To evaluate the performance of machine learning (ML) algorithms and to compare them with logistic regression for the prediction of risk of cardiovascular diseases (CVDs), chronic kidney disease ...(CKD), diabetes (DM), and hypertension (HTN) and in a prospective cohort study using simple clinical predictors.
We conducted analyses in a population-based cohort study in Asian adults (n = 6,762). Five different ML models were considered—single-hidden-layer neural network, support vector machine, random forest, gradient boosting machine, and k-nearest neighbor—and were compared with standard logistic regression.
The incidences at 6 years of CVD, CKD, DM, and HTN cases were 4.0%, 7.0%, 9.2%, and 34.6%, respectively. Logistic regression reached the highest area under the receiver operating characteristic curve for CKD (0.905 0.88, 0.93) and DM (0.768 0.73, 0.81) predictions. For CVD and HTN, the best models were neural network (0.753 0.70, 0.81) and support vector machine (0.780 0.747, 0.812), respectively. However, the differences with logistic regression were small (less than 1%) and nonsignificant. Logistic regression, gradient boosting machine, and neural network were systematically ranked among the best models.
Logistic regression yields as good performance as ML models to predict the risk of major chronic diseases with low incidence and simple clinical predictors.
•Low-dimensional settings include low number of events and predictors.•In such settings, logistic regression yields as good performance as ML models.•ML techniques may not be warranted in such cases.
The pore size of porous scaffold plays a critical role in bone regeneration, but its mechanism and optimal value remain unclear. This study investigated the effect of pore size on hydromechanical ...properties of porous scaffold and its correlation with cellular response and bone regeneration. Porous titanium scaffolds with similar porosity and different pore sizes (400, 650, 850, and 1100 μm) were fabricated by selective laser melting. Their hydromechanical properties were derived by computational fluid dynamics analysis. The MC3T3 cells were dynamic seeded and cultured on the scaffolds to evaluate the cellular response. The rabbit distal femoral condyle implantation models were used to assess the bone ingrowth. Results indicated that the permeability, flow velocity, and the inflow of fluid linearly increased with the pore size. The wall shear stress evaluated from 400 to 650 μm and then dropped. These changes induced various performances in cell penetration, adhesion, proliferation, and differentiation, and finally induced best bone ingrowth in scaffold with pore size of 650 μm. This study provided a new understanding of the effect of pore size on bone regeneration of porous scaffold from the perspective of hydromechanics and indicated the potential of combining computational simulation and laboratory experiments in future studies.
Display omitted
•Increasing the pore size linearly increased the permeability, flow velocity, and inflow of fluid.•The shear stress first increased and then decreased with the increase in pore size.•The pore size significantly affected the cell penetration, adhesion, proliferation, differentiation, and bone ingrowth.•The hydromechanical properties closely correlated with the cellular response and bone regeneration
To examine the relationship between neurocognitive function (NCF) and quality of life (QOL) in patients with brain metastases after whole-brain radiotherapy.
A total of 208 patients from the ...whole-brain radiotherapy arm of a Phase III trial (PCI-P120-9801), who underwent regular NCF and QOL (ADL activities of daily living and FACT-Br Functional Assessment of Cancer Therapy-Brain-specific) testing, were analyzed. Spearman's rank correlation was calculated between NCF and QOL, using each patient's own data, at each time point. To test the hypothesis that NCF declines before QOL changes, the predictive effect of NCF from previous visits on QOL was studied with a linear mixed-effects model. Neurocognitive function or QOL deterioration was defined relative to each patient's own baseline. Lead or lag time, defined as NCF deterioration before or after the date of QOL decline, respectively, was computed.
At baseline, all NCF tests showed statistically significant correlations with ADL, which became stronger at 4 months. A similar observation was made with FACT-Br. Neurocognitive function scores from previous visits predicted ADL (p < 0.05 for seven of eight tests) or FACT-Br. Scores on all eight NCF tests deteriorated before ADL decline (net lead time 9-153 days); and scores on six of eight NCF tests deteriorated before FACT-Br (net lead time 9-82 days).
Neurocognitive function and QOL are correlated. Neurocognitive function scores from previous visits are predictive of QOL. Neurocognitive function deterioration precedes QOL decline. The sequential association between NCF and QOL decline suggests that delaying NCF deterioration is a worthwhile treatment goal in brain metastases patients.
The Phanerozoic granites in northeast China bear key information for studying the tectonic evolution and crustal growth or reworking in the Central Asian Orogenic Belt (CAOB). The Daqing granitic ...batholith widely outcrops as a high-level intrusion in the Xing'an-Mongolia Orogenic Belt, southeastern CAOB. Three types of enclaves in granites have been identified: (1) mafic magmatic enclaves (MMEs), (2) volcanic xenoliths, and (3) biotite-rich enclaves. The batholith is mainly composed of peraluminous biotite granite and granodiorite with SiO2=63.95-69.48 wt.%, A/CNK=1.15-1.27, and 2.54 to 4.30 wt.% of normative corundum. They exhibit remarkable enrichment in large ion lithophile elements (LILEs; K, Rb, Th, and Pb) and depletion in high field strength elements (HFSEs; Nb, Ta, and Ti), P, Eu, and Sr, as well as relatively enriched Sr-Nd isotopes (87Sr/86Sri=0.70530-0.70576, εNd(t) = -0.1-+0.2). Zircon U-Pb dating suggests that this batholith was emplaced in the Early Permian (ca. 283-282 Ma), consistent with a period of intensive magmatic activities in northern Inner Mongolia. The Nb/Ta ratios of MMEs (17.6-20.1) are higher than those of the host granites (11.4-12.5), together with the reaction rims where biotite crystals cluster around the amphibole cores, suggesting magma mixing between mantle- and crust-derived melts. Zircons from a biotite-rich enclave define a protolith age of ca. 320 Ma and an anatectic age at ca. 281 Ma. Whole-rock Sr-Nd isotopic modeling and zircon Hf isotopes reveal that the batholith was mainly produced by remelting of newly accreted continental crust with minor addition of mantle-derived materials. The geochemical compositions imply that their precursor magmas originated from a relatively high crustal level (<5 kbar) with crystallization temperatures ranging from 800 to 850°C. We suggest that the Daqing peraluminous granitoids were derived from partial melting of newly accreted crustal materials at a relatively shallow crustal depth, associated with a ridge subduction-related heat source. Such mantle-derived magmas through a slab tear window resulting from ridge subduction provide not only the heat for the widespread crustal remelting and therefore maturity but also juvenile materials for crustal growth.
Display omitted
•A new snail shell structure is used as the prototype of the optimal design.•The multi-objectives are changed into one objective through the weight coefficient.•The designed labyrinth ...seal has good load-bearing and heat insulation performance.
Labyrinth seals are widely used in turbomachinery devices to prevent the leakage of fluid medium in seal structures. Under high parametrical environments, the heat generated by frictional resistance of the turbulence between the stator and rotor walls not only increases the temperature of the fluid and the solid, but also reduces the operation efficiency of the device. Therefore, it is necessary to implement the fluid–solid-thermal multi-field analysis of labyrinth seals and propose a comprehensive optimization design of the load-bearing and heat insulation. In the present study, inspired by a bionic snail shell, a sandwich structure with load-bearing material on both upper and lower surfaces and the heat insulation material in the middle is obtained through the genetic algorithm optimization. Then, the luid-solid-thermal multi-field analysis of the labyrinth seal with the sandwich structure is carried out. It is verified that the layer arrangement and layer thickness ratio of the bionic structure ensured the optimal performance of load-bearing and heat insulation. After changing the layer arrangement or layer thickness ratio, the comprehensive performance of load-bearing and heat insulation is obviously weakened. The result shows that the bionic multilayer sandwich structure can be applied to the labyrinth seal to achieve optimal comprehensive performance of load-bearing and heat insulation.
A mild, versatile and efficient method for the silver(I)-catalyzed oxidative decarboxylative gem-difluoromethylenation has been developed. The radical cascade reaction comprises the addition of an ...oxidatively generated difluoromethylene radical to the isonitrile functionality and subsequent homolytic aromatic substitution. It provides a novel and efficient access to the C-CF2 bond formation.
Background Clinical practice guidelines recommend using equations for estimating glomerular filtration rate (GFR) in chronic kidney disease (CKD) management and research. The MDRD (Modification of ...Diet in Renal Disease) Study and CKD-EPI (CKD Epidemiology Collaboration) equations originally were derived from a North American population and had an ethnic coefficient adjustment for African Americans. A Chinese coefficient for the MDRD Study equation subsequently was determined, but this has not been externally validated. We compared the accuracy of the equations, evaluated the ethnic coefficients, and assessed the equations for disease staging in a multiethnic Asian population with CKD. Study Design A diagnostic test study comparing the Asian coefficient (and subgroups)–modified MDRD Study and CKD-EPI equations and a cross-sectional study assessing disease staging. Setting & Participants 232 outpatients (52% men; 40.5% Chinese, 32% Malay, and 27.5% Indian/other) with stable CKD. Index Test Asian and ethnicity-based modifications of the MDRD Study and CKD-EPI equations. Reference Test Measured GFR using 3-sample plasma clearance of technetium-99m diethylenetriaminepentaacetic acid (99m Tc-DTPA), calculated using the slope-intercept method, with body surface area normalization (du Bois) and Brochner-Mortensen correction. Results Overall, the CKD-EPI equation is more accurate than the MDRD Study equation throughout the GFR range, with improved bias (median difference of estimated GFR − measured GFR) and root mean square error ( P <0.001). CKD-EPI versus MDRD Study equation: bias, 1.1 ± 13.8 vs −1.0 ± 15.2 mL/min/1.73 m2 ; precision, 12.1 vs 12.2 mL/min/1.73 m2 . Ethnic coefficients did not improve estimates of GFR significantly. The correctness of staging was improved using the CKD-EPI equation. Limitations All participants had CKD, but few were of European descent. The reference GFR technique was different from the original studies. Conclusions The CKD-EPI is more accurate than the MDRD Study equation, particularly at higher GFRs. Therefore, we recommend adopting the CKD-EPI equation without ethnic adjustment for estimating GFR in multiethnic Asian patients with CKD.
We propose a multithreshold change plane regression model which naturally partitions the observed subjects into subgroups with different covariate effects. The underlying grouping variable is a ...linear function of observed covariates and thus multiple thresholds produce change planes in the covariate space. We contribute a novel two‐stage estimation approach to determine the number of subgroups, the location of thresholds, and all other regression parameters. In the first stage we adopt a group selection principle to consistently identify the number of subgroups, while in the second stage change point locations and model parameter estimates are refined by a penalized induced smoothing technique. Our procedure allows sparse solutions for relatively moderate‐ or high‐dimensional covariates. We further establish the asymptotic properties of our proposed estimators under appropriate technical conditions. We evaluate the performance of the proposed methods by simulation studies and provide illustrations using two medical data examples. Our proposal for subgroup identification may lead to an immediate application in personalized medicine.
Thresholding variable plays a crucial role in subgroup identification for personalized medicine. Most existing partitioning methods split the sample based on one predictor variable. In this paper, we ...consider setting the splitting rule from a combination of multivariate predictors, such as the latent factors, principle components, and weighted sum of predictors. Such a subgrouping method may lead to more meaningful partitioning of the population than using a single variable. In addition, our method is based on a change point regression model and thus yields straight forward model‐based prediction results. After choosing a particular thresholding variable form, we apply a two‐stage multiple change point detection method to determine the subgroups and estimate the regression parameters. We show that our approach can produce two or more subgroups from the multiple change points and identify the true grouping with high probability. In addition, our estimation results enjoy oracle properties. We design a simulation study to compare performances of our proposed and existing methods and apply them to analyze data sets from a Scleroderma trial and a breast cancer study.