BMC Medical Imaging[JOURNAL]
Sun
200 papers
Predictive value of a combined nomogram model integrating ultrasound and cytology for 2025 ATA risk stratification of lymph node negative papillary thyroid carcinoma.
Qin X, Liu X, He F … +2 more , Xiao W, Ye X
BMC Med Imaging · 2026 May · PMID 42215886 · Full text
OBJECTIVES: This study aimed to predict the 2025 American Thyroid Association (ATA) risk stratification of lymph node negative (N0) papillary thyroid carcinoma (PTC) patients using preoperative ultrasonic and cytological...
OBJECTIVES: This study aimed to predict the 2025 American Thyroid Association (ATA) risk stratification of lymph node negative (N0) papillary thyroid carcinoma (PTC) patients using preoperative ultrasonic and cytological features to provide guidance for clinical treatment strategies. METHODS: In this multicenter, retrospective study, 500 N0 PTC patients who underwent total thyroidectomy with lymph node dissection were included across 2 institutions from September 2018 to February 2024. Patients were categorized into low, intermediate-high recurrence risk groups based on the 2025 ATA risk stratification system. Univariate logistic regression analysis was performed to assess the relationship between ultrasonic and cytological features and 2025 ATA risk stratification. Significant features (P < 0.05) were then incorporated into a multivariate logistic regression model to identify independent predictors of risk stratification. A Nomogram was constructed using predictors from the final multivariate logistic regression. RESULTS: Papillary like arrangement, Escape like arrangement, Nucleolus, Size, Echo, Margin, and Extracapsular extension (ECE) were identified as independent predictors of 2025 ATA risk stratification. A Nomogram model was developed based on these predictors, demonstrating good discrimination with a C-index of 0.799. The calibration curve further demonstrated excellent consistency between the predicted and actual preoperative 2025 ATA risk stratifications. Additionally, the Nomogram displayed a C-index of 0.778 in the testing cohort. CONCLUSIONS: Risk stratification in N0 PTC patients correlates with factors such as Papillary arrangement, Escape like arrangement, Nucleolus, Size, Echo, Margin, and ECE, emphasizing the necessity of closely monitoring patients presenting with these risk factors. Additionally, the Nomogram model integrating seven preoperative risk factors specifically tailored for solitary N0 PTC patients was devised, showcasing notable predictive accuracy for preoperative 2025 ATA risk stratification.
Preoperative prediction of visceral pleural invasion in clinical stage T1 invasive lung adenocarcinoma using intratumoral and peritumoral radiomics: a multicenter development, validation, and comparison study.
Dong H, Yang Z, Qiu Y … +8 more , Qiu F, Wang X, Xie Z, Yang J, Shou L, Guan X, Ye X, Xu X
BMC Med Imaging · 2026 May · PMID 42210170 · Full text
BACKGROUND: Accurate preoperative identification of visceral pleural invasion (VPI) is crucial for surgical planning and lymph node dissection in patients with early-stage lung adenocarcinoma. This study aims to develop...
BACKGROUND: Accurate preoperative identification of visceral pleural invasion (VPI) is crucial for surgical planning and lymph node dissection in patients with early-stage lung adenocarcinoma. This study aims to develop and independently validate a predictive model for VPI in clinical stage T1 invasive lung adenocarcinoma, based on intratumoral and peritumoral radiomic features. Additionally, it initiates a preliminary exploration into the interpretability of the optimal model. METHODS: This retrospective analysis gathered imaging and clinical data from 316 patients diagnosed with potentially pleural-invading invasive lung adenocarcinoma based on computed tomography (CT) images across three medical institutions. A total of 109 patients from Center 1 and 99 patients from Center 2 were randomly assigned to a training set (146 patients) and a validation set (62 patients), in a 7:3 ratio, respectively. And 108 patients from Center 3 constituted an independent external test set. Feature selection was executed using univariate and multivariate analyses, Spearman correlation, minimal redundancy maximum relevance (mRMR), and least absolute shrinkage and selection operator (LASSO) techniques, to develop of clinical CT semantics, intratumoral radiomics, peritumoral radiomics, and combined intratumoral-peritumoral radiomics models. The influential features of the optimal model were subsequently ranked and evaluated using the SHapley additive explanation (SHAP). RESULTS: The intratumoral model, exhibiting the lowest Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC), indicated a superior model fitting in the training set. It also showed the highest Matthews Correlation Coefficient (MCC) and Diagnostic Odds Ratio (DOR) in both validation and external test sets, demonstrating the best predictive capacity. The Area Under the Curve (AUC) values for the intratumoral model across training, validation, and external test sets were 0.837 (95% CI: 0.812-0.860), 0.825 (95% CI: 0.794-0.858), and 0.819 (95% CI: 0.791-0.845) respectively; these values surpassed those of the peritumoral, combined models, and clinical CT semantic models. SHAP analysis results showed Small-Area-Low-Gray-Level-Emphasis emerges as the pivotal feature. CONCLUSION: Intratumoral radiomic features possess significant predictive value for VPI in clinical stage T1 invasive lung adenocarcinoma, aiding physicians in formulating precise clinical strategies.
Cardiac echocardiographic analysis with multi-scale effective fusion module: a novel stroke prediction approach.
Xie J, Tan D, Zheng T … +2 more , Chen Y, Wei L
BMC Med Imaging · 2026 May · PMID 42210140 · Full text
BACKGROUND: Stroke remains a leading cause of death and disability, and early risk stratification is critical for prevention. Existing models based only on clinical factors may miss subtle cardiac structural and hemodyna...
BACKGROUND: Stroke remains a leading cause of death and disability, and early risk stratification is critical for prevention. Existing models based only on clinical factors may miss subtle cardiac structural and hemodynamic information visible on echocardiography. We aimed to develop a multimodal stroke prediction model integrating multi-view echocardiographic images and clinical indicators. METHODS: In this retrospective study, 712 hypertensive patients (10,992 echocardiographic images; 27 clinical variables) were included. Long-axis, short-axis, and apical four-chamber views were analyzed. We developed a Multi-Scale Effective Fusion (MSEF) module combining Global Feature Fusion, Multi-Feature Reconstruction, Channel Attention, and Positional Attention to improve multi-scale feature representation. Imaging features were integrated with clinical variables to build multimodal models. Model performance was evaluated on validation and test sets using accuracy, precision, recall, and F1 score. RESULTS: The MSEF-based imaging model outperformed comparator fusion variants and achieved an accuracy of 76.8% and an F1 score of 64.7% on the test set. After integrating clinical indicators, performance further improved, with a test accuracy of 80.2% and an F1 score of 72.1%. CONCLUSIONS: The proposed MSEF-based multimodal framework improves stroke risk prediction by effectively combining echocardiographic and clinical information, and may support earlier risk identification and clinical decision-making.
Extracellular volume fraction derived from spectral CT for liver function reserve evaluation and complication prediction in clinically stable liver cirrhosis.
Zhang L, Sun Y, Zhang X … +2 more , Zhang Z, Ji Y
BMC Med Imaging · 2026 May · PMID 42210133 · Full text
OBJECTIVES: To investigate the value of spectral CT-derived extracellular volume fraction (fECV) in assessing liver functional reserve and predicting cirrhosis-related complications in clinically stable cirrhotic patient...
OBJECTIVES: To investigate the value of spectral CT-derived extracellular volume fraction (fECV) in assessing liver functional reserve and predicting cirrhosis-related complications in clinically stable cirrhotic patients. MATERIALS AND METHODS: This retrospective study enrolled clinically stable patients with cirrhosis who underwent contrast-enhanced spectral CT examinations, as well as a control group without major diseases. Iodine concentrations in the liver and aorta were measured on equilibrium phase images to calculate the fECV. The diagnostic performance of fECV for Child-Pugh classification was evaluated using receiver operating characteristic (ROC) curve analysis. The predictive ability of fECV for cirrhosis-related complications was assessed using the Kaplan-Meier time-dependent ROC curves, and Cox proportional hazards regression models. RESULTS: A total of 116 patients with cirrhosis (mean age, 60.20 ± 12.11 years) and 38 control subjects (mean age, 60.16 ± 8.93 years) were enrolled in this study. The area under the ROC curve (AUC) of fECV for differentiating Child-Pugh classes was 0.852-0.881 (all p < 0.001). The high fECV group (≥38.48%) demonstrated a significantly higher cumulative risk of complications (p < 0.001). fECV showed superior predictive performance for long-term (3-year) complications (AUC = 0.813, p < 0.001). Furthermore, fECV was identified as an independent risk factor for cirrhosis-related complications (Model 1: HR = 1.073, p = 0.004; Model 2: HR = 1.078, p < 0.001). CONCLUSION: fECV is a promising imaging biomarker for liver functional reserve assessment and complication prediction in cirrhosis.
Correction: An efficient Alzheimer's disease detection by NV classifier with BWTDL approach using MRI image.
Dwarkanth PS, Anitha R, Babu GH … +5 more , Ahammad SKH, Karamti H, Kumar A, Oza AD, Moges DM
BMC Med Imaging · 2026 May · PMID 42204484 · Full text
Abstract loading — click title to view on PubMed.
Correction: In vivo study of intravoxel incoherent motion imaging in assessing obesity-related kidney injury.
Xia X, Huang Y, Qin Y … +5 more , Song J, Cui W, Cheng Z, Feng Y, Cai X
BMC Med Imaging · 2026 May · PMID 42204483 · Full text
Abstract loading — click title to view on PubMed.
The value of sound touch elastography in the noninvasive staging of fibrosis in IgA nephropathy.
Chen RR, Lan ZK, Huang WJ … +3 more , Wei GY, Meng XY, Ge XY
BMC Med Imaging · 2026 May · PMID 42204462 · Full text
OBJECTIVE: To investigate the value of Sound Touch Elastography (STE) in fibrosis grading of IgA Nephropathy (IgAN), focusing on the diagnostic performance of the mean elasticity (STE) alone and combined with conventiona...
OBJECTIVE: To investigate the value of Sound Touch Elastography (STE) in fibrosis grading of IgA Nephropathy (IgAN), focusing on the diagnostic performance of the mean elasticity (STE) alone and combined with conventional ultrasound parameters. METHODS: A total of 129 biopsy-proven IgAN patients underwent renal conventional ultrasound and STE.Based on the Oxford MEST-C T-score(reflecting the percentage of tubular atrophy/interstitial fibrosis area), patients were classified as T0 (≤ 25%), T1 (26-50%), and T2 (> 50%). Non-parametric tests were used to compare intergroup differences, correlations and diagnostic performance were assessed by Spearman and receiver operating characteristic (ROC) analyses. RESULTS: STE correlated significantly with T-score (r = 0.61, P < 0.001), eGFR (r = - 0.71, P < 0.001), and Scr (r = 0.67, P < 0.001). STE differed significantly across T0, T1, and T2 groups, as well as between S0 and S1 groups (all P < 0.001). Within-session measurement reliability was excellent (ICC = 0.990, 95% CI: 0.988-0.993). The combined model integrating STE and renal length (RL) achieved an AUC of 0.863 (95% CI: 0.795-0.932) for discriminating T0 from T1-2 and 0.826 (95% CI: 0.751-0.901) for discriminating T2 from T0-1, outperforming STE alone. This model provided high specificity (92.9%) and PPV (95.1%) for ruling in fibrosis, and high sensitivity (87.0%) and NPV (88.1%) for excluding severe fibrosis. CONCLUSION: STE, particularly when combined with renal length, demonstrated good diagnostic performance for the noninvasive assessment of renal fibrosis severity in IgA nephropathy, especially for identifying significant (≥ T1) and severe (T2) fibrosis. The distinction between adjacent moderate and severe fibrosis stages remains challenging, and the proposed thresholds require external validation. These findings support the potential role of STE as a noninvasive tool to complement fibrosis evaluation in IgAN.
Distinct neural substrates of disease severity and cognitive impairment in COPD: a multimodal MRI study.
Wang W, Xue Y, Wang G … +5 more , Zhao M, Xiao Y, Huang W, Yang Y, Liu Z
BMC Med Imaging · 2026 May · PMID 42204460 · Full text
BACKGROUND: The neural substrates distinguishing Chronic Obstructive Pulmonary Disease (COPD) patients with mild cognitive impairment (MCI) from those with preserved cognition are not fully understood. This study aims to...
BACKGROUND: The neural substrates distinguishing Chronic Obstructive Pulmonary Disease (COPD) patients with mild cognitive impairment (MCI) from those with preserved cognition are not fully understood. This study aims to investigate the distinct neural substrates associated with disease severity and cognitive impairment in COPD using multimodal MRI. METHODS: Thirty-three COPD patients with MCI (COPD-MCI), 30 cognitively normal patients (COPD-nMCI), and 34 healthy controls underwent structural and functional MRI. Voxel-Based Morphometry (VBM) and seed-based Functional Connectivity (FC) analyses identified gray matter volume (GMV) and network alterations. Diagnostic utility was evaluated using ROC analysis. RESULTS: Common to all COPD patients, GMV atrophy and reduced FC in the sensorimotor network (left postcentral gyrus/inferior parietal lobule) were observed, correlating with disease duration (r = -0.523) and distinguished patients from healthy controls (cross-validated AUC = 0.746). Notably, the MCI subgroup was characterized by focal atrophy in the left thalamus and functional decoupling within the visual association network. Visual network FC served as a robust discriminator between the two COPD subgroups (cross-validated AUC = 0.833) and correlated with MoCA scores (r = 0.585). Additionally, reduced fronto-striatal connectivity was observed in the MCI group, while its integrity correlated with lung function in the nMCI group (r = 0.666). CONCLUSIONS: Our findings suggest distinct substrates underlying clinical heterogeneity in COPD. Sensorimotor alterations likely reflect a generalized disease trait of chronic respiratory burden, whereas the integrity of higher-order associative networks, involving the thalamus, visual, and fronto-striatal circuits, acts as a critical checkpoint against cognitive decline. These findings offer novel biomarkers for the early detection of COPD-related cognitive impairment.
Photoacoustic imaging of oxygen saturation in the radial arterial wall and perivascular tissue in patients with systemic lupus erythematosus.
Zhong X, Pan Y, Liu L … +10 more , Zheng Y, Wang H, Liu X, Lin X, Peng G, Yu M, Sheng Y, Xu J, Luo S, Liu Y
BMC Med Imaging · 2026 May · PMID 42192526 · Full text
OBJECTIVES: Systemic lupus erythematosus (SLE) is associated with vasculitis of varying severity. Conventional ultrasonography faces challenges in accurately assessing this condition, whereas the development of photoacou...
OBJECTIVES: Systemic lupus erythematosus (SLE) is associated with vasculitis of varying severity. Conventional ultrasonography faces challenges in accurately assessing this condition, whereas the development of photoacoustic (PA) imaging offers a promising approach. This study aims to evaluate the clinical utility of PA imaging in assessing vasculitis changes in patients with SLE. METHODS: A total of 102 patients diagnosed with SLE were enrolled as the case group and subdivided into two subgroups: the non-active group and the active group, according to the systemic lupus erythematosus disease activity index 2000 (SLEDAI-2000) scoring criteria. 35 healthy volunteers were recruited as the control group. PA was performed to measure the oxygen saturation of the bilateral radial arterial wall and perivascular tissues. RESULTS: (1) Active SLE patients showed significantly elevated oxygen saturation (SO) levels of radial artery wall and perivascular tissues compared to nonactive patients and healthy controls (p < 0.05). (2) Correlation analysis showed that the average oxygen saturation of bilateral radial artery wall and perivascular tissues was positively correlated with SLEDAI-2000 (r = 0.571) and negatively correlated with complement 3 (C3), complement 4 (C4), and serum albumin (r=-0.453, -0.493, -0.313, respectively). Univariate and multivariate regression analysis showed that SLEDAI-2000 was independently associated with SO. (3) The SO was associated with disease activity status in SLE patients, with an area under the curve (AUC) of 79.6%. CONCLUSIONS: Photoacoustic imaging effectively detects the oxygen saturation changes in the radial arterial wall and perivascular tissues in patients with SLE. These findings suggest its potential as a tool for evaluating SLE-related vasculitis.
Exploring the value of peritumoral brain zone for classification of two malignant brain tumors based on the MRI interpretable models.
Zhao E, Yang YF, Zhang H … +6 more , Yang YY, Shi Y, Li B, Song X, Lou S, Yang C
BMC Med Imaging · 2026 May · PMID 42192376 · Full text
OBJECTIVES: This study aims to develop and validate a novel multimodal interpretable artificial intelligence model capable of fusing radiomics features and imaging features to accurately classify primary central nervous...
OBJECTIVES: This study aims to develop and validate a novel multimodal interpretable artificial intelligence model capable of fusing radiomics features and imaging features to accurately classify primary central nervous system lymphoma (PCNSL) and glioblastoma (GBM), and to explore the impact of including the tumor and different peritumoral brain zones (PBZs) on the diagnostic performance of the model. MATERIALS AND METHODS: A retrospective cohort of 242 patients with PCNSL or GBM, along with their multi-sequence MRI scans and clinical data, was enrolled from two medical centers. After image preprocessing, tumor and PBZ regions were segmented as volumes of interest (VOIs), and imaging features were evaluated. Based on different combinations of three image sequences and four VOIs, a total of seven models (T1WI-Tumor-Model, T2WI-Tumor-Model, T1CE-Tumor-Model, T1CE-Tumor-Edema-Model, T1CE-Tumor-PBZ10mm-Model, T1CE-Tumor-PBZ20mm-Model, and Multimodal-Radiomics-Model) were established. Receiver operating characteristic (ROC) curves, sensitivity, specificity, calibration curves, and Decision Curve Analysis (DCA) were used to assess the models' performance and clinical application value. Additionally, the best model was analyzed for interpretability using SHapley Additive exPlanations (SHAP). RESULTS: The T1CE-Tumor-Edema-Model and Multimodal-Radiomics-Model demonstrated superior predictive performance among the seven models, with AUCs of 0.91 and 0.94 in the external validation cohort, respectively. The SHAP interpretation of the results revealed that the Rad-Score and Clinical-Rad-Score features contributed the most to the models' decision-making process. CONCLUSION: A radiomics framework based on interpretable machine learning is proposed. Based on this framework, models combining different ranges of peritumoral brain regions with tumors were constructed for the first time, and radiomics features were fused with imaging features to improve classification accuracy and interpretability in distinguishing between PCNSL and GBM.
F‑FDG PET/CT in metastatic chordoma: a retrospective analysis of imaging features and potential clinical relevance.
Song L, Hua P, Wei F … +1 more , Zhang W
BMC Med Imaging · 2026 May · PMID 42192335 · Full text
BACKGROUND: Systematic data on F-fluorodeoxyglucose positron emission tomography/computed tomography (F-FDG PET/CT) in metastatic chordoma are scarce. This study aimed to evaluate its imaging characteristics and potentia...
BACKGROUND: Systematic data on F-fluorodeoxyglucose positron emission tomography/computed tomography (F-FDG PET/CT) in metastatic chordoma are scarce. This study aimed to evaluate its imaging characteristics and potential relevance to clinical assessment and management. METHODS: In this single-center retrospective analysis, 21 patients with pathologically confirmed chordoma and prior treatment underwent F-FDG PET/CT for suspected recurrence or metastasis. Metastatic disease was diagnosed per a composite standard (biopsy, imaging progression, or characteristic multimodal findings). Relevant clinical and histopathological data were collected. Images were independently reviewed by two experienced nuclear medicine physicians for metabolic activity and whole-body disease assessment. They assessed metabolic activity at the primary site and performed whole-body evaluation. For each metastatic lesion, maximum standardized uptake value (SUVmax) and size were measured; CT features were also documented. Interobserver agreement for key assessments was formally evaluated. RESULTS: Metastatic disease was identified in 11 of 21 patients (52.4%). Metastases were found in bone (7 patients), soft tissue (8 patients), and lung (5 patients). Site-specific metabolic patterns emerged: pulmonary metastases had lower FDG avidity (median SUVmax 2.3) correlated with size, whereas bone and soft-tissue avidity (SUVmax 3.8-3.9) was size-independent. Notably, in 4 out of 21 patients (19.0%), PET/CT detected metastases outside the field of view of conventional imaging. Interobserver agreement was perfect for metastatic status and excellent for total lesion counts. A case illustrated differential treatment response linked to baseline metabolic avidity. CONCLUSIONS: F-FDG PET/CT offers a reproducible whole-body assessment for chordoma, enabling the detection of occult metastases and revealing metabolic heterogeneity of potential clinical relevance. These hypothesis-generating findings suggest a potential role in surveillance and personalized management, warranting further prospective validation before clinical implementation.
Prediction of myelosuppression in cervical cancer after concurrent chemoradiotherapy by CT radiomics-based model.
Sun C, Chu A, Liu S … +8 more , Song R, Yang J, Liu X, Gan L, Wang Y, Liu Z, Wang X, Li M
BMC Med Imaging · 2026 May · PMID 42192333 · Full text
BACKGROUND: Concurrent chemoradiotherapy (CCRT) was highly effective in treating cervical cancer (CC) but raised the risk of bone marrow suppression. However, models to predict the risk of myelosuppression in CC patients...
BACKGROUND: Concurrent chemoradiotherapy (CCRT) was highly effective in treating cervical cancer (CC) but raised the risk of bone marrow suppression. However, models to predict the risk of myelosuppression in CC patients after CCRT based on computed tomography (CT) imaging histology are immature. METHODS: The Region of interest (ROI) of the CT images was segmented, and radiomic features were extracted. Next, important features were further selected. The training and test sets were split 7:3, and 8 machine learning algorithms were applied to classify the features in the training set. The model's performance was assessed in the test set, and the best algorithm was chosen. The selected algorithm predicted the radiomic feature score. The clinical features were compared between mild and severe groups, and a clinical model was constructed using the best algorithm, and predicted clinical feature scores. Finally, logistic regression models were used to identify independent prognostic factors for myelosuppression, and nomograms were drawn. RESULTS: 14 important radiomics features were selected. The random forest (RF) algorithm was considered the best machine learning method in both the classification imaging model and the clinical classification model (0.735 (95% CI = 0.624-0.846), accuracy = 0.719 (95% CI = 0.618-0.802), sensitivity = 0.703 (95% CI = 0.507-0.845), recall = 0.703 (95% CI = 0.507-0.845), and F1 score = 0.782 (95% CI = 0.513-1.000)). A logistic regression model built from the predictions of the RF al-gorithm showed that both the rad_score and Clinical_score could serve as independent factors (p value < 0.05). Furthermore, the nomogram constructed based on these two scores were found to have moderate predictive performance (AUC = 0.724 (95% CI = 0.650-0.798)). CONCLUSION: A CT-based radiomics model combined with clinical characteristics demonstrated favourable predictive performance in forecasting bone marrow suppression among cervical cancer patients undergoing concurrent chemoradiotherapy. However, further multicentre studies are required to validate its clinical utility.
CT-based nomogram for estimating visceral pleural invasion in solid pulmonary adenocarcinoma nodules with pleural contact: a retrospective single-center study.
Xu R, Tang J, Yuan Q … +5 more , Ding M, Zhang J, Yao W, Zhang D, Zhao H
BMC Med Imaging · 2026 May · PMID 42192325 · Full text
BACKGROUND: Visceral pleural invasion (VPI) is an adverse prognostic factor in lung adenocarcinoma. Accurate preoperative estimation of VPI risk in solid tumors with pleural contact may provide useful supplementary infor...
BACKGROUND: Visceral pleural invasion (VPI) is an adverse prognostic factor in lung adenocarcinoma. Accurate preoperative estimation of VPI risk in solid tumors with pleural contact may provide useful supplementary information during preoperative assessment. METHODS: This was a retrospective study of 162 patients. All patients had surgically resected, pathologically confirmed solid lung adenocarcinoma in pleural contact between November 2018 and August 2025. Patients were classified into VPI-positive and VPI-negative groups based on postoperative pathology. The patients were randomly divided into a training cohort and a testing cohort at a ratio of 7:3 for internal validation. Multivariable logistic regression was performed to identify independent risk factors for VPI. A nomogram prediction model was developed based on the multivariable analysis. Its predictive performance was then evaluated using receiver operating characteristic (ROC) curves, calibration curves, and decision curve analysis (DCA). RESULTS: Of the 162 patients with solid lung adenocarcinoma nodules, 58 had VPI confirmed by pathology. Multivariate logistic regression analysis identified spiculation, pleural indentation, and pleural contact length as independent risk factors for predicting VPI. A nomogram based on these three CT features showed good discriminative performance in both the training and testing cohorts, with AUCs of 0.901 (95% CI: 0.846-0.956) and 0.864 (95% CI: 0.737-0.937), respectively. Calibration curves showed the predictions matched observations well. Decision curve analysis suggested potential utility in preoperative risk assessment. CONCLUSION: Among patients with solid lung adenocarcinoma nodules with pleural contact, we developed a nomogram based on routinely assessed CT semantic features to estimate the preoperative probability of VPI. The model is intended as a practical supplementary tool for preoperative VPI risk estimation.
Comparison of systolic and diastolic CT-FFR for myocardial ischemia diagnosis.
Guo W, Sheng A, Wang Y … +17 more , Lu Y, Zhao S, Yin L, Zhao Y, Xu H, Zhang H, Qiao G, Shen L, Pang Y, Yin J, Yao Z, Li C, Yang S, Yan C, He W, He F, Zeng M
BMC Med Imaging · 2026 May · PMID 42192315 · Full text
BACKGROUND: CT-derived fractional flow reserve (CT-FFR) is a powerful tool for identifying hemodynamic ischemia. Coronary CT angiography (CCTA) images with the best quality in a cardiac cycle are conventionally reconstru...
BACKGROUND: CT-derived fractional flow reserve (CT-FFR) is a powerful tool for identifying hemodynamic ischemia. Coronary CT angiography (CCTA) images with the best quality in a cardiac cycle are conventionally reconstructed in clinical practice. To compare the diagnostic performance of machine learning (ML)-based CT-FFR between systolic and diastolic phases in identifying myocardial ischemia, using invasive fractional flow reserve (FFR) as a reference standard. METHODS: From December 2020 to October 2021, consecutive coronary artery disease (CAD) patients who underwent coronary computed tomography angiography (CCTA) and invasive FFR were prospectively enrolled. Prospective electrocardiographic (ECG)-triggered scan from 30% to 80% of R-R interval was applied on image extraction of all the patients. CT-FFR was implemented in systolic and diastolic phases for each target vessel. RESULTS: Ninety-six patients (mean age 62 ± 8, 31.3% female) with 98 vessels were successfully simulated in both systolic and diastolic phases. The area under the curve (AUC) was significantly higher in diastolic CT-FFR (0.885 vs. 0.790, p < 0.001). The limits of agreement (LoA) range was narrower in diastolic CT-FFR (-0.134 to 0.148), compared with systolic CT-FFR (-0.234 to 0.273). Diastolic CT-FFR performed better than systolic CT-FFR in sensitivity (81.5% vs. 66.7%), specificity (83.1% vs. 71.8%), and accuracy (82.7% vs. 70.4%, all p < 0.05). The LoA range was narrower for diastolic CT-FFR across all stenosis degree groups, with diagnostic performance better in sensitivity (84.0% vs. 68.0%) and accuracy (81.4% vs. 58.1%, p < 0.001) among severe stenosis group. CONCLUSIONS: In ML-based CT-FFR methods, diastolic CT-FFR showed better diagnostic performance and narrower LoA than systolic CT-FFR. TRIAL REGISTRATION: The study has been registered through the Ethics Committee of Zhongshan Hospital, Fudan University, with the clinical trial number B2020-088R. The registration date is May 15, 2020.
Leveraging the non-contrast CT component of PET/CT: an AI-driven delta-radiomics approach to monitor treatment response in metastatic breast cancer.
Kahraman EG, Unal OU, Taskaynatan H … +3 more , Ozdemir O, Budak E, Selver MA
BMC Med Imaging · 2026 May · PMID 42185986 · Full text
PURPOSE: 18 F-FDG PET/CT is the standard modality for monitoring treatment response in metastatic breast cancer. This study aims to evaluate the predictive value of delta-radiomics derived solely from the low-dose, non-c...
PURPOSE: 18 F-FDG PET/CT is the standard modality for monitoring treatment response in metastatic breast cancer. This study aims to evaluate the predictive value of delta-radiomics derived solely from the low-dose, non-contrast CT component acquired during routine PET/CT imaging-without requiring an additional dedicated CT examination or extra contrast administration-for monitoring response to CDK4/6 inhibitors in de novo metastatic hormone receptor-positive (HR+)/HER2-negative breast cancer. METHODS: This retrospective study included 33 patients with bone-predominant metastatic breast cancer. Delta radiomic features were extracted from the non-contrast CT component of paired baseline and follow-up 18 F-FDG PET/CT scans. Patients were stratified into Responders (Complete or Partial Response) and Non-Responders (Stable or Progressive Disease) based on standard PERCIST criteria. We developed an integrated machine learning model using logistic regression with elastic net regularization, validated via leave-one-out cross-validation (LOOCV). RESULTS: The cohort consisted of 25 Responders and 8 Non-Responders. Non-Responders exhibited distinct longitudinal increases in Delta_Pct_shape_Elongation and Delta_Pct_firstorder_90Percentile compared to Responders. The integrated model, combining these features with clinical variables, achieved an Area Under the Curve (AUC) of 0.930, significantly outperforming the baseline clinical-only model (AUC = 0.775). While the default threshold prioritized sensitivity (96.0%) with limited specificity (25.0%), post-hoc threshold optimization maximizing the Youden index demonstrated a highly balanced performance, achieving 88.0% sensitivity and 87.5% specificity. CONCLUSIONS: Delta radiomics analysis of the routinely acquired non-contrast CT component of PET/CT provides substantial incremental prognostic value over standard clinical variables. This approach demonstrates the potential of utilizing existing low-dose CT data as a cost-effective, supportive biomarker for the early prediction of therapeutic resistance.
Metabolic dysfunction-associated steatotic liver disease is associated with myocardial ischemia in patients with suspected or known coronary artery disease, especially in patients without obstructive coronary artery disease.
Tian Y, Yu W, Zhang F … +8 more , Wang J, Shao X, Liu B, Yang X, Wan P, Chen Y, Li S, Wang Y
BMC Med Imaging · 2026 May · PMID 42185825 · Full text
BACKGROUND: To investigate the relationship between metabolic dysfunction-associated steatotic liver disease (MASLD) and myocardial ischemia in patients with suspected or known coronary artery disease (CAD). METHODS: Thi...
BACKGROUND: To investigate the relationship between metabolic dysfunction-associated steatotic liver disease (MASLD) and myocardial ischemia in patients with suspected or known coronary artery disease (CAD). METHODS: This retrospective study enrolled 281 patients with suspected or known CAD who underwent single-photon emission computed tomography myocardial perfusion imaging (SPECT-MPI) at the Third Affiliated Hospital of Soochow University from January 1, 2022, to December 31, 2023. The mean CT values of the liver and spleen, coronary artery calcification score (CACS), and epicardial fat volume (EFV) were acquired through non-enhanced computed tomography (CT). Myocardial ischemia is defined by SDS ≥ 2 detected by MPI. Obstructive CAD is defined as the degree of coronary artery lumen narrowing ≥ 50%. MASLD was defined as mean liver/spleen CT density ratio < 1 or mean liver CT density ≤ 40HU. Clinical data such as age, gender, body mass index (BMI), diabetes mellitus (DM) and hypertension were collected. RESULTS: Of the whole 281 patients, 140 (49.8%) had myocardial ischemia, and 107 (38.1%) had MASLD. Compared to the non-ischemic group, the myocardial ischemic group had higher EFV (137.11 cm³ versus 123.12 cm³) and BMI (25.31 kg/m² versus 24.40 kg/m²), a higher proportion of DM (32.0% versus 19.2%) and MASLD (56.4% versus 19.9%) (all P < 0.05). Multivariate regression analysis indicated that MASLD was the independent risk factor for myocardial ischemia (OR = 4.97, 95% CI: 2.68 ~ 9.24, P < 0.001). Smooth curve fitting results showed that the liver/spleen mean CT density ratio was linearly correlated with the risk of myocardial ischemia. MASLD is an independent risk factor for myocardial ischemia regardless of the presence of obstructive CAD. (obstructive CAD: OR = 5.04, 95% CI: 1.81 ~ 14.00, P = 0.002; without obstructive CAD: OR = 7.43, 95% CI: 2.69 ~ 20.52, P < 0.001). Higher OR value in patients without obstructive CAD indicates that the correlation is more significant. CONCLUSION: MASLD is significantly associated with myocardial ischemia in suspected or known CAD patients, especially in patients without obstructive CAD, independent of traditional risk factors.
A clinical-radiomics model based on multiparametric MRI for discriminating solitary primary spinal tumors from solitary spinal metastases.
Wang C, Xu J, Zhang M … +2 more , Hao D, Lang N
BMC Med Imaging · 2026 May · PMID 42185785 · Full text
OBJECTIVE: To develop and validate a clinical-radiomics model based on multiparametric MRI for differentiating solitary primary spinal tumors from solitary spinal metastases. METHODS: This dual-center retrospective study...
OBJECTIVE: To develop and validate a clinical-radiomics model based on multiparametric MRI for differentiating solitary primary spinal tumors from solitary spinal metastases. METHODS: This dual-center retrospective study included 510 patients with pathologically confirmed spinal tumors, randomly split into training (n = 328), internal validation (n = 82), and external test (n = 100) cohorts. Radiomics and deep learning features were extracted from T1-weighted, T2-weighted, T2-weighted fat-suppressed, and T1-weighted contrast-enhanced sequences. Three models were constructed and compared: a radiomics model (Rad-M), a deep learning-radiomics fusion model (DRad-M), and a clinical-radiomics model (CRad-M) that integrated radiomics features with patient age. The performance of seven machine learning classifiers was evaluated for each model. RESULTS: The CRad-M, utilizing a logistic regression (LR) classifier, demonstrated superior performance, achieving areas under the curve (AUCs) of 0.909 and 0.824 on the internal and external test sets, respectively. It significantly outperformed both the Rad-M and DRad-M models (all p < 0.05). The incorporation of deep learning features did not yield a significant improvement over the radiomics-only model. Calibration and decision curve analyses confirmed the robust clinical utility of the CRad-M. CONCLUSION: The proposed LR-based CRad-M is an effective non-invasive tool for the preoperative differentiation of solitary primary spinal tumors and solitary spinal metastases, with its performance enhanced by the integration of clinical data (age) alongside radiomic features.
Evaluating the grade and IDH mutation status of gliomas by using intravoxel incoherent motion (IVIM) MRI and DSC perfusion MRI.
Tacyildiz C, Aslan K, İncesu L … +2 more , Yergin Tacyildiz S, Genç B
BMC Med Imaging · 2026 May · PMID 42177397 · Full text
BACKGROUND: This study aimed to evaluate the diagnostic performance of intravoxel incoherent motion (IVIM) imaging and DSC perfusion MRI to predict the IDH mutation status of gliomas and differentiate high-grade gliomas...
BACKGROUND: This study aimed to evaluate the diagnostic performance of intravoxel incoherent motion (IVIM) imaging and DSC perfusion MRI to predict the IDH mutation status of gliomas and differentiate high-grade gliomas (HGG) from low-grade gliomas (LGG). METHODS: In this retrospective study, IVIM parameters, including perfusion fraction (f), tissue diffusion (D), and pseudodiffusion (D*), were obtained using a double exponential IVIM model in 61 patients with a pathological diagnosis of glioma. Imaging was performed on a 3 Tesla MRI scanner using only four b-values (0, 50, 400, and 800 s/mm²). The diagnostic performance of rCBV, f, D, D*, ADC, and age parameters in distinguishing HGG from LGG and predicting IDH mutation status in gliomas was evaluated using ROC curve analysis, and their diagnostic powers were compared using the DeLong test. RESULTS: rCBV, f, and age were higher in patients with HGG, whereas D and ADC were higher in patients with LGG (p < 0.001). rCBV showed the highest AUC for HGG-LGG differentiation (AUC = 0.983), without statistically significant superiority over f, D, or ADC by DeLong testing. rCBV, f, and age were higher in patients with IDH-wild type (IDH-WT) gliomas, whereas D and ADC values were higher in IDH-mutant type (IDH-MT) gliomas (p < 0.05). D showed the highest observed AUC for predicting IDH mutation status (AUC = 0.894) without statistically significant superiority over rCBV, f, or ADC by DeLong testing. CONCLUSIONS: IVIM D obtained from a simplified four-b-value protocol may provide complementary information for predicting IDH mutation status in gliomas. However, these preliminary findings require validation in larger, multicenter cohorts.
Noninvasive prediction of Ki67 proliferation index in breast cancer based on integrated photoacoustic imaging.
Mo S, Wang M, Li G … +8 more , Tian H, Wu H, Tang S, Zou X, Shao M, Xu J, Huang Z, Dong F
BMC Med Imaging · 2026 May · PMID 42174497 · Full text
BACKGROUND: Predicting Ki67 expression is crucial for understanding tumor proliferation and guiding personalized breast cancer treatment. Non-invasive methods remain limited, underscoring the need for innovative imaging-...
BACKGROUND: Predicting Ki67 expression is crucial for understanding tumor proliferation and guiding personalized breast cancer treatment. Non-invasive methods remain limited, underscoring the need for innovative imaging-based approaches to enhance molecular subtyping and clinical decisions. PURPOSE: This study aimed to develop a predictive model integrating photoacoustic/ultrasound (PA/US) imaging data and clinical variables to differentiate high and low Ki67 expression levels in breast cancer. It sought to identify imaging and clinical factors associated with Ki67 expression, contributing to the molecular subtyping of breast cancer. METHODS AND MATERIALS: In this study, 336 breast tumors were analyzed and divided into high Ki67 expression (≥14%) and low Ki67 expression (<14%) groups. The samples were randomly split into training and test sets at a 7:3 ratio. Statistical methods included t‑tests and rank‑sum tests, with independent predictors identified through univariate and multivariate logistic regression analyses. Four predictive models were developed: Model A (clinical factors), Model B (clinical factors combined with ultrasound features), Model C (combining clinical, ultrasound, and photoacoustic oxygen saturation [PA‑SO₂]), and Model D (clinical factors combined with PA‑SO₂). RESULTS: Using univariate and multivariate logistic regression analysis, four independent predictive factors were identified: histological grade, axillary lymph node status (ALN), intratumoral color Doppler flow imaging (Inter CDFI), and PA‑SO₂. Based on these factors, four logistic regression models were constructed for predicting high vs. low Ki67 expression: Model A (clinical factors only): histological grade + ALN; Model B (clinical + ultrasound): Model A + Inter CDFI; Model C (comprehensive model): Model B + PA‑SO₂; Model D (comprehensive model): Model A + PA‑SO₂. In the test set, the areas under the receiver operating characteristic curve (AUCs) with 95% confidence intervals were as follows: Model A, 0.781 (0.696-0.866); Model B, 0.779 (0.691-0.867); Model C, 0.823 (0.739-0.908), and Model D, 0.807 (0.721-0.892). Model C demonstrated the highest diagnostic efficiency for distinguishing Ki67 expression levels. CONCLUSION: This study developed a predictive model incorporating histological grade, ALN, Inter CDFI, and PA‑SO₂ to estimate Ki67 expression levels in breast cancer. This model provides a valuable tool for early prognosis, aiding molecular classification and facilitating the prompt initiation of personalized treatment strategies.
Y-Polyp: research on devices for detecting colorectal polyps with limited samples.
Wan J, Wang S, Zhu W … +3 more , Chen B, Li Z, Wang L
BMC Med Imaging · 2026 May · PMID 42174488 · Full text
BACKGROUNDS: Colonoscopy plays a crucial role in preventing the malignant transformation of colorectal polyps, with early diagnosis and detection of colorectal cancer being effective approaches to reducing incidence and...
BACKGROUNDS: Colonoscopy plays a crucial role in preventing the malignant transformation of colorectal polyps, with early diagnosis and detection of colorectal cancer being effective approaches to reducing incidence and mortality rates among patients. With the rise of neural networks, research on computer-aided detection of colorectal polyps has garnered increasing attention. However, existing computer-aided diagnostic systems are constrained by insufficient training sample sizes, making it difficult to train high-performance systems. METHODS: This paper proposes the Y-Polyp model for colorectal polyp detection. By employing a parallel strategy with multidimensional attention mechanisms, this model enables convolutional kernels to learn more flexible attention across four spatial dimensions, thus fully capturing target features from limited data samples. Additionally, the model filters conflicting information in spatial domains, suppressing inconsistent features and addressing inconsistencies among features at different scales. To validate the effectiveness of the Y-Polyp model, extensive experiments are conducted and evaluated using a Kvasir-SEG dataset. RESULTS: The experimental results show that the Y-Polyp model significantly improves the accuracy in the detection of colorectal polyps. Specifically, the overall precision of the polyp detection model has increased by 3%, while the recall rate and average precision have increased by 2.7%, 2% and 2.7%, respectively. CONCLUSIONS: This method exhibits strong generalization ability and robustness, adapting to various colonic structures and changes in target appearance, and the proposed colorectal polyp detection approach thus possesses wide applicability and reliability.
About
- Frequency
- Sun
- Papers found
- 200
- RSS feed
- Subscribe