This survey paper describes a literature review of deep learning (DL) methods for cyber security applications. A short tutorial-style description of each DL method is provided, including deep ...autoencoders, restricted Boltzmann machines, recurrent neural networks, generative adversarial networks, and several others. Then we discuss how each of the DL methods is used for security applications. We cover a broad array of attack types including malware, spam, insider threats, network intrusions, false data injection, and malicious domain names used by botnets.
Abstract This expert consensus statement summarizes the available data regarding the prognostic value of CAC in the asymptomatic population and its ability to refine individual risk prediction, ...addresses the limitations identified in the current traditional risk factor-based treatment strategies recommended by the 2013 ACC/AHA Prevention guidelines including use of the Pooled Cohort Equations (PCE), and the US Preventive Services Task Force (USPSTF) Recommendation Statement for Statin Use for the Primary Prevention of Cardiovascular Disease in Adults. It provides CAC based treatment recommendations both within the context of the shared decision making model espoused by the 2013 ACC/AHA Prevention guidelines and independent of these guidelines.
The ISCHEMIA trial (International Study of Comparative Health Effectiveness With Medical and Invasive Approaches) postulated that patients with stable coronary artery disease (CAD) and moderate or ...severe ischemia would benefit from revascularization. We investigated the relationship between severity of CAD and ischemia and trial outcomes, overall and by management strategy.
In total, 5179 patients with moderate or severe ischemia were randomized to an initial invasive or conservative management strategy. Blinded, core laboratory-interpreted coronary computed tomographic angiography was used to assess anatomic eligibility for randomization. Extent and severity of CAD were classified with the modified Duke Prognostic Index (n=2475, 48%). Ischemia severity was interpreted by independent core laboratories (nuclear, echocardiography, magnetic resonance imaging, exercise tolerance testing, n=5105, 99%). We compared 4-year event rates across subgroups defined by severity of ischemia and CAD. The primary end point for this analysis was all-cause mortality. Secondary end points were myocardial infarction (MI), cardiovascular death or MI, and the trial primary end point (cardiovascular death, MI, or hospitalization for unstable angina, heart failure, or resuscitated cardiac arrest).
Relative to mild/no ischemia, neither moderate ischemia nor severe ischemia was associated with increased mortality (moderate ischemia hazard ratio HR, 0.89 95% CI, 0.61-1.30; severe ischemia HR, 0.83 95% CI, 0.57-1.21;
=0.33). Nonfatal MI rates increased with worsening ischemia severity (HR for moderate ischemia, 1.20 95% CI, 0.86-1.69 versus mild/no ischemia; HR for severe ischemia, 1.37 95% CI, 0.98-1.91;
=0.04 for trend,
=NS after adjustment for CAD). Increasing CAD severity was associated with death (HR, 2.72 95% CI, 1.06-6.98) and MI (HR, 3.78 95% CI, 1.63-8.78) for the most versus least severe CAD subgroup. Ischemia severity did not identify a subgroup with treatment benefit on mortality, MI, the trial primary end point, or cardiovascular death or MI. In the most severe CAD subgroup (n=659), the 4-year rate of cardiovascular death or MI was lower in the invasive strategy group (difference, 6.3% 95% CI, 0.2%-12.4%), but 4-year all-cause mortality was similar.
Ischemia severity was not associated with increased risk after adjustment for CAD severity. More severe CAD was associated with increased risk. Invasive management did not lower all-cause mortality at 4 years in any ischemia or CAD subgroup. Registration: URL: https://www.clinicaltrials.gov. Unique identifier: NCT01471522.
Artificial intelligence (AI) has transformed key aspects of human life. Machine learning (ML), which is a subset of AI wherein machines autonomously acquire information by extracting patterns from ...large databases, has been increasingly used within the medical community, and specifically within the domain of cardiovascular diseases. In this review, we present a brief overview of ML methodologies that are used for the construction of inferential and predictive data-driven models. We highlight several domains of ML application such as echocardiography, electrocardiography, and recently developed non-invasive imaging modalities such as coronary artery calcium scoring and coronary computed tomography angiography. We conclude by reviewing the limitations associated with contemporary application of ML algorithms within the cardiovascular disease field.
The role of myocardial viability assessment in identifying patients with ischemic cardiomyopathy who will benefit from surgical revascularization is controversial. This study assessed myocardial ...viability and its relationship to long-term outcomes in 601 patients with ischemic cardiomyopathy who were assigned to surgical revascularization plus medical therapy or medical therapy alone.
Sex beyond the genitalia Joel, Daphna; Berman, Zohar; Tavor, Ido ...
Proceedings of the National Academy of Sciences,
12/2015, Volume:
112, Issue:
50
Journal Article
Peer reviewed
Open access
Whereas a categorical difference in the genitals has always been acknowledged, the question of how far these categories extend into human biology is still not resolved. Documented sex/gender ...differences in the brain are often taken as support of a sexually dimorphic view of human brains (“female brain” or “male brain”). However, such a distinction would be possible only if sex/gender differences in brain features were highly dimorphic (i.e., little overlap between the forms of these features in males and females) and internally consistent (i.e., a brain has only “male” or only “female” features). Here, analysis of MRIs of more than 1,400 human brains from four datasets reveals extensive overlap between the distributions of females and males for all gray matter, white matter, and connections assessed. Moreover, analyses of internal consistency reveal that brains with features that are consistently at one end of the “maleness-femaleness” continuum are rare. Rather, most brains are comprised of unique “mosaics” of features, some more common in females compared with males, some more common in males compared with females, and some common in both females and males. Our findings are robust across sample, age, type of MRI, and method of analysis. These findings are corroborated by a similar analysis of personality traits, attitudes, interests, and behaviors of more than 5,500 individuals, which reveals that internal consistency is extremely rare. Our study demonstrates that, although there are sex/gender differences in the brain, human brains do not belong to one of two distinct categories: male brain/female brain.
This study evaluated the added predictive value of combining clinical information and myocardial perfusion single-photon emission computed tomography (SPECT) imaging (MPI) data using machine learning ...(ML) to predict major adverse cardiac events (MACE).
Traditionally, prognostication by MPI has relied on visual or quantitative analysis of images without objective consideration of the clinical data. ML permits a large number of variables to be considered in combination and at a level of complexity beyond the human clinical reader.
A total of 2,619 consecutive patients (48% men; 62 ± 13 years of age) who underwent exercise (38%) or pharmacological stress (62%) with high-speed SPECT MPI were monitored for MACE. Twenty-eight clinical variables, 17 stress test variables, and 25 imaging variables (including total perfusion deficit TPD) were recorded. Areas under the receiver-operating characteristic curve (AUC) for MACE prediction were compared among: 1) ML with all available data (ML-combined); 2) ML with only imaging data (ML-imaging); 3) 5-point scale visual diagnosis (physician MD diagnosis); and 4) automated quantitative imaging analysis (stress TPD and ischemic TPD). ML involved automated variable selection by information gain ranking, model building with a boosted ensemble algorithm, and 10-fold stratified cross validation.
During follow-up (3.2 ± 0.6 years), 239 patients (9.1%) had MACE. MACE prediction was significantly higher for ML-combined than ML-imaging (AUC: 0.81 vs. 0.78; p < 0.01). ML-combined also had higher predictive accuracy compared with MD diagnosis, automated stress TPD, and automated ischemic TPD (AUC: 0.81 vs. 0.65 vs. 0.73 vs. 0.71, respectively; p < 0.01 for all). Risk reclassification for ML-combined compared with visual MD diagnosis was 26% (p < 0.001).
ML combined with both clinical and imaging data variables was found to have high predictive accuracy for 3-year risk of MACE and was superior to existing visual or automated perfusion assessments. ML could allow integration of clinical and imaging data for personalized MACE risk computations in patients undergoing SPECT MPI.
Display omitted
Fluorine-18 flurpiridaz is a novel positron emission tomography (PET) myocardial perfusion imaging tracer.
This study sought to assess the diagnostic efficacy of flurpiridaz PET versus ...technetium-99m–labeled single photon emission computed tomography SPECT for the detection and evaluation of coronary artery disease (CAD), defined as ≥50% stenosis by quantitative invasive coronary angiography (ICA). Flurpiridaz safety was also evaluated.
In this phase III prospective multicenter clinical study, 795 patients with known or suspected CAD from 72 clinical sites in the United States, Canada, and Finland were enrolled. A total of 755 patients were evaluable, and the mean age was 62.3 ± 9.5 years, 31% were women, 55% had body mass index ≥30 kg/m2, and 71% had pharmacological stress. Patients underwent 1-day rest-stress (pharmacological or exercise) flurpiridaz PET and 1- or 2-day rest-stress Tc-99m–labeled SPECT and ICA. Images were read by 3 experts blinded to clinical and ICA data.
Sensitivity of flurpiridaz PET (for detection of ≥50% stenosis by ICA) was 71.9% (95% confidence interval CI: 67.0% to 76.3%), significantly (p < 0.001) higher than SPECT (53.7% 95% CI: 48.5% to 58.8%), while specificity did not meet the prespecified noninferiority criterion (76.2% 95% CI: 71.8% to 80.1% vs. 86.6% 95% CI: 83.2% to 89.8%; p = NS). Receiver-operating characteristic curve analysis demonstrated superior discrimination of CAD by flurpiridaz PET versus SPECT in the overall population, in women, obese patients, and patients undergoing pharmacological stress testing (p < 0.001 for all). Flurpiridaz PET was superior to SPECT for defect size (p < 0.001), image quality (p < 0.001), diagnostic certainty (p < 0.001), and radiation exposure (6.1 ± 0.4 mSv vs. 13.4 ± 3.2 mSv; p < 0.001). Flurpiridaz PET was safe and well tolerated.
Flurpiridaz PET myocardial perfusion imaging shows promise as a new tracer for CAD detection and assessment of women, obese patients, and patients undergoing pharmacological stress testing. A second phase III Food and Drug Administration trial is ongoing. (A Phase 3 Multi-center Study to Assess PET Imaging of Flurpiridaz F 18 Injection in Patients with CAD; NCT01347710)
Display omitted
Objectives This study examined a large cohort to assess whether progression of coronary artery calcium (CAC) was associated with all-cause mortality, and which among 3 different methods to assess CAC ...progression provided the best estimate of risk. Background Serial assessment of CAC scores has been proposed as a method to follow progression of coronary artery disease, and it has been suggested that excessive CAC progression may be a useful noninvasive predictor of the patient's risk of future events. However, the optimal method to measure calcium progression has not been well established. Methods The study sample consisted of 4,609 consecutive asymptomatic individuals referred by primary physicians for CAC measurement with electron beam tomography, who underwent repeat screening. Three general statistical approaches were taken: 1) the absolute difference between follow-up and baseline CAC score; 2) percent annualized differences between follow-up and baseline CAC score; and 3) difference between square root of baseline and square root of follow-up CAC score >2.5 (the “SQRT method”). Results The average interscan time was 3.1 years, and there were 288 deaths. Progression of CAC was significantly associated with mortality regardless of the method used to assess progression (p < 0.0001). After adjusting for baseline score, age, sex, and time between scans, the best CAC progression model to predict mortality was the SQRT method (hazard ratio HR: 3.34; 95% confidence interval CI: 2.65 to 4.21; p < 0.0001), followed by a >15% yearly increase (HR: 2.98; 95% CI: 2.20 to 4.95; p < 0.0001). Progression was very limited and did not predict mortality in patients with baseline CAC = 0. Conclusions The CAC progression added incremental value in predicting all-cause mortality over baseline score, time between scans, demographics, and cardiovascular risk factors. Serial assessment may have clinical value in assessing plaque progression and future cardiovascular risk.