Areas Under The ROC Curve Research Articles

Prediction of cardiovascular adverse events in newly diagnosed multiple myeloma: Development and validation of a risk score prognostic model.

Multiple myeloma (MM) is the second most common hematological malignancy, and the treatments markedly elevate the survival rate of the patients in recent years. However, the prevalence of cardiovascular adverse events (CVAEs) in MM had been increasing recently. CVAEs in MM patients are an important problem that we should focus on. Clinical tools for prognostication and risk-stratification are needed. This is a retrospective study that included patients who were newly diagnosed with multiple myeloma (NDMM) in Shanghai Changzheng Hospital and Affiliated Jinhua Hospital, Zhejiang University School of Medicine from June 2018 to July 2020.A total of 253 patients from two medical centers were divided into training cohort and validation cohort randomly. Univariable analysis of the baseline factors was performed using CVAEs endpoints. Multivariable analysis identified three factors for a prognostic model that was validated in internal validation cohorts. Factors independently associated with CVAEs in NDMM were as follows: age>61 years old, high level of baseline office blood pressure, and left ventricular hypertrophy (LVH). Age contributed 2 points, and the other two factors contributed 1 point to a prognostic model. The model distinguished the patients into three groups: 3-4 points, high risk; 2 points, intermediate risk; 0-1 point, low risk. These groups had significant difference in CVAEs during follow-up days in both training cohort (p<0.0001) and validation cohort (p=0.0018). In addition, the model had good calibration. The C-indexes for the prediction of overall survival of CVAEs in the training and validation cohorts were 0.73 (95% CI, 0.67-0.79) and 0.66 (95% CI, 0.51-0.81), respectively. The areas under the receiver operating characteristic curve (AUROCs) of the 1-year CVAEs probability in the training and validation cohorts were 0.738 and 0.673, respectively. The AUROCs of the 2-year CVAE probability in the training and validation cohorts were 0.722 and 0.742, respectively. The decision-curve analysis indicated that the prediction model provided greater net benefit than the default strategies of providing assessment or not providing assessment for all patients. A prognostic risk prediction model for predicting CVAEs risk of NDMM patients was developed and internally validated. Patients at increased risk of CVAEs can be identified at treatment initiation and be more focused on cardiovascular protection in the treatment plan.

Extracellular volume fraction obtained by dual-energy CT depicting the etiological differences of liver fibrosis.

To assess etiological differences in extracellular volume fraction (ECV) and evaluate its influence on staging performance. A total of 166 patients with normal liver (n = 14) and chronic liver disease related to viral hepatitis (n = 71), alcohol (n = 44), and nonalcoholic steatohepatitis (NASH) (n = 37) underwent dual-energy CT (DECT) of the liver (5-min equilibrium-phase images) between January 2020 and July 2022. The iodine densities of the parenchyma and aorta were measured and ECV was calculated. Comparisons of ECV between each etiology and METAVIR fibrosis stage were statistically analyzed (p < 0.05). ECV in each etiology and all patients significantly increased with higher fibrosis stage (p < 0.001) and showed a strong or moderate correlation with fibrosis stage (Spearman's ρ; all patients, 0.701; viral hepatitis, 0.638; alcoholic, 0.885; NASH, 0.791). In stages F2-F4, ECV in alcoholic liver disease was significantly larger than those for viral hepatitis and NASH (p < 0.05); however, no significant difference in stage F1 was found among the three etiologies. The cutoff values and areas under the receiver operating characteristic curve (AUC-ROCs) for discriminating fibrosis stage (≥ F1- ≥ F4) were higher for alcohol (cutoff values and AUC-ROC; 20.1% and 0.708 for ≥ F1, 23.8% and 0.990 for ≥ F2, 24.3% and 0.968 for ≥ F3, and 26.6% and 0.961 for ≥ F4, respectively) compared with those for the others. ECV in alcoholic liver disease is higher than that in other etiologies in the advanced stages of fibrosis, and etiological differences in ECV affect the staging performance of fibrosis.

Developing a predictive model for clinically significant prostate cancer by combining age, PSA density, and mpMRI

PurposeThe study aimed to construct a predictive model for clinically significant prostate cancer (csPCa) and investigate its clinical efficacy to reduce unnecessary prostate biopsies.MethodsA total of 847 patients from institute 1 were included in cohort 1 for model development. Cohort 2 included a total of 208 patients from institute 2 for external validation of the model. The data obtained were used for retrospective analysis. The results of magnetic resonance imaging were obtained using Prostate Imaging Reporting and Data System version 2.1 (PI-RADS v2.1). Univariate and multivariate analyses were performed to determine significant predictors of csPCa. The diagnostic performances were compared using the receiver operating characteristic (ROC) curve and decision curve analyses.ResultsAge, prostate-specific antigen density (PSAD), and PI-RADS v2.1 scores were used as predictors of the model. In the development cohort, the areas under the ROC curve (AUC) for csPCa about age, PSAD, PI-RADS v2.1 scores, and the model were 0.675, 0.823, 0.875, and 0.938, respectively. In the external validation cohort, the AUC values predicted by the four were 0.619, 0.811, 0.863, and 0.914, respectively. Decision curve analysis revealed that the clear net benefit of the model was higher than PI-RADS v2.1 scores and PSAD. The model significantly reduced unnecessary prostate biopsies within the risk threshold of > 10%.ConclusionsIn both internal and external validation, the model constructed by combining age, PSAD, and PI-RADS v2.1 scores exhibited excellent clinical efficacy and can be utilized to reduce unnecessary prostate biopsies.

Development of a nomogram for predicting 90-day mortality in patients with sepsis-associated liver injury

The high mortality rate in sepsis patients is related to sepsis-associated liver injury (SALI). We sought to develop an accurate forecasting nomogram to estimate individual 90-day mortality in SALI patients. Data from 34,329 patients were extracted from the public Medical Information Mart for Intensive Care (MIMIC-IV) database. SALI was defined by total bilirubin (TBIL) > 2 mg/dL and the occurrence of an international normalized ratio (INR) > 1.5 in the presence of sepsis. Logistic regression analysis was performed to establish a prediction model called the nomogram based on the training set (n = 727), which was subsequently subjected to internal validation. Multivariate logistic regression analysis showed that SALI was an independent risk factor for mortality in patients with sepsis. The Kaplan‒Meier curves for 90-day survival were different between the SALI and non-SALI groups after propensity score matching (PSM) (log rank: P < 0.001 versus P = 0.038), regardless of PSM balance. The nomogram demonstrated better discrimination than the sequential organ failure assessment (SOFA) score, logistic organ dysfunction system (LODS) score, simplified acute physiology II (SAPS II) score, and Albumin–Bilirubin (ALBI) score in the training and validation sets, with areas under the receiver operating characteristic curve (AUROC) of 0.778 (95% CI 0.730–0.799, P < 0.001) and 0.804 (95% CI 0.713–0.820, P < 0.001), respectively. The calibration plot showed that the nomogram was sufficiently successful to predict the probability of 90-day mortality in both groups. The DCA of the nomogram demonstrated a higher net benefit regarding clinical usefulness than SOFA, LODS, SAPSII, and ALBI scores in the two groups. The nomogram performs exceptionally well in predicting the 90-day mortality rate in SALI patients, which can be used to assess the prognosis of patients with SALI and may assist in guiding clinical practice to enhance patient outcomes.

Sequential Organ Failure Assessment (SOFA) Score for predicting mortality in patients with sepsis in Vietnamese intensive care units: a multicentre, cross-sectional study

ObjectivesTo compare the accuracy of the Sequential Organ Failure Assessment (SOFA) and Acute Physiology and Chronic Health Evaluation II (APACHE II) Scores in predicting mortality among intensive care unit (ICU)...

Quelle est la meilleure méthode de délimitation du volume tumoral total des gliomes avant traitement en TEP à la [18F]-FDOPA ? Une étude contrôlée par biopsie

The definition of tumor boundaries in newly diagnosed glioma is crucial for optimal local therapies and/or non-invasive tumor characterization. The widespread use of [18F]FDOPA PET in this setting is hampered by the lack of a histologically validated standardized method. This study aimed to determine the best way to delineate the overall tumor volume in untreated gliomas regarding diagnostic performance and reliability. We retrospectively analyzed 33 biopsy sites and 33 biopsy sites and 33 contralateral virtual biopsies, from fourteen patients. Six sets of 10 minutes images were reconstructed starting at three different time points (10, 20, and 30 minutes post-injection) using two sets of features for the iterative algorithm. Four reference regions (“hemispheric,” “banana-shaped,” “striatum” and “white matter”) were drawn independently by four readers. A step-by-step selection was applied based on areas under the ROC curve (AUC) for the identification of tumor tissue, intraclass correlation coefficients (ICC) across reconstruction features, and inter-reader agreement. The best cut-off value was calculated as maximizing Youden's index. Tumor infiltration was confirmed in all biopsy samples. TargetSUVmax-to-StriatumSUVmean (TSRmean) and TargetSUVmax-to-StriatumSUVmax (TSRmax) on images performed at 20 minutes provided the best overall performances with AUCs = 0.91, inter-algorithm ICCs = 0.99 and inter-reader Fleiss’ kappas = 0.99 for both. Optimal threshold values of TSRmean20–30 = 0.90 and TSRmax20–30 = 0.68 provided sensitivities of 85% and 83%, respectively, and specificities of 86% for both. This study completes the current EANM guidelines and is a step forward in standardizing the delineation methods in untreated glioma patients. TSRmean20–30 or TSRmax20–30 provided the best combination of accuracy and reliability.

The development and validation of a circulating tumor cells-based integrated model for improving the indeterminate lung solid nodules diagnosis.

There is a risk of over investigation and delayed treatment in the work up of solid nodules. Thus, the aim of our study was to develop and validate an integrated model that estimates the malignant risk for indeterminate pulmonary solid nodules (IPSNs). Patients included in this study with IPSNs who was diagnosed malignant or benign by histopathology. Univariate and multivariate logistic regression were used to build integrated model based on clinical, circulating tumor cells (CTCs) and radiomics features. The performance of the integrated model was estimated by applying receiver operating characteristic (ROC) analysis, and tested in different nodules size and intermediate risk IPSNs. Net reclassification index (NRI) was applied to quantify the additional benefit derived from the integrated model. The integrated model yielded areas under the ROC curves (AUCs) of 0.83 and 0.76 in internal and external set, respectively, outperforming CTCs (0.70, P=0.001; 0.68, P=0.128), the Mayo clinical model (0.68, P<0.001; 0.55, P=0.007), the and radiomics model (0.72, P=0.002; 0.67, P=0.050) in both validation sets. Robust performance with high sensitivity up to 98% was also maintained in IPSNs with different solid size and intermediate risk probability. The performance of the integrated model was comparable with positron emission tomography/computed tomography (PET-CT) examination (P=0.308) among the participants with established PET-CT records. NRI demonstrated that the integrated model provided net reclassification of at least 10% on the external validation set compared with single CTCs test. The integrated model could complement conventional risk models to improve the diagnosis of IPSNs, which is not inferior to PET-CT and could help to guide clinician's decision-making on clinically specific population.

Nomogram for tumour response based on prospective cohorts of hepatocellular carcinoma patients receiving immunotherapy combined with targeted therapy: development and validation.

Although immunotherapy combined with targeted therapy can be effective for hepatocellular carcinoma (HCC), not all HCC patients respond to this treatment. Models for predicting tumour response in HCC patients receiving immunotherapy combined with targeted therapy are lacking. A total of 221 HCC patients from two independent prospective cohorts were retrospectively reviewed. The patients were randomly divided into training and validation cohorts at a ratio of 7:3. Standard clinical data were collected from each patient, including age, sex, hepatitis B infection status, laboratory tests, and immune target-related adverse events (itrAEs). Tumour responses were evaluated using the Response Evaluation Criteria in Solid Tumours (RECIST) v1.1 guidelines. ItrAEs were assessed based on the Common Terminology Criteria for Adverse Events version 4.0. The nomogram for tumour response prediction was constructed based on the results of the multivariate logistic regression analysis, areas under the receiver operating characteristic curves (AUROCs) were used to determine the sensitivity and specificity of the model, and calibration plots and Hosmer-Lemeshow chi-square tests were performed to assess the calibration of the model. In the multivariate logistic regression analysis, a solitary tumour (P=0.006), neutropenia (P=0.003) and hypertension (P=0.042) independently predicted objective response (OR). A nomogram for OR was established with AUROCs of 0.734, 0.675, 0.730, and 0.707 in the training, validation, first-line and second-line treatment sets, respectively. Tumour sizes less than 5 cm (P=0.005), a solitary tumour (P=0.037), prognostic nutritional indices greater than or equal to 54.3 (P=0.037), neutropenia (P=0.004) and fatigue (P=0.041) independently predicted disease control (DC). A nomogram for DC was established with AUROCs of 0.804, 0.667, and 0.768 in the training, first-line and second-line treatment sets, respectively. All the Hosmer-Lemeshow tests and calibration curves showed acceptable calibration. The current provides clinicians with new insights into selecting patients for immunotherapy combined with targeted therapy and contributes to the development of immunotherapy for HCC. It is necessary to expand the scale of our research and perform prospective studies to verify our findings.

Evaluation of in silico pathogenicity prediction tools for the classification of small in-frame indels

BackgroundThe use of in silico pathogenicity predictions as evidence when interpreting genetic variants is widely accepted as part of standard variant classification guidelines. Although numerous algorithms have been developed and evaluated for classifying missense variants, in-frame insertions/deletions (indels) have been much less well studied.MethodsWe created a dataset of 3964 small (< 100 bp) indels predicted to result in in-frame amino acid insertions or deletions using data from gnomAD v3.1 (minor allele frequency of 1–5%), ClinVar and the Deciphering Developmental Disorders (DDD) study. We used this dataset to evaluate the performance of nine pathogenicity predictor tools: CADD, CAPICE, FATHMM-indel, MutPred-Indel, MutationTaster2021, PROVEAN, SIFT-indel, VEST-indel and VVP.ResultsOur dataset consisted of 2224 benign/likely benign and 1740 pathogenic/likely pathogenic variants from gnomAD (n = 809), ClinVar (n = 2882) and, DDD (n = 273). We were able to generate scores across all tools for 91% of the variants, with areas under the ROC curve (AUC) of 0.81–0.96 based on the published recommended thresholds. To avoid biases caused by inclusion of our dataset in the tools’ training data, we also evaluated just DDD variants not present in either gnomAD or ClinVar (70 pathogenic and 81 benign). Using this subset, the AUC of all tools decreased substantially to 0.64–0.87. Several of the tools performed similarly however, VEST-indel had the highest AUCs of 0.93 (full dataset) and 0.87 (DDD subset).ConclusionsAlgorithms designed for predicting the pathogenicity of in-frame indels perform well enough to aid clinical variant classification in a similar manner to missense prediction tools.

Deep-Learning Models for Detection and Localization of Visible Clinically Significant Prostate Cancer on Multi-Parametric MRI.

Deep learning for diagnosing clinically significant prostate cancer (csPCa) is feasible but needs further evaluation in patients with prostate-specific antigen (PSA) levels of 4-10 ng/mL. To explore diffusion-weighted imaging (DWI), alone and in combination with T2-weighted imaging (T2WI), for deep-learning-based models to detect and localize visible csPCa. Retrospective. One thousand six hundred twenty-eight patients with systematic and cognitive-targeted biopsy-confirmation (1007 csPCa, 621 non-csPCa) were divided into model development (N=1428) and hold-out test (N=200) datasets. DWI with diffusion-weighted single-shot gradient echo planar imaging sequence and T2WI with T2-weighted fast spin echo sequence at 3.0-T and 1.5-T. The ground truth of csPCa was annotated by two radiologists in consensus. A diffusion model, DWI and apparent diffusion coefficient (ADC) as input, and a biparametric model (DWI, ADC, and T2WI as input) were trained based on U-Net. Three radiologists provided the PI-RADS (version 2.1) assessment. The performances were determined at the lesion, location, and the patient level. The performance was evaluated using the areas under the ROC curves (AUCs), sensitivity, specificity, and accuracy. A P value <0.05 was considered statistically significant. The lesion-level sensitivities of the diffusion model, the biparametric model, and the PI-RADS assessment were 89.0%, 85.3%, and 90.8% (P=0.289-0.754). At the patient level, the diffusion model had significantly higher sensitivity than the biparametric model (96.0% vs. 90.0%), while there was no significant difference in specificity (77.0%. vs. 85.0%, P=0.096). For location analysis, there were no significant differences in AUCs between the models (sextant-level, 0.895 vs. 0.893, P=0.777; zone-level, 0.931 vs. 0.917, P=0.282), and both models had significantly higher AUCs than the PI-RADS assessment (sextant-level, 0.734; zone-level, 0.863). The diffusion model achieved the best performance in detecting and localizing csPCa in patients with PSA levels of 4-10 ng/mL. 3 TECHNICAL EFFICACY: Stage 2.

Clinical study of the value of shear wave elastography in evaluating the degree of liver fibrosis in children.

To explore the clinical application of shear wave elastography (SWE) in evaluating the degree of liver fibrosis in children. To explore the value of SWE in assessing liver fibrosis in children, the correlation between elastography values and the METAVIR grade of liver fibrosis in children with biliary system or liver diseases was studied. Children with significant liver enlargement were enrolled, and the fibrosis grade was analyzed to explore the value of SWE in assessing the degree of liver fibrosis in the presence of significant liver enlargement. A total of 160 children with bile system or liver diseases were recruited. The areas under the receiver operating characteristic curve (AUROCs) for liver biopsy from stage F1 to F4 were 0.990, 0.923, 0.819, and 0.884. According to the degree of liver fibrosis at liver biopsy, there was a high correlation between the SWE value and the degree of liver fibrosis (correlation coefficient 0.74). There was no significant correlation between the Young's modulus value of the liver and the degree of liver fibrosis (correlation coefficient 0.16). Supersonic SWE can generally accurately evaluate the degree of liver fibrosis in children with liver disease. However, When the liver is significantly enlarged, SWE can only evaluate liver stiffness based on Young's modulus values, and the degree of liver fibrosis must still be determined by pathologic biopsy.

Performance of 2D-shear wave elastography in autoimmune hepatitis-primary biliary cholangitis overlap syndrome.

To evaluate the diagnostic values of liver stiffness (LS) measured by 2D-SWE, fibrosis index based on the four factors (FIB-4), aspartate aminotransferase to platelet ratio index (APRI), and GGT to PLT ratio (GPR) for assessing liver fibrosis and high-risk esophageal varices (EVs) in patients with autoimmune hepatitis-primary biliary cholangitis (AIH-PBC) overlap syndrome. Data of 141 patients were retrospectively collected. Liver fibrosis was staged according to the Scheuer scoring system. The Spearman correlation coefficient was used for correlation analysis. Receiver operating characteristic (ROC) curves were plotted to evaluate the diagnostic performance. LS and FIB-4 were positively correlated with the fibrosis stage (r = 0.555 and 0.198, respectively). LS had significantly higher areas under the ROC curves (AUROCs) values than FIB-4 for predicting advanced fibrosis (0.818 vs. 0.567, P < 0.001), cirrhosis (0.879 vs. 0.637, P < 0.001), whereas LS and FIB-4 similarly predicted significant fibrosis (0.748 vs. 0.638, P = 0.071) and high-risk EVs (0.731 vs. 0.659, P = 0.303). The optimal cut-off values of 2D-SWE for detecting significant fibrosis, advanced fibrosis, cirrhosis, and high-risk EVs were 8.7kPa, 12.8kPa, 14.0kPa, and 11.0kPa, respectively. LS values were influenced by fibrosis stage, serum GGT, albumin, and total bilirubin levels. The overall concordance rate of the liver stiffness vs. Scheuer stages was 49.65%. 2D-SWE shows significantly greater diagnostic accuracy than serum fibrosis indexes for diagnosing advanced fibrosis and cirrhosis in patients with AIH-PBC overlap syndrome.

Comparison of Four Intrapartum Cardiotocography Classifications for Predicting Neonatal Acidemia at Birth.

To compare diagnostic values of four intrapartum cardiotocography (CTG) classifications in predicting neonatal acidemia at birth. Retrospective case-control study. Forty-three CTG traces with an umbilical artery pH < 7.00 (study group) and 43 traces with a pH ≥ 7.00 (control group) were analyzed. Inclusion criteria were singleton pregnancy, cephalic presentation, admission to labour ward during active phase of first stage of labour, and gestational age 37+0 to 41+6 weeks. Exclusion criteria were suspected intrauterine growth restriction, oligohydramnios, polyhydramnios, pregestational or gestational insulin-dependent diabetes mellitus, and preeclampsia. Last 30-60 minutes before delivery of CTG traces was classified retrospectively according to four classification systems-International Federation of Gynecology and Obstetrics (FIGO), Royal College of Obstetricians and Gynaecologists (RCOG), National Institute of Child Health and Human Development (NICHD), and the 5-tier system by Parer and Ikeda. Predictive value of each classification for neonatal acidemia was assessed using receiver operating characteristics (ROC) analysis. FIGO, RCOG, and NICHD classifications predicted neonatal acidemia with areas under the ROC curves (AUC) of 0.73, 95% confidence interval (CI) 0.63-0.84; 0.72, 95% CI 0.60-0.83; and 0.69, 95% CI 0.57-0.80, respectively. The five-tier system by Parer and Ikeda had significantly better predictive value with an AUC of 0.96, 95% CI 0.91-1.00. The 5-tier classification system proposed by Parer and Ikeda for assessing CTG in labour was superior to FIGO, RCOG, and NICHD intrapartum CTG classifications in predicting severe neonatal acidemia at birth.

GIS-Based Landslide Susceptibility Modeling: A Comparison between Best-First Decision Tree and Its Two Ensembles (BagBFT and RFBFT)

This study aimed to explore and compare the application of current state-of-the-art machine learning techniques, including bagging (Bag) and rotation forest (RF), to assess landslide susceptibility with the base classifier best-first decision tree (BFT). The proposed two novel ensemble frameworks, BagBFT and RFBFT, and the base model BFT, were used to model landslide susceptibility in Zhashui County (China), which suffers from landslides. Firstly, we identified 169 landslides through field surveys and image interpretation. Then, a landslide inventory map was built. These 169 historical landslides were randomly classified into two groups: 70% for training data and 30% for validation data. Then, 15 landslide conditioning factors were considered for mapping landslide susceptibility. The three ensemble outputs were estimated with a receiver operating characteristic (ROC) curve and statistical tests, as well as a new approach, the improved frequency ratio accuracy. The areas under the ROC curve (AUCs) for the training data (success rate) of the three algorithms were 0.722 for BFT, 0.869 for BagBFT, and 0.895 for RFBFT. The AUCs for the validating groups (prediction rates) were 0.718, 0.834, and 0.872, respectively. The frequency ratio accuracy of the three models was 0.76163 for the BFT model, 0.92220 for the BagBFT model, and 0.92224 for the RFBFT model. Both BagBFT and RFBFT ensembles can improve the accuracy of the BFT base model, and RFBFT was relatively better. Therefore, the RFBFT model is the most effective approach for the accurate modeling of landslide susceptibility mapping (LSM). All three models can improve the identification of landslide-prone areas, enhance risk management ability, and afford more detailed information for land-use planning and policy setting.

A novel non-invasive model for the prediction of advanced liver fibrosis in chronic hepatitis B patients with NAFLD.

Thereare still lack of non-invasive models to evaluate liver fibrosis in chronic hepatitis B (CHB) patients with nonalcoholic fatty liver disease (NAFLD). We aimed to establish a predictive model for advanced fibrosis in these patients. A total of 504 treatment-naive CHB patients with NAFLD who underwent liver biopsy were enrolled and randomly divided into a training set (n=336) and a validation set (n=168). Receiver operating characteristic (ROC) curve was used to compare predicting accuracy for the different models. One hundred fifty-six patients (31.0%) had advanced fibrosis. In the training set, platelet, prothrombin time, type 2 diabetes, HBeAg positivity and globulin were significantly associated with advanced fibrosis by multivariable analysis. A predictive model namely PPDHG for advanced fibrosis was developed based on these parameters. The areas under the ROC curve (AUROC) of PPDHG with an optimal cut-off value of -0.980 in predicting advanced fibrosis was 0.817 (95% confidence interval 0.772 to 0.862), with a sensitivity of 81.82% and a specificity of 66.81%. The predicting accuracy of PPDHG for advanced fibrosis was significantly superior to AST to platelet ratio index (APRI), fibrosis-4 score (FIB-4) and NAFLD fibrosis score (NFS). Further analysis revealed that the AUROC of PPDHG remained significantly higher than FIB-4 and NFS indexes, while it was comparable with APRI for predicting advanced fibrosis in the validation set. PPDHG had a better predicting performance than established models for advanced fibrosis in CHB patients with NAFLD. The application of PPDHG can reduce the necessary for liver biopsy in these patients.

Evaluating Histological Subtypes Classification of Primary Lung Cancers on Unenhanced Computed Tomography Based on Random Forest Model.

Lung cancer is the leading cause of cancer-related death in many countries, and an accurate histopathological diagnosis is of great importance in subsequent treatment. The aim of this study was to establish the random forest (RF) model based on radiomic features to automatically classify and predict lung adenocarcinoma (ADC), lung squamous cell carcinoma (SCC), and small cell lung cancer (SCLC) on unenhanced computed tomography (CT) images. Eight hundred and fifty-two patients (mean age: 61.4, range: 29-87, male/female: 536/316) with preoperative unenhanced CT and postoperative histopathologically confirmed primary lung cancers, including 525 patients with ADC, 161 patients with SCC, and 166 patients with SCLC, were included in this retrospective study. Radiomic features were extracted, selected, and then used to establish the RF classification model to analyse and classify primary lung cancers into three subtypes, including ADC, SCC, and SCLC according to histopathological results. The training (446 ADC, 137 SCC, and 141 SCLC) and testing cohorts (79 ADC, 24 SCC, and 25 SCLC) accounted for 85% and 15% of the whole datasets, respectively. The prediction performance of the RF classification model was evaluated by F1 scores and the receiver operating characteristic (ROC) curve. On the testing cohort, the areas under the ROC curve (AUC) of the RF model in classifying ADC, SCC, and SCLC were 0.74, 0.77, and 0.88, respectively. The F1 scores achieved 0.80, 0.40, and 0.73 in ADC, SCC, and SCLC, respectively, and the weighted average F1 score was 0.71. In addition, for the RF classification model, the precisions were 0.72, 0.64, and 0.70; the recalls were 0.86, 0.29, and 0.76; and the specificities were 0.55, 0.96, and 0.92 in ADC, SCC, and SCLC. The primary lung cancers were feasibly and effectively classified into ADC, SCC, and SCLC based on the combination of RF classification model and radiomic features, which has the potential for noninvasive predicting histological subtypes of primary lung cancers.

The value of enhanced CT features and texture-signatures in assessing the inflammatory infiltration of pancreatic ductal adenocarcinoma.

To explore the predictive value of computed tomography (CT) imaging features and CT-based texture analysis in assessing inflammatory infiltration in pancreatic ductal adenocarcinoma (PDAC). A total of 43 patients with PDAC confirmed by surgical pathology were included in the study. The clinical, radiological, surgical, and pathological features of the patients were analyzed retrospectively using the chi-square test or Spearman's correlation. Receiver operating characteristic (ROC) curves were utilized to assess the overall predictive ability of the tumor enhancement degree on triphasic contrast-enhanced CT images for the inflammatory infiltration degree in PDAC. Furthermore, all CT data were uploaded to the RadCloud platform for region of interest (ROI) delineation and feature extraction. Then, the Variance Threshold and SelectKBest algorithms were used to find the optimal CT features. Binary logistic regression was employed to analyze the selected features in all three contrast-enhanced CT phases, and regression equations were formulated. ROC analysis was performed to evaluate the predictive effectiveness of each equation. The analysis revealed a statistically significant correlation between the degree of differentiation and radiological findings such as necrosis and cystic degeneration, vascular invasion, and the presence of ascites (P < 0.05). The enhancement degree of the tumor in both the arterial and venous phases was significantly correlated with the inflammatory infiltration degree (P < 0.05); however, the areas under the ROC curve (AUCs) of arterial and venous enhancement were 0.570 and 0.542, respectively. Regression equations based on the texture features of triphasic contrast-enhanced tumors were formulated, and their AUCs were 0.982, 0.643, and 0.849. Conventional radiological features are not significantly correlated with the degree of inflammatory infiltration in PDAC. The enhancement degrees in both the arterial phase and venous phase were statistically correlated with the inflammatory infiltration level but had poor predictive value. The texture features of PDAC on contrast-enhanced CT may show a better assessment value, especially in the arterial phase.

Combined application of multiple biomarkers for early auxiliary diagnosis of silicosis.

Silicosis is an important industrial health problem for those workers exposed to silica. The present study aimed to investigate the sensitivity and specificity of combined detection of biomarkers in early auxiliary diagnosis of silicosis, the risk factors of silicosis were also studied. The study sample comprised 65 workers who had clinical silicosis and 70 matched control subjects who were exposed to silica but did not have clinical silicosis. The levels of superoxide dismutase, malondialdehyde, interleukin 6 (IL-6), tumor necrosis factor-alpha, and cholinesterases in the serum of 135 subjects were measured. After completing the biochemical assays, a logistic regression model based on the above biochemical determination results was established, and the receiver operating characteristic curve was used for judging the discrimination ability of different statistical indexes. The expression levels of MDA, IL-6, and TNF-alpha in serum samples of patients with stage I silicosis and MDA and IL-6 in serum samples of patients with stage II silicosis were all significantly higher. Results from logistic regression analysis showed that ChEs were protective factors for silicosis, while age, chronic respiratory symptoms, IL-6, and MDA were risk factors. The areas under the ROC curve (AUC) were 0.86 (IL-6), 0.81 (MDA), and 0.65 (TNF-alpha or ChEs). AUC-ROC = 0.90 (95%CI:0.84-0.95). The diagnostic efficiency of IL-6 combined with MDA and TNF-alpha was better than that of any single biomarker.

Developing a deep learning model to predict epilepsy recurrence in patients with focal cortical dysplasia type III.

A sizable number of patients with focal cortical dysplasia (FCD) type III-related refractory epilepsy continue to experience seizures postsurgically. Deep learning models can automatically assess complex medical image characteristics and predict prognosis with higher efficiency. This study sought to determine whether T2-weighted fluid attenuated inversion recovery (T2W FLAIR) images could predict prognosis of FCD type III-related refractory epilepsy using a deep learning approach. Magnetic resonance imaging (MRI) images of 266 patients with FCD type III diagnosed between 2015 and 2019 were included in this retrospective analysis. A deep learning algorithm utilizing a convolutional neural network (CNN) was trained to classify T2W FLAIR images according to Engel's classification. The preprocessed original image and the region of interest (ROI) outlined by clinicians were input into our neural network separately and then together. Precision, sensitivity, specificity, receiver operating characteristic (ROC) curves, and areas under the ROC curves (AUCs) were computed as part of the statistical analyses of the network performance with varied inputs of the network model assessed. The overall performance met the following metrics when the original image only was input: AUC of 96.22%, sensitivity of 84.47%, and specificity of 97.21%. The metrics were as follows when the ROI only was input: area under the ROC curve of 94.76%, sensitivity of 84.92%, and specificity of 96.24%. For the combined inputs, the metrics were as follows: AUC of 97.17%, sensitivity of 90.86%, and specificity of 96.63%. Deep learning used with conventional MRI can effectively predict the recurrence conditions of epilepsy. Artificial intelligence may help the design of clinical management and enable more precise and individualized prediction for postsurgical prognosis of FCD type III-related refractory epilepsy.

Accuracy of the Sysmex UF-5000 analyzer for urinary tract infection screening and pathogen classification.

The screening performance of urine flow cytometry parameters (e.g., white blood cell and bacteria) for urinary tract infection (UTI) has been widely recognized. The majority of previous studies, however, investigated the screening performance of Sysmex UF-1000i urine flow cytometer. This study aimed to investigate the screening performance of Sysmex UF-5000 analyzer, a third-generation urinary flow cytometer, for UTI and its novel parameter named Gram flag for discriminating gram-positive and negative pathogens. Urine specimens sent to the clinical microbiology laboratory of our hospital for bacterial culture between September 13, 2021, and November 15, 2021, were prospectively and consecutively collected. The Sysmex UF-5000 analyzer was used to determine urine white blood cell (WBC) and bacteria simultaneously. A chemical strip was used to assess urine nitrate. UTI was defined as positive urine bacterial culture > 104 CFU /ml. The receiver operating characteristics (ROC) curve, nomogram, decision tree, and decision curve were used to determine the screening performance of urine WBC, nitrate, and bacterial. A total of 246 UTIs and 425 non-UTIs were enrolled. The areas under the ROC curve (AUCs) for WBC and bacterial were 0.74 and 0.86, respectively. The decision curve showed that urine bacteria had a higher benefit than WBC. The nomogram indicated that urine bacterial had the largest effect on the probability of UTI. The sensitivity and specificity of the decision tree were 0.69 and 0.95, respectively. The flag of Gram-negative had a positive predictive value (PPV) of 0.93 in patients with urine bacteria > 1367 /μl. Therefore, we conclude that urine bacteria determined by the Sysmex UF-5000 had higher screening performance and greater benefit than WBC. The decision tree can be used to improve the screening performance of routine urinary parameters. The flag of Gram-negative is a reliable indicator to confirm gram-negative bacteria infection in UTI patients.