Conventional Multivariate Regression Research Articles

A nationwide longitudinal investigation on the role of prenatal exposure to infectious diseases on the onset of chronic conditions in children and adolescents in Brazil.

In utero exposure to infections might set the stage for a chain of events leading to a wide spectrum of long-term health outcomes observed in children and adolescents. This proposal aims to investigate whether syphilis, zika, dengue and chikungunya during pregnancy can increase the risk of the offspring developing a non-infectious chronic condition during childhood and adolescence. 1) Estimate the risk of non-infectious chronic conditions associated to syphilis, zika, dengue and chikungunya during pregnancy and when appropriate, explore if the risk varies by timing during pregnancy when the infection is acquired (first, second or third trimester) and severity (such as severe or mild dengue); 2) Investigate whether in uterus exposure to maternal infection affects the growth pattern of children and adolescents; 3) Examine the extent to which the relationship between maternal infection and non-infectious chronic outcomes are mediated by intrauterine growth restriction and preterm birth. We will compare health outcomes and growth trajectories of children and adolescents born to mothers with and without specific infections during pregnancy using conventional multivariable regression in the whole study population, in a within sibship design, using the subgroup of offspring with at least one sibling who is not exposed to the infection, and negative control outcome. Then we will decompose the direct and mediated effects (by preterm birth and small for gestational age) of maternal infection on chronic disorders. The results from this study will advance our understanding of the relationship between infections during pregnancy and chronic disorders, with widespread implications enabling targeting of critical points along the path from in utero exposure to outcomes to avoid or mitigate illness and disability over the life course.

Comparing AI/ML approaches and classical regression for predictive modeling using large population health databases: Applications to COVID-19 case prediction

BackgroundResearch comparing artificial intelligence and machine learning (AI/ML) methods with classical statistical methods applied to large population health databases is limited. ObjectivesThis retrospective cohort study aimed to compare the predictive performance of AI/ML algorithms against conventional multivariate logistic regression models using linked health administrative data. MethodsUsing Ontario's population health databases, we created a cohort of residents of the city of Ottawa, Ontario, who underwent a PCR test for COVID-19 between March 10, 2020, and May 13, 2021. Using demographic, socio-economic and health data (including COVID-19 PCR test results and available, symptom data), we developed predictive models for the purpose of COVID-19 case identification using the following approaches: classical multivariate logistic regression (LR); deep neural network (DNN); random forest (RF); and gradient boosting trees (GBT). Model performance comparisons were made using the area under the curve (AUC) swarm plot for 10-fold cross-validation. ResultsThe cohort consisted of n = 351,248 Ottawa residents tested for COVID-19 during the study period. Among whom, a total of n = 883,879 unique COVID-19 tests were performed (2.6 % positive test results). Inclusion of COVID-19 symptoms data in the analysis improved model performance and variable predictive value across all tested models (p < 0.0001), with the 10-fold cross-validation AUC increasing to near or over 0.7 in all models when symptoms data were included. In various pairwise comparisons, the GBT method had the highest predictive ability (AUC = 0.796 ± 0.017), significantly outperforming multivariate logistic regression and the other AI/ML approaches. ConclusionsConventional multivariate regression-based models are better than some and worse than other machine learning algorithms to provide good predictive accuracy in a moderate dataset with a reasonable number of features. However, whenever possible, the AI/ML GBT approach should be considered.

A nationwide longitudinal investigation on the role of prenatal exposure to infectious diseases on the onset of chronic conditions in children and adolescents in Brazil

Background In utero exposure to infections might set the stage for a chain of events leading to a wide spectrum of long-term health outcomes observed in children and adolescents. This proposal aims to investigate whether syphilis, zika, dengue and chikungunya during pregnancy can increase the risk of the offspring developing a non-infectious chronic condition during childhood and adolescence. Objectives 1) Estimate the risk of non-infectious chronic conditions associated to syphilis, zika, dengue and chikungunya during pregnancy and explore if the risk varies by timing during pregnancy when the infection is acquired (first, second or third trimester) and severity (such as severe or mild dengue); 2) Investigate whether in uterus exposure to maternal infection affects the growth pattern of children and adolescents; 3) Examine the extent to which the relationship between maternal infection and non-infectious chronic outcomes are mediated by intrauterine growth restriction and preterm birth. Methods We will compare health outcomes and growth trajectories of children and adolescents born to mothers with and without specific infections during pregnancy using conventional multivariable regression in the whole study population, in a within sibship design, using the subgroup of offspring with at least one sibling who is not exposed to the infection, and negative control outcome. Then we will decompose the direct and mediated effects (by preterm birth and small for gestational age) of maternal infection on chronic disorders. Results and Conclusions The results from this study will advance our understanding of the relationship between infections during pregnancy and chronic disorders, with widespread implications enabling targeting of critical points along the path from in utero exposure to outcomes to avoid or mitigate illness and disability over the life course.

Proof‐of‐concept of a data‐driven approach to estimate the associations of comorbid mental and physical disorders with global health‐related disability

AbstractObjectiveThe standard method of generating disorder‐specific disability scores has lay raters make rankings between pairs of disorders based on brief disorder vignettes. This method introduces bias due to differential rater knowledge of disorders and inability to disentangle the disability due to disorders from the disability due to comorbidities.MethodsWe propose an alternative, data‐driven, method of generating disorder‐specific disability scores that assesses disorders in a sample of individuals either from population medical registry data or population survey self‐reports and uses Generalized Random Forests (GRF) to predict global (rather than disorder‐specific) disability assessed by clinician ratings or by survey respondent self‐reports. This method also provides a principled basis for studying patterns and predictors of heterogeneity in disorder‐specific disability. We illustrate this method by analyzing data for 16 disorders assessed in the World Mental Health Surveys (n = 53,645).ResultsAdjustments for comorbidity decreased estimates of disorder‐specific disability substantially. Estimates were generally somewhat higher with GRF than conventional multivariable regression models. Heterogeneity was nonsignificant.ConclusionsThe results show clearly that the proposed approach is practical, and that adjustment is needed for comorbidities to obtain accurate estimates of disorder‐specific disability. Expansion to a wider range of disorders would likely find more evidence for heterogeneity.

Unsupervised Clustering Analysis of Regimen and HLA Characteristics Identifies Subgroups with Improved Relapse-Free Survival in Children with ALL and AML Undergoing Umbilical Cord Blood Transplantation

Introduction. Umbilical cord blood transplantation (UCBT) is an efficacious therapeutic intervention for pediatric acute lymphoblastic leukemia (ALL) and acute myeloid leukemia (AML). However, contention persists regarding how best to achieve long-term relapse-free survival (RFS). Differences in clinical practice can be attributed to an incompletely understood interplay between graft-versus-leukemia (GvL) effect and graft-versus-host disease (GvHD), as well as an incomplete understanding of how UCBTs differ from those of other stem cell sources. Human leukocyte antigen (HLA) matching is a critical determinant of post-transplant outcome, but factors such as GvHD prophylaxis and total body irradiation (TBI) usage also play a key role. Objective. To use unsupervised clustering algorithms to examine the interplay between HLA matching, immunosuppressive regimens, and conditioning factors with greater power than that conferred by conventional multivariate regression analyses. Methodology. Data pertaining to UCBTs performed for pediatric ALL (n = 908) and AML (n = 417) between 2004 and 2022 was obtained from the Eurocord/EBMT registry. ALL and AML UCBTs were separately subjected to a partitioning around medoids clustering algorithm, which identified k = 15 ALL clusters and k = 11 AML clusters. These clusters were used to form hypotheses for post-hoc testing, which were tested using Kaplan-Meier and competing risks analyses. Results. Clusters of both diseases were frequently divided by usage of ATG, which was administered in 77% of 8/8 ABCDR matched UCBTs. ALL and AML patients who did not receive ATG experienced greater RFS than recipients of ATG, and this appeared to be driven by elimination of relapse mortality in the late (≥5 year) post-transplant setting (Figs 1-2). ALL clusters were frequently divided by use of total body irradiation (TBI), and those featuring TBI regimens conferred RFS comparable or superior to non-TBI regimens; this appeared to be driven by a reduction in relapse mortality within the first three years post-UCBT (P = 0.007). In AML but not ALL, both HLA-C and KIR match levels appeared to be relevant (P = 0.04), despite typing only having been performed for 70% and 24% of AML UCBTs, respectively. Patients receiving a 7/8 ABCDR-matched graft with a single HLA-C mismatch experienced significantly poorer RFS than 8/8 matched UCBTs (P = 0.04), but a concomitant single co-mismatch at HLA-A or HLA-B appeared to restore RFS to levels comparable to 8/8 matched grafts. Conclusions. ATG is over-used in ALL and AML patient-graft pairings where risk of non-relapse mortality (NRM) is low, and this may be creating an avoidable risk of relapse. The conflicting results in published literature may be explained by follow-up time, as the effect does not become apparent until 5 years post-UCBT and many studies do not follow patients beyond this. Incorporation of TBI into ALL myeloablative regimens prevents relapse mortality in the early to mid-post UCBT setting, and this effect was independent of HLA matching. These findings align with the known genetic and biological determinants of these diseases and redouble the importance of typing at HLA-C in AML; administration of TBI in ALL; and a measured approach to the usage of ATG.

RADT-41. PREDICTION OF TREATMENT OUTCOME IN GAMMA KNIFE RADIOSURGERY OF BENIGN MENINGIOMAS

Abstract Objective This study aimed to assess the predictability of machine learning (ML) ensemble models for the success of Gamma Knife radiosurgery (GKRS) on non-skull base benign meningiomas. METHODS A total of 331 non-skull base benign meningiomas treated primarily with GKRS between 1997 and 2015 were retrospectively analyzed. When a tumor remained not larger than 125% of the initial volume at the last imaging follow-up and no significant change in the high signal intensity area around the tumor, it was classified as 'treatment success (TS).' Univariate correlation and multivariate logistic regression analyses were performed to identify statistically significant prognostic factors. Four ensemble algorithms were trained to predict the outcome of GKRS. The importance of features related to TS was analyzed. RESULTS The median MR imaging follow-up period was 8.6 years. The tumor control and signal change rates were 91.5% and 16.3%, respectively, and the TS rate was 77.6%. The conventional multivariate logistic regression model showed 83.7% accuracy and 37.8% specificity. The ML models had accuracies between 78.0% and 82.2% and specificities between 30.3% and 80.7%, and they were similar to those of conventional logistic regression analyses. The smaller brain volume absorbed more than 14 Gy, and the absence of signal change before GKRS was a favorable prognostic factor for TS in all models. The bagging algorithm with selected features showed the best specificity of 80.7% and the most significant AUC of 0.723. CONCLUSION A bagging ensemble algorithm with selected features had better specificity with similar accuracy than conventional multivariate logistic regression in predicting GKRS outcomes for non-skull base meningiomas.

Modeling stream baseflow nitrate concentration in an agricultural watershed using neural network and bootstrap method

The evaluation of water resource vulnerability to nonpoint source pollution in the presence of uncertainty remains a crucial concern. To enhance the accuracy of the assessment and pinpoint the key factors that affect watershed water quality, the integration of an artificial neural network (ANN) into the evaluation process is imperative. The research involved the collection of streams baseflow samples from thirty-eight sites between 1994 and 1999 in the Tomorrow-Waupaca River Watershed in central Wisconsin, USA, with a focus on examining the relationship between nitrate concentrations and a range of environmental and land use variables extracted from the watershed GIS database. This study utilized ANN methodology, combined with a bootstrap technique that employed a random resampling approach from a single input dataset, to simulate monthly stream baseflow nitrate concentrations. The effectiveness and predictive ability of the ANN model were assessed by comparing it to conventional multivariate regression methods. Through the use of ANN, more precise outcomes can be achieved while taking into account the uncertainty associated with the analysis. The findings demonstrated that the ANN outperformed both the multivariate linear regression and nonlinear quadratic response surface models in explaining the variance of stream nitrate and in external prediction consistency. This study also highlighted several key variables, such as the areal percentage of agricultural land and grassland, stream order, and the average slope of the groundwater flow path, that significantly impacted the stream baseflow nitrate concentrations in this watershed that was dominated by dairy farming and groundwater. Of these variables, the percentage of agricultural land emerged as the most significant factor.

Regional flood frequency analysis using data-driven models (M5, random forest, and ANFIS) and a multivariate regression method in ungauged catchments

Flooding is recognized worldwide joined of the most expensive natural hazards. To adopt proper structural and nonstructural measurements for controlling and mitigating the rising flood risk, the availability of streamflow values along a river is essential. This raises concerns in the hydrological assessment of poorly gauged or ungauged catchments. In this regard, several flood frequency analysis approaches have been conducted in the literature including index flow method (IFM), square grids method (SGM), hybrid method (HM), as well as the conventional multivariate regression method (MRM). While these approaches are often based on assumptions that simplify the complex nature of the hydrological system, they might not be able to address uncertainties associated with the complexity of the system. One of the powerful tools to deal with this issue is data-driven model that can be easily adopted in complex systems. The objective of this research is to utilize three different data-driven models: random forest (RF), adaptive neuro-fuzzy inference system (ANFIS), and M5 decision tree algorithm to predict peak flow associated with various return periods in ungauged catchments. Results from each data-driven model were assessed and compared with the conventional multivariate regression method. Results revealed all the three data-driven models performed better than the multivariate regression method. Among them, the RF model not only demonstrated the superior performance of peak flow prediction compared to the other algorithms but also provided insight into the complexity of the system through delivering a mathematical formulation.

High-Resolution Precipitation Modeling in Complex Terrains Using Hybrid Interpolation Techniques: Incorporating Physiographic and MODIS Cloud Cover Influences

The inclusion of physiographic and atmospheric influences is critical for spatial modeling of orographic precipitation in complex terrains. However, attempts to incorporate cloud cover frequency (CCF) data when interpolating precipitation are limited. CCF considers the rain shadow effect during interpolation to avoid an overly strong relationship between elevation and precipitation in areas at equivalent altitudes across rain shadows. Conventional multivariate regression or geostatistical methods assume the precipitation–explanatory variable relationship to be steady, even though this relation is often non-stationarity in complex terrains. This study proposed a novel spatial mapping approach for precipitation that combines regression-kriging (RK) to leverage its advantages over conventional multivariate regression and the spatial autocorrelation structure of residuals via kriging. The proposed hybrid model, RK (GT + CCF), utilized CCF and other physiographic factors to enhance the accuracy of precipitation interpolation. The implementation of this approach was examined in a mountainous region of southern Syria using in situ monthly precipitation data from 57 rain gauges. The RK model’s performance was compared with conventional multivariate regression models (CMRMs) that used geographical and topographical (GT) factors and CCF as predictors. The results indicated that the RK model outperformed the CMRMs with a root mean square error of &lt;8 mm, a mean absolute percentage error range of 5–15%, and an R2 range of 0.75–0.96. The findings of this study showed that the incorporation of MODIS–CCF with physiographic variables as covariates significantly improved the interpolation accuracy by 5–20%, with the largest improvement in modeling precipitation in March.

Mendelian randomization on the association of obesity with vitamin D: Guangzhou Biobank Cohort Study.

Mendelian randomization (MR) analyses from the West provide evidence that obesity causes lower 25-hydroxyvitamin D [25(OH)D]. As Asian populations are prone to metabolic disorders at a lower body mass index (BMI), whether the association remains in Asian is unclear. We studied whether obesity causes vitamin D deficiency using MR analysis in Chinese. We used data from the Guangzhou Biobank Cohort Study. A genetic score including seven BMI-related single-nucleotide polymorphisms (n = 15,249) was used as the instrumental variable (IV) for BMI. Two-stage least square regression and conventional multivariable linear regression in 2,036 participants with vitamin D data were used to analyze association of BMI with vitamin D. Proportion of variation explained by the genetic score was 0.7% and the first stage F-statistic for MR analysis was 103. MR analyses showed that each 1 kg/m2 higher BMI was associated with lower 25(OH)D by -2.35 (95% confidence interval (CI) -4.68 to -0.02) nmol/L. In conventional multivariable linear regression, higher BMI was also associated with lower 25(OH)D (β = -0.26 nmol/L per 1 kg/m2 increase in BMI, 95% CI -0.46 to -0.06). Sensitivity analyses using two-sample IV analysis and leave-one-out method showed similar results. We have first shown by MR and conventional multivariable linear regression that higher BMI causes vitamin D deficiency in Chinese. Our findings highlight the importance of weight control and suggest that vitamin D supplementation may be needed in individuals with overweight or obesity.

Effect of preoperative statins on respiratory complications after coronary artery bypass grafting.

Limited data exist on the effect of preoperative statin therapy on postoperative respiratory complications. Machine learning algorithms (MLA) can process large, heterogenous data, and have immensely improved the ability for risk prediction. In this study, we sought to examine the role of preoperative statins on respiratory complications in patients undergoing coronary artery bypass grafting(CABG) using MLA. The study population contained the data of patients who underwent CABG between the years 2015 and 2019 (n = 5638). Three hundred and thirty-seven independent variables were recorded and the data was randomly split with stratified sampling into training and testing data with 20% of the data (1113 records) reserved for model testing. Various models including linear models, Random forest, SVM, and XGboostwere trained to predict the incidence of postoperative respiratory complications. Forty-seven important features were found to impact model prediction (p ≤ .05) using the global surrogate model method. A conventional multivariable linear regression model was then used to identify predictors of respiratory complications. One thousand three hundred sixty-two(24.5%) patients developed a respiratory complication in our series. The respiratory complication was seen in 561 (29.7%) of the patients who were not on statin compared to only 801 (21.8%) who were on a statin, p < .0001.The area under the curvefor receiver operating characteristiccurve using statins and respiratory complications was 0.706. Statins showed positive feature importance in all the MLA models. MLA showed that statins impacted the prediction of respiratory complications in all the models studied. The study confirmed that preoperative statins reduced the risk of respiratory complications by 21%.

Unraveling the effects of hydrological connectivity and landscape characteristics on reservoir water quality

Dam construction and reservoir operation altered the landscape and hydrological process of reservoir bays, affecting reservoir water quality. However, many landscape and hydrological connectivity metrics are highly correlated and may introduce redundancies and misleading results when use conventional multivariate regression techniques. Knowledge concerning the pure effects of landscape and hydrological connectivity metrics are crucial for understanding nonpoint pollution processes and guiding water quality protection strategies in reservoirs. Based on water quality monitoring data for six years (2015–2020) from 66 reservoir bays of the Danjiangkou Reservoir in China during both flood discharge periods (FDPs) and water storage periods (WSPs), machine learning approaches (boosted regression trees and random forest) were conducted to decipher the effects of hydrological connectivity and landscape characteristics on water quality. The results showed that landscape composition, landscape configuration, and topography had the combined importance of 46.69%, 31.48%, and 10.14% on the overall water quality changes during the FDPs, respectively. However, landscape configuration was the largest importance factor controlling overall water quality for the WSPs with the combined importance of 38.57%. For the FDPs, the top two importance variables of overall water quality variation were the proportions of shrub and agricultural land in the reservoir bay. For the WSPs, the top two importance variables were flow length and index of connectivity. For specific water quality parameters, the highest importance factor controlling the variation in total nitrogen, total phosphorus, ammonia nitrogen, nitrate and chlorophyll a were landscape configuration, while the landscape composition had the highest importance on the variation in permanganate index, suggesting landscape characteristics affected water quality with specific responses to distinct water quality parameters. These findings emphasize the distinct roles of landscape and hydrological characteristics on water quality and provide important information for the efficient formulation of water quality protection strategies in reservoirs.

Detection of Water pH Using Visible Near-Infrared Spectroscopy and One-Dimensional Convolutional Neural Network

pH is an important parameter for water quality detection. This study proposed a novel calibration regression strategy based on a one-dimensional convolutional neural network (1D-CNN) for water pH detection using visible near-infrared (Vis-NIR) spectroscopy. Two groups of Vis-NIR spectral analysis experiments of water pH detection were employed to evaluate the performance of 1D-CNN. Two conventional multivariate regression calibration methods, including partial least squares (PLS) and least squares support vector machine (LS-SVM), were introduced for comparative analysis with 1D-CNN. The successive projections algorithm (SPA) was adopted to select the feature variables. In addition, the learning mechanism of 1D-CNN was interpreted through visual feature maps by convolutional layers. The results showed that the 1D-CNN models obtained the highest prediction accuracy based on full spectra for the two experiments. For the spectrophotometer experiment, the root mean square error of prediction (RMSEP) was 0.7925, and the determination coefficient of prediction () was 0.8515. For the grating spectrograph experiment, the RMSEP was 0.5128 and the was 0.9273. The convolutional layers could automatically preprocess the spectra and effectively extract the spectra features. Compared with the traditional regression methods, 1D-CNN does not need complex spectra pretreatment and variable selection. Therefore, 1D-CNN is a promising regression approach, with higher prediction accuracy and better modeling convenience for rapid water pH detection using Vis-NIR spectroscopy.

Deep learning aided evaluation for electromechanical properties of complexly structured polymer nanocomposites

In the present work, two types of deep neural networks (DNNs) were employed to establish the structure–property relationship of polymer nanocomposites. The trained DNNs based on multiscale analysis results can not only overcome the limitations of the conventional clustering density-based model and multivariate regression models but also exhibit superior performance in evaluating the electromechanical properties of polypropylene matrix composites, wherein spherical SiC nanoparticles were randomly distributed and dispersed. A simple graph convolution network showed better capability than a complex artificial neural network, despite fewer features considered; this implies that the graph convolution network is more appropriate and user-friendly for evaluating the effect of nanoparticle distribution and agglomeration. In addition, the trained graph convolution network can effectively provide mechanical and electrical properties corresponding to large representative volume element (RVE) without a loss of accuracy. The present study demonstrates that deep learning techniques can be put to practical use for the design of next-generation polymer nanocomposite materials.

Effect of Intraoperative Neuromonitoring on the Risks of Recurrent Laryngeal Nerve Injury During Thyroidectomy: A Doubly Robust Approach.

To evaluate the relationship between the use intraoperative neuromonitoring (IONM) during thyroidectomy and the risk of recurrent laryngeal nerve (RLN) injury. The role of IONM in reducing RLN injury during thyroidectomy remains controversial. Several studies on this topic apply conventional multivariable regression to adjust for confounding. However, estimates from this method may be biased due to model misspecification, especially with a rare outcome such as RLN injury. We used a pooled dataset created by linking the 2016-2019 National Surgical Quality Improvement Project General Participant User File with the corresponding Targeted-Thyroidectomy file. The primary outcome was RLN injury rates, and the secondary outcomes were operating time and postoperative length of stay. A doubly robust (DR) estimator, in the form of an inverse-probability-weighted regression adjustment model, was used to estimate the effect of the use of IONM on the risk of RLN injury. Sensitivity analyses was performed. Twenty-four thousand three hundred seventy patients were evaluated, out of which 15,836 (70%) patients had IONM during thyroidectomy, and RLN injury occurred in 1498 (6.2%) cases. Rates of RLN injury increase with increasing age and BMI and are higher in patients with a cancer diagnosis, previous neck operation, total thyroidectomy, and node dissection. Doubly robust model suggests that the use of IONM was associated with a significant reduction in overall rate of RLN injury [risk ratio 0.77, confidence interval (CI), 0.68-0.87, P <0.001], and postoperative length of stay [-2.5 hours (CI, -4.18 to -0.81h), P =0.004]. However, IONM use was associated with an increase in operating time [15.41 minutes (CI, 13.29-17.54 minutes), P <0.0001]. Sensitivity analyses revealed that our estimates are largely robust to confounding. In a balanced cohort of patients undergoing thyroidectomy from multiple sites and surgeons participating in National Surgical Quality Improvement Project, the use of IONM during thyroidectomy was associated with reduction in RLN injury.

Propensity Score and Instrumental Variable Techniques in Observational Transplantation Studies: An Overview and Worked Example Relating to Pre-Transplant Cardiac Screening.

Inferring causality from observational studies is difficult due to inherent differences in patient characteristics between treated and untreated groups. The randomised controlled trial is the gold standard study design as the random allocation of individuals to treatment and control arms should result in an equal distribution of known and unknown prognostic factors at baseline. However, it is not always ethically or practically possible to perform such a study in the field of transplantation. Propensity score and instrumental variable techniques have theoretical advantages over conventional multivariable regression methods and are increasingly being used within observational studies to reduce the risk of confounding bias. An understanding of these techniques is required to critically appraise the literature. We provide an overview of propensity score and instrumental variable techniques for transplant clinicians, describing their principles, assumptions, strengths, and weaknesses. We discuss the different patient populations included in analyses and how to interpret results. We illustrate these points using data from the Access to Transplant and Transplant Outcome Measures study examining the association between pre-transplant cardiac screening in kidney transplant recipients and post-transplant cardiac events.

Predicting pathologic complete response in locally advanced rectal cancer patients after neoadjuvant therapy: a machine learning model using XGBoost

PurposeWatch and wait strategy is a safe and effective alternative to surgery in patients with locally advanced rectal cancer (LARC) who have achieved pathological complete response (pCR) after neoadjuvant therapy (NAT); present restaging methods do not meet clinical needs. This study aimed to construct a machine learning (ML) model to predict pCR preoperatively.MethodsLARC patients who received NAT were included to generate an extreme gradient boosting-based ML model to predict pCR. The group was divided into a training set and a tuning set at a 7:3 ratio. The SHapley Additive exPlanations value was used to quantify feature importance. The ML model was compared with a nomogram model developed using independent risk factors identified by conventional multivariate logistic regression analysis.ResultsCompared with the nomogram model, our ML model improved the area under the receiver operating characteristics from 0.72 to 0.95, sensitivity from 43 to 82.2%, and specificity from 87.1 to 91.6% in the training set, the same trend applied to the tuning set. Neoadjuvant radiotherapy, preoperative carbohydrate antigen 125 (CA125), CA199, carcinoembryonic antigen level, and depth of tumor invasion were significant in predicting pCR in both models.ConclusionOur ML model is a potential alternative to the existing assessment tools to conduct triage treatment for patients and provides reference for clinicians in tailoring individual treatment: the watch and wait strategy is used to avoid surgical trauma in pCR patients, and non-pCR patients receive surgical treatment to avoid missing the optimal operation time window.

The ‘just’ management of urban air pollution? A geospatial analysis of low emission zones in Brussels and London

The increasing evidence base and public concern on the health effects of exposure to high levels of air pollution, combined with stricter environmental legislation, are forcing local governments to take drastic measures. One of the policy instruments, the low emission zone (LEZ), specifically targets a reduction in emissions from vehicles, a key source in urban environments. It is a contested instrument, with supporters who think it is a fair “polluter pays” instrument that especially benefits more deprived communities, while opponents fear an unequal social impact on people's accessibility and finances. This study wants to add a data-driven perspective to the discussion by simultaneously analysing the unequal exposure to air pollution and the unequal accessibility impact, in a comparative study of the LEZs in London and Brussels. The analysis combines a conventional multivariate regression analysis with a geographically weighted regression (GWR) modelling to define the local spatial variation in the relationships, which is of particular concern when considering an explicitly spatial problem and solution. The study shows that GWR is a promising method in distributional environmental justice research through identifying parts of the city where effects are more unequal, as such facilitating customized policy instruments and targeted support.

Theory and practice of propensity score analysis.

Propensity score analysis has been widely used in observational studies to make a causal inference. This study introduces three assumptions for causal inferences-conditional exchangeability, positivity, and consistency-and five steps for propensity score (PS) analysis-1) construct appropriate PS models, 2) check overlap in PS, 3) apply appropriate weighting (inverse probability of treatment weighting, standardized mortality ratio weighting, matching weights, and overlap weights) or matching methods according to the target of inference, 4) check the balance of covariates, and 5) estimate the effect of exposure appropriately. Finally, the advantages of PS analyses over conventional multivariable regression are discussed.

Innovative use of spatial regression models to predict the effects of green infrastructure on land surface temperatures

Understanding the complex and dynamic interplay and cumulative effects of green infrastructure (GI) and urban form on land surface temperatures (LST) is important to design and implement heat mitigation strategies. Past research has mostly employed two-dimensional (2D) indicators, simple correlations and conventional regression models using coarse-level analytical approaches that obviate spatial autocorrelation effects. For the first time, this study applies a holistic approach to evaluate GI and urban settings as complex dynamic systems. The objectives of this paper are to: (1) develop novel ‘spatially-based’ predictive models that account for spatial dependencies; (2) implement a fine-scale analytical unit (<50 m) for a more precise and accurate analysis; (3) incorporate the ‘multi-temporal’ diurnal and seasonal variations into predictions; and (4) propose the novel combination of 2D and 3D morphological, compositional and configurational parameters of GI and urban form derived from very high resolution (VHR) remotely-sensed data (<2m), using Sydney metropolitan region as case study. Results show a strong spatial association of LST at fine scale (<50 m) and spatial autocorrelation among residuals in traditional models. Spatial error model (SEM) exhibits a superior performance over conventional multivariate regression, however, results presented significant heteroscedasticity caused by the large temperature variability in certain areas, although this problem was partially solved. Future studies should incorporate unmeasured factors related to material-specific properties (i.e. albedo, emissivity), and capture the thermal variation within urban areas by segmenting datasets into zones with relatively homogenous thermal and physical properties. Overall, ground imperviousness mostly defines the LST profile of a place, with a relative warming effect of 0.23 °C and 0.61 °C during day; and 0.18 °C and 0.41 °C at night per 10% of area increment in winter and summer, respectively. The same increment in the proportion of water and trees contributes to a maximum LST reduction of 0.42–0.85 °C in summer, and 0.25–1.17 °C in winter; however, this causes an increase of nocturnal LST between 0.12 °C and 0.30 °C throughout the year. In general, the cooling effects from GI do not outweigh the warming effects from man-made surfaces. Compared to abundance, the spatial configuration of trees is less influential on LST. Ground sky view factor (GSVF), altitude and distance to coast are of relative importance in defining LST profiles. These results used to numerically simulate different greening scenarios at neighbourhood scale for Sydney; illustrating the potential of spatial models to define heat mitigation scenarios to inform urban design and planning policies.