BMC Medical Imaging最新文献

{"title":"Automated 3D segmentation of the hyoid bone in CBCT using nnU-Net v2: a retrospective study on model performance and potential clinical utility.","authors":"Ismail Gümüssoy, Emre Haylaz, Suayip Burak Duman, Fahrettin Kalabalik, Seyda Say, Ozer Celik, Ibrahim Sevki Bayrakdar","doi":"10.1186/s12880-025-01797-9","DOIUrl":"https://doi.org/10.1186/s12880-025-01797-9","url":null,"abstract":"<p><strong>Objective: </strong>This study aimed to identify the hyoid bone (HB) using the nnU-Net based artificial intelligence (AI) model in cone beam computed tomography (CBCT) images and assess the model's success in automatic segmentation.</p><p><strong>Methods: </strong>CBCT images of 190 patients were randomly selected. The raw data was converted to DICOM format and transferred to the 3D Slicer Imaging Software (Version 4.10.2; MIT, Cambridge, MA, USA). HB was labeled manually using the 3D Slicer. The dataset was divided into training, validation, and test sets in a ratio of 8:1:1. The nnU-Net v2 architecture was utilized to process the training and test datasets, generating the algorithm weight factors. To assess the model's accuracy and performance, a confusion matrix was employed. F1-score, Dice coefficient (DC), 95% Hausdorff distance (95% HD), and Intersection over Union (IoU) metrics were calculated to evaluate the results.</p><p><strong>Results: </strong>The model's performance metrics were as follows: DC = 0.9434, IoU = 0.8941, F1-score = 0.9446, and 95% HD = 1.9998. The receiver operating characteristic (ROC) curve was generated, yielding an AUC value of 0.98.</p><p><strong>Conclusion: </strong>The results indicated that the nnU-Net v2 model achieved high precision and accuracy in HB segmentation on CBCT images. Automatic segmentation of HB can enhance clinicians' decision-making speed and accuracy in diagnosing and treating various clinical conditions.</p><p><strong>Clinical trial number: </strong>Not applicable.</p>","PeriodicalId":9020,"journal":{"name":"BMC Medical Imaging","volume":"25 1","pages":"217"},"PeriodicalIF":2.9,"publicationDate":"2025-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144538220","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Comparison of accuracy of CT parameters across chest, low-dose lung, and abdominal CT in diagnosing steatotic liver disease.","authors":"Yan Li, Jiahao Wang, Gengyu Xu, Yixin Si, Kaiyao Huang, Yinquan Ye, Yun Peng, Yuanyuan Liu","doi":"10.1186/s12880-025-01791-1","DOIUrl":"https://doi.org/10.1186/s12880-025-01791-1","url":null,"abstract":"<p><strong>Objective: </strong>This study aims to determine the accuracy of computed tomography (CT) parameters obtained from three various scanning protocols (chest CT, low-dose lung CT, and abdominal CT) in diagnosing steatotic liver disease (SLD).</p><p><strong>Materials and methods: </strong>This retrospective study included 234 individuals who underwent chest CT, low-dose lung CT, or abdominal CT. SLD presence or absence was confirmed through ultrasound in all participants. Two radiologists independently measured the CT attenuation values of the liver (CT_L) and spleen (CT_S). The differences (CT_L-S) and ratios (CT_L/S) between liver and spleens values were calculated. Independent sample t-tests or Mann-Whitney U tests were used to compare CT_S, CT_L, CT_L-S, and CT_L/S between SLD and control groups. One-way analysis of covariance or Kruskal-Wallis H tests were conducted to compare the parameters across scanning protocols. Receiver operating characteristic (ROC) analysis was performed.</p><p><strong>Results: </strong>The parameters (CT_L, CT_L-S, and CT_L/S) were significantly lower in the SLD group than in the control group across all scanning protocols (P < 0.001). In the control group, significant differences in CT_S and CT_L were observed among the three scanning protocols (P < 0.05), while non-statistically significant differences were found for CT_L-S or CT_L/S across protocols in either group (P > 0.05). The ROC analysis revealed abdominal CT_L as the most accurate diagnostic marker for SLD (area under the curve [AUC]: 0.894, sensitivity: 90.91%, specificity: 83.33%). CT_L-S maintained stable diagnostic performance across protocols (AUC range: 0.813-0.816). The low-dose protocol achieved the best performance for CT_L/S (AUC: 0.810), demonstrating high specificity (94.34%) despite moderate sensitivity (64.81%).</p><p><strong>Conclusion: </strong>Both chest CT and low-dose CT-derived parameters demonstrated diagnostic accuracy comparable to that of abdominal CT in assessing SLD, suggesting their potential as viable alternatives in specific clinical scenarios.</p>","PeriodicalId":9020,"journal":{"name":"BMC Medical Imaging","volume":"25 1","pages":"232"},"PeriodicalIF":2.9,"publicationDate":"2025-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144538223","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Computed tomography-based radiomics predicts prognostic and treatment-related levels of immune infiltration in the immune microenvironment of clear cell renal cell carcinoma.","authors":"Shiyan Song, Wenfei Ge, Xiaochen Qi, Xiangyu Che, Qifei Wang, Guangzhen Wu","doi":"10.1186/s12880-025-01749-3","DOIUrl":"https://doi.org/10.1186/s12880-025-01749-3","url":null,"abstract":"&lt;p&gt;&lt;strong&gt;Objectives: &lt;/strong&gt;The composition of the tumour microenvironment is very complex, and measuring the extent of immune cell infiltration can provide an important guide to clinically significant treatments for cancer, such as immune checkpoint inhibition therapy and targeted therapy. We used multiple machine learning (ML) models to predict differences in immune infiltration in clear cell renal cell carcinoma (ccRCC), with computed tomography (CT) imaging pictures serving as a model for machine learning. We also statistically analysed and compared the results of multiple typing models and explored an excellent non-invasive and convenient method for treatment of ccRCC patients and explored a better, non-invasive and convenient prediction method for ccRCC patients.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Methods: &lt;/strong&gt;The study included 539 ccRCC samples with clinicopathological information and associated genetic information from The Cancer Genome Atlas (TCGA) database. The Single Sample Gene Set Enrichment Analysis (ssGSEA) algorithm was used to obtain the immune cell infiltration results as well as the cluster analysis results. ssGSEA-based analysis was used to obtain the immune cell infiltration levels, and the Boruta algorithm was further used to downscale the obtained positive/negative gene sets to obtain the immune infiltration level groupings. Multifactor Cox regression analysis was used to calculate the immunotherapy response of subgroups according to Tumor Immune Dysfunction and Exclusion (TIDE) algorithm and subgraph algorithm to detect the difference in survival time and immunotherapy response of ccRCC patients with immune infiltration. Radiomics features were screened using LASSO analysis. Eight ML algorithms were selected for diagnostic analysis of the test set. Receiver operating characteristic (ROC) curve was used to evaluate the performance of the model. Draw decision curve analysis (DCA) to evaluate the clinical personalized medical value of the predictive model.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Results: &lt;/strong&gt;The high/low subtypes of immune infiltration levels obtained by optimisation based on the Boruta algorithm were statistically different in the survival analysis of ccRCC patients. Multifactorial immune infiltration level combined with clinical factors better predicted survival of ccRCC patients, and ccRCC with high immune infiltration may benefit more from anti-PD-1 therapy. Among the eight machine learning models, ExtraTrees had the highest test and training set ROC AUCs of 1.000 and 0.753; in the test set, LR and LightGBM had the highest sensitivity of 0.615; LR, SVM, ExtraTrees, LightGBM and MLP had higher specificities of 0.789, 1.000, 0.842, 0.789 and 0.789, respectively; and LR, ExtraTrees and LightGBM had the highest accuracy of 0. 719, 0.688 and 0.719 respectively. Therefore, the CT-based ML achieved good predictive results in predicting immune infiltration in ccRCC, with the ExtraTrees machine learning algorithm being optimal.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Con","PeriodicalId":9020,"journal":{"name":"BMC Medical Imaging","volume":"25 1","pages":"213"},"PeriodicalIF":2.9,"publicationDate":"2025-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144538234","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A nomogram based on multiparametric magnetic resonance imaging radiomics for prediction of acute pancreatitis activity.","authors":"Ting-Ting Liu, You-Qiang Hu, Ning-Jun Yu, Xue-Ying Zhang, Dong-Lin Jiang, Jiang Luo, Yong Chen, Di Tao, Xing-Hui Li, Xiao-Ming Zhang","doi":"10.1186/s12880-025-01778-y","DOIUrl":"10.1186/s12880-025-01778-y","url":null,"abstract":"<p><strong>Purpose: </strong>In acute pancreatitis (AP), disease activity is defined as the reversible manifestation of the disease. The aim of this study was to develop a nomogram for predicting disease activity in AP based on multiparametric magnetic resonance imaging (MRI) radiomics.</p><p><strong>Methods: </strong>This retrospective study included 310 patients with first-episode AP from two medical centers in China. Patients from the first medical center were randomly divided into a training cohort (n = 122) and an internal validation cohort (n = 123) in a 5:5 ratio. Patients from the second medical center were used as the external independent validation cohort (n = 65). Radiomics features were extracted from multiparametric MRI images based on pancreatic parenchymal regions. The least absolute shrinkage and selection operator (LASSO) was used for feature screening, logistic regression was used to establish radiomic feature, and statistically significant laboratory parameters were incorporated to construct the nomogram. The area under the receiver operator characteristic curve assessed the predictive performance of the nomogram. Furthermore, decision curve analysis (DCA) was used to assess the clinical utility of the nomogram, and the disease activity was validated against follow-up clinical outcomes (e.g., organ failure progression, ICU admission) and imaging-confirmed changes within one-week after MRI.</p><p><strong>Results: </strong>The AUCs of the radiomic signature were 0.808 (training cohort), 0.789 (internal validation cohort), and 0.783 (external validation cohort). Radiomic signature, extrapancreatic inflammation on MRI (EPIM) scores, and WBC count were identified as independent risk factors for the activity of AP and were therefore included in the nomogram. The AUC of the nomogram were 0.881 (training cohort), 0.922 (internal validation cohort) and 0.912 (external validation cohort). Additionally, the nomogram model obtained the greatest net benefit, according to the results of decision curves Based on the follow-up results, we also found that AP patients with higher disease activity were more likely to experience exacerbations.</p><p><strong>Conclusions: </strong>This nomogram can accurately predict the activity of AP patients, thus providing objective monitoring of the patient's course and potentially improving patient prognosis.</p>","PeriodicalId":9020,"journal":{"name":"BMC Medical Imaging","volume":"25 1","pages":"241"},"PeriodicalIF":2.9,"publicationDate":"2025-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144538217","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Automatic recognition and differentiation of pulmonary contusion and bacterial pneumonia based on deep learning and radiomics.","authors":"Tie Deng, Junbang Feng, Xingyan Le, Yuwei Xia, Feng Shi, Fei Yu, Yiqiang Zhan, Xinghua Liu, Chuanming Li","doi":"10.1186/s12880-025-01802-1","DOIUrl":"10.1186/s12880-025-01802-1","url":null,"abstract":"<p><strong>Background: </strong>In clinical work, there are difficulties in distinguishing pulmonary contusion(PC) from bacterial pneumonia(BP) on CT images by the naked eye alone when the history of trauma is unknown. Artificial intelligence is widely used in medical imaging, but its diagnostic performance for pulmonary contusion is unclear. In this study, artificial intelligence was used for the first time to identify lung contusion and bacterial pneumonia, and its diagnostic performance was compared with that of manual.</p><p><strong>Methods: </strong>In this retrospective study, 2179 patients between April 2016 and July 2022 from two hospitals were collected and divided into a training set, an internal validation set, an external validation set. PC and BP were automatically recognized, segmented using VB-net and radiomics features were automatically extracted. Four machine learning algorithms including Decision Trees, Logistic Regression, Random Forests and Support Vector Machines(SVM) were using to built the models. De-long test was used to compare the performance among models. The best performing model and four radiologists diagnosed the external validation set, and compare the diagnostic efficacy of human and artificial intelligence.</p><p><strong>Results: </strong>VB-net automatically detected and segmented PC and BP. Among the four machine learning models we've built, De-long test showed that SVM model had the best performance, with AUC, accuracy, sensitivity, and specificity of 0.998 (95% CI: 0.995-1), 0.980, 0.979, 0.982 in the training set, 0.891 (95% CI: 0.854-0.928), 0.979, 0.750, 0.860 in the internal validation set, 0.885 (95% CI: 0.850-0.920), 0.903, 0.976, 0.794 in the external validation set. The diagnostic ability of the SVM model was superior to that of human (P < 0.05).</p><p><strong>Conclusion: </strong>Our VB-net automatically recognizes and segments PC and BP in chest CT images. SVM model based on radiomics features can quickly and accurately differentiate between them with higher accuracy than experienced radiologist.</p>","PeriodicalId":9020,"journal":{"name":"BMC Medical Imaging","volume":"25 1","pages":"234"},"PeriodicalIF":2.9,"publicationDate":"2025-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144538222","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Comparison of cardiac magnetic resonance and speckle tracking echocardiography in cardiac evaluation of children with acute myocarditis with preserved left ventricular function.","authors":"Fatos Alkan, Onur Bircan, Fatma Can, Yuksel Pabuscu, Senol Coskun","doi":"10.1186/s12880-025-01772-4","DOIUrl":"10.1186/s12880-025-01772-4","url":null,"abstract":"<p><p>This study aimed to evaluate the reliability and efficacy of speckle tracking echocardiography (STE) compared to cardiac magnetic resonance (CMR) in assessing left ventricular function and segmental involvement in patients with acute myocarditis and preserved left ventricular systolic function. We analyzed conventional echocardiography, two-dimensional STE, and CMR findings in 33 pediatric patients (aged 0-18 years) diagnosed with acute myocarditis. The STE results were compared with CMR findings. The mean patient age was 14.67 years (± 2.88), with 13 (39.4%) females and 20 (60.6%) males. The mean ejection fraction (EF) was 68.54% (± 6.54), and the mean fractional shortening (FS) was 38.20% (± 5.34). All patients had an EF greater than 55%, with no detected wall motion abnormalities. Mild pleural effusion was observed in 4 (12.1%) patients. A significantly reduced left ventricular global longitudinal strain (LV-GLS) pattern was detected in 45.4% (n = 15) of patients, with an average LV-GLS value of -18.12% (± 3.5). The LV-GLS reduction was uniformly distributed across all left ventricular segments (LV-GLSAP2: -18.1 ± 3.85, LV-GLSAP3: -17.33 ± 4.34, LV-GLSAP4: -18.88 ± 4.20). STE measurements showed a mean left ventricular end-diastolic volume of 71.54 ± 24.41 and an end-systolic volume of 37.62 ± 16.42, with a mean EF of 48.52 ± 9.39%. CMR identified widespread myocardial contrast enhancement in 25 (75.7%) patients. When comparing STE to CMR, using an LV-GLS cut-off value of -18%, the sensitivity and specificity for diagnosing myocarditis were 52% and 63%, respectively. Lowering the cut-off to -16% reduced sensitivity to 40% but increased specificity to 75%. No significant association was found between decreased LV-GLS values (<-18%) and late gadolinium enhancement on CMR or regional edema (p > 0.05). Our findings suggest that STE is a valuable diagnostic tool for detecting cardiac involvement, particularly in focal cases of pediatric acute myocarditis with normal EF. While CMR remains the gold standard, STE provides a practical, accessible alternative for monitoring disease progression in suspected myocarditis cases.</p>","PeriodicalId":9020,"journal":{"name":"BMC Medical Imaging","volume":"25 1","pages":"243"},"PeriodicalIF":2.9,"publicationDate":"2025-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144538224","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Non-invasive assessment of liver function in cirrhosis using iodine density difference between the portal vein and hepatic vein on spectral CT: correlation with Child-Pugh grades.","authors":"Lu Yu, Zhengping Song, Zhoulei Li, Wen Lv, Fei Yang, Xuehua Li, Canhui Sun","doi":"10.1186/s12880-025-01768-0","DOIUrl":"10.1186/s12880-025-01768-0","url":null,"abstract":"<p><strong>Objective: </strong>To explore the feasibility and clinical significance of assessing liver function damage in patients with post-hepatitis cirrhosis using spectral CT by measuring the iodine density difference and CT value difference between the portal vein and hepatic vein.</p><p><strong>Methods: </strong>A study was conducted involving 65 patients with post-hepatitis cirrhosis (30 with Child-Pugh grade A, 28 with grade B, and 7 with grade C) and 82 healthy controls. All underwent dual-phase enhanced spectral CT scans of the upper abdomen. Post-processing with IntelliSpace Portal software yielded iodine density (ID), 45 keV virtual monoenergetic (VMI), and conventional CT (HU) images. The mean iodine densities of the portal vein (ID_P) and hepatic vein (ID_V) were measured, and the difference in ID (ID_d-value), VMI (VMI_d-value), and HU (HU_d-value) between the portal vein and hepatic vein was calculated. These values were compared between the control group and different Child-Pugh grades of cirrhosis, as well as between the compensated cirrhosis group (Child-Pugh grade A) and the decompensated cirrhosis group (grade B/C). The ability to diagnose decompensated cirrhosis was evaluated by plotting ROC curves.</p><p><strong>Results: </strong>The ID_d-value, VMI_d-value, and HU _d-value were significantly higher in the cirrhosis group compared to the control group and were also higher in the grade B/C group compared to the grade A group (all p<0.05). Among all patients with cirrhosis, the AUROC values for diagnosing grade B/C using ID_d-value, VMI_d-value, HU_d-value, NID_d-value, NVMI _d-value, and NHU_d-value were 0.774, 0.798, 0.736, 0.775, 0.774, and 0.757, respectively (all p < 0.05).</p><p><strong>Conclusions: </strong>The iodine density difference (ID_d-value) and CT value difference (VMI_d-value and HU _d-value) between the portal vein and hepatic vein, as measured by spectral CT, demonstrate a significant positive correlation with the Child-Pugh classification of liver function in cirrhotic patients. These quantitative parameters provide a simple and non-invasive approach for assessing liver function in cirrhotic patients.</p><p><strong>Clinical trial number: </strong>Not applicable.</p>","PeriodicalId":9020,"journal":{"name":"BMC Medical Imaging","volume":"25 1","pages":"254"},"PeriodicalIF":2.9,"publicationDate":"2025-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144538231","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Predicting abnormal epicardial adipose tissue in psoriasis patients by integrating radiomics from non-contrast chest CT with serological biomarkers.","authors":"Rui Han, Juan Hou, Ping Xia, Yan Xing, Wenya Liu","doi":"10.1186/s12880-025-01755-5","DOIUrl":"10.1186/s12880-025-01755-5","url":null,"abstract":"<p><strong>Background: </strong>Psoriasis patients frequently present with cardiovascular comorbidities, which maybe associated with abnormal epicardial adipose tissue (EAT). This study aimed to evaluate the predictive value of radiomics features derived from non-contrast chest CT (NCCT) combined with serological parameters for identifying abnormal EAT in psoriasis.</p><p><strong>Methods: </strong>In this retrospective case-control study, we enrolled consecutive psoriasis patients who underwent chest NCCT between September 2021 and February 2024, along with a matched healthy control group. Psoriasis patients were stratified into mild-to-moderate (PASI ≤ 10) and severe (PASI > 10) groups based on the Psoriasis Area and Severity Index (PASI). Using TIMESlice, we extracted EAT volume, CT values, and 86 radiomics features. The cohort was randomly divided into a training (70%) and test (30%) set. LASSO regression selected radiomic features to calculate the Rad_Score. Serum uric acid (UA) and C-reactive protein (CRP) levels were collected. We compared EAT volume, CT values, Rad_Score, UA, and CRP between groups and developed three models: Model A (UA, CRP, EAT CT values), Model B (Rad_Score), and Model C (UA, CRP, EAT CT values, Rad_Score). Model accuracy was evaluated using ROC curves (P < 0.05).</p><p><strong>Results: </strong>The study included 77 psoriasis patients and 76 matched controls. Psoriasis patients had higher UA and CRP levels than controls (both P < 0.001). EAT CT value was higher in psoriasis (P = 0.020), with no volume difference. Eight radiomics features and Rad_Score significantly differed between groups (P < 0.001), and Rad_Score also higher in severe group than that in mild-to-moderate group (P < 0.001). Model C showed the highest AUC in both sets: training 0.947 and test 0.895, indicating superior predictive performance.</p><p><strong>Conclusions: </strong>Combining radiomics features, EAT CT values, UA, and CRP in a predictive model accurately predicts EAT abnormalities in psoriasis, potentially improving cardiovascular comorbidity diagnosis.</p><p><strong>Clinical trial number: </strong>Not applicable.</p>","PeriodicalId":9020,"journal":{"name":"BMC Medical Imaging","volume":"25 1","pages":"240"},"PeriodicalIF":2.9,"publicationDate":"2025-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144538243","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Deep learning for automated segmentation of radiation-induced changes in cerebral arteriovenous malformations following radiosurgery.","authors":"Hsing-Hao Ho, Huai-Che Yang, Wen-Xiang Yang, Cheng-Chia Lee, Hsiu-Mei Wu, I-Chun Lai, Ching-Jen Chen, Syu-Jyun Peng","doi":"10.1186/s12880-025-01796-w","DOIUrl":"https://doi.org/10.1186/s12880-025-01796-w","url":null,"abstract":"<p><strong>Background: </strong>Despite the widespread use of stereotactic radiosurgery (SRS) to treat cerebral arteriovenous malformations (AVMs), this procedure can lead to radiation-induced changes (RICs) in the surrounding brain tissue. Volumetric assessment of RICs is crucial for therapy planning and monitoring. RICs that appear as hyper-dense areas in magnetic resonance T2-weighted (T2w) images are clearly identifiable; however, physicians lack tools for the segmentation and quantification of these areas. This paper presents an algorithm to calculate the volume of RICs in patients with AVMs following SRS. The algorithm could be used to predict the course of RICs and facilitate clinical management.</p><p><strong>Methods: </strong>We trained a Mask Region-based Convolutional Neural Network (Mask R-CNN) as an alternative to manual pre-processing in designating regions of interest. We also applied transfer learning to the DeepMedic deep learning model to facilitate the automatic segmentation and quantification of AVM edema regions in T2w images.</p><p><strong>Results: </strong>The resulting quantitative findings were used to explore the effects of SRS treatment among 28 patients with unruptured AVMs based on 139 regularly tracked T2w scans. The actual range of RICs in the T2w images was labeled manually by a clinical radiologist to serve as the gold standard in supervised learning. The trained model was tasked with segmenting the test set for comparison with the manual segmentation results. The average Dice similarity coefficient in these comparisons was 71.8%.</p><p><strong>Conclusions: </strong>The proposed segmentation algorithm achieved results on par with conventional manual calculations in determining the volume of RICs, which were shown to peak at the end of the first year after SRS and then gradually decrease. These findings have the potential to enhance clinical decision-making.</p><p><strong>Trial registration: </strong>Not applicable.</p>","PeriodicalId":9020,"journal":{"name":"BMC Medical Imaging","volume":"25 1","pages":"218"},"PeriodicalIF":2.9,"publicationDate":"2025-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144538240","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Development and validation of a risk nomogram predicting pneumothorax requiring chest tube placement post-percutaneous CT-guided lung biopsy.","authors":"Masha Bondarenko, Jianxiang Zhang, Ulysis Hugo Baal, Brian Lam, Gunvant Chaudhari, Yoo Jin Lee, Jamie Schroeder, Maya Vella, Brian Haas, Thienkhai Vu, Kimberly Kallianos, Jonathan Liu, Shravan Sridhar, Brett Elicker, Jae Ho Sohn","doi":"10.1186/s12880-025-01794-y","DOIUrl":"https://doi.org/10.1186/s12880-025-01794-y","url":null,"abstract":"<p><strong>Background: </strong>Pneumothorax requiring chest tube after CT-guided transthoracic lung biopsy presents added clinical risk and costs to the healthcare system. Identifying high-risk patients can prompt alternative biopsy modes and/or better preparation for more focused post-procedural care. We aimed to develop and externally validate a risk nomogram for pneumothorax requiring chest tube placement following CT-guided lung biopsy, leveraging quantitative emphysema algorithm.</p><p><strong>Methods: </strong>This two-center retrospective study included patients who underwent CT-guided lung biopsy from between 1994 and 2023. Data from one hospital was set aside for validation (n = 613). Emphysema severity was quantified and categorized to 3-point scale using a previously published algorithm based on 3×3×3 kernels and Hounsfield thresholding, and a risk calculator was developed using forward variable selection and logistic regression. The model was validated using bootstrapping and Harrell's C-index.</p><p><strong>Results: </strong>2,512 patients (mean age, 64.47 years ± 13.38 [standard deviation]; 1250 men) were evaluated, of whom 157 (6.7%) experienced pneumothorax complications requiring chest tube placement. After forward variable selection to reduce the covariates to maximize clinical usability, the risk score was developed using age over 60 (OR 1.80 [1.15-2.93]), non-prone patient position (OR 2.48 [1.63-3.75]), and severe emphysema (OR 1.99 [1.35-2.94]). The nomogram showed a mean absolute error of 0.5% in calibration and Harrell's C-index of 0.664 in discrimination in the internal cohort.</p><p><strong>Conclusion: </strong>The developed nomogram predicts age over 60, non-prone position during biopsy, and severe emphysema to be most predictive of pneumothorax requiring chest tube placement following CT-guided lung biopsy.</p>","PeriodicalId":9020,"journal":{"name":"BMC Medical Imaging","volume":"25 1","pages":"220"},"PeriodicalIF":2.9,"publicationDate":"2025-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144538251","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}