Conventional Regression Analysis Research Articles

Estimating the True Effect of Lifestyle Risk Factors for Myopia: A Longitudinal Study of UK Children.

Lifestyle risk factors are implicated in driving the current surge in myopia prevalence yet, paradoxically, known risk factors explain little of the variation in refractive error in the population. Here, we applied "instrumental variable" (IV) methods designed to avoid reverse causation and decrease confounding bias, to gauge lifestyle risk factor effect sizes. Three myopia risk factors (time outdoors, time reading, and sleep duration) were assessed in participants of the Avon Longitudinal Study of Parents and Children: a cross-sectional sample of 2302 children aged 15years old and a longitudinal sample of 3086 children followed from age 7 to 15years. Seven IVs were considered jointly: dog ownership, cat ownership, bedtime variability, birth order, and polygenic scores quantifying genetic predisposition to spend additional time outdoors, years in fulltime education, and time asleep overnight. Risk factor effect sizes were 4-fold to 9-fold higher in the IV analyses compared with conventional regression analyses. In IV analyses, one extra hour spent outdoors every day during childhood was associated with a shift toward hyperopia of +0.53 to +0.94 diopters (D), whereas 1 extra hour spent reading every day was associated with a shift toward myopia of -0.44 to -0.88 D. There was inconsistent evidence that sleep duration influenced refractive error. Myopia risk factor effects were underestimated up to 9-fold in conventional analyses in this sample, compared with IV analyses. We speculate that the effects of lifestyle risk factors for myopia have been underestimated in past studies.

Modelling of Steady-State Seepage of an Embankment Dam Using Teaching-Learning Based Optimization Algorithm

The goal of the this study is to investigate the applicability of the teaching-learning based optimization (TLBO) algorithm for modeling seepage in embankment dams. The input parameters selected for the models to be built are the values of permeability (ks), van Genuchten's suitability parameters α and n, whose effect on seepage has been investigated over the years due to their uncertainties. The validity of the TLBO was compared with that of conventional regression analysis (CRA) methods. Both methods were utilized with different regression forms. The parameters chosen as input are modeled as random variables with a log-normal distribution, and total discharge (Q) was obtained. Four statistical indices, that is, root mean square error, mean absolute error, average relative error and coefficient of determination, were used to evaluate the performance of the models. The equations obtained using TLBO algorithms can predict the total discharge in embankment dams better than CRA. In addition, the reliability of TLBO has been demonstrated by conducting analyses using the outputs of CRA as a benchmark.

Predicting the Recurrence of Ovarian Cancer Based on Machine Learning.

Recurrence is the main factor for poor prognosis in ovarian cancer, but few prognostic biomarkers were reported. In this study, we used machine learning methods based on multiple biomarkers to develop a specific prediction model for the recurrence of ovarian cancer. A total of 277 ovarian cancer patients were enrolled in this study and randomly classified into training and testing cohorts. The prediction information was obtained through 47 clinical parameters using six supervised clustering machine learning algorithms, including K-Nearest Neighbor (K-NN), Decision Tree (DT), Random Forest (RF), Adaptive Boosting (AdaBoost), Gradient Boosting Machine (GBM), and Extreme Gradient Boosting (XGBoost). In predicting the recurrence of ovarian cancer, machine learning algorithm was superior to conventional logistic regression analysis. In this study, XGBoost showed the best performance in predicting the recurrence of ovarian cancer, with an accuracy of 0.95. In addition, neoadjuvant chemotherapy, Monocyte ratio (MONO%), Hematocrit (HCT), Prealbumin (PAB), Aspartate aminotransferase (AST), and carbohydrate antigen 125 (CA125) are the most important biomarkers to predict the recurrence of ovarian cancer. The machine learning techniques can achieve a more accurate assessment of the recurrence of ovarian cancer, which can help clinicians make decisions, and develop personalized treatment strategies.

Machine learning and fluorosensing for estimation of maize nitrogen status at early growth-stages

Potential of mobile fluorescence sensor measurements have been in focus for quantifying plant nitrogen (N) variability early in the crop growing season. Real time estimation of such N status indicators at field scale would enable precision management of N fertilizers. In standard practice, linear regression analysis involves the use of several fluorescence channels and indices as predictive variables for estimating plant nitrogen content. Considering the multi-collinearity between these predictor variables, the conventional regression analysis (multiple linear regression) often fails to deliver a good range of prediction accuracies. Hence, machine learning regression techniques are utilized in this study to estimate N status indicators i.e., %N, above ground biomass, and N uptake at V6 and V9 growth stages of maize across three site-years in 2012 and 2013 crop growing seasons. The Multiplex®3 (FORCE-A) portable active fluorescence system was used to capture fluorescence information. Derived indices including four N balance indices (NBI_R, NBI_B, NBI_B, and NBI1), two chlorophyll indices (CHL and CHL1), and one flavonoid index (FLAV) were used as predictors. The independent site data were first utilized in a Support Vector Regression (SVR) model to assess the training and test accuracies in estimation of N status indicators considering a comparative analysis between V6 and V9 growth stages. The current research also involved assessing how well the machine learning-trained model could be applied to a different dataset and validated its performance in a cross-site experimental setting. Subsequently, cross-site comparisons of nitrogen status estimates were conducted to recommend the selection of machine learning strategies. These strategies include (1) Partial Least Square Regression, (2) Support Vector Regression, (3) Gaussian Process Regression, (4) Random Forest Regression, and (5) Artificial Neural Network. The comparative investigation demonstrated promising accuracy in estimating plant nitrogen content, above-ground biomass, and nitrogen uptake at the V6 stages of maize, with correlation coefficients in the moderate range (r = 0.72 ± 0.03) and Root Mean Square Error. Superior prediction accuracies were obtained at V9 growth stages than at V6. Among the various machine learning models assessed, Support Vector Regression consistently outperformed the others across the three site-year datasets, delivering error estimates well within acceptable ranges and demonstrating the lowest RMSE values for maize nitrogen indicators.

Exploring the impact of thiol collectors system on copper sulfide flotation through machine learning-driven modeling

Collector selection is a critical step in flotation, as it has a direct impact on product quality, flotation recovery, and selectivity. Collectors can consist of different components, and their effectiveness can vary depending on the type of ore being processed. The general practice in both literature and in industry is to use a mixture of collectors rather than a single collector. However, the use of a collector mixture introduces several complex issues. It is challenging to determine the specific effects of each collector on different minerals, as well as to understand the synergistic effects of mixed collectors in flotation. This study presents a novel investigation focusing on the impact of blends of NAX, AEROPHINE® 3422, and AERO® MX 5149, in varying dosages and combinations, on the flotation performance of Kupferschiefer copper ore. Kinetics flotation tests were conducted using a mechanical flotation cell with various combinations and dosages of listed collectors. For this investigation, different predictive models such as machine-learning (ML) and conventional regression analyses were developed. For model construction, a database including the results of comprehensive experimental results was constructed. The best performing model was selected considering statistical performance indicators and their performance on unseen data. A sensitivity analysis was conducted on the model to justify contributions of collectors on the copper recovery and grade. The results showed that the ML-based models provide compatible results with the expert opinions and have higher statistical performance than conventional modelling tools. According to the experimental results and models’ findings, it has shown that AEROPHINE® 3422 (a blend of isopropyl ethyl thionocarbamate and dithiophosphinate) was the most influential collector for the copper recovery. In addition, two ternary graphs were generated from the modeled data to formulate mixtures for different grades and recovery priorities.

The nonlinear impact of climate change on inland waterway transportation in the Upper Mississippi—Illinois River Region

The U.S. inland waterway system has played a critical role in promoting economic growth by efficiently transporting agricultural and manufacturing commodities along major riverine systems, such as the Mississippi River. However, the system faces escalating challenges due to the growing disruptions caused by climate change. For instance, the drought of the Mississippi River during the fall of 2022 resulted in an unprecedented increase in barge shipping rates. While the industry generally recognizes that climate change can significantly impact agricultural supply chains, it remains unclear how various environmental factors, including changes in water levels and temperature, affect inland waterway operations. This study aims to fill this research gap by conducting an empirical assessment of the impact of climate change-related factors on the operations of the inland waterway system in the Upper Mississippi—Illinois River Region. Unlike previous studies using conventional regression analysis, this study investigates the nonlinear relationship between changes in barge shipping rates and various environmental factors using the Gradient Boosting Decision Trees machine-learning method. Our analysis identifies the 29 ft gage height as a pivotal threshold influencing barge rates; rates tend to rise as the gage height decreases, and surpassing this threshold can also lead to increased shipping rates. This new approach enhances the model’s predictive capacity, allowing for a better understanding of the nonlinear effects of climate change on barge rates and productivity. It also enables planners and operational agencies to better understand the uncertainty of environmental conditions on the variations in barge rates and productivity performance. These findings provide valuable insights for decision-makers to understand the threshold effect of environmental conditions on inland waterway operations and facilitate the creation of effective adaptive strategies to mitigate future risks and consequences of climate change-induced disruptions.

The complex network patterns of human migration at different geographical scales: network science meets regression analysis

Migration’s influence in shaping population dynamics in times of impending climate and population crises exposes its crucial role in upholding societal cohesion. As migration impacts virtually all aspects of life, it continues to require attention across scientific disciplines. This study aims to bridge the gap between theoretical understanding and practical application by integrating network analysis and regression methodologies within Migration Studies. In the study we employ network analysis to elucidate migration patterns at various geographical scales-city, country, and global. Additionally, regression analysis is discussed on an exploratory level, where we focus on the underlying factors driving migration, and identifying the key independent variables to enhance predictive accuracy. The study exposes distinct migration network structure and its features, and the consequences these have on conventional regression analysis applications. We conclude on the importance of methodological coherence and disciplinary integration, and highlight the avenues for enhancing the predictive power of migration models.

From Realizable Dreams to Sustainable Reality: An Empirical Study on the Relationship between Internship Experience and Future Career Prospects among Business Students

The objective of this study is to establish the relationship between Internship Experience and Future Career Prospects among Business Students at Mbarara University of Science and Technology. Using a quantitative approach and a cross-sectional survey design, data was gathered from 100 alumni. Using an open-ended questionnaire, Mbarara University of Science and Technology business students' internship experiences and future career prospects were surveyed quantitatively. In addition, a conventional linear regression analysis was performed. The study's findings demonstrated that among the Mbarara University of Science and Technology's business students, internship experience had a strong positive and significant influence on their future career prospects. The study relates to the ongoing discussion about how business students' internship experience influences their chances for future employment.

Navigating the confluence of econometrics and data science: Implications for economic analysis and policy

This paper explores the transformative integration of econometrics and data science, a synergy poised to redefine empirical research within economics. By merging traditional econometric methods with advanced data science techniques, such as machine learning algorithms and big data analytics, this interdisciplinary approach enables a deeper, more nuanced understanding of complex economic phenomena. We delve into the theoretical foundations underlying this integration, highlighting how machine learning algorithms like random forests and neural networks complement conventional regression analysis, thereby enhancing model complexity and predictive accuracy. The paper further discusses methodological advancements, including handling high-dimensional data, incorporating unstructured data through natural language processing, and the evolution of model selection processes empowered by machine learning. Practical applications are thoroughly examined across three pivotal areas: economic forecasting and policy analysis, financial markets and risk management, and social economic analysis and public policy, showcasing the significant contributions of this convergence to economic forecasting, policy formulation, and the assessment of public interventions. This comprehensive exploration underscores the potential of combining econometrics and data science to offer more precise and actionable insights for policymakers, researchers, and practitioners in the field of economics.

Beyond regression: Unpacking research of human complex systems with qualitative comparative analysis

This study underscores the effectiveness of Qualitative Comparative Analysis (QCA) when compared to conventional regression analysis (CRA) in the investigation of complex human systems. Utilizing historical secondary cross-national data from Lipset & Man (1960) spanning 18 countries, where CRA may be impractical, the research emphasizes the superior performance of QCA, specifically utilizing both crisp set QCA and fuzzy set QCA. The dataset includes variables such as democracy survival and its precursors, such as gross national product per capita, urbanization, literacy rate, and industrial labor force. In contrast to conventional regression results indicating an insignificant relationship between democracy survival and its antecedents, crisp set QCA identifies two distinct combinations of antecedents associated with high levels of democracy survival, albeit with limited solution coverage. Meanwhile, fuzzy set qualitative comparative analysis (fsQCA) reveals five combinations of antecedents linked to robust democracy survival, providing adequate solution coverage and consistency. These findings suggest that fsQCA excels in capturing the intricacies of real-life scenarios in human complex systems, offering more robust empirical solutions compared to crisp set QCA and conventional regression. As a result, researchers may find value in integrating fsQCA into their new projects focused on human complex systems.

A tutorial for conducting intersectional multilevel analysis of individual heterogeneity and discriminatory accuracy (MAIHDA)

Intersectional multilevel analysis of individual heterogeneity and discriminatory accuracy (I-MAIHDA) is an innovative approach for investigating inequalities, including intersectional inequalities in health, disease, psychosocial, socioeconomic, and other outcomes. I-MAIHDA and related MAIHDA approaches have conceptual and methodological advantages over conventional single-level regression analysis. By enabling the study of inequalities produced by numerous interlocking systems of marginalization and oppression, and by addressing many of the limitations of studying interactions in conventional analyses, intersectional MAIHDA provides a valuable analytical tool in social epidemiology, health psychology, precision medicine and public health, environmental justice, and beyond. The approach allows for estimation of average differences between intersectional strata (stratum inequalities), in-depth exploration of interaction effects, as well as decomposition of the total individual variation (heterogeneity) in individual outcomes within and between strata.Specific advice for conducting and interpreting MAIHDA models has been scattered across a burgeoning literature. We consolidate this knowledge into an accessible conceptual and applied tutorial for studying both continuous and binary individual outcomes. We emphasize I-MAIHDA in our illustration, however this tutorial is also informative for understanding related approaches, such as multicategorical MAIHDA, which has been proposed for use in clinical research and beyond. The tutorial will support readers who wish to perform their own analyses and those interested in expanding their understanding of the approach. To demonstrate the methodology, we provide step-by-step analytical advice and present an illustrative health application using simulated data. We provide the data and syntax to replicate all our analyses.

Estimating the economic effects of US state and local fiscal policy: A synthetic control method matching‐regression approach

AbstractIn this paper, we advance the empirical literature on US state and local fiscal policymaking by using the synthetic control method (SCM) to create pairwise matches for states in subsequent regression analysis of the relationships between state and local fiscal policies and several state economic outcomes. Additional contributions include the use of principal component analysis to construct broader narratives of state economic performance and to reduce the dimensionality of the characteristics used in SCM matching, while the regressions also include variables to control for post‐matching economic shocks. Compared to conventional regression analysis, the SCM matching‐regression approach better addresses potential endogeneity, reduces interpolation bias, and creates fiscal policy measures that better reflect policy differences. The SCM‐matched regressions produce more statistically significant relationships between state and local fiscal variables and economic outcomes than do the conventional unmatched regressions, suggesting improved identification of state and local fiscal policy effects on economic outcomes. Robust relationships found include negative economic effects of the own‐source revenue burden and property taxes. Consistent with the existing literature, the estimated fiscal policy effects are quantitatively small and unlikely to drive differences in state economic performance.

Machine learning study using 2020 SDHS data to determine poverty determinants in Somalia

Extensive research has been conducted on poverty in developing countries using conventional regression analysis, which has limited prediction capability. This study aims to address this gap by applying advanced machine learning (ML) methods to predict poverty in Somalia. Utilizing data from the first-ever 2020 Somalia Demographic and Health Survey (SDHS), a cross-sectional study design is considered. ML methods, including random forest (RF), decision tree (DT), support vector machine (SVM), and logistic regression, are tested and applied using R software version 4.1.2, while conventional methods are analyzed using STATA version 17. Evaluation metrics, such as confusion matrix, accuracy, precision, sensitivity, specificity, recall, F1 score, and area under the receiver operating characteristic (AUROC), are employed to assess the performance of predictive models. The prevalence of poverty in Somalia is notable, with approximately seven out of ten Somalis living in poverty, making it one of the highest rates in the region. Among nomadic pastoralists, agro-pastoralists, and internally displaced persons (IDPs), the poverty average stands at 69%, while urban areas have a lower poverty rate of 60%. The accuracy of prediction ranged between 67.21% and 98.36% for the advanced ML methods, with the RF model demonstrating the best performance. The results reveal geographical region, household size, respondent age group, husband employment status, age of household head, and place of residence as the top six predictors of poverty in Somalia. The findings highlight the potential of ML methods to predict poverty and uncover hidden information that traditional statistical methods cannot detect, with the RF model identified as the best classifier for predicting poverty in Somalia.

Prediction of the axial compression capacity of stub CFST columns using machine learning techniques.

Concrete-filled steel tubular (CFST) columns have extensive applications in structural engineering due to their exceptional load-bearing capability and ductility. However, existing design code standards often yield different design capacities for the same column properties, introducing uncertainty for engineering designers. Moreover, conventional regression analysis fails to accurately predict the intricate relationship between column properties and compressive strength. To address these issues, this study proposes the use of two machine learning (ML) models-Gaussian process regression (GPR) and symbolic regression (SR). These models accept a variety of input variables, encompassing geometric and material properties of stub CFST columns, to estimate their strength. An experimental database of 1316 specimens was compiled from various research papers, including circular, rectangular, and double-skin stub CFST columns. In addition, a dimensionless output variable, referred to as the strength index, is introduced to enhance model performance. To validate the efficiency of the introduced models, predictions from these models are compared with those from two established standard codes and various ML algorithms, including support vector regression optimized with particle swarm optimization (PSVR), artificial neural networks, XGBoost (XGB), CatBoost (CATB), Random Forest, and LightGBM models. Through performance metrics, the CATB, GPR, PSVR and XGB models emerge as the most accurate and reliable models from the evaluation results. In addition, simple and practical design equations for the different types of CFST columns have been proposed based on the SR model. The developed ML models and proposed equations can predict the compressive strength of stub CFST columns with reliable and accurate results, making them valuable tools for structural engineering. Furthermore, the Shapley additive interpretation (SHAP) technique is employed for feature analysis. The results of the feature analysis reveal that section slenderness ratio and concrete strength parameters negatively impact the compressive strength index.

New findings in prognostic factor assessment for adenocarcinoma of transverse colon: a comparison study between competing-risk and COX regression analysis.

Competing-risk analysis was used to accurately assess prognostic factors for cancer-specific death in patients with adenocarcinoma of transverse colon (ATC), and the results were compared with those from a conventional Cox regression analysis. Patients diagnosed with ATC between 2000 and 2019 were selected from the Surveillance, Epidemiology, and End Results database. The crude mortality rates of patients with ATC were calculated and their differences were tested using the Gray's test, respectively. In performing multivariate analysis, the Cox regression model and the subdistribution hazard function (SD) in competing risk analysis were utilized, respectively. This study included 21,477 eligible patients. The SD model indicated that age, etc. are actual independent prognostic factors. In contrast to previous recognition, the results of the Cox regression showed false-positives for sex and Carcinoembryonic antigen, and underestimated point-estimates in the stage and American Joint Committee on Cancer stage due to competing events. A detailed comparison of treatment revealed that the larger surgical scopes were prognostic risk factors compared with the smaller scope of local tumor excision, partial colectomy, or segmental resection. Patients treated with external proton beam radiotherapy had an increased risk compared with those with no radiotherapy and internal radiotherapy. After comparing the results of the two methods and mitigating the significant bias introduced by Cox regression, we found independent factors that really affect the prognosis of ATC. On the other hand, in terms of ATC, a larger surgical scope and external proton beam radiotherapy may not improve the long-term survival of patients. Therefore, when faced with ATC patients, these differences should be noted and treated differently from common colorectal cancer patients. Thus, clinicians are able to give more targeted treatment plans and prognostic assessments.

PARTIALLY FUNCTIONAL LINEAR QUANTILE REGRESSION WITH MEASUREMENT ERRORS.

Ignoring measurement errors in conventional regression analyses can lead to biased estimation and inference results. Reducing such bias is challenging when the error-prone covariate is a functional curve. In this paper, we propose a new corrected loss function for a partially functional linear quantile model with function-valued measurement errors. We establish the asymptotic properties of both the functional coefficient and the parametric coefficient estimators. We also demonstrate the finite-sample performance of the proposed method using simulation studies, and illustrate its advantages by applying it to data from a children obesity study.

A recall-optimised machine learning framework for small data improves risk stratification for Hirschsprung's disease

ObjectiveTo improve the selection of patients with suspected Hirschsprung's disease for rectal biopsy with a recall-optimised machine learning framework that strongly penalises missed diagnoses, identify the most important clinical features, and investigate if this methodology is superior to the current practice of conventional logistic regression analysis. Study designThe study data comprised the medical records of 178 children older than one month who underwent a rectal biopsy for the evaluation of Hirschsprung's disease. Each medical record contained the clinical features recorded before the rectal biopsy and the biopsy result. However, only 20 of the 178 children were diagnosed with Hirschsprung's disease. Thus, as is frequently the case for rare diseases, the dataset was small and imbalanced. We present a machine learning framework that, based on these data, produces a champion machine learning model for predicting the presence of Hirschsprung's disease in patients. Notably, the machine learning framework is recall-optimised, i.e., it strongly penalises missed diagnoses. This is achieved using a combination of synthetic minority over-sampling techniques, Bayesian hyper-parameter optimisation, and bootstrapping-based prediction thresholding for a competing set of models comprised of a logistic regression model, a random forest classifier, and a gradient-boosted classifier. Finally, the machine learning framework evaluates the performance of these three competing models and conventional logistic regression with 5-fold stratified cross-validation using an 80:20 train-test split. Model performance is ranked by the average recall across the cross-validation folds. The model with the highest recall is selected as the champion. The model with the highest accuracy is chosen in case of a tie. ResultsThis study revealed that about 35% of our cohort's children without Hirschsprung's disease could have been spared a rectal biopsy while incurring no missed diagnoses using our machine learning framework. Given our aim of avoiding missed diagnosis, the champion model was vastly superior to conventional logistic regression, i.e., the status quo, which missed 40% of HD-positive cases. Moreover, the champion model pointed to a new hierarchy for the importance of clinical features associated with Hirschsprung's disease. Among all features, “gross abdominal distention” was the most important. ConclusionThere is considerable scope for a stricter selection when referring constipated children for a rectal biopsy to diagnose or exclude Hirschsprung's disease. This study demonstrates that a recall-optimised pipeline based on classical supervised machine learning is superior to the conventional statistical and heuristic approaches used today, also for small and imbalanced datasets. This finding opens a path to better care for patients with rare diseases while alleviating pressures on healthcare systems.

Machine Learning in the Hyperspectral Classification of Glycaspis brimblecombei (Hemiptera Psyllidae) Attack Severity in Eucalyptus

Assessing different levels of red gum lerp psyllid (Glycaspis brimblecombei) can influence the hyperspectral reflectance of leaves in different ways due to changes in chlorophyll. In order to classify these levels, the use of machine learning (ML) algorithms can help process the data faster and more accurately. The objectives were: (I) to evaluate the spectral behavior of the G. brimblecombei attack levels; (II) find the most accurate ML algorithm for classifying pest attack levels; (III) find the input configuration that improves performance of the algorithms. Data were collected from a clonal eucalyptus plantation (clone AEC 0144—Eucalyptus urophilla) aged 10.3 months old. Eighty sample evaluations were carried out considering the following severity levels: control (no shells), low infestation (N1), intermediate infestation (N2), and high infestation (N3), for which leaf spectral reflectances were obtained using a spectroradiometer. The spectral range acquired by the equipment was 350 to 2500 nm. After obtaining the wavelengths, they were grouped into representative interval means in 28 bands. Data were submitted to the following ML algorithms: artificial neural networks (ANN), REPTree (DT) and J48 decision trees, random forest (RF), support vector machine (SVM), and conventional logistic regression (LR) analysis. Two input configurations were tested: using only the wavelengths (ALL) and using the spectral bands (SB) to classify the attack levels. The output variable was the severity of G. brimblecombei attack. There were differences in the hyperspectral behavior of the leaves for the different attack levels. The highest attack level shows the greatest distinction and the highest reflectance values. LR and SVM show better accuracy in classifying the severity levels of G. brimblecombei attack. For the correct classification percentage, the RL and SVM algorithms performed better, both with accuracy above 90%. Both algorithms achieved F-score values close to 0.90 and above 0.8 for Kappa. The entire spectral range guaranteed the best accuracy for both algorithms.

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.