Effective treatment of imbalanced datasets in health care using modified SMOTE coupled with stacked deep learning algorithms

Artificial neural network prediction of postoperative complications in papillary thyroid microcarcinoma based on preoperative ultrasonographic features

OBJECTIVE

To predict post-thyroidectomy complications in papillary thyroid microcarcinoma (PTMC) patients using a deep learning model based on preoperative ultrasonographic features. This study addresses the global rise in PTMC incidence and the challenges in treatment decision-making with high-resolution ultrasonography.

METHOD

This study enrolled 1638 patients with clinically staged cN0 PTMC who received surgical treatment from 1997 to 2019 at Beijing Friendship Hospital. Deep learning model was developed using fully connected neural network. Feature selection included 1000 iterations of Bootstrap sampling and Recursive Feature Elimination (RFE) to identify the top 10 features. Data preprocessing involved normalization and imputation for missing values. SMOTE addressed class imbalance. The model was trained and tested on random data split, with performance metrics including Accuracy (ACC), Area Under the Curve (AUC), Sensitivity (SEN), and Specificity (SPE), visualized through a ROC curve and confusion matrix.

RESULTS

The fully connected deep neural network model demonstrated high accuracy (ACC 0.81), Area Under the Curve (AUC 0.74), sensitivity (SEN 0.65), and specificity (SPE 0.83) and visualized by ROC curve and confusion matrix. These results highlight the model's reliability and potential as an effective tool in predicting postoperative complications and assisting in clinical decision-making for PTMC patients.

CONCLUSION

This study highlights the potential of deep learning in enhancing medical predictions and personalized healthcare. Despite promising results, limitations include a single-center data source and unconsidered factors like lifestyle and genetics. Future research should expand data sources, include more influencing factors, and refine algorithms to improve accuracy and applicability in thyroid cancer treatment.

Robust and consistent biomarker candidates identification by a machine learning approach applied to pancreatic ductal adenocarcinoma metastasis

BACKGROUND

Machine Learning (ML) plays a crucial role in biomedical research. Nevertheless, it still has limitations in data integration and irreproducibility. To address these challenges, robust methods are needed. Pancreatic ductal adenocarcinoma (PDAC), a highly aggressive cancer with low early detection rates and survival rates, is used as a case study. PDAC lacks reliable diagnostic biomarkers, especially metastatic biomarkers, which remains an unmet need. In this study, we propose an ML-based approach for discovering disease biomarkers, apply it to the identification of a PDAC metastatic composite biomarker candidate, and demonstrate the advantages of harnessing data resources.

METHODS

We utilised primary tumour RNAseq data from five public repositories, pooling samples to maximise statistical power and integrating data by correcting for technical variance. Data were split into train and validation sets. The train dataset underwent variable selection via a 10-fold cross-validation process that combined three algorithms in 100 models per fold. Genes found in at least 80% of models and five folds were considered robust to build a consensus multivariate model. A random forest model was constructed using selected genes from the train dataset and tested in the validation set. We also assessed the goodness of prediction by recalibrating a model using only the validation data. The biological context and relevance of signals was explored through enrichment and pathway analyses using QIAGEN Ingenuity Pathway Analysis and GeneMANIA.

RESULTS

We developed a pipeline that can detect robust signatures to build composite biomarkers. We tested the pipeline in PDAC, exploiting transcriptomics data from different sources, proposing a composite biomarker candidate comprised of fifteen genes consistently selected that showed very promising predictive capability. Biological contextualisation revealed links with cancer progression and metastasis, underscoring their potential relevance. All code is available in GitHub.

CONCLUSION

This study establishes a robust framework for identifying composite biomarkers across various disease contexts. We demonstrate its potential by proposing a plausible composite biomarker candidate for PDAC metastasis. By reusing data from public repositories, we highlight the sustainability of our research and the wider applications of our pipeline. The preliminary findings shed light on a promising validation and application path.

Prediction of likelihood of conception in dairy cows using milk mid-infrared spectra collected before the first insemination and machine learning algorithms

Accurate and ex-ante prediction of cows' likelihood of conception (LC) based on milk composition information could improve reproduction management on dairy farms. Milk composition is already routinely measured by mid-infrared (MIR) spectra, which are known to change with advancing stages of pregnancy. For lactating cows, MIR spectra may also be used for predicting the LC. Our objectives were to classify the LC at first insemination using milk MIR spectra data collected from calving to first insemination and to identify the spectral regions that contribute the most to the prediction of LC at first insemination. After quality control, 4,866 MIR spectra, milk production, and reproduction records from 3,451 Holstein cows were used. The classification accuracy and area under the curve (AUC) of 6 models comprising different predictors and 3 machine learning methods were estimated and compared. The results showed that partial least square discriminant analysis (PLS-DA) and random forest had higher prediction accuracies than logistic regression. The classification accuracy of good and poor LC cows and AUC in herd-by-herd validation of the best model were 76.35 ± 10.60% and 0.77 ± 0.11, respectively. All wavenumbers with values of variable importance in the projection higher than 1.00 in PLS-DA belonged to 3 spectral regions, namely from 1,003 to 1,189, 1,794 to 2,260, and 2,300 to 2,660 cm-1. In conclusion, the model can predict LC in dairy cows from a high productive TMR system before insemination with a relatively good accuracy, allowing farmers to intervene in advance or adjust the insemination schedule for cows with a poor predicted LC.

Predictive modelling and identification of key risk factors for stroke using machine learning

Strokes are a leading global cause of mortality, underscoring the need for early detection and prevention strategies. However, addressing hidden risk factors and achieving accurate prediction become particularly challenging in the presence of imbalanced and missing data. This study encompasses three imputation techniques to deal with missing data. To tackle data imbalance, it employs the synthetic minority oversampling technique (SMOTE). The study initiates with a baseline model and subsequently employs an extensive range of advanced models. This study thoroughly evaluates the performance of these models by employing k-fold cross-validation on various imbalanced and balanced datasets. The findings reveal that age, body mass index (BMI), average glucose level, heart disease, hypertension, and marital status are the most influential features in predicting strokes. Furthermore, a Dense Stacking Ensemble (DSE) model is built upon previous advanced models after fine-tuning, with the best-performing model as a meta-classifier. The DSE model demonstrated over 96% accuracy across diverse datasets, with an AUC score of 83.94% on imbalanced imputed dataset and 98.92% on balanced one. This research underscores the remarkable performance of the DSE model, compared to the previous research on the same dataset. It highlights the model's potential for early stroke detection to improve patient outcomes.

The transition from genomics to phenomics in personalized population health

Modern health care faces several serious challenges, including an ageing population and its inherent burden of chronic diseases, rising costs and marginal quality metrics. By assessing and optimizing the health trajectory of each individual using a data-driven personalized approach that reflects their genetics, behaviour and environment, we can start to address these challenges. This assessment includes longitudinal phenome measures, such as the blood proteome and metabolome, gut microbiome composition and function, and lifestyle and behaviour through wearables and questionnaires. Here, we review ongoing large-scale genomics and longitudinal phenomics efforts and the powerful insights they provide into wellness. We describe our vision for the transformation of the current health care from disease-oriented to data-driven, wellness-oriented and personalized population health.

Unveiling Adolescent Suicidality: Holistic Analysis of Protective and Risk Factors Using Multiple Machine Learning Algorithms

Adolescent suicide attempts are on the rise, presenting a significant public health concern. Recent research aimed at improving risk assessment for adolescent suicide attempts has turned to machine learning. But no studies to date have examined the performance of stacked ensemble algorithms, which are more suitable for low-prevalence conditions. The existing machine learning-based research also lacks population-representative samples, overlooks protective factors and their interplay with risk factors, and neglects established theories on suicidal behavior in favor of purely algorithmic risk estimation. The present study overcomes these shortcomings by comparing the performance of a stacked ensemble algorithm with a diverse set of algorithms, performing a holistic item analysis to identify both risk and protective factors on a comprehensive data, and addressing the compatibility of these factors with two competing theories of suicide, namely, The Interpersonal Theory of Suicide and The Strain Theory of Suicide. A population-representative dataset of 173,664 Norwegian adolescents aged 13 to 18 years (mean = 15.14, SD = 1.58, 50.5% female) with a 4.65% rate of reported suicide attempt during the past 12 months was analyzed. Five machine learning algorithms were trained for suicide attempt risk assessment. The stacked ensemble model significantly outperformed other algorithms, achieving equal sensitivity and a specificity of 90.1%, AUC of 96.4%, and AUCPR of 67.5%. All algorithms found recent self-harm to be the most important indicator of adolescent suicide attempt. Exploratory factor analysis suggested five additional risk domains, which we labeled internalizing problems, sleep disturbance, disordered eating, lack of optimism regarding future education and career, and victimization. The identified factors provided stronger support for The Interpersonal Theory of Suicide than for The Strain Theory of Suicide. An enhancement to The Interpersonal Theory based on the risk and protective factors identified by holistic item analysis is presented.

Fusion machine learning model predicts CAD-CAM ceramic colors and the corresponding minimal thicknesses over various clinical backgrounds

OBJECTIVES

This study has developed and optimized a machine learning model to accurately predict the final colors of CAD-CAM ceramics and determine their required minimum thicknesses to cover different clinical backgrounds.

METHODS

A total of 120 ceramic specimens (2 mm, 1 mm and 0.5 mm thickness; n = 10) of four CAD-CAM ceramics - IPS e.max, IPS ZirCAD, Upcera Li CAD and Upcera TT CAD - were studied. The CIELab coordinates (L*, a* and b*) of each specimen were obtained over seven different clinical backgrounds (A1, A2, A3.5, ND2, ND7, cobalt-chromium alloy (CC) and medium precious alloy (MPA)) using a digital spectrophotometer. The color difference (ΔE) and lightness difference (ΔL) results were submitted to 39 different models. The prediction results from the top-performing models were used to develop a fusion model via the Stacking integrated learning method for best-fitting prediction. The SHapley Additive exPlanation (SHAP) was performed to interpret the feature importance.

RESULTS

The fusion model, which combined the ExtraTreesRegressor (ET) and XGBRegressor (XGB) models, demonstrated minimal prediction errors (R2 = 0.9) in the external testing sets. Among the investigated variables, thickness and background colors (CC and MPA) majorly influenced the final color of restoration. To achieve perfect aesthetic restoration (ΔE<2.6), at least 1.9 mm IPS ZirCAD or 1.6 mm Upcera TT CAD were required to cover the CC background, while two tested glass-ceramics did not meet the requirements even with thicknesses over 2 mm.

SIGNIFICANCE

The fusion model provided a promising tool for automate decision-making in material selection with minimal thickness over various clinical background.

Breast cancer diagnosis and management guided by data augmentation, utilizing an integrated framework of SHAP and random augmentation

Recent research indicates that early detection of breast cancer (BC) is critical in achieving favorable treatment outcomes and reducing the mortality rate associated with it. With the difficulty in obtaining a balanced dataset that is primarily sourced for the diagnosis of the disease, many researchers have relied on data augmentation techniques, thereby having varying datasets with varying quality and results. The dataset we focused on in this study is crafted from SHapley Additive exPlanations (SHAP)-augmentation and random augmentation (RA) approaches to dealing with imbalanced data. This was carried out on the Wisconsin BC dataset and the effectiveness of this approach to the diagnosis of BC was checked using six machine-learning algorithms. RA synthetically generated some parts of the dataset while SHAP helped in assessing the quality of the attributes, which were selected and used for the training of the models. The result from our analysis shows that the performance of the models used generally increased to more than 3% for most of the models using the dataset obtained by the integration of SHAP and RA. Additionally, after diagnosis, it is important to focus on providing quality care to ensure the best possible outcomes for patients. The need for proper management of the disease state is crucial so as to reduce the recurrence of the disease and other associated complications. Thus the interpretability provided by SHAP enlightens the management strategies in this study focusing on the quality of care given to the patient and how timely the care is.

A Federated Learning Approach to Breast Cancer Prediction in a Collaborative Learning Framework

Breast cancer continues to pose a substantial worldwide public health concern, necessitating the use of sophisticated diagnostic methods to enable timely identification and management. The present research utilizes an iterative methodology for collaborative learning, using Deep Neural Networks (DNN) to construct a breast cancer detection model with a high level of accuracy. By leveraging Federated Learning (FL), this collaborative framework effectively utilizes the combined knowledge and data assets of several healthcare organizations while ensuring the protection of patient privacy and data security. The model described in this study showcases significant progress in the field of breast cancer diagnoses, with a maximum accuracy rate of 97.54%, precision of 96.5%, and recall of 98.0%, by using an optimum feature selection technique. Data augmentation approaches play a crucial role in decreasing loss and improving model performance. Significantly, the F1-Score, a comprehensive metric for evaluating performance, turns out to be 97%. This study signifies a notable advancement in the field of breast cancer screening, fostering hope for improved patient outcomes via increased accuracy and reliability. This study highlights the potential impact of collaborative learning, namely, in the field of FL, in transforming breast cancer detection. The incorporation of privacy considerations and the use of diverse data sources contribute to the advancement of early detection and the treatment of breast cancer, hence yielding significant benefits for patients on a global scale.

Machine learning (ML) techniques to predict breast cancer in imbalanced datasets: a systematic review

Knowledge discovery in databases (KDD) is crucial in analyzing data to extract valuable insights. In medical outcome prediction, KDD is increasingly applied, particularly in diseases with high incidence, mortality, and costs, like cancer. ML techniques can develop more accurate predictive models for cancer patients' clinical outcomes, aiding informed healthcare decision-making. However, cancer prediction modeling faces challenges because of the unbalanced nature of the datasets, where there is a small minority category of patients with a cancer diagnosis compared to a majority category of cancer-free patients. Imbalanced datasets pose statistical hurdles like bias and overfitting when developing accurate prediction models. This systematic review focuses on breast cancer prediction articles published from 2008 to 2023. The objective is to examine ML methods used in three critical steps of KDD: preprocessing, data mining, and interpretation which address the imbalanced data problem in breast cancer prediction. This work synthesizes prior research in ML methods for breast cancer prediction. The findings help identify effective preprocessing strategies, including balancing and feature selection methods, robust predictive models, and evaluation metrics of those models. The study aims to inform healthcare providers and researchers about effective techniques for accurate breast cancer prediction.

SKCV: Stratified K-fold cross-validation on ML classifiers for predicting cervical cancer

Cancer is the unregulated development of abnormal cells in the human body system. Cervical cancer, also known as cervix cancer, develops on the cervix’s surface. This causes an overabundance of cells to build up, eventually forming a lump or tumour. As a result, early detection is essential to determine what effective treatment we can take to overcome it. Therefore, the novel Machine Learning (ML) techniques come to a place that predicts cervical cancer before it becomes too serious. Furthermore, four common diagnosis testing namely, Hinselmann, Schiller, Cytology, and Biopsy have been compared and predicted with four common ML models, namely Support Vector Machine (SVM), Random Forest (RF), K-Nearest Neighbors (K-NNs), and Extreme Gradient Boosting (XGB). Additionally, to enhance the better performance of ML models, the Stratified k-fold cross-validation (SKCV) method has been implemented over here. The findings of the experiments demonstrate that utilizing an RF classifier for analyzing the cervical cancer risk, could be a good alternative for assisting clinical specialists in classifying this disease in advance.