Area Under The Curve Values Research Articles

Lateralizing value of interictal epileptiform discharges and other parameters in hypothalamic hamartoma.

Hypothalamic hamartomas (HHs) are associated with pharmacoresistant epilepsy. Stereotactic radiofrequency thermocoagulation (SRT) shows promise as a disconnecting intervention. Although magnetic resonance imaging (MRI) is typically used to determine the attachment and intervention side, it presents challenges in cases of bilaterally attached HH, where the epileptogenic side is unclear. The lateralizing potential of electroclinical parameters in such cases remains uncertain. This retrospective study evaluates the lateralization value of specific parameters, particularly in patients with unilateral HH, to improve future diagnostics and treatment approaches for bilateral HH. Four lateralizing parameters-semiology, ictal electroencephalography (EEG), and interictal epileptiform discharges during awake (IEDsw) and sleep states (IEDss)-were assessed for correlation with HH attachment side using Spearman's ρ. We calculated areas under the curves (AUCs) and cutoffs for left and right IEDs prognostic lateralizing value, plotting differences between IEDsright and IEDsleft in a receiver-operating characteristic(ROC) curve to establish the required preponderance of unilateral IEDss to differentiate between left and right HHs. Binomial logistic regression was employed to predict the HH attachment side. We included 25 patients (2-55 years of age) with mainly unilateral (n = 22) HHs who underwent SRT and presurgical evaluation. All parameters correlated with HH attachment side (semiology R = -.62, p = .005; ictal EEG R = .51, p = .047; IEDs R = .55, p = .018; IEDw, R = .61, p = .018). AUC values for right and left IEDs were .76 (p = .047) and .85 (p = .019), respectively, with cutoffs of .34 and .15. The AUC for "IEDsright-IEDsleft" was .98 (p = .0018) with a cutoff of .16. IEDss and semiology were significant predictors, achieving 88% correct lateralization. IEDss are promising biomarkers for HH lateralization in unilateral HH. The predominance of unilateral IEDss suggests ipsilateral HH. Even in cases with predominantly bilateral IEDss, a slight preponderance of unilateral IEDss can indicate the attachment side. In addition, combining IEDss and semiology provides a predictive model for HH lateralization.

Radiomics and dosiomics approaches to estimate lung function after stereotactic body radiation therapy in patients with lung tumors.

Lung function assessment is essential for determining the optimal treatment strategy for radiation therapy in patients with lung tumors. This study aimed to develop radiomics and dosiomics approaches to estimate pulmonary function test (PFT) results in post-stereotactic body radiation therapy (SBRT). Sixty-four patients with lung tumors who underwent SBRT were included. Models were created to estimate the PFT results at 0-6months (Cohort 1) and 6-24months (Cohort 2) after SBRT. Radiomics and dosiomics features were extracted from the computed tomography (CT) images and dose distributions, respectively. To estimate the PFT results, Models A (dose-volume histogram [DVH] + radiomics features) and B (DVH + radiomics + dosiomics features) were created. In the PFT results, the forced expiratory volume in one second (FEV1) and forced vital capacity (FVC) were estimated using each model, and the ratio of FEV1 to FVC (FEV1/FVC) was calculated. The Pearson's correlation coefficient (Pearson r) and area under the curve (AUC) for FEV1/FVC (< 70%) were calculated. The models were evaluated by comparing them with the conventional calculation formulae (Conventional). The Pearson r (FEV1/FVC) values were 0.30, 0.64, and 0.69 for Conventional and Models A and B (Cohort 2), respectively, and the AUC (FEV1/FVC < 70%) values were 0.63, 0.80, and 0.78, respectively. This study demonstrates the possibility of estimating lung function after SBRT using radiomics and dosiomics features based on planning CT images and dose distributions.

Development and Validation of a Predictive Model of Prostate Screening Compliance: A Nationwide Population-Based Study.

Prostate cancer (PCa) is the second most common cancer in men worldwide, with significant incidence and mortality, particularly in Mexico, where diagnosis at advanced stages is common. Early detection through screening methods such as digital rectal examination and prostate-specific antigen testing is essential to improve outcomes. Despite current efforts, compliance with prostate screening (PS) remains low due to several barriers. This study aims to develop and validate a predictive model for PCa screening compliance in Mexican men. Retrospective observational design with data from the Mexican Health and Aging Study (MHAS). Participants were men aged 50-69 from three cohorts: development/internal validation, temporal validation, and external validation. Key predictors were identified using relaxed Least Absolute Shrinkage and Selection Operator (LASSO) regression, and model performance was assessed using the area under the curve (AUC) from receiver operating characteristic (ROC) analyses, along with calibration and decision curve analysis (DCA). A web nomogram was also developed. The final model included seven key predictors. AUC values indicated good predictive performance: 0.783 for the training subgroup, 0.722 for the test subgroup, 0.748 for the time cohort, and 0.756 for the external cohort, with sensitivities of 73.5%. The DCA demonstrated the superior clinical utility of the model compared to the reference strategies. The predictive model developed for performance to PCa screening is robust across different cohorts and highlights critical factors influencing performance. The accompanying web-based nomogram enhances clinical applicability and supports interventions aimed at improving PCa screening rates among Mexican men.

Predictive Diagnostic Power of Anthropometric Indicators for Metabolic Syndrome: A Comparative Study in Korean Adults

Background/Objectives: Metabolic syndrome (MetS) is a cluster of risk factors that significantly increase the risk of cardiovascular disease, including type 2 diabetes, etc. Assessing the predictive diagnostic power of anthropometric indicators for MetS is crucial for the early identification and prevention of related health issues. This study focuses on the Korean adult population while providing insights that may be applicable to broader global contexts. Therefore, this study aimed to compare the predictive diagnostic capabilities of various anthropometric indicators, including body mass index (BMI), waist-to-height ratio (WHtR), muscle mass-to-fat mass ratio (MFR), muscle mass-to-waist ratio (MWR), and body shape index (ABSI), in relation to MetS in Korean adults. Methods: Data from 13,725 participants of the Korea National Health and Nutrition Examination Survey (2008–2011) were analyzed. The diagnostic power of each indicator was assessed using the receiver operating characteristic (ROC) curve analysis, and the area under the curve (AUC) values were compared. Participants were classified into normal (NG) and abnormal (AG) groups based on established cutoff values, and logistic regression analysis was performed to evaluate the odds of MetS in each group. Results: WHtR showed the highest AUC values (0.792 for men and 0.768 for women), indicating superior diagnostic accuracy compared to the other indicators (p &lt; 0.001). Logistic regression analysis indicated that both the unadjusted and adjusted odds ratios (OR) for MetS were significantly higher in the AG than in the NG across all indicators (p &lt; 0.001). Specifically, the adjusted OR (95% confidence interval) for WHtR in the AG was 6.793 (5.929–7.784) for men and 4.665 (4.151–5.423) for women, representing the highest values among all indicators (p &lt; 0.001). Conclusion: Among the various anthropometric indicators, WHtR is the most reliable and practical for predicting MetS in Korean adults. It is useful for early intervention and prevention in both clinical and public health settings.

Identification of biomarkers for knee osteoarthritis through clinical data and machine learning models

Knee osteoarthritis (KOA) represents a progressive degenerative disorder characterized by the gradual erosion of articular cartilage. This study aimed to develop and validate biomarker-based predictive models for KOA diagnosis using machine learning techniques. Clinical data from 2594 samples were obtained and stratified into training and validation datasets in a 7:3 ratio. Key clinical features were identified through differential analysis between KOA and control groups, combined with least absolute shrinkage and selection operator (LASSO) regression. The SHapley Additive Planning (SHAP) method was employed to rank feature importance quantitatively. Based on these rankings, predictive models were constructed using Logistic Regression (LR), Random Forest (RF), eXtreme Gradient Boosting (xGBoost), Naive Bayes (NB), Support Vector Machine (SVM), and Decision Tree (DT) algorithms. Models were developed for subsets of variables, including the top 5, top 10, top 15, and all identified features. Receiver operating characteristic (ROC) curves were applied to compare diagnostic performance across models. Additionally, a risk stratification framework for KOA prediction was designed using recursive partitioning analysis (RPA). Using difference analysis and LASSO, 44 critical clinical features were identified. Among these, age, plasma prothrombin time, gender, body mass index (BMI), and prothrombin time and international normalized ratio (PTINR) emerged as the top five features, with SHAP values of 0.1990, 0.0981, 0.0471, 0.0433, and 0.0422, respectively. Machine learning analysis demonstrated that these variables provided robust diagnostic performance for KOA. In the training set, area under the curve (AUC) values for LR, RF, xGBoost, NB, SVM, and DT models were 0.947, 0.961, 0.892, 0.952, 0.885, and 0.779, respectively. Similarly, in the validation dataset, these models achieved AUC values of 0.961, 0.943, 0.789, 0.957, 0.824, and 0.76. Among them, RF consistently exhibited superior diagnostic accuracy for KOA. Additionally, RPA analysis indicated a higher prevalence of KOA among individuals aged 54 years and older. The integration of the top five clinical variables significantly enhanced the diagnostic accuracy for KOA, particularly when employing the RF model. Moreover, the RPA model offered valuable insights to assist clinicians in refining prognostic assessments and optimizing clinical decision-making processes.

Noninvasive identification of HER2 status by integrating multiparametric MRI-based radiomics model with the vesical imaging-reporting and data system (VI-RADS) score in bladder urothelial carcinoma.

HER2 expression is crucial for the application of HER2-targeted antibody-drug conjugates. This study aims to construct a predictive model by integrating multiparametric magnetic resonance imaging (mpMRI) based multimodal radiomics and the Vesical Imaging-Reporting and Data System (VI-RADS) score for noninvasive identification of HER2 status in bladder urothelial carcinoma (BUC). A total of 197 patients were retrospectively enrolled and randomly divided into a training cohort (n = 145) and a testing cohort (n = 52). The multimodal radiomics features were derived from mpMRI, which were also utilized for VI-RADS score evaluation. LASSO algorithm and six machine learning methods were applied for radiomics feature screening and model construction. The optimal radiomics model was selected to integrate with VI-RADS score to predict HER2 status, which was determined by immunohistochemistry. The performance of predictive model was evaluated by receiver operating characteristic curve with area under the curve (AUC). Among the enrolled patients, 110 (55.8%) patients were demonstrated with HER2-positive and 87 (44.2%) patients were HER2-negative. Eight features were selected to establish radiomics signature. The optimal radiomics signature achieved the AUC values of 0.841 (95% CI 0.779-0.904) in the training cohort and 0.794 (95%CI 0.650-0.938) in the testing cohort, respectively. The KNN model was selected to evaluate the significance of radiomics signature and VI-RADS score, which were integrated as a predictive nomogram. The AUC values for the nomogram in the training and testing cohorts were 0.889 (95%CI 0.840-0.938) and 0.826 (95%CI 0.702-0.950), respectively. Our study indicated the predictive model based on the integration of mpMRI-based radiomics and VI-RADS score could accurately predict HER2 status in BUC. The model might aid clinicians in tailoring individualized therapeutic strategies.

Quantifying intratumoral heterogeneity within sub-regions to predict high-grade patterns in clinical stage I solid lung adenocarcinoma

BackgroundThis study aims to quantify intratumoral heterogeneity (ITH) using preoperative CT image and evaluate its ability to predict pathological high-grade patterns, specifically micropapillary and/or solid components (MP/S), in patients diagnosed with clinical stage I solid lung adenocarcinoma (LADC).MethodsIn this retrospective study, we enrolled 457 patients who were postoperatively diagnosed with clinical stage I solid LADC from two medical centers, assigning them to either a training set (n = 304) or a test set (n = 153). Sub-regions within the tumor were identified using the K-means method. Both intratumoral ecological diversity features (hereafter referred to as ITH) and conventional radiomics (hereafter referred to as C-radiomics) were extracted to generate ITH scores and C-radiomics scores. Next, univariate and multivariate logistic regression analyses were employed to identify clinical-radiological (Clin-Rad) features associated with the MP/S (+) group for constructing the Clin-Rad classification. Subsequently, a hybrid model which presented as a nomogram was developed, integrating the Clin-Rad classification and ITH score. The performance of models was assessed using the receiver operating characteristic (ROC) curves, and the area under the curve (AUC), accuracy, sensitivity, and specificity were determined.ResultsThe ITH score outperformed both C-radiomics scores and Clin-Rad classification, as evidenced by higher AUC values in the training set (0.820 versus 0.810 and 0.700, p = 0.049 and p = 0.031, respectively) and in the test set (0.805 versus 0.771 and 0.732, p = 0.041 and p = 0.025, respectively). Finally, the hybrid model consistently demonstrated robust predictive capabilities in identifying presence of MP/S components, achieving AUC of 0.830 in the training set and 0.849 in the test set (all p < 0.05).ConclusionThe ITH derived from sub-region within the tumor has been shown to be a reliable predictor for MP/S (+) in clinical stage I solid LADC.

Utilizing machine-learning techniques on MRI radiomics to identify primary tumors in brain metastases

ObjectiveTo develop a machine learning-based clinical and/or radiomics model for predicting the primary site of brain metastases using multiparametric magnetic resonance imaging (MRI).Materials and methodsA total of 202 patients (87 males, 115 females) with 439 brain metastases were retrospectively included, divided into training sets (brain metastases of lung cancer [BMLC] n = 194, brain metastases of breast cancer [BMBC] n = 108, brain metastases of gastrointestinal tumor [BMGiT] n = 48) and test sets (BMLC n = 50, BMBC n = 27, BMGiT n = 12). A total of 3,404 quantitative image features were obtained through semi-automatic segmentation from MRI images (T1WI, T2WI, FLAIR, and T1-CE). Intra-class correlation coefficient (ICC) was used to examine segmentation stability between two radiologists. Radiomics features were selected using analysis of variance (ANOVA), recursive feature elimination (RFE), and Kruskal–Wallis test. Three machine learning classifiers were used to build the radiomics model, which was validated using five-fold cross-validation on the training set. A comprehensive model combining radiomics and clinical features was established, and the diagnostic performance was compared by area under the curve (AUC) and evaluated in an independent test set.ResultsThe radiomics model differentiated BMGiT from BMLC (13 features, AUC = 0.915 ± 0.071) or BMBC (20 features, AUC = 0.954 ± 0.064) with high accuracy, while the classification between BMLC and BMBC was unsatisfactory (11 features, AUC = 0.729 ± 0.114). However, the combined model incorporating radiomics and clinical features improved the predictive performance, with AUC values of 0.965 for BMLC vs. BMBC, 0.991 for BMLC vs. BMGiT, and 0.935 for BMBC vs. BMGiT.ConclusionThe machine learning-based radiomics model demonstrates significant potential in distinguishing the primary sites of brain metastases, and may assist screening of primary tumor when brain metastasis is suspected whereas history of primary tumor is absent.

Reliable Performance of mALBI Grade‐Based Risk Models for Predicting the Prognosis of Patients With Hepatocellular Carcinoma Receiving Atezolizumab Plus Bevacizumab as First‐Line Treatment: Comparative Analysis of 13 Risk Models

ABSTRACTAimThis study aimed to compare the prognostic performance of the risk models for patients with hepatocellular carcinoma (HCC) receiving atezolizumab and bevacizumab (Atez/Bev) as first‐line treatment.MethodsAmong 449 patients included in this retrospective multicenter study, we compared the prognostic performance of 13 risk models for the 12‐month and 18‐month survival status using area under the curve (AUC), net reclassification improvement (NRI), and relative integrated discrimination improvement (IDI) analysis. We also constructed a calibration plot to assess the fitness of each model.ResultsRegarding the analysis of the 12‐month survival status, none of the risk models demonstrated AUC values higher than the modified albumin‐bilirubin (mALBI) grade. In the NRI analysis, only the IMmunotherapy with AFP, BCLC staging, mALBI, and DCP evaluation (IMABALI‐De score) exhibited a statistically significant improvement compared with the mALBI grade (p = 0.009). While the modified albumin‐bilirubin grade (mALF) score and prognostic nutritional index (PNI) did not exhibit significant differences compared to the mALBI grade (p = 0.3 and 0.2, respectively), the remaining risk models were inferior to the mALBI grade. In the relative IDI analysis, none of the risk models showed a significant improvement compared with the mALBI grade. The calibration plot of the PNI was unsatisfactory. The results for the 18‐month survival status were consistent with those for the 12‐month survival status. A time‐dependent ROC analysis demonstrated that both the mALBI grade and the mALBI‐based model showed consistent performance over time.ConclusionsThe mALBI grade, as well as the IMABALI‐De and mALF scores (both of which are risk models based on mALBI grade), exhibited reliable performance in predicting the prognosis of patients with HCC.

The Validity of a Smartphone-Based Application for Assessing Cognitive Function in the Elderly.

Background/Objectives: The early detection of individuals at risk of cognitive impairment is a clinical imperative. With the recent advancement of digital devices, smartphone application-based cognitive assessment is considered a promising tool for cognitive screening and monitoring inside and outside the clinic. This study examined whether a smartphone-based cognitive assessment, Brain OK, was valid for evaluating cognitive performance and identifying people at risk of cognitive impairment. Methods: We recruited 88 study participants aged over 60. They completed two cognitive tests with the Montreal Cognitive Assessment (MoCA), a validated paper-and-pencil cognitive screening tool, and Brain OK, a smartphone-based cognitive testing application. To examine convergent validity, we conducted analyses of Spearman correlations between MoCA and BrainOK, a Bland-Atman plot with regression analysis, and the area under the curve (AUC). Results: There was a significant positive association between Brain OK and the MoCA total score, with a coefficient of 0.9044 (SE = 0.057, t = 15.750, p < 0.001). The Bland-Altman plot represented a reasonable level of agreement between the two tests. We conducted the AUC analysis of Brain OK to compare the cognitively normal and impaired groups. The AUC value for the Brain OK score of 13.5 was the highest at 0.941. The sensitivity and specificity were 0.958 and 0.925, respectively. Conclusions: The smartphone app-based Brain OK test was feasible for assessing cognitive function and acceptable for identifying subjects with cognitive impairment. The results suggest Brain OK complements traditional in-person cognitive assessments and may help enhance cognitive health dialogue between doctors and patients.

Weakly supervised deep learning-based classification for histopathology of gliomas: a single center experience

Multiple artificial intelligence systems have been created to facilitate accurate and prompt histopathological diagnosis of tumors using hematoxylin-eosin-stained slides. We aimed to investigate whether weakly supervised deep learning can aid in glioma diagnosis. We analyzed 472 whole slide images (WSIs) from 226 patients in West China Hospital (WCH) and 1604 WSIs from 880 patients in The Cancer Genome Atlas (TCGA). We utilized the OpenSlide library to load WSIs, segmented them into small patches using the DeepZoom module, and then normalized the color using the Reinhard method. A weakly supervised deep learning model was developed using ResNet-50 combined with an attention mechanism. We investigated the performance of the model by calculating area under the curve (AUC) in a ten-fold cross-validation setting. Heatmap visualizations showed the prediction mechanism of the model. The results were promising, with high AUC values for differentiating grades of astrocytomas, oligodendrogliomas, all gliomas, and glioma types in the TCGA dataset (0.9419, 0.8659, 0.9904, and 0.9298, respectively), and in the WCH cohort (0.9048, 0.7423, 0.9510, and 0.7098, respectively). The model demonstrated a strong ability to infer IDH status in the TCGA dataset (AUC = 0.9488). The weakly supervised deep learning model proved to be an effective and reliable tool for neuropathological diagnosis, making it an attractive auxiliary tool.

MicroRNA Biomarkers and Host Response Pathways in Severe Pulmonary Hemorrhagic Syndrome due to Leptospirosis: A Multi-Omics Study.

Severe pulmonary hemorrhagic syndrome (SPHS) remains a fatal complication of leptospirosis with poorly understood mechanisms and an urgent need for effective biomarkers. A nested case-control analysis was conducted using blood specimens from two previous Thai leptospirosis cohorts. Candidate microRNAs were initially discovered through a global profiling of 798 serum microRNAs in five SPHS and seven non-SPHS patients, and then validated using real-time polymerase chain reactions in 168 patients. Pathways enriched from microRNA targets were compared to those from an integrated transcriptomic-proteomic analysis. Proteins pertaining to the key resulting pathway were measured to validate significance and reveal correlation with microRNA biomarkers. Serum microRNA profiling revealed a total of 81 significantly expressed microRNAs, of which seven were selected for further validation in the whole cohort of 168 leptospirosis patients, including 28 in SPHS and 140 nonSPHS groups. Among the selected microRNAs, miR-5010-3p and miR-147b-3p had significantly higher expression in SPHS group compared to nonSPHS group, with consistently higher expression after adjusting for age, sex, days of illness, comorbidity, smoking status or recruitment site. The two had area under the curve (AUC) values of 0.76 (95% CI: 0.67-0.85) and 0.70 (95% CI: 0.56-0.81) for discriminating SPHS, respectively. These microRNAs also exhibited consistent AUC values in patients tested before chest radiograph shadows manifested. Combination of miR-5010-3p with miR-548ai and miR-224-5p, as selected by Bayesian Model Average algorithm, substantially boosts the AUC value to 0.86 (95% CI: 0.77-0.94). The miRNA biomarkers also enhanced the predictive values of a previously validated clinical model, increasing AUC value from 0.87 to 0.92 with a significant reclassification net index. Multi-omics pathway analysis incorporating microRNA targets and transcriptomic-proteomic data suggested TNF signaling as among the key pathways. In validation, seven out of ten pathway proteins were significantly different between groups, with principal components correlated with severity and miR-5010-3p. MiR-5010-3p and miR-147b-3p are novel biomarkers with good predictability and potential relevance with TNF signaling pathway, an important host response mechanism in leptospirosis SPHS.

Integrating multi-omics features enables non-invasive early diagnosis and treatment response prediction of diffuse large B-cell lymphoma.

Multi-omics features of cell-free DNA (cfDNA) can effectively improve the performance of non-invasive early diagnosis and prognosis of cancer. However, multimodal characterization of cfDNA remains technically challenging. We developed a comprehensive multi-omics solution (COMOS) to specifically obtain an extensive fragmentomics landscape, presented by breakpoint characteristics of nucleosomes, CpG islands, DNase clusters and enhancers, besides typical methylation, copy number alteration of cfDNA. The COMOS was tested on 214 plasma samples of diffuse large B-cell lymphoma (DLBCL) and matched healthy controls. For early diagnosis, COMOS improved the area under the curve (AUC) value to.993 compared with the individual omics model, with a sensitivity of 95% at 98% specificity. Detection sensitivity achieved 91% at 99% specificity in early-stage patients, while the AUC values of the individual omics model were 0.942, 0.968, 0.989, 0.935, 0.921, 0.781 and 0.917, respectively, with lower sensitivity and specificity. In the treatment response cohort, COMOS yielded a superior sensitivity of 88% at 86% specificity (AUC, 0.903). COMOS has achieved excellent performance in early diagnosis and treatment response prediction. Our study provides an effectively improved approach with high accuracy for the diagnosis and prognosis of DLBCL, showing great potential for future clinical application. A comprehensive multi-omics solution to specifically obtain an extensive fragmentomics landscape, presented by breakpoint characteristics of nucleosomes, CpG islands, DNase clusters and enhancers, besides typical methylation, copy number alteration of cfDNA. Integrated model of cfDNA multi-omics could be used for non-invasive early diagnosis of DLBCL. Integrated model of cfDNA multi-omics could effectively evaluate the efficacy of R-CHOP before DLBCL treatment.

Artificial Intelligence Models Could Enhance the Diagnostic Accuracy (DA) of Fecal Immunochemical Test (FIT) in the Detection of Colorectal Adenoma in a Screening Setting.

This study evaluated the diagnostic accuracy (DA) for colorectal adenomas (CRA), screened by fecal immunochemical test (FIT), using five artificial intelligence (AI) models: logistic regression (LR), support vector machine (SVM), neural network (NN), random forest (RF), and gradient boosting machine (GBM). These models were tested together with clinical features categorized as low-risk (lowR) and high-risk (highR). The colorectal neoplasia (CRN) screening cohort of 5,090 patients included 222 CRA patients and 264 non-CRA patients. Three consecutive fecal samples from each individual were analyzed by two fecal occult blood (FOB) assays. Five AI models including clinical features of CRN patients and CV test results were used to test the DA for CRA measured by receiving operating characteristic (ROC) curves. In conventional ROC analysis, the area under the curve (AUC) values for different AI models ranged from 0.659 and 0.691 (for AIs with LR and SVM), while the highest AUC values were reached by NN, RF, and GBM models (0.809, 0.840, and 0.858, respectively). In the hierarchical summary ROC (HSROC) analysis, the AUC values were as follows: i) with lowR variables, AUC=0.508; ii) with highR variables, AUC=0.566 and iii) with all AI models, AUC= 0.789. The differences in AUC values were: between i) and ii) p=0.008; between i) and iii) p<0.0001 and between ii) and iii) p<0.0001. In detection of CRA, the AI models proved to be superior to the diagnostic features without AI. This is the first study to report that DA in the diagnosis of CRA can be enhanced by AI models that include clinical data of the patients and results of FIT test.

Differentiating low- and high-proliferative soft tissue sarcomas using conventional imaging features and radiomics on MRI

Abstract Background Soft-tissue sarcomas are rare tumors of the soft tissue. Recent diagnostic studies mainly dealt with conventional image analysis and included only a few cases. This study investigated whether low- and high-proliferative soft tissue sarcomas can be differentiated using conventional imaging and radiomics features on MRI. Methods In this retrospective study, soft tissue sarcomas were separated into two groups according to their proliferative activity: high-proliferative (Ki-67 ≥ 20%) and low-proliferative soft tissue sarcomas (Ki-67 &lt; 20%). Several radiomics features, and various conventional imaging features on MRI like tumor heterogeneity, peritumoral edema, peritumoral contrast-enhancement, percentage of ill-defined tumor margins, Apparent Diffusion Coefficient (ADC) values, and area under the curve (AUC) in contrast dynamics were collected. These imaging features were independently compared with the two mentioned groups. Results 118 sarcoma cases were included in this study. Metastases were more prevalent in high-proliferative soft tissue sarcomas (p &lt; 0.001), and time till metastasis negatively correlated with the Ki-67 proliferation index (k -0.43, p = 0.021). Several radiomics features representing intratumoral heterogeneity differed significantly between both groups, especially in T2-weighted (T2w) and contrast-enhanced T1-weighted (CE-T1w) sequences. Peritumoral contrast enhancement and edema were significantly more common in soft tissue sarcomas with a high Ki-67 index (p &lt; 0.001). Tumor configuration, heterogeneity, and ill-defined margins were commonly seen in high-proliferative soft tissue sarcomas (p = 0.001–0.008). Diffusion restriction (ADC values) and contrast dynamics (AUC values) did not present significant differences between low- and high-proliferative soft tissue sarcomas. Conclusions Several radiomics and conventional imaging features indicate a higher Ki-67 proliferation index in soft tissue sarcomas and can therefore be used to distinguish between low- and high-proliferative soft tissue sarcomas.

Radiomics models based on thoracic and upper lumbar spine in chest LDCT to predict low bone mineral density

This study aims to develop and validate different radiomics models based on thoracic and upper lumbar spine in chest low-dose computed tomography (LDCT) to predict low bone mineral density (BMD) using quantitative computed tomography (QCT) as standard of reference. A total of 905 participants underwent chest LDCT and paired QCT BMD examination were retrospectively included from August 2018 and June 2019. The patients with low BMD (n = 388) and the normal (n = 517) were randomly divided into a training set (n = 622) and a validation set (n = 283). Radiomics features (RFs) were extracted from the single and consecutive vertebrae in chest LDCT images to construct the single vertebra RFs models, mixed RFs models and Radscore models, respectively. The performance of these models was evaluated by the area under the curve (AUC) of receiver operator characteristic curve, using QCT as standard of reference. The Radscore models, mixed RFs models, and single vertebra RFs models yielded the AUC values ranging from 0.809 to 0.906, 0.792 to 0.883, and 0.731 to 0.884 for predicting low BMD in the validation set, respectively. For predicting low BMD, the Radscore model of L1-L2 vertebrae yielded the highest AUC of 0.906, and of T1-T3 yielded the lowest AUC of 0.809 (P < 0.05), respectively. However, there was no significant difference among the AUC values of three Radscore models constructed on the vertebrae of T4-T6 (AUC = 0.855), T7-T9 (AUC = 0.845), and T10-T12 (AUC = 0.871) for predicting low BMD in the validation set (P > 0.1). The Radscore model of L1-L2 have potential to serve as an important tool for predicting and screening low BMD from normal in chest LDCT images.

Diagnostic Value and Short-Term Prognosis Assessment of Copeptin in Non-ST-Elevation Acute Coronary Syndrome.

This study explored the early diagnosis and prognostic value of copeptin in non-ST-segment elevation acute coronary syndrome (NSTE-ACS). 171 patients with chest pain or myocardial ischemia symptoms were enrolled. Patients with NSTE-ACS were further divided into the non-ST-elevation myocardial infarction (NSTEMI) and unstable angina (UA). All NSTE-ACS patients were followed up to record the occurrence of Major Adverse Cardiovascular Events (MACEs). Serum copeptin concentration in the NSTE-ACS group was significantly higher than that in the control group. The Area under the curve (AUC) value of copeptin in the diagnosis of NSTE-ACS was 0.798. The combined AUC value of copeptin and hypersensitive troponin I (hs-TnI) to NSTE-ACS increased to 0.930. In addition, copeptin and hs-TnI have been proven to be independent risk factors for MACEs in patients with NSTE-ACS. The use of copeptin in combination with conventional myocardial markers contributes to the early diagnosis and short-term prognosis assessment of NSTE-ACS.

Impact of retraining and data partitions on the generalizability of a deep learning model in the task of COVID-19 classification on chest radiographs.

This study aimed to investigate the impact of different model retraining schemes and data partitioning on model performance in the task of COVID-19 classification on standard chest radiographs (CXRs), in the context of model generalizability. Two datasets from the same institution were used: Set A (9860 patients, collected from 02/20/2020 to 02/03/2021) and Set B (5893 patients, collected from 03/15/2020 to 01/01/2022). An original deep learning (DL) model trained and tested in the task of COVID-19 classification using the initial partition of Set A achieved an area under the curve (AUC) value of 0.76, whereas Set B yielded a significantly lower value of 0.67. To explore this discrepancy, four separate strategies were undertaken on the original model: (1) retrain using Set B, (2) fine-tune using Set B, (3) regularization, and (4) repartition of the training set from Set A 200 times and report AUC values. The model achieved the following AUC values (95% confidence interval) for the four methods: (1) 0.61 [0.56, 0.66]; (2) 0.70 [0.66, 0.73], both on Set B; (3) 0.76 [0.72, 0.79] on the initial test partition of Set A and 0.68 [0.66, 0.70] on Set B; and (4) on repartitions of Set A. The lowest AUC value (0.66 [0.62, 0.69]) of the Set A repartitions was no longer significantly different from the initial 0.67 achieved on Set B. Different data repartitions of the same dataset used to train a DL model demonstrated significantly different performance values that helped explain the discrepancy between Set A and Set B and further demonstrated the limitations of model generalizability.

Non-enhancing margin and pial invasion in MRI can predict IDH status in glioma patients.

The presence of isocitrate dehydrogenase (IDH) mutations and 1p/19q codeletion significantly influences the diagnosis and prognosis of patients with lower-grade gliomas (LGGs). The ability to predict these molecular signatures preoperatively can inform surgical strategies. This study sought to establish an interpretable imaging feature set for predicting molecular signatures and overall survival in LGGs. A cohort of 113 patients with grade 2 or 3 glioma (66 with mutated IDH and 47 with wild-type IDH) was analyzed. The feature set, chief complaints, and onset symptoms were integrated into a logistic regression model to predict IDH mutation and 1p/19q codeletion statuses. Receiver operator characteristic (ROC) and area under the curve (AUC) analyses were performed. The predictive model was externally validated using a public database from The Cancer Genome Atlas (TCGA). Smooth non-enhancing margin and pial invasion were significant predictors of IDH mutation, with odds ratio (OR) values of 3.55 (P = 0.03) and 7.89 (P = 1.0 × 10-3), respectively. Using the Visually Accessible Rembrandt Images (VASARI) feature set alone to predict IDH mutation status yielded an AUC value of 0.83, which increased to 0.85 and 0.87 when incorporating clinical information and onset symptoms for predicting IDH mutation and 1p/19q codeletion, respectively. Gliomas with IDH mutations were more likely to exhibit smooth non-enhancing margins and pial invasion. In clinical practice, imaging prediction allows for the assessment of IDH mutation to shift from a postoperative outcome to a preoperative guidance indicator, facilitating more precise treatment for patients with LGGs.

Predicting the risk of pulmonary embolism in patients with tuberculosis using machine learning algorithms

BackgroundThis study aimed to develop predictive models with robust generalization capabilities for assessing the risk of pulmonary embolism in patients with tuberculosis using machine learning algorithms.MethodsData were collected from two centers and categorized into development and validation cohorts. Using the development cohort, candidate variables were selected via the Recursive Feature Elimination (RFE) method. Five machine learning algorithms, logistic regression (LR), random forest (RF), extreme gradient boosting (XGBoost), light gradient boosting machine (LightGBM), and support vector machine (SVM), were utilized to construct the predictive models. Model performance was evaluated through nested cross-validation and area under the curve (AUC) metrics, supplemented by interpretations using Shapley Additive explanations (SHAP) and line charts of AUC values. Models were subjected to external validation using an independent validation group, facilitating the early identification and management of pulmonary embolism risks in tuberculosis patients.ResultsData from 694 patients were used for model development, and 236 patients from the validation group met the enrollment criteria. The optimal subset of variables identified included D-dimer, smoking status, dyspnea, age, sex, diabetes, platelet count, cough, fibrinogen, hemoglobin, hemoptysis, hypertension, chronic obstructive pulmonary disease (COPD), and chest pain. The RF model outperformed others, achieving an AUC of 0.839 (95% CI 0.780–0.899) and maintaining the highest average performance in external fivefold cross-validation (AUC: 0.906 ± 0.041).ConclusionsThe RF model demonstrates high and consistent effectiveness in predicting pulmonary embolism risk in tuberculosis patients.