Cardiovascular disease detection using machine learning and carotid/femoral arterial imaging frameworks in rheumatoid arthritis patients

The study proposes a novel machine learning (ML) paradigm for cardiovascular disease (CVD) detection in individuals at medium to high cardiovascular risk using data from a Greek cohort of 542 individuals with rheumatoid arthritis, or diabetes mellitus, and/or arterial hypertension, using conventional or office-based, laboratory-based blood biomarkers and carotid/femoral ultrasound image-based phenotypes. Two kinds of data (CVD risk factors and presence of CVD-defined as stroke, or myocardial infarction, or coronary artery syndrome, or peripheral artery disease, or coronary heart disease) as ground truth, were collected at two-time points: (i) at visit 1 and (ii) at visit 2 after 3 years. The CVD risk factors were divided into three clusters (conventional or office-based, laboratory-based blood biomarkers, carotid ultrasound image-based phenotypes) to study their effect on the ML classifiers. Three kinds of ML classifiers (Random Forest, Support Vector Machine, and Linear Discriminant Analysis) were applied in a two-fold cross-validation framework using the data augmented by synthetic minority over-sampling technique (SMOTE) strategy. The performance of the ML classifiers was recorded. In this cohort with overall 46 CVD risk factors (covariates) implemented in an online cardiovascular framework, that requires calculation time less than 1 s per patient, a mean accuracy and area-under-the-curve (AUC) of 98.40% and 0.98 (p < 0.0001) for CVD presence detection at visit 1, and 98.39% and 0.98 (p < 0.0001) at visit 2, respectively. The performance of the cardiovascular framework was significantly better than the classical CVD risk score. The ML paradigm proved to be powerful for CVD prediction in individuals at medium to high cardiovascular risk.

Independent and interactive effects of kidney stone formation and conventional risk factors for chronic kidney disease: a follow-up study of Japanese men

PURPOSE

To assess the impact of interactions between kidney stone formation and conventional risk factors on incident chronic kidney disease (CKD).

METHODS

A total of 11,402 subjects (men 30-69 years of age, Japanese) without CKD at baseline were observed over an average period of 4 years. Cox proportional hazards regression models were used to calculate hazard ratios (HRs) with 95% confidence intervals (CIs) to determine the association between incident CKD, kidney stone formation, and conventional risk factors (diabetes mellitus, hypertension, overweight/obesity, dyslipidemia, and hyperuricemia/gout). We also examined the interactions of renal stones and the conventional risk factors for CKD.

RESULTS

In total, 2301 men (20.2%) developed incident CKD during the follow-up period. After multivariable adjustment, kidney stones were found to increase the risk of incident CKD (HR 1.16; 95% CI 1.03-1.32). Kidney stone formers with hypertension, dyslipidemia, or hyperuricemia/gout presented a greater risk for incident CKD than those who had either kidney stones or other risk factors. However, no significant interactions between kidney stones and other risk factors were found to increase CKD risk. On the other hand, a negative interactive effect between kidney stones and overweight/obesity was observed, leading to reversed risk of incident CKD in coexistence of both factors.

CONCLUSION

Kidney stones were linked to a higher risk for the development of CKD. However, no positive interactive effects were identified between renal stones and conventional risk factors on the risk of incident CKD.

Cardiovascular/stroke risk prevention: A new machine learning framework integrating carotid ultrasound image-based phenotypes and its harmonics with conventional risk factors

MOTIVATION

Machine learning (ML)-based stroke risk stratification systems have typically focused on conventional risk factors (CRF) (AtheroRisk-conventional). Besides CRF, carotid ultrasound image phenotypes (CUSIP) have shown to be powerful phenotypes risk stratification. This is the first ML study of its kind that integrates CUSIP and CRF for risk stratification (AtheroRisk-integrated) and compares against AtheroRisk-conventional.

METHODS

Two types of ML-based setups called (i) AtheroRisk-integrated and (ii) AtheroRisk-conventional were developed using random forest (RF) classifiers. AtheroRisk-conventional uses a feature set of 13 CRF such as age, gender, hemoglobin A1c, fasting blood sugar, low-density lipoprotein, and high-density lipoprotein (HDL) cholesterol, total cholesterol (TC), a ratio of TC and HDL, hypertension, smoking, family history, triglyceride, and ultrasound-based carotid plaque score. AtheroRisk-integrated system uses the feature set of 38 features with a combination of 13 CRF and 25 CUSIP features (6 types of current CUSIP, 6 types of 10-year CUSIP, 12 types of quadratic CUSIP (harmonics), and age-adjusted grayscale median). Logistic regression approach was used to select the significant features on which the RF classifier was trained. The performance of both ML systems was evaluated by area-under-the-curve (AUC) statistics computed using a leave-one-out cross-validation protocol.

RESULTS

Left and right common carotid arteries of 202 Japanese patients were retrospectively examined to obtain 404 ultrasound scans. RF classifier showed higher improvement in AUC (~57%) for leave-one-out cross-validation protocol. Using RF classifier, AUC statistics for AtheroRisk-integrated system was higher (AUC = 0.99,p-value<0.001) compared to AtheroRisk-conventional (AUC = 0.63,p-value<0.001).

CONCLUSION

The AtheroRisk-integrated ML system outperforms the AtheroRisk-conventional ML system using RF classifier.

The role of the socio-economic environment on medical outcomes after ST-segment elevation myocardial infarction

Background

According to the World Health Organization, coronary artery disease (CAD), including ST-segment elevation myocardial infarction (STEMI), is the most common cause of death worldwide as well as in Europe and Austria. There is valid data on the impact of conventional risk factors on the medical outcomes for STEMI patients. However, only few studies examine the role of the socio-economic environment for medical outcomes. The main task of this study is to investigate if the socio-economic environment of patients who underwent percutaneous coronary intervention (PCI) after STEMI has an impact on the distribution of risk factors and medical outcomes.

Methods

The study focuses on the population of the City of Vienna, Austria, and includes 870 STEMI patients, who underwent PCI at the General Hospital of Vienna (AKH Wien) between 2008 and 2012. The following data were collected: conventional risk factors (hypertension, hyperlipidemia, diabetes, overweight, smoking, family history and vascular disease) and socio-economic indicators of the patient’s residential district (number of residents, income pre-tax, residents per general practitioner, residents per internal specialist, compulsory education only, academic degree and rate of unemployment). Cox regressions were performed to evaluate the impact of socio-economic environment and conventional risk factors on survival.

Results

Most of the conventional risk factors show a significant difference between deceased and surviving patients. The study revealed significant differences across districts in relation to the socio-economic background of STEMI patients. Surprisingly, medical outcomes, as measured by the survival of patients, are significantly related to a patient’s district of residence (p-Value = 0.028) but not in a systematic way as far as the socio-economic environment of these districts is concerned.

Conclusions

The study provides intuitive evidence for a hitherto understudied Central European context on the link between socio-economic environment and conventional risk factors at population level and the link between conventional risk factors and survival both at the population at the individual level. While this is in line with previous evidence and suggestive of the incorporation of measures of socio-economic status (SES) into policy & guidelines toward the management of CAD, more data on the SES – STEMI nexus are needed at individual level.

Genetic risk analysis of coronary artery disease in Pakistani subjects using a genetic risk score of 21 variants

BACKGROUND AND AIMS

Conventional coronary artery disease (CAD) risk factors like age, gender, blood lipids, hypertension and smoking have been the basis of CAD risk prediction algorithms, but provide only modest discrimination. Genetic risk score (GRS) may provide improved discrimination over and above conventional risk factors. Here we analyzed the genetic risk of CAD in subjects from Pakistan, using a GRS of 21 variants in 18 genes and examined whether the GRS is associated with blood lipid levels.

METHODS

625 (405 cases and 220 controls) subjects were genotyped for variants, NOS3 rs1799983, SMAD3 rs17228212, APOB rs1042031, LPA rs3798220, LPA rs10455872, SORT1 rs646776, APOE rs429358, GLUL rs10911021, FTO rs9939609, MIA3 rs17465637, CDKN2Ars10757274, DAB2IP rs7025486, CXCL12 rs1746048, ACE rs4341, APOA5 rs662799, CETP rs708272, MRAS rs9818870, LPL rs328, LPL rs1801177, PCSK9 rs11591147 and APOE rs7412 by TaqMan and KASPar allele discrimination techniques.

RESULTS

Individually, the single SNPs were not associated with CAD except APOB rs1042031 and FTO rs993969 (p = 0.01 and 0.009 respectively). However, the combined GRS of 21 SNPs was significantly higher in cases than controls (19.37 ± 2.56 vs. 18.47 ± 2.45, p = 2.9 × 10^-5), and compared to the bottom quintile, CAD risk in the top quintile of the GRS was 2.96 (95% CI 1.71-5.13). Atherogenic blood lipids showed significant positive association with GRS.

CONCLUSIONS

The GRS was quantitatively associated with CAD risk and showed association with blood lipid levels, suggesting that the mechanism of these variants is likely to be, in part at least, through creating an atherogenic lipid profile in subjects carrying high numbers of risk alleles.

Risk prediction-Homocysteine in Coronary Heart Disease

Majority of patients who experience a Coronary Heart disease event have one or more of the conventional risk factors for atherosclerosis and so do many people who have not yet experienced such an event. Thus predictive models based on conventional risk factors have lower than the desired accuracy, providing a stimulus to search for new factors to predict accurately the risk of CHD. In this regard newer risk factors like homocysteine, Lp(a), insulin resistances are the important ones and are called as 'novel risk factors'. The study was undertaken to find the prediction of CHD risk by homocysteine in comparison with other conventional risk factors. The data obtained suggests a very high sensitivity, specificity and accuracy with above 90% positive prediction value for homocysteine in CHD patients when compared to commonest conventional risk factors.