Respective Receiver Operating Characteristic Research Articles

Prediction of late atrial fibrillation recurrence after catheter ablation using machine learning

Abstract Background Late recurrence of atrial fibrillation (LRAF) occurring 3 up to 12 months following catheter ablation is considered a relevant clinical problem (rate of occurrence estimated at 20%-45%). Hence, it is significant to define LRAF predictors which would allow for a more suitable selection of patients susceptible to AF recurrence and which could increase the effectiveness of the ablation procedure. Purpose The aim of the present study is to develop a machine-learning model of arrhythmic recurrence within one year from catheter ablation. Methods This study group comprised 201 patients with paroxysmal, persistent and long standing persistent AF who underwent pulmonary vein isolation performed by cryoballoon ablation and radiofrequency ablation using the CARTO-mapping (122 cryoablation and 79 radiofrequency) in a single center study between May 2016 and March 2018. The biomarker concentrations were measured at baseline and after ablation in a 24-h follow-up. First, five different supervised machine-learning models (decision tree, logistic regression, random forest, XGBoost, support vector machines) were developed for predicting AF recurrence within one year following catheter ablation. Next, SHapley Additive exPlanations (SHAP) were derived to explain the predictions using 82 parameters from clinical, laboratory and procedural variables collected from each patient. The models were trained and validated using stratified 5-fold cross-validation and a feature selection was performed with the permutation importance. Results During the 1-year follow-up LRAF occurred in 57 patients (28.3%). The respective receiver operating characteristic (ROC) curves for this model calculated for the Validation set are presented in Figure 1. It is observed that the XGBoost model with the 12 variables achieves the highest AUC of 0.75 (95% confidence interval: 0.7395, 0.7653). The high values of the top 2 features (early recurrence of atrial fibrillation (ERAF) and TSH) contribute to predicting LRAF most, while low values of the top 3 features (RBC, hs-Troponin T, HDL cholesterol) contribute positively to this class (Figure 2). Conclusion The machine learning model predicting the 1-year risk of atrial fibrillation recurrence after catheter ablation has achieved good predictive ability. The above model based on the easily available clinical and laboratory data will support the clinical decision-making for selecting good responders to the ablation procedure.

Tissue Characteristics of Endometrial Carcinoma Analyzed by Quantitative Synthetic MRI and Diffusion-Weighted Imaging.

This study investigates the association of T1, T2, proton density (PD) and the apparent diffusion coefficient (ADC) with histopathologic features of endometrial carcinoma (EC). One hundred and nine EC patients were prospectively enrolled from August 2019 to December 2020. Synthetic magnetic resonance imaging (MRI) was acquired through one acquisition, in addition to diffusion-weighted imaging (DWI) and other conventional sequences using 1.5T MRI. T1, T2, PD derived from synthetic MRI and ADC derived from DWI were compared among different histopathologic features, namely the depth of myometrial invasion (MI), tumor grade, cervical stromal invasion (CSI) and lymphovascular invasion (LVSI) of EC by the Mann-Whitney U test. Classification models based on the significant MRI metrics were constructed with their respective receiver operating characteristic (ROC) curves, and their micro-averaged ROC was used to evaluate the overall performance of these significant MRI metrics in determining aggressive histopathologic features of EC. EC with MI had significantly lower T2, PD and ADC than those without MI (p = 0.007, 0.006 and 0.043, respectively). Grade 2-3 EC and EC with LVSI had significantly lower ADC than grade 1 EC and EC without LVSI, respectively (p = 0.005, p = 0.020). There were no differences in the MRI metrics in EC with or without CSI. Micro-averaged ROC of the three models had an area under the curve of 0.83. Synthetic MRI provided quantitative metrics to characterize EC with one single acquisition. Low T2, PD and ADC were associated with aggressive histopathologic features of EC, offering excellent performance in determining aggressive histopathologic features of EC.

PRM-MS Quantitative Analysis of Isomeric N-Glycopeptides Derived from Human Serum Haptoglobin of Patients with Cirrhosis and Hepatocellular Carcinoma.

Currently, surveillance strategies have inadequate performance for cirrhosis and early detection of hepatocellular carcinoma (HCC). The glycosylation of serum haptoglobin has shown to have significant differences between cirrhosis and HCC, thus can be used for diagnosis. We performed a comprehensive liquid chromatography—parallel reaction monitoring—mass spectrometry (LC-PRM-MS) approach, where a targeted parallel reaction monitoring (PRM) strategy was coupled to a powerful LC system, to study the site-specific isomerism of haptoglobin (Hp) extracted from cirrhosis and HCC patients. We found that our strategy was able to identify a large number of isomeric N-glycopeptides, mainly located in the Hp glycosylation site Asn207. Four N-glycopeptides were found to have significant changes in abundance between cirrhosis and HCC samples (p < 0.05). Strategic combinations of the significant N-glycopeptides, either with alpha-fetoprotein (AFP) or themselves, better estimate the areas under the curve (AUC) of their respective receiver operating characteristic (ROC) curves with respect to AFP. The combination of AFP with the isomeric sialylated fucosylated N-glycopeptides Asn207 + 5-6-1-2 and Asn207 + 5-6-1-3, resulted with an AUC value of 0.98, while the AUC value for AFP alone was 0.85. When comparing cirrhosis vs. early HCC, the isomeric N-glycopeptide Asn207 + 5-6-0-1 better estimated AUC with respect to AFP (AUCAFP = 0.81, and AUCAsn207 + 5-6-0-1 = 0.88, respectively).

The Importance of Nasal Provocation Testing in the Diagnosis of Dermatophagoides pteronyssinus-Induced Allergic Rhinitis.

Allergen identification is the first step for allergen-specific immunotherapy (AIT) of allergic rhinitis (AR). Currently, the diagnosis of AR is based mainly on the positive results of the skin prick test (SPT) and/or serum specific immunoglobulin E (sIgE) measurement. However, the results of these two tests may not always directly correlate with AR. To investigate the importance of nasal provocation testing (NPT) in the diagnosis of Dermatophagoides pteronyssinus (Der p)-induced AR. Rhinitis patients willing to undergo AIT (n = 171) were enrolled. The correlations of Der p SPT, sIgE, NPT, and clinical symptom severity were assessed. NPT-positive responses were more common in patients with higher SPT and sIgE levels. The optimal cut-off value for a NPT-positive response for SPT was 5.5 mm and for sIgE was 2.77 kUA/L, based on the respective receiver operating characteristic (ROC) curves. The area under the curve (AUC) of the ROCs was 0.814 (SPT only) and 0.794 (sIgE only) and increased to 0.828 with the combination of SPT and sIgE. The Der p-NPT concentration was inversely correlated with SPT and sIgE levels (r = -0.477, P < .001, and r = -0.461, P < .001, respectively), but none was correlated with the total nasal symptom score. For patients who are willing to receive Der p AIT, NPT is a useful and safe test to confirm diagnosis prior to treatment initiation, especially in patients with lower levels of Der p SPT (< 5.5 mm) or sIgE (< 2.77 kUA/L).

A new rapid screening program based on risk scores for COVID-19 patients

We aimed at establishing a new COVID-19 risk scores, serving as a guide for rapidly screening the COVID-19 patients in order to reduce the risk of COVID-19 hospital-related transmission. As the COVID-19 disease is breaking out across the world, hospital-related transmission is one of the main factors accountable for the spread of COVID-19. For COVID-19 prevention it is urgent to establish a fast and efficient screening strategy for the COVID-19 patients. We analyzed 335 patients (including 124 patients with COVID-19). Five significant clinical attributes were selected as the components for establishing a COVID-19 risk score system, and every attribute was assigned a specific score according to their respective odds ratio values. We also compared three different screening schemes (Scheme I: temperature higher than 37.2 °C on admission, Scheme II: exposure to a source of transmission within 14 days in addition to fever, Scheme III: our new COVID-19 risk score) in terms of their respective receiver operating characteristic (ROC) curves, so as to evaluate their respective screening effectiveness. Five significant risk factors, which were exposed to a source of transmission (9 points), cluster onset (6 points), history of fever or temperature higher than 37.2 °C on admission (4 points), cough (1 point) and other atypical symptoms (1 point), were ultimately selected from many candidates to construct the new rapid COVID-19 screening program. Based on the screening scheme, the patients were quickly divided into three subgroups according to their respective COVID-19 risk scores: low risk (≤ 6 points, risk < 10%), medium risk (7–13 points) and high risk (≥ 14 points, risk > 80%). When the score of 10 points was selected as a cut-off point for differentiating the patients with COVID-19 from all of the other patients, the sensitivity was 93.6%, with a specificity of 86.3%. The area under the ROC curve (AUC) of COVID-19 risk score system was 0.96 (P = 0.000), much higher than the AUCs of Scheme I (0.56, P = 0.000) and Scheme II (0.85, P = 0.000), respectively. Our COVID-19 risk score system can help the clinicians effectively and rapidly identify and differentiate the patients with COVID-19 infections, to be mainly used in those areas where COVID-19 still exhibits epidemiological characteristics.Electronic supplementary materialThe online version of this article (10.1007/s11739-020-02534-6) contains supplementary material, which is available to authorized users.

Accuracy of Oral 67Gallium Citrate Scintigraphy in assessment of inflammatory activity of Crohn's disease.

The main goal in Crohn´s disease (CD) is a sustained suppression of inflammatory activity associated with mucosa healingin endoscopic evaluation. During clinical routine, there are small numbers of good markers to monitor inflammatory activity under treatment. We postulated that Oral 67Gallium Citrate Scintigraphy is able to mark inflammatory disease in mucosa and deep inflammation in CD, when used in oral form. Measure the accuracy of Oral 67Gallium Citrate Scintigraphy in intestinal inflammatory activity of Crohn´s disease. In a prospective consecutive cross-sectional study from January 2018 to June 2019, theileocolonic region of32 patients with CD were studied by dividing into four regions of interest (ROI) from the ileum to the rectum.A total of128 intestinal segmentswere analyzed in cluster data. Accuracy values of Oral 67Gallium Scintigraphy and colonoscopy tests were evaluated with the histological reference test. Values of the respective receiver operating characteristic (ROC) curves were obtained and compared. The reliability between the tests wasevaluated by Kappa statistical with the segment-level analyses using variance adjustments. All statistical analyses were performed with a test significance level of 0.05. The study population included 32 patients with CD(10 men, 22 women; average age 39years). Diseasetime was five years on average. Anti-TNFs treatment was found in 71%. The most found phenotype of the Montreal classification was L3.Differences in ROC curves for colonoscopy (0.94) and Oral 67Ga Scintigraphy (0.96) did not show significant value (p = 0.32). The sensitivity of scintigraphy to detect intestinal inflammatory activity in CD was 64%, specificity of96% and accuracy of 84%. A high agreement was found between oral scintigraphy and histological measurements with kappa = 0.64. Oral 67Ga Scintigraphy had similar accuracy and agreement compared to colonoscopy in the identification of inflammatory activity in Crohn´s Disease. This new approach may be useful and less invasive for long term follow-ups.

Adequacy of the Fogging Test in the Detection of Clinically Significant Hyperopia in School-Aged Children

Purpose To evaluate the efficacy of the “fogging test,” performed with a +2 diopters (D) lens, in the exclusion of clinically significant hyperopia in school-aged children. Methods We studied 54 children between 5 and 11 years of age, with 10/10 best-corrected bilateral visual acuity (VA) without significant degree of correction. VA was assessed in each eye with a “bilateral” +2 D sphere over-refraction followed by cycloplegic retinoscopy. The capacity of the test to detect hyperopia of ≥+2 D and ≥+1.5 D was evaluated by examining the respective receiver operating characteristic (ROC) curves and sensitivity and specificity values for different cutoff values of VA. Results For the detection of hyperopia ≥+2 D, the area under the ROC curve (AUC) was 0.955 (p ≤ 0.001). The VA cutoff with best discriminative capacity was ≥5/10, with a sensitivity of 100%, specificity of 79%, positive predictive value (PPV) of 57%, and negative predictive value (NPV) of 100%. In respect of ≥+1.5 D hyperopia, the test capacity was lower (AUC = 0.832; p ≤ 0.001). The best VA cutoff was also of ≥5/10, with a PPV of 81% and a NPV of 85%. Conclusion The accuracy of the test was high for the evaluation of ≥+2 D hyperopia but lower for ≥+1.5 D hyperopia. For the detection of ≥+2 D hyperopia, the VA cutoff of <5/10 may permit the exclusion of clinically significant hyperopia in selected children, without the need for cycloplegia. For the same cutoff, the PPV was low, meaning that in children with ≥5/10 VA cycloplegic refraction remains obligatory.

A new prostate cancer biopsy reporting system with prognostic potential.

e16572 Background: Presently, prostate biopsy (PBx) results typically report the highest Gleason Grade Group in the PBx as the single metric used to gauge the clinical aggressiveness of tumor and dictate treatment. However, using that single parameter alone limits the clinician by lack of consideration of the entire PBx. Intuitively we presumed that a PBx showing multiple cores of cancer would represent more aggressive disease. Herein, we propose the Weighted Gleason Grade Group (WGGG), a novel scoring system that synthesizes histopathologic data and cancer volume into a single numeric value representing the entire PBx, allowing the urologist to more accurately predict adverse pathologic outcomes of radical prostatectomy (RP). Methods: We studied 246 men who underwent RP after standard multi-core PBx. The highest Gleason score of each PBx was converted to its final GGG. The WGGG was calculated by summing the Gleason score of each positive core and normalized for a 12 core total. RP pathology was studied to determine adverse pathological outcomes, specifically, extraprostatic extension (EPE), positive surgical margins (PSM), seminal vesical invasion (SVI), pathological up-grading and any adverse feature. We then studied the ability of conventional GGG versus WGGG to have ‘predicted’ the risk of these adverse features comparing their respective receiver operating characteristic (ROC) areas under the curve (AUC) for each, as well as any adverse feature. Results: The AUC for the WGGG vs. GGG was significantly higher for predicting EPE (AUC 0.782 vs. 0.697, respectively; z = -3.29; p = .0009), PSM (AUC 0.644 vs. 0.563, respectively; z = -3.00; p = .0027), SVI (AUC 0.829 vs. 0.713, respectively; z = -3.05; p = .0023) and pathologic up-grading (AUC 0.584 vs. 0.349, respectively; z = -3.13; p = .002). Finally, the AUC for any adverse feature was 0.637 for WGGG versus 0.556 for GGG; z = -3.29; p = .001. Conclusions: The WGGG, by providing a metric reflecting the entirety of the PBx, is more informative than conventional single GGG alone in ‘predicting’ adverse pathologic outcomes on radical prostatectomy specimens. We are now studying if % cancer/ PBx core will improve WGGG performance.

Change Detection in Streaming Multivariate Data Using Likelihood Detectors

Change detection in streaming data relies on a fast estimation of the probability that the data in two consecutive windows come from different distributions. Choosing the criterion is one of the multitude of questions that need to be addressed when designing a change detection procedure. This paper gives a log-likelihood justification for two well-known criteria for detecting change in streaming multidimensional data: Kullback-Leibler (K-L) distance and Hotelling's T-square test for equal means (H). We propose a semiparametric log-likelihood criterion (SPLL) for change detection. Compared to the existing log-likelihood change detectors, SPLL trades some theoretical rigor for computation simplicity. We examine SPLL together with K-L and H on detecting induced change on 30 real data sets. The criteria were compared using the area under the respective Receiver Operating Characteristic (ROC) curve (AUC). SPLL was found to be on the par with H and better than K-L for the nonnormalized data, and better than both on the normalized data.

Magnetic Resonance for Quantitative Assessment of Liver Steatosis: A New Potential Tool to Monitor Antiretroviral-Drug-Related Toxicities

There is an increasing need for new diagnostic tools to monitor antiretroviral drug-related toxicities. Magnetic resonance (MR) imaging and MR spectroscopy are non-invasive diagnostic methods used in the detection and quantification of liver fat. The aim of this study was to compare sensitivity and specificity of different MR techniques in the quantitative assessment of liver steatosis, using liver biopsy as the reference standard, in patients with and without HIV infection. Sequentially evaluated patients with suspected steatosis who were referred for liver biopsy at our tertiary care site were eligible. MR liver fat content (LFC) was estimated by T2-weighted and fat-suppressed T2-weighted spin-echo, dual-phase T1-weighted gradient-echo, multiecho gradient-echo and (1)H spectroscopy. Association between LFC and histological steatosis percentage was calculated by using univariate linear regressions and Pearson's coefficient. Respective receiver operating characteristic (ROC) curves were used to compare specificity and sensitivity of MR methods in diagnosis (cutoff 5%) and in quantitative evaluation (cutoff 33%) of steatosis. A total of 28 patients were identified: 12 refused or had contraindications for liver biopsy and 16 had biopsies plus MR. LFC and histological steatosis percentage were strongly associated (fat-suppressed r=0.86 [P<0.001], dual-phase r=0.88 [P<0.001], multiecho r=0.95 [P<0,001] and spectroscopy r=0.84 [P=0.01]). MR techniques had high sensitivity and specificity in diagnosis and quantitative assessment of steatosis (areas under ROC curves ranging from 0.88 to 0.98). This pilot study confirms that MR may be a sensitive non-invasive alternative to biopsy for the quantitative assessment of liver fat and a potential end point to monitor antiretroviral-drug-related toxicities.

SELDI-TOF biomarker signatures for cystic fibrosis, asthma and chronic obstructive pulmonary disease

Objectives: The aim of this work was to establish protein profiles in serum and nasal epithelial cells of cystic fibrosis individuals in comparison with controls, asthma and chronic obstructive pulmonary disease patients for specific biomarker signatures identification. Design and methods: Protein extracts were analyzed by Surface Enhanced Laser Desorption/Ionization Time-Of-Flight Mass-Spectrometry (SELDI-TOF-MS). Results: The mass spectra revealed a set of peaks with differential expression in serum and nasal cells among the different groups studied, resulting into peak signatures representative/specific of each pathology. Logistic regressions were applied to those peaks; sensitivity, specificity, Youden's indexes and area under the curve (AUC) of the respective receiver operating characteristic (ROC) curves were compared. Discussion: Multivariate analysis demonstrated that combination of peaks has a better predictive value than the individual ones. These protein signatures may serve as diagnostic/prognostic markers for the studied diseases with common clinical features, or as follow-up assessment markers of therapeutic interventions.

99mTc-Hexamethylpropylene Amineoxime Leukocyte Scintigraphy in Acute Pancreatitis: An Alternative to Contrast-Enhanced Computed Tomography?

Contrast-enhanced computed tomography (CECT) is the most efficient imaging technique for the diagnosis and staging of acute pancreatitis (AP); its use, however, may be unfeasible in some patients as a consequence of the drawbacks of intravenous (IV) contrast material. The aim of this study was to test the utility of labeled leukocyte scintigraphy (LLS) as an alternative imaging technique to CECT for the staging of AP. Sixty-six patients with AP were prospectively studied. All patients underwent CECT and pancreatic LLS using (99m)Tc-hexamethylpropylene amineoxime as leukocyte label within a time interval of 2 days, in the early phase of AP. In addition, all patients had their serum C-reactive protein (CRP) concentration measured within 48-72 h after admission. CECT images were analyzed for Balthazar's grade of pancreatitis and for the presence or absence of pancreatic necrosis. Scintigraphic activity of 3-4 h planar images was scored on a 0-2 scale in relation to physiological liver uptake. LLS score was significantly related (p < 0.001) to both components of CECT (grade of pancreatitis and pancreatic necrosis). LLS and serum CRP showed similar results for detecting the most severe pancreatic damage as showed by their respective receiver operating characteristic (ROC) curves. Sensitivities and specificities of LLS score of 2 were, respectively, 62% and 96% for the detection of grade D-E pancreatitis and 90% and 89% for the detection of pancreatic necrosis. Scintigraphic score of 2 increased the likelihood of grade D-E pancreatitis from 32% (pretest probability) to 87% (posttest probability) (likelihood ratio: 13.9) and that of pancreatic necrosis from 16% to 60% (likelihood ratio: 8.4). Our results show that leukocytes are related to the severity of local pancreatic damage in AP. Thus, LLS is a potential alternative technique to CECT for staging AP.