","DOIUrl":"10.1186/s13244-025-01930-w","url":null,"abstract":"<p><strong>Objectives: </strong>Altered body fat and muscle mass in Crohn's disease (CD) have been linked to adverse disease course and outcomes. Prediction of treatment response or remission (RoR) of small bowel CD (SBCD) to biologic therapy remains challenging. We aimed to establish the prognostic value of body composition parameters measured using MR enterography (MRE) for RoR at 1 year in patients with SBCD commencing biologic therapy.</p><p><strong>Methods: </strong>Participants were identified from those recruited to a prospective, multicentre study investigating the predictive ability of motility MRI for 1 year RoR in patients starting biologic therapy for active SBCD (MOTILITY trial). Myopenia, skeletal muscle:fat and visceral:subcutaneous fat were measured from baseline MRE. RoR at 1 year was judged using a composite of clinical and morphological MRE parameters. We compared the likelihood of RoR in patients with and without myopenia or low skeletal muscle:fat using logistic regression models.</p><p><strong>Results: </strong>Ninety-six participants were included (mean age 38.2 years; 40 (42%) female). There were 34 (35%) responders. There was no significant difference in RoR at 1 year between those patients with and without skeletal muscle myopenia (OR: 0.85, 95% CI: 0.27, 2.66, p-value: 0.78), or those with or without low skeletal muscle:fat (OR: 0.71, 95% CI: 0.19, 2.71, p-value: 0.62).</p><p><strong>Conclusions: </strong>Body composition parameters demonstrated no value for predicting therapeutic RoR in patients commencing biologic therapy for SBCD.</p><p><strong>Critical relevance statement: </strong>Prediction of response to biologic therapy in small bowel Crohn's disease (SBCD) remains challenging. Body composition parameters cannot predict biologic therapeutic response or remission for SBCD reliably.</p><p><strong>Key points: </strong>Altered body fat and muscle mass in Crohn's disease have been linked to adverse outcomes. Prediction of response to biologic therapy in small bowel CD (SBCD) would be useful for treatment optimisation. Body composition parameters measured using MRI cannot reliably predict biological therapeutic response or remission for SBCD.</p>","PeriodicalId":13639,"journal":{"name":"Insights into Imaging","volume":"16 1","pages":"61"},"PeriodicalIF":4.1,"publicationDate":"2025-03-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11923306/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143663340","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Ozden Camurdan, Toygar Tanyel, Esma Aktufan Cerekci, Deniz Alis, Emine Meltem, Nurper Denizoglu, Mustafa Ege Seker, Ilkay Oksuz, Ercan Karaarslan
{"title":"Annotation-efficient, patch-based, explainable deep learning using curriculum method for breast cancer detection in screening mammography.","authors":"Ozden Camurdan, Toygar Tanyel, Esma Aktufan Cerekci, Deniz Alis, Emine Meltem, Nurper Denizoglu, Mustafa Ege Seker, Ilkay Oksuz, Ercan Karaarslan","doi":"10.1186/s13244-025-01922-w","DOIUrl":"10.1186/s13244-025-01922-w","url":null,"abstract":"<p><strong>Objectives: </strong>To develop an efficient deep learning (DL) model for breast cancer detection in mammograms, utilizing both weak (image-level) and strong (bounding boxes) annotations and providing explainable artificial intelligence (XAI) with gradient-weighted class activation mapping (Grad-CAM), assessed by the ground truth overlap ratio.</p><p><strong>Methods: </strong>Three radiologists annotated a balanced dataset of 1976 mammograms (cancer-positive and -negative) from three centers. We developed a patch-based DL model using curriculum learning, progressively increasing patch sizes during training. The model was trained under varying levels of strong supervision (0%, 20%, 40%, and 100% of the dataset), resulting in baseline, curriculum 20, curriculum 40, and curriculum 100 models. Training for each model was repeated ten times, with results presented as mean ± standard deviation. Model performance was also tested on an external dataset of 4276 mammograms to assess generalizability.</p><p><strong>Results: </strong>F1 scores for the baseline, curriculum 20, curriculum 40, and curriculum 100 models were 80.55 ± 0.88, 82.41 ± 0.47, 83.03 ± 0.31, and 83.95 ± 0.55, respectively, with ground truth overlap ratios of 60.26 ± 1.91, 62.13 ± 1.2, 62.26 ± 1.52, and 64.18 ± 1.37. In the external dataset, F1 scores were 74.65 ± 1.35, 77.77 ± 0.73, 78.23 ± 1.78, and 78.73 ± 1.25, respectively, maintaining a similar performance trend.</p><p><strong>Conclusion: </strong>Training DL models with a curriculum method and a patch-based approach yields satisfactory performance and XAI, even with a limited set of densely annotated data, offering a promising avenue for deploying DL in large-scale mammography datasets.</p><p><strong>Critical relevance: </strong>This study introduces a DL model for mammography-based breast cancer detection, utilizing curriculum learning with limited, strongly labeled data. It showcases performance gains and better explainability, addressing challenges of extensive dataset needs and DL's \"black-box\" nature.</p><p><strong>Key points: </strong>Increasing numbers of mammograms for radiologists to interpret pose a logistical challenge. We trained a DL model leveraging curriculum learning with mixed annotations for mammography. The DL model outperformed the baseline model with image-level annotations using only 20% of the strong labels. The study addresses the challenge of requiring extensive datasets and strong supervision for DL efficacy. The model demonstrated improved explainability through Grad-CAM, verified by a higher ground truth overlap ratio. He proposed approach also yielded robust performance on external testing data.</p>","PeriodicalId":13639,"journal":{"name":"Insights into Imaging","volume":"16 1","pages":"60"},"PeriodicalIF":4.1,"publicationDate":"2025-03-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11923329/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143663336","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Adrian P Brady, Graciano Paulo, Boris Brkljacic, Christian Loewe, Martina Szucsich, Monika Hierath
{"title":"Current status of radiologist staffing, education and training in the 27 EU Member States.","authors":"Adrian P Brady, Graciano Paulo, Boris Brkljacic, Christian Loewe, Martina Szucsich, Monika Hierath","doi":"10.1186/s13244-025-01925-7","DOIUrl":"10.1186/s13244-025-01925-7","url":null,"abstract":"<p><p>This second article of a series of three publications summarises the radiologist situation regarding staffing as well as education and training as analysed by The European Union Radiation, Education, Staffing & Training (EU-REST) study. Despite certain limitations posed by the dependence on survey responses, the results demonstrate that, for both workforce and education/training, considerable heterogeneity exists between Member States, which will impact healthcare delivery and the level of knowledge, skills, and competencies available. The number of radiologists per million inhabitants varies from 51 to 270. 16 out of 27 Member States have Radiologist numbers below the EU average of 127, and 45% of Radiologists in Europe are over 51 years old (in 2022). Clear guidance and metrics about workforce availability for the professions involved in the use of ionising radiation are needed to secure and improve the quality of healthcare delivery in Europe. Although the main scope of the EU-REST study was education, training and workforce availability, an attempt was made to characterise the numbers of pieces of medical imaging and radiotherapy equipment. CRITICAL RELEVANCE STATEMENT: Clear guidance and metrics on radiologist staffing and education/training are needed to address workforce shortages and harmonise education and training standards across the EU-27. KEY POINTS: The article describes the radiologist situation regarding staffing and radiation protection education in the EU Member States. Radiologist staffing and training vary considerably across the EU-27. The fact that more than half of the EU Member States have radiologist numbers below the EU average, and the large proportion of radiologists over 51 years of age, show that clear guidance and metrics are needed to ensure future quality of radiological care.</p>","PeriodicalId":13639,"journal":{"name":"Insights into Imaging","volume":"16 1","pages":"59"},"PeriodicalIF":4.1,"publicationDate":"2025-03-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11910488/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143634004","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"RadLex and SNOMED CT integration: a pilot study for standardising radiology classification.","authors":"Merit Marquis, Igor Bossenko, Peeter Ross","doi":"10.1186/s13244-025-01935-5","DOIUrl":"10.1186/s13244-025-01935-5","url":null,"abstract":"<p><strong>Background: </strong>Effective communication and information exchange across diverse platforms are critical in healthcare data systems. However, the presence of multiple coding systems and varying standards creates discrepancies and misalignments, highlighting the need for innovative solutions to address these challenges.</p><p><strong>Objective: </strong>The study aimed to develop a technical and semantic interoperability method specifically for radiology procedures, utilising the terminology management tool TermX to facilitate efficient data exchange and utilisation in healthcare.</p><p><strong>Results: </strong>The study resulted in a revised RadLex data model using SNOMED CT, accompanied by a mapping guide and a classification system for X-ray and angiography procedures. This classification system consists of nineteen distinct properties, each defined by specific value sets derived from SNOMED CT terminology. A total of 380 concepts were utilised to describe the 622 procedures examined comprehensively.</p><p><strong>Conclusion: </strong>Through twelve design cycles involving in-depth analysis and iterative refinement, the mapping of angiography and X-ray procedures was successfully achieved, culminating in the creation and validation of a universal model that enhances both primary and secondary data collection. The efficacy and innovation of this system pave the way for further advancements in healthcare interoperability.</p><p><strong>Critical relevance statement: </strong>The innovative integration achieved in this study for standardising radiology classification promises to improve data management practices and enhance patient care outcomes through increased interoperability within the healthcare sector.</p><p><strong>Key points: </strong>A universal radiology procedure model to enhance capture would be valuable. A tool to facilitate technical and semantic interoperability for efficient data exchange in healthcare was created. This system could pave the way for futher advancements in healthcare interoperability.</p>","PeriodicalId":13639,"journal":{"name":"Insights into Imaging","volume":"16 1","pages":"58"},"PeriodicalIF":4.1,"publicationDate":"2025-03-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11906921/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143624437","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}