DeLong Test Research Articles

A comparative study on the diagnostic value of 18F-PSMA PET/CT PRIMARY score and PSMA expression score for clinically significant prostate cancer

Objective: To compare the diagnostic value of fluorine 18-labelled prostate-specific membrane antigen (PSMA) PET/CT PRIMARY score and PSMA expression score for clinically significant prostate cancer (csPCa). Methods: The data of 70 patients with prostate cancer who underwent radical prostatectomy at Beijing Hospital from February 1, 2019 to February 29, 2024 were retrospectively analyzed. All patients underwent whole body 18F-PSMA PET/CT examination before surgery and pathological large sections of prostate specimens were made after surgery. According to the pathological section results, the lesions were divided into csPCa group (Gleason scores≥7) and non-csPCa group (Gleason scores=6 or benign). The PRIMARY score and PSMA expression score were used to evaluate prostatic lesions on PSMA PET/CT images, respectively. The differences in PRIMARY score, PSMA expression score, maximum standardized uptake value (SUVmax), lesion-to-background ratios (LBR), PSMA uptake characteristics, and location distribution were compared between csPCa group and non-csPCa group. Multivariate logistic regression analysis was performed to determine the correlation factors for the incidence of csPCa. By plotting the receiver operator characteristic (ROC) curve and calculating the area under the curve (AUC), the optimal diagnostic threshold for csPCa of each factor was determined. The differences in AUC were compared using the Delong test. Quantitative data was represented as M (Q1, Q3). Results: The 70 patients aged 70 (64, 75) years with a total of 108 lesions, including 83 lesions (76.9%) in the csPCa group and 25 lesions (23.1%) in the non-csPCa group. There were differences between csPCa group and non-csPCa group in PRIMARY score [4 (3, 5) vs 2 (1, 4)], PSMA expression score [2 (2, 2) vs 1 (1, 2)], SUVmax [9.10 (5.70, 15.80) vs 5.40 (3.35, 6.90)], LBR [2.86 (2.09, 4.53) vs 1.96 (1.42, 2.58)], and proportion of focal uptake patterns [74.7% (62/83) vs 32.0% (8/25)] (all P<0.05). There was no statistically significant difference in the location distribution between the two groups (P>0.05). Multivariate logistic regression analysis indicated the higher the PRIMARY score, the higher the risk of developing csPCa (OR=1.863, 95%CI: 1.360-2.552). ROC curves revealed that the AUCs of the PRIMARY score and PSMA expression score for csPCa were 0.751 (95%CI: 0.659-0.829) and 0.697 (95%CI: 0.601-0.781), respectively, without statistically difference (Z=1.438, P>0.05). The cut-off values for diagnosing csPCa were 3 score for PRIMARY score and 2 score for PSMA expression score, respectively. Conclusions: The higher the 18F-PSMA PET/CT PRIMARY score, the higher the risk of developing csPCa. The PRIMARY score has good diagnostic efficacy for csPCa.

Identifying Primary Sites of Spinal Metastases: Expert-Derived Features vs. ResNet50 Model Using Nonenhanced MRI.

The spinal column is a frequent site for metastases, affecting over 30% of solid tumor patients. Identifying the primary tumor is essential for guiding clinical decisions but often requires resource-intensive diagnostics. To develop and validate artificial intelligence (AI) models using noncontrast MRI to identify primary sites of spinal metastases, aiming to enhance diagnostic efficiency. Retrospective. A total of 514 patients with pathologically confirmed spinal metastases (mean age, 59.3 ± 11.2 years; 294 males) were included, split into a development set (360) and a test set (154). Noncontrast sagittal MRI sequences (T1-weighted, T2-weighted, and fat-suppressed T2) were acquired using 1.5 T and 3 T scanners. Two models were evaluated for identifying primary sites of spinal metastases: the expert-derived features (EDF) model using radiologist-identified imaging features and a ResNet50-based deep learning (DL) model trained on noncontrast MRI. Performance was assessed using accuracy, precision, recall, F1 score, and the area under the receiver operating characteristic curve (ROC-AUC) for top-1, top-2, and top-3 indicators. Statistical analyses included Shapiro-Wilk, t tests, Mann-Whitney U test, and chi-squared tests. ROC-AUCs were compared via DeLong tests, with 95% confidence intervals from 1000 bootstrap replications and significance at P < 0.05. The EDF model outperformed the DL model in top-3 accuracy (0.88 vs. 0.69) and AUC (0.80 vs. 0.71). Subgroup analysis showed superior EDF performance for common sites like lung and kidney (e.g., kidney F1: 0.94 vs. 0.76), while the DL model had higher recall for rare sites like thyroid (0.80 vs. 0.20). SHapley Additive exPlanations (SHAP) analysis identified sex (SHAP: -0.57 to 0.68), age (-0.48 to 0.98), T1WI signal intensity (-0.29 to 0.72), and pathological fractures (-0.76 to 0.25) as key features. AI techniques using noncontrast MRI improve diagnostic efficiency for spinal metastases. The EDF model outperformed the DL model, showing greater clinical potential. Spinal metastases, or cancer spreading to the spine, are common in patients with advanced cancer, often requiring extensive tests to determine the original tumor site. Our study explored whether artificial intelligence could make this process faster and more accurate using noncontrast MRI scans. We tested two methods: one based on radiologists' expertise in identifying imaging features and another using a deep learning model trained to analyze MRI images. The expert-based method was more reliable, correctly identifying the tumor site in 88% of cases when considering the top three likely diagnoses. This approach may help doctors reduce diagnostic time and improve patient care. 3 TECHNICAL EFFICACY: Stage 2.

Transition zone-based prostate-specific antigen density for differentiating clinically significant prostate cancer in PI-RADS score 3 lesions

We intended to investigate the potential of several transitional zone (TZ) volume-related variables for the detection of clinically significant prostate cancer (csPCa) among lesions scored as Prostate Imaging Reporting and Data System (PI-RADS) category 3. Between September 2018 and August 2023, patients who underwent mpMRI examination and scored as PI-RADS 3 were queried from our institution. The diagnostic performances of prostate-specific antigen density (PSAD), TZ-adjusted PSAD (TZPSAD), and TZ-ratio (TZ volume/whole gland prostate volume) were analyzed. We calculated the sensitivity and specificity for each variable, the overall accuracy was evaluated with the area under the receiver operating characteristic curve (AUC). The best cutoff value was determined by the Youden index, and differences between diagnostic performances were compared with the Delong test. A total of 154 TZ lesions were included, of them 28 were diagnosed with csPCa. The AUC for PSAD, TZPSAD, and TZ-ratio were 0.644 (95% CI 0.538–0.751), 0.714 (95% CI 0.607–0.821), and 0.746 (95% CI 0.636–0.856), with corresponding optimal cutoff values of 0.11 ng/ml/ml, 0.21 ng/ml/ml, and 0.66 ng/ml/ml, respectively. PSAD was significantly inferior to either TZPSAD (P = 0.03) or TZ-ratio (P = 0.03). The combination model including TZPSAD, TZ-ratio, and age generated an AUC of 0.838 (95% CI 0.759–0.917), significantly higher than use of these variables alone, with P values of 0.001 and 0.035 for the TZPSAD and TZ-ratio, respectively. TZPSAD and TZ-ratio were found to be independent predictors for differentiating csPCa among TZ lesions categorized as PI-RADS score 3. Furthermore, by combining these two variables with others, the diagnostic performance can be improved significantly.

Superior performance in classification of breast cancer molecular subtype and histological factors by radiomics based on ultrafast MRI over standard MRI: evidence from a prospective study.

To compare the performance of ultrafast MRI with standard MRI in classifying histological factors and subtypes of invasive breast cancer among radiologists with varying experience. From October 2021 to November 2022, this prospective study enrolled 225 participants with 233 breast cancers before treatment (NCT06104189 at clinicaltrials.gov). Tumor segmentation on MRI was performed independently by two readers (R1, dedicated breast radiologist; R2, radiology resident). We extracted 1618 radiomic features and four kinetic features from ultrafast and standard images, respectively. Logistic regression algorithms were adopted for prediction modeling, following feature selection by the least absolute shrinkage and selection operator. The performance of predicting histological factors and subtypes was evaluated using the area under the receiver-operating characteristic curve (AUC). Performance differences between MRI methods and radiologists were assessed using the DeLong test. Ultrafast MRI outperformed standard MRI in predicting HER2 status (AUCs [95% CI] of ultrafast MRI vs standard MRI; 0.87 [0.83-0.91] vs 0.77 [0.64-0.90] for R1 and 0.88 [0.83-0.91] vs 0.77 [0.69-0.84] for R2) (all P < 0.05). Both ultrafast MRI and standard MRI showed comparable performance in predicting hormone receptors. Ultrafast MRI exhibited superior performance to standard MRI in classifying subtypes. The classification of the luminal subtype for both readers, the HER2-overexpressed subtype for R2, and the triple-negative subtype for R1 was significantly better with ultrafast MRI (P < 0.05). Ultrafast MRI-based radiomics holds promise as a noninvasive imaging biomarker for classifying hormone receptors, HER2 status, and molecular subtypes compared to standard MRI, regardless of radiologist experience.

Deep learning–assisted diagnosis of acute mesenteric ischemia based on CT angiography images

PurposeAcute Mesenteric Ischemia (AMI) is a critical condition marked by restricted blood flow to the intestine, which can lead to tissue necrosis and fatal outcomes. We aimed to develop a deep learning (DL) model based on CT angiography (CTA) imaging and clinical data to diagnose AMI.MethodsA retrospective study was conducted on 228 patients suspected of AMI, divided into training and test sets. Clinical data (medical history and laboratory indicators) was included in a multivariate logistic regression analysis to identify the independent factors associated with AMI and establish a clinical factors model. The arterial and venous CTA images were utilized to construct DL model. A Fusion Model was constructed by integrating clinical factors into the DL model. The performance of the models was assessed using receiver operating characteristic (ROC) curves and decision curve analysis (DCA).ResultsAlbumin and International Normalized Ratio (INR) were associated with AMI by univariate and multivariate logistic regression (P &amp;lt; 0.05). In the test set, the area under ROC curve (AUC) of the clinical factor model was 0.60 (sensitivity 0.47, specificity 0.86). The AUC of the DL model based on CTA images reached 0.90, which was significantly higher than the AUC values of the clinical factor model, as confirmed by the DeLong test (P &amp;lt; 0.05). The Fusion Model also showed exceptional performance in terms of AUC, accuracy, sensitivity, specificity, and precision, with values of 0.96, 0.94, 0.94, 0.95, and 0.98, respectively. DCA indicated that the Fusion Model provided a greater net benefit than those of models based solely on imaging and clinical information across the majority of the reasonable threshold probabilities.ConclusionThe incorporation of CTA images and clinical information into the model markedly enhances the diagnostic accuracy and efficiency of AMI. This approach provides a reliable tool for the early diagnosis of AMI and the subsequent implementation of appropriate clinical intervention.

The development and evaluation of nine obesity-based indices for gallstones in U.S. adults.

Gallstones have gradually become a highly prevalent digestive disease worldwide. This study aimed to investigate the association of nine different obesity-related indicators (BRI, RFM, BMI, WC, LAP, CMI, VAI, AIP, TyG) with gallstones and to compare their predictive properties for screening gallstones. Data for this study were obtained from the National Health and Nutrition Examination Survey (NHANES) for the 2017-2020 cycle, and weighted logistic regression analyses with multi-model adjustment were conducted to explore the association of the nine indicators with gallstones. Subject working curves were analyzed to assess the screening ability of the nine indicators. In addition, the association between the most predictive indicator and gallstones was investigated with smooth curve fitting, and differences in risk across populations were explored with subgroup analyses. In total, 3654 participants were involved in the final analysis and 383 (10.48%) carried gallstones. The results of weighted multifactorial logistic regression analysis indicated that BRI, RFM, BMI, WC, LAP, and CMI were independent risk factors for gallstones. The ORs for the highest quartile were 4.13 for RFM, 3.13 for BRI, 2.85 for BMI, 2.86 for WC, 2.45 for LAP, and 1.49 for CMI. The area under the ROC curve for RFM was 0.70. The Delong test compared the performance of different ROCs and revealed that the difference between the area under the curve of RFM and the other metrics was significant (P < 0.05). Smooth curve fitting suggested a linear positive correlation between RFM and gallstones (LLR > 0.05), especially in women, non-Hispanic White, insufficient physical activity, hypertensive, and diabetic populations. RFM, BRI, BMI, WC, LAP, and CMI were essential indicators for recognizing gallstones. By comparison, we realized that RFM was a better predictor of gallstones.

Clinical significance of PCT, CRP, IL-6, NLR, and TyG Index in early diagnosis and severity assessment of acute pancreatitis: A retrospective analysis

To evaluate the clinical utility of PCT, CRP, IL-6, NLR, and TyG index in improving the early diagnosis and severity assessment of acute pancreatitis (AP). This retrospective study included 137 AP patients and 30 healthy controls from Hunan Provincial People’s Hospital (January 2021–September 2023). Univariate and multivariate logistic regression analyses assessed the associations between biomarkers and severe acute pancreatitis (SAP). Receiver operating characteristic (ROC) curves, DeLong test, and Bonferroni correction were used to evaluate predictive performance. Model robustness was validated via 5-fold cross-validation. PCT, CRP, IL-6, NLR, and TyG index levels were significantly elevated in AP patients compared to controls (P < 0.001) and correlated with disease severity (P < 0.05). CRP and NLR levels differed significantly among mild, moderate, and severe AP (P < 0.01). Alcohol consumption and hyperlipidemia were significantly linked to AP severity (P for trend < 0.0001). Multivariate analysis identified hyperlipidemia (OR = 3.030, P = 0.040), CRP (OR = 1.011, P < 0.001), and NLR (OR = 1.078, P = 0.020) as independent SAP predictors. The combined model of CRP + NLR + TyG achieved the highest AUC (0.882, sensitivity = 77.2%, specificity = 88.5%), though it was not significantly better than CRP + NLR or CRP + TyG models (P > 0.05). 5-fold cross-validation confirmed consistent performance (mean AUC = 0.817 ± 0.118). PCT, CRP, IL-6, NLR, and TyG index are valuable in diagnosing and assessing AP prognosis. Hyperlipidemia, CRP, and NLR are reliable independent predictors of SAP. Combining multiple biomarkers enhances diagnostic precision and provides guidance for personalized treatment strategies in AP.

Baseline [18F]FDG PET/CT radiomics for predicting interim efficacy in follicular lymphoma treated with first-line R-CHOP

ObjectiveTo investigate the predictive value of machine learning-based PET/CT radiomics and clinical risk factors in predicting interim efficacy in patients with follicular lymphoma (FL).MethodsThis study retrospectively analyzed data from 97 patients with FL diagnosed via histopathological examination between July 2012 and November 2023. Lesion segmentation was performed using LIFEx software, and radiomics features were extracted through the uAI Research Portal (uRP) platform, including first-order features, shape features, and texture features. Fourteen filters were applied to the raw images to extract higher-order features from the derived images. Univariate analysis was employed to identify clinical risk factors, and correlation coefficients, MRMR, and LASSO algorithms were used for dimensionality reduction and selection of radiomics features. Finally, a logistic regression machine learning model was developed to predict the interim efficacy of FL using a five-fold cross-validation strategy. Model performance was assessed using the area under the receiver operating characteristic (ROC) curve, accuracy, and the Delong test to compare AUC differences.ResultAmong the 97 patients, 42 (43.30%) achieved complete response (CR) for interim efficacy, while 55 (56.70%) had non-complete response (non-CR). A total of 2264 radiomics features were extracted from the images. Seven clinical risk factors and ten radiomics features associated with interim efficacy were selected to construct the clinical, radiomics, and radiomics-clinical combined models. Among the three logistic regression machine learning models developed, the radiomics-clinical combined model demonstrated the best performance, achieving a mean AUC of 0.849 (95% CI, 0.676–1.000) and an accuracy of 0.795, outperforming the other two models.ConclusionOur preliminary results demonstrate that a radiomics-clinical combined model, based on baseline [18F]FDG PET/CT radiomics features and clinical risk factors, may contribute to predicting interim efficacy in FL patients.

MRI Radiomics Analysis in the Diagnostic Differentiation of Malignant Soft Tissue Myxoid Sarcomas From Benign Soft Tissue Musculoskeletal Myxomas.

Differentiation of benign myxomas and malignant myxoid sarcomas can be difficult with an overlapping spectrum of morphologic MR findings. To assess the diagnostic utility of MRI radiomics in the differentiation of musculoskeletal myxomas and myxoid sarcomas. Retrospective. A total of 523 patients were included; histologically proven myxomas (N = 201) and myxoid sarcomas (N = 322), randomly divided (70:30) into training:test subsets. T1-weighted (T1W), T2-weighted fat-suppressed (fluid-sensitive), and T1-weighted post-contrast (T1W + C) sequences at 1.0 T, 1.5 T, or 3.0 T. Seven semantic (qualitative) tumor features were assessed in each case. Manual 3D tumor segmentations performed with radiomics features extracted from T1W, fluid-sensitive, and T1W + C acquisitions. Models were constructed based on radiomic features from individual sequences and from their combination, both with and without the addition of qualitative tumor features. Intraclass correlation evaluated in 60 cases segmented by three readers. Features with intraclass correlation <0.7 excluded from further analysis. Boruta feature selection and Random Forest modeling performed using the training-dataset, with resultant models used to assess class discrimination (myxoma vs. myxoid sarcoma) in the test dataset. Radiomics score defined as probability class = myxoma. Logistic regression modeling employed to estimate performance of the radiomics score. Area under the receiver operating characteristic curve (AUC) was used to assess diagnostic performance, and DeLong's test to assess performance between constructed models. A P-value <0.05 was considered significant. Four qualitative semantic features showed significant predictive power in class discrimination. Radiomic models demonstrated excellent differentiation of myxomas from myxoid sarcomas: AUC of 0.9271 (T1W), 0.9049 (fluid-sensitive), and 0.9179 (T1W + C). Incorporation of multiparametric data or semantic features did not significantly improve model performance (P ≥ 0.08) compared to radiomic models derived from any individual MRI sequence alone. MRI radiomics appears to be accurate in the differentiation of myxomas from myxoid sarcomas. Classification performance did not improve when incorporating qualitative features or multiparametric imaging data. Accurately distinguishing between benign soft tissue myxomas and malignant myxoid sarcomas is essential for guiding appropriate management but remains challenging with conventional MRI interpretation. This study utilized radiomics, a method that extracts quantitative mathematically derived features from images, to develop predictive models based on routine MRI examination. Analyzing over 500 cases, MRI radiomics demonstrated excellent diagnostic accuracy in differentiating between benign myxomas and malignant myxoid sarcomas, highlighting the potential of the technique, as a powerful non-invasive tool that could complement current diagnostic approaches, and enhance clinical decision-making in patients with soft tissue myxoid tumors of the musculoskeletal system. 3 TECHNICAL EFFICACY: Stage 2.

Location-specific net water uptake and malignant cerebral edema in acute anterior circulation occlusion ischemic stroke.

Early identification of malignant cerebral edema (MCE) in patients with acute ischemic stroke is crucial for timely interventions. We aimed to identify regions critically associated with MCE using the Alberta Stroke Program Early Computed Tomography Score (ASPECTS) to evaluate the association between location-specific-net water uptake (NWU) and MCE. This multicentre, retrospective cohort study included patients with acute ischemic stroke following large anterior circulation occlusion. The ASPECTS was determined by RAPID ASPECTS software. ASPECTS-NWU and Region-NWU were calculated automatically by comparing the Hounsfield units values in the ischemic and contralateral regions. Critical ASPECTS MCE regions and Region-NWU were evaluated by multivariate logistic regression and the areas under the receiver operating characteristic curves (AUCs). The study included 513 patients. Multivariate analysis showed that the ASPECTS insula (OR=2.49; 95% CI, 1.44-4.31) and M5 (OR=1.59; 95% CI, 1.11-3.41) regions were significantly associated with MCE. After adjustment, only the insula (OR=2.34; 95% CI, 1.23-4.45) was independently associated with MCE. Univariable ROC analysis found AUCs for Insula-NWU (AUC, 0.70; 95% CI, 0.65- 0.76)and ASPECTS-NWU (AUC, 0.64; 95% CI, 0.58-0.70) .The Insula-NWU had better diagnostic power than ASPECTS-NWU (DeLong test; P=0.01). A multivariate regression model that combined the NIHSS, ASPECTS, insula involvement, and Insula-NWU had good discriminatory power (AUC=0.80; 95% CI, 0.74-0.86) and better diagnostic power than Insula-NWU (DeLong test; P<0.01). Brief statement directed to the stated purpose or hypothesis; no references should be cited.The insula region is critical for MCE, and Insula-NWU has better prediction efficacy than ASPECTS-NWU. This method does not rely on advanced imaging, facilitating rapid assessment in emergencies. ASPECTS = the Alberta Stroke Program Early Computed Tomography Score; AUC= the areas under the receiver operating characteristic curve; CT=computed tomography; CTP=CT perfusion; HU = hounsfield unit; MCE = malignant cerebral edema; NCCT=non-contrast Computed Tomography; NWU = net water uptake; ROC = receiver operating characteristic curve.

Prediction of pre-eclampsia using maternal hemodynamic parameters at 12 + 0 to 15 + 6 weeks.

To compare the maternal hemodynamic profile at 12 + 0 to 15 + 6 weeks' gestation in women who subsequently developed pre-eclampsia (PE) and those who did not, and to assess the screening performance of maternal hemodynamic parameters for PE in combination with the Fetal Medicine Foundation (FMF) triple test, including maternal factors (MF), mean arterial pressure (MAP), uterine artery pulsatility index and placental growth factor. This was a prospective case-control study involving Chinese women with a singleton pregnancy who underwent preterm PE screening at 11 + 0 to 13 + 6 weeks' gestation using the FMF triple test, between February 2020 and February 2023. Women identified as being at high risk (≥ 1:100) for preterm PE by the FMF triple test were matched 1:1 with women identified as low risk (< 1:100) for maternal age ± 3 years, maternal weight ± 5 kg and date of screening ± 14 days. Two-dimensional transthoracic echocardiography was performed at 12 + 0 to 15 + 6 weeks to evaluate maternal hemodynamic parameters (heart rate (HR), stroke volume (SV), cardiac output (CO) and systemic vascular resistance (SVR)). Maternal hemodynamic parameters were expressed as multiples of the median (MoM) values, determined by linear regression models to adjust for gestational age (GA) and MF. The distribution of log10 MoM values of maternal hemodynamic parameters in cases of PE and unaffected pregnancies, and the association between these hemodynamic parameters and GA at delivery, were assessed. The risks of preterm PE (delivery before 37 weeks) and any-onset PE (delivery at any time) were reassessed using Bayes' theorem after maternal hemodynamic parameters were added to the FMF triple test. The screening performance for preterm PE and any-onset PE was determined by the area under the receiver-operating-characteristics curve (AUC) and detection rate at a 10% fixed false-positive rate (FPR). Differences in AUC (ΔAUC) were assessed using DeLong's test. A total of 743 cases were analyzed, of whom 39 (5.2%) subsequently developed PE, including 29 (3.9%) cases of preterm PE and 10 (1.3%) cases of term PE. Mean log10 SVR MoM was significantly higher in cases of preterm PE and any-onset PE compared with unaffected pregnancies. Mean log10 SV MoM and log10 CO MoM were significantly lower in cases of preterm PE and any-onset PE compared with unaffected pregnancies. Mean log10 HR MoM was not significantly different between the study groups. Mean log10 CO MoM and log10 SVR MoM were not significantly correlated with GA at delivery in preterm PE and any-onset PE. For the prediction of preterm PE and any-onset PE, adding CO or SVR or replacing MAP with CO and SVR in the FMF triple test achieved an identical or greater AUC compared with the FMF triple test, but ΔAUC was not significantly different. In addition, adding CO or SVR or replacing MAP by CO and SVR in the FMF triple test did not improve the detection rate for preterm PE and any-onset PE at a fixed FPR of 10%. Women with preterm PE or any-onset PE exhibited increased SVR and decreased CO before the clinical manifestations of PE became apparent. These changes may serve as early indicators of cardiovascular maladaptation. However, assessment of maternal hemodynamics at 12 + 0 to 15 + 6 weeks does not enhance the screening performance for preterm PE and any-onset PE of these parameters. The FMF triple test remains superior to other biomarker combinations for predicting PE. © 2025 The Authors. Ultrasound in Obstetrics & Gynecology published by John Wiley & Sons Ltd on behalf of International Society of Ultrasound in Obstetrics and Gynecology.

End-to-End CT Radiomics-Based Pipeline for Predicting Renal Interstitial Fibrosis Grade in CKD Patients.

Non-invasive assessment of renal fibrosis in patients with chronic kidney disease (CKD) remains a clinical challenge. This study aims to integrate radiomics and clinical factors to develop an end-to-end pipeline for predicting interstitial fibrosis (IF) in CKD patients. This retrospective study included 80 patients with CKD, with 53 patients in training set and 27 patients in test set. All patients underwent renal computed tomography (CT) scans and biopsy. Patients were classified into two groups based on their renal IF grade: mild-moderate and severe. Radiomics features were extracted from the automatically segmented right renal region on CT images, and univariate analysis along with multiple Least Absolute Shrinkage and Selection Operator (LASSO) was employed to construct the radiomics signature. Subsequently, logistic regression models were developed to create the radiomics model and the combined model. The predictive performance of both models was evaluated through discrimination, calibration, and decision curve analysis (DCA), and a nomogram was constructed for the model demonstrating superior performance. The combined model significantly outperformed the radiomics model, achieving a cross-validated AUC of 0.935±0.041 in the training set, compared to 0.804±0.024 for the radiomics model. In the test set, the combined model outperformed the radiomics model, with an AUC of 0.918 [95% CI 0.799-1] vs. 0.764 [95% CI 0.549-0.979], p=0.031 (DeLong test, Statistic: -2.152). Calibration curves and DCA indicated that the combined model demonstrated good calibration and better clinical net benefit. This end-to-end workflow could serve as a potential non-invasive tool to predict renal IF grade (mild-moderate vs. severe) in CKD patients.

Predictive value of a constructed artificial neural network model for microvascular invasion in hepatocellular carcinoma: A retrospective study.

Microvascular invasion (MVI) is a significant risk factor for recurrence and metastasis following hepatocellular carcinoma (HCC) surgery. Currently, there is a paucity of preoperative evaluation approaches for MVI. To investigate the predictive value of texture features and radiological signs based on multiparametric magnetic resonance imaging in the non-invasive preoperative prediction of MVI in HCC. Clinical data from 97 HCC patients were retrospectively collected from January 2019 to July 2022 at our hospital. Patients were classified into two groups: MVI-positive (n = 57) and MVI-negative (n = 40), based on postoperative pathological results. The correlation between relevant radiological signs and MVI status was analyzed. MaZda4.6 software and the mutual information method were employed to identify the top 10 dominant texture features, which were combined with radiological signs to construct artificial neural network (ANN) models for MVI prediction. The predictive performance of the ANN models was evaluated using area under the curve, sensitivity, and specificity. ANN models with relatively high predictive performance were screened using the DeLong test, and the regression model of multilayer feedforward ANN with backpropagation and error backpropagation learning method was used to evaluate the models' stability. The absence of a pseudocapsule, an incomplete pseudocapsule, and the presence of tumor blood vessels were identified as independent predictors of HCC MVI. The ANN model constructed using the dominant features of the combined group (pseudocapsule status + tumor blood vessels + arterial phase + venous phase) demonstrated the best predictive performance for MVI status and was found to be automated, highly operable, and very stable. The ANN model constructed using the dominant features of the combined group can be recommended as a non-invasive method for preoperative prediction of HCC MVI status.

"The M-APNE score: an objective screening tool for OSA highlighting the area under the inspiratory flow-volume curve".

Polysomnography (PSG) is resource-intensive but remains the gold standard for diagnosing Obstructive Sleep Apnea (OSA). We aimed to develop a screening tool to better allocate resources by identifying individuals at higher risk for OSA, overcoming limitations of current tools that may under-diagnose based on self-reported symptoms. A total of 884 patients (490 diagnosed with OSA) were included, which was divided into the training, validation, and test sets. Using multivariate logistic regression analyses, we developed a scoring system incorporating male sex, age, sawtooth pattern, area under the inspiratory flow-volume curve (AreaFI), and neck circumference to objectively identify patients at higher risk of OSA. Sensitivity and specificity were evaluated using area under the curve (AUC) metrics. The M-APNE Score was compared to other non-symptom-based tools, the No-Apnea Score and the Symptomless Multivariable Apnea Prediction (sMVAP) model, using the Delong test. The M-APNE Score showed sensitivity rates of 79.3% in the training set, 70.8% in the test, and 80% in the validation set. ROC analysis for M-APNE score yielded AUCs of 0.82 in the training, 0.76 in the test, 0.82 in the validation set. The discriminative accuracy of M-APNE Score were found to be better than the No-Apnea Score (AUC = 0.82 vs. 0.76, p < 0.001) and the sMVAP (AUC = 0.82 vs. 0.75, p = 0.001) in the training set. Hosmer Lemeshow test indicated good calibration for M-Apne Score (p = 0.46). The M-APNE Score is a robust and objective tool for OSA screening, potentially reducing classification errors and improving accuracy.

Advancing presurgical non-invasive spread through air spaces prediction in clinical stage IA lung adenocarcinoma using artificial intelligence and CT signatures.

To accurately identify spread through air spaces (STAS) in clinical stage IA lung adenocarcinoma, our study developed a non-invasive and interpretable biomarker combining clinical and radiomics features using preoperative CT. The study included a cohort of 1,325 lung adenocarcinoma patients from three centers, which was divided into four groups: a training cohort (n = 930), a testing cohort (n = 238), an external validation 1 cohort (n = 93), and 2 cohort (n = 64). We collected clinical characteristics and semantic features, and extracted radiomics features. We utilized the LightGBM algorithm to construct prediction models using the selected features. Quantifying the contribution of radiomics features of CT to prediction model using Shapley additive explanations (SHAP) method. The models' performance was evaluated using metrics such as the area under the receiver operating characteristic curve (AUC), negative predictive value (NPV), positive predictive value (PPV), sensitivity, specificity, calibration curve, and decision curve analysis (DCA). In the training cohort, the clinical model achieved an AUC value of 0.775, the radiomics model achieved an AUC value of 0.836, and the combined model achieved an AUC value of 0.837. In the testing cohort, the AUC values of the models were 0.743, 0.755, and 0.768. In the external validation 1 cohort, the AUC values of the models were 0.717, 0.758, and 0.765, while in the external validation 2 cohort, 0.725, 0.726 and 0.746. The DeLong test results indicated that the combined model outperformed the clinical model (p < 0.05). DCA indicated that the models provided a net benefit in predicting STAS. The SHAP algorithm explains the contribution of each feature in the model, visually demonstrating the impact of each feature on the model's decisions. The combined model has the potential to serve as a biomarker for predicting STAS using preoperative CT scans, determining the appropriate surgical strategy, and guiding the extent of resection.

Machine-Learning Based Computed Tomography Radiomics Nomgram For Predicting Perineural Invasion In Gastric Cancer.

The aim of this study was to develop and validate predictive models for perineural invasion (PNI) in gastric cancer (GC) using clinical factors and radiomics features derived from contrast-enhanced computed tomography (CE-CT) scans and to compare the performance of these models. This study included 205 GC patients, who were randomly divided into a training set (n=143) and a validation set (n=62) in a 7:3 ratio. Optimal radiomics features were selected using the least absolute shrinkage and selection operator (LASSO) algorithm. A radiomics model was constructed utilizing the optimal among five machine learning filters, and a radiomics score (rad-score) was computed for each participant. A clinical model was built based on clinical factors identified through multivariate logistic regression. Independent clinical factors were combined with the radscore to create a combined radiomics nomogram. The discrimination ability of the models was evaluated by receiver operating characteristic (ROC) curves and the DeLong test. Independent predictive factors of the clinical model included tumor T stage, N stage, and tumor differentiation, with AUC values of 0.777 and 0.809 in the training and validation sets. The radiomics model was constructed using the support vector machine (SVM) classifier with the best AUC (0.875 in the training set and 0.826 in the validation set). The combined radiomics nomogram, which combines independent clinical predictors and the rad-score, demonstrated better predictive performance (AUC=0.889 in the training set; AUC=0.885 in the validation set). The nomogram integrating independent clinical predictors and CE-CT radiomics was constructed to predict PNI in GC. This model demonstrated favorable performance and could potentially assist in prognosis evaluation and clinical decision-making for GC patients.

Prognostic value of multi-PLD ASL radiomics in acute ischemic stroke.

Early prognosis prediction of acute ischemic stroke (AIS) can support clinicians in choosing personalized treatment plans. The aim of this study is to develop a machine learning (ML) model that uses multiple post-labeling delay times (multi-PLD) arterial spin labeling (ASL) radiomics features to achieve early and precise prediction of AIS prognosis. This study enrolled 102 AIS patients admitted between December 2020 and September 2024. Clinical data, such as age and baseline National Institutes of Health Stroke Scale (NIHSS) score, were collected. Radiomics features were extracted from cerebral blood flow (CBF) images acquired through multi-PLD ASL. Features were selected using least absolute shrinkage and selection operator regression, and three models were developed: a clinical model, a CBF radiomics model, and a combined model, employing eight ML algorithms. Model performance was assessed using receiver operating characteristic curves and decision curve analysis (DCA). Shapley Additive exPlanations was applied to interpret feature contributions. The combined model of extreme gradient boosting demonstrated superior predictive performance, achieving an area under the curve (AUC) of 0.876. Statistical analysis using the DeLong test revealed its significant outperformance compared to both the clinical model (AUC = 0.658, p < 0.001) and the CBF radiomics model (AUC = 0.755, p = 0.002). The robustness of all models was confirmed through permutation testing. Furthermore, DCA underscored the clinical utility of the combined model. The prognostic prediction of AIS was notably influenced by the baseline NIHSS score, age, as well as texture and shape features of CBF. The integration of clinical data and multi-PLD ASL radiomics features in a model offers a secure and dependable approach for predicting the prognosis of AIS, particularly beneficial for patients with contraindications to contrast agents. This model aids clinicians in devising individualized treatment plans, ultimately enhancing patient prognosis.

Liver Elastography for Liver Fibrosis Stratification: A Comparison of Three Techniques in a Biopsy-Controlled MASLD Cohort.

Background: The aim of this study was to investigate the accuracy in fibrosis staging of a novel shear wave elastography (SWE) device (S-Shearwave Imaging by Samsung) and a previously validated 2D-SWE by Supersonic Imagine (SSI) in patients with biopsy proven metabolic dysfunction-associated steatotic liver disease (MASLD). Methods: This prospective study included 75 consecutive patients with MASLD who underwent liver biopsy for suspected MASH. All patients underwent S-Shearwave Imaging by Samsung and 2D-SWE with SSI on the same day of liver biopsy. Fibrosis was histologically assessed using the METAVIR classification system. Agreement between the equipment was assessed with the Pearson coefficient. A receiver operator characteristic curve (ROC) analysis with the Youden index was used to establish thresholds for fibrosis staging. Results: A good correlation was found between S-Shearwave Imaging by Samsung and 2D-SWE with SSI (Pearson's R = 0.68; p < 0.01). At multivariate regression analysis, S-Shearwave Imaging was associated with advanced fibrosis (≥F3) independently from age, diabetes and platelets (OR 2.94, CI 1.69-5.11, p < 0.01). The fibrosis diagnostic accuracy of both S-Shearwave Imaging and 2D-SWE was good to optimal with AUROCs of 0.81 and 0.70 for significant fibrosis (≥F2), 0.94 and 0.91 for severe fibrosis (≥F3), respectively. The accuracy of S-Shearwave is not significantly different from Fibroscan and Agile3+ (DeLong test p value 0.16 and 0.15, respectively) while is slightly better than 2D-SWE, FIB4 and NFS (DeLong test p value < 0.05). For S-Shearwave Imaging by Samsung, the best cut-off values for diagnosing fibrosis ≥F2, ≥F3 were, respectively, 7.9 kPa (Sens 74.4%, Spec 87.5%) and 8.1 kPa (Sens 95.6%, Spec 78.8%). For 2D-SWE by SSI, the best cut-off values for diagnosing fibrosis ≥F2, ≥F3 were, respectively, 7.2 kPa (Sens 55.8%, Spec 84.4%) and 7.6 kPa (Sens 82.6%, Spec 84.6%). Conclusion: S-Shearwave Imaging is a useful and reliable non-invasive technique for staging liver fibrosis in patients with MASLD. Its diagnostic accuracy is non-inferior to other shear wave elastography techniques (TE and 2D-SWE by SSI).

Clinical decision instruments for predicting mortality in patients with cirrhosis seeking emergency department care.

Clinical decision instruments (CDIs) could be useful to aid risk stratification and disposition of emergency department (ED) patients with cirrhosis. Our primary objective was to derive and internally validate a novel Cirrhosis Risk Instrument for Stratifying Post-Emergency department mortality (CRISPE) for the outcomes of 14- and 30-day post-ED mortality. Secondarily, we externally validated the existing Model for End-Stage Liver Disease (MELD) scores for explicit use in ED patients and prediction of the same outcomes. A cohort of 2093 adults with cirrhosis, at 16 sites in a statewide health system, was analyzed for 119 candidate variables available at ED disposition. LASSO with 10-fold cross-validation was used in variable selection for 14-day (CRISPE-14) and 30-day (CRISPE-30) logistic regression models. Area under the receiver operating characteristic curve (AUROC) was calculated for each variant of the CRISPE and MELD scores and compared via Delong's test. Predictions were compared to actual ED disposition for predictive value and reclassification statistics. Median (interquartile range [IQR]) characteristics of the cohort were age 62 (53-70) years and MELD 3.0 13.0 (8.0-20.0). Mortality was 4.3% and 8.5% at 14 and 30 days, respectively. CRISPE-14 and CRISPE-30 outperformed each MELD variant, achieving AUROC of 0.824 (95% CI: 0.781-0.866) and 0.829 (0.796-0.861), respectively. MELD 3.0 AUROCs were 0.724 (0.667-0.781) and 0.715 (0.672-0.781), respectively. Compared to ED disposition, CRISPE-14, CRISPE-30, and MELD 3.0 significantly improved positive and negative predictive value and net reclassification index at multiple cutoffs. Applying CRISPE-30 (cutoff 4.5) favorably reclassified one net ED disposition for mortality for every 12 patients, while MELD 3.0 net reclassified one disposition per 84 patients. CDIs may be useful in risk-stratifying ED patients with cirrhosis and aiding disposition decision making. The novel CRISPE CDI showed powerful performance and requires external validation, while the existing MELD 3.0 score has moderate performance and is now externally-validated in an ED population for short-term mortality.

Ultrafast dynamic contrast-enhanced breast MRI with quantitative perfusion parameters in differentiating breast cancer: a study focusing on triple-negative and HER2 positive breast cancer.

In the realm of breast cancer diagnosis and treatment, accurately discerning molecular subtypes is of paramount importance, especially when aiming to avoid invasive tests. The updated guidelines for diagnosing and treating HER2 positive advanced breast cancer, as presented at the 2021 National Breast Cancer Conference and the Annual Meeting of the Chinese Society of Clinical Oncology, highlight the significance of this approach. A new generation of drug-antibody combinations has emerged, expanding the array of treatment options for HER2 positive advanced breast cancer and significantly improving patient survival rates. Triple-negative breast cancer (TNBC), on the other hand, may indicate survival outcomes following multi-agent adjuvant chemotherapy. DISCO is a more recent DCE MRI technique that has achieved high spatial and temporal resolution and minimized image artifacts in cases like malignant focal liver lesions, enhanced focal breast lesions, and intracranial aneurysms. To employ the method mentioned above to differentiate between triple-negative and non-triple-negative as well as HER2 positive and HER2 negative cancer lesions, and to assess the value of quantitative and semi-quantitative parameters in molecularly typing breast cancer. All participants were scanned with a 3.0-T MR scanner (GE SIGNA™ Premier) using a 16-channel phased-array body coil. Each participant underwent a DISCO DCE-MRI with a scan time of approximately 1 minute and 40 seconds. The ROIs were outlined with the GenIQ software, avoiding regions with blood vessels, susceptibility artifacts, hemorrhage, and necrosis. We evaluated four quantitative parameters (Ktrans, kep, ve, vp) and four semi-quantitative parameters (TTP, MAX Conc, AUC, MAX Slope). The carcinomas were segregated into respective subgroups (HER2+ vs HER2-, TNBC vs non-TNBC, HER2+ vs TNBC) and we compared the eight parameters across these groups. The AUC of the models was compared using DeLong's test as per the ROC analysis. We analyzed a total of 96 female patients, revealing significant differences in the semi-quantitative parameters such as TTP, MAX Conc, AUC, and MAX Slope among different groups. HER2-positive versus HER2-negative exhibited significant differences in quantitative parameters (Ktrans: 0.22 min-1 vs. 0.43 min-1, kep: 0.11 min-1 vs. 0.35 min-1, vp: 0.01 vs. 0.04, all P < 0.05). TNBC versus non-TNBC revealed statistical variations in quantitative parameters (Ktrans: 1.03 min-1 vs. 0.15 min-1, kep: 0.61 min-1 vs. 0.19 min-1, vp: 0.18 vs. 0.01, all P < 0.05). Additionally, HER2-positive compared to TNBC demonstrated significant differences in quantitative parameters (Ktrans: 0.22 min-1 vs. 1.03 min-1, kep: 0.11 min-1 vs. 0.61 min-1, vp: 0.01 vs. 0.18, all P < 0.05). As per ROC analysis, Ktrans, kep, vp, TTP, and MAX Conc effectively differentiated TNBC from non-TNBC, with TTP being the strongest determinant for TNBC. Furthermore, these parameters successfully distinguished between HER2 positive and HER2 negative, with kep being particularly effective in identifying HER2. Importantly, Ktrans, kep, vp, TTP, and MAX Conc were effective in discriminating HER2 positive from TNBC, with kep and TTP exhibiting notable efficacy in this context. Our study suggests that DISCO DCE-MRI derived parameters could serve as reliable quantitative biomarkers for differentiating between TNBC and HER2 positive breast cancer.