Construction Of Predictive Models Research Articles

Application of Raman spectroscopy and machine learning for Candida auris identification and characterization.

Currently, combating Candida auris infections and transmission is challenging due to the lack of efficient identification and characterization methods for this species. To address these challenges, our study presents a novel approach that utilizes Raman spectroscopy and artificial intelligence to achieve precise identification and characterization of C. auris at the singe-cell level. It can accurately identify a single cell from the four C. auris clades. Additionally, we developed machine learning models designed to detect antifungal resistance in C. auris cells and differentiate between two distinct phenotypes based on the single-cell Raman spectrum. We also constructed prediction models for detecting aggregating and filamentous cells in C. auris, both of which are closely linked to its virulence. These results underscore the merits of Raman spectroscopy in the identification and characterization of C. auris, promising improved outcomes in the battle against C. auris infections and transmission.

Predicting cardiovascular death 1-year after transcatheter aortic valve implantation: an artificial-intelligence based risk model

Abstract Background Prediction of cardiovascular death after transcatheter aortic valve implantation (TAVI) remains difficult. Currently, more lower risk patients are considered for TAVI and predicting not only immediate survival, but more longer-term outcomes seems important. Most developed risk tools use standard statistical approaches and focus mainly on short-term (30 days post-TAVI) outcomes. Purpose The aim of this project was to identify important predictors for worse outcomes after aortic valve replacement using machine learning features. Furthermore, we want to develop and artificial intelligence (AI) -model to predict cardiovascular death 1 year after TAVI. Methods For our analysis we used 725 patients with severe AS that underwent TAVI between 2008 and 2022 and were included in a prospective single centre observational study. Leveraging an automated machine learning framework, we systematically evaluated various machine learning algorithms including Random Forest, AdaBoost, Neural Networks, Randomized Decision Trees and and Gradient Boosting. Results We constructed prediction models utilizing three feature subsets: the top 10 most important features, the top 15 most important features (Figure 1), and all available features (n=66). Our analyses yielded area under curve (AUC) values of 0.71, 0.83, and 0.92, respectively (Figure 2). The best model was obtained by Randomized Decision Trees. NYHA classification, left ventricular ejection fraction, Charlson comorbidity index, posterior wall thickness and myocardial infarction 90-days before TAVI appeared as the most influential predictors for cardiovascular death within one year after TAVI. Conclusions Our model was able to identify the most important features indicating high risk for patients undergoing TAVI. Using the 15-feature model, we were able to accurately predict cardiovascular death 1 year after TAVI. AI-models to predict risk can be used to better inform/select patients undergoing TAVI.

Proteomic and serological markers for diagnosing cardia gastric cancer and precursor lesions in a Chinese population

Cardia gastric cancer (CGC) is prevalent in East Asia, and noninvasive, cost-effective screening methods are needed. This study investigated the diagnostic value of serum pepsinogen (PG), gastrin-17 (G-17), Helicobacter pylori (H. pylori) antibodies, and proteomic profiling for CGC and precancerous lesions. We conducted a case-control study involving biopsy-confirmed patients with CGC (n = 60), low-grade intraepithelial neoplasia (CLGD, n = 60), high-grade intraepithelial neoplasia (CHGD, n = 64), and healthy controls (n = 120) matched for age and sex from high-incidence areas in China. Serological markers including PGI, PGII, G-17, and H. pylori were measured using ELISA and Western blot, while plasma protein markers were assessed using Olink® technology. The VSOLassoBag algorithm and nine machine learning (ML) algorithms were employed to identify crucial features and construct predictive models. Various evaluation metrics, including the area under the receiver-operating-characteristic curve (AUC), were utilized to compare predictive performance. Elevated PGII levels, decreased PGR, and H. pylori infection were significantly associated with an increased risk of CGC and precancerous lesions (P for trend < 0.05). The eXtreme Gradient Boosting (XGBoost) model performed best in discriminative ability among the 9 ML models. Following feature reduction based on predictive performance, a final explainable XGBoost model was developed, incorporating five protein biomarkers (CDHR2, ICAM4, PTPRM, CDC27, and FLT1). This model exhibited excellent performance in distinguishing individuals with CGC and precancerous lesions from healthy controls (AUC = 0.931 for CGC, 0.867 for CHGD, and 0.763 for CLGD), surpassing the traditional serological marker-based model. This study underscores the diagnostic potential of serological markers and proteomic profiling in the detection of CGC. Further validation and exploration of combined biomarker approaches are warranted to enhance early diagnosis and improve outcomes in high-risk populations.

Construction and interpretation of machine learning-based prognostic models for survival prediction among intestinal-type and diffuse-type gastric cancer patients

BackgroundGastric cancer is one of the most common malignant tumors worldwide, with high incidence and mortality rates, and it has a complex etiology and complex pathological features. Depending on the tumor type, gastric cancer can be classified as intestinal-type and diffuse-type gastric cancer, each with distinct pathogenic mechanisms and clinical presentations. In recent years, machine learning techniques have been widely applied in the medical field, offering new perspectives for the diagnosis, treatment, and prognosis of gastric cancer patients.MethodsThis study recruited 2158 gastric cancer patients and constructed prognostic prediction models for both intestinal-type and diffuse-type gastric cancer. Clinical pathological data were collected from patients, and machine learning algorithms were used for feature selection and model construction. The performance of the models was validated with training and testing datasets. The Shapley additive explanations (SHAP) values were used to interpret the model predictions and identify the main factors that influence patient survival.ResultsIn the prognostic model for intestinal-type gastric cancer, the gradient boosting decision tree (GBDT) model demonstrated the best performance, with key features including pTNM, CA125, tumor size, CA199, and PALB. Similarly, in the prognostic model for diffuse-type gastric cancer, the GBDT model was utilized, with key features comprising pTNM, Borrmann type IV disease, lymphocyte (LYM), lactate dehydrogenase (LDH), potassium (K), perineural invasion (PNI), tumor size, and whole stomach location. Risk stratification analysis revealed that the prognosis of high-risk patients was significantly worse than that of low-risk patients.ConclusionMachine learning shows great potential in predicting survival outcomes of gastric cancer patients, providing strong support for the development of personalized treatment plans.

Assessing substrate scope of the cyclodehydratase LynD by mRNA display-enabled machine learning models.

Many of the biosynthetic pathways for ribosomal synthesized and post-translationally modified peptide (RiPP) natural products make use of multi-domain enzymes with separate recruitment and catalysis domains that separately bind and modify peptide substrates. This "division of labor" allows RiPP enzymes to use relatively open and promiscuous active sites to perform chemistry at multiple residues within a peptide substrate seemingly regardless of the surrounding context. Defining, measuring, and predicting the seemingly broad substrate promiscuity of RiPPs necessitates high throughput assays, capable of assessing activity against very large libraries of peptides. Using mRNA display, a high throughput peptide display technology, we examine the substrate promiscuity of the RiPP cyclodehydratase, LynD. The vast substrate profiling that can be done with mRNA display enables the construction of deep learning models for accurate prediction of substrate processing by LynD. These models further inform on epistatic interactions involved in enzymatic processing. This work will facilitate the further elucidation of other RiPP enzymes and enable their use in the modification of mRNA display libraries for selection of modified peptide-based inhibitors and therapeutics.

Research on Gold Price Prediction Based on LSTM Modeling

Gold plays a pivotal role in asset allocation, and the construction of gold price prediction models represents a complex yet rewarding task within the field of finance. The problem of international gold price is addressed in this paper forecasting by proposing a standard Long-Short Term Memory (LSTM) model and introducing bi-directional LSTM (Bi-LSTM) networks and multivariate analysis to compare the forecasting accuracy of the relevant models. This comparison is based on the daily gold price of the London Bullion Market Association (LBMA) for 2013-2022. This methodology takes into account various factors that influence the gold market and incorporates them as input sources into the Multi-factor LSTM model, thereby enhancing the interpretability of the LSTM model. Correlation analysis and the Granger test are employed to analyze these influencing factors. However, the results of this study suggest that while the Multi-factor LSTM model may be more accurate in predicting outliers, it tends to increase the number of small errors. On the other hand, the inclusion of Bi-LSTM networks can improve the overall accuracy of gold price prediction.

Machine Learning Approaches for Predicting Power Conversion Efficiency in Organic Solar Cells: A Comprehensive Review

Organic solar cells (OSCs), renowned for their lightweight, cost efficiency, and adaptability nature, stand out as a promising option for developing renewable energy. Improving the power conversion efficiency (PCE) of OSCs is essential, and researchers are delving into novel materials to achieve this. Traditional approaches are often laborious and costly, highlighting the need for predictive modeling. Machine learning (ML), especially via quantitative structure–property relationship (QSPR) models, is streamlining material development, with a goal to exceed a 20% PCE. In this review, the application of ML in OSCs is explored, and recent studies utilizing ML approaches for PCE prediction are reviewed, encompassing empirical functions, ML algorithms, self‐devised ML frameworks, and the combination with automated experimental technologies. First, the benefits of ML in predicting PCE for OSCs are addressed. Second, the development of high‐efficiency predictive models for both fullerene and nonfullerene acceptors is delved into. The impact of various ML algorithm models on PCE prediction is then assessed, taking into account the construction of predictive models. Moreover, the quality of databases and the selection of descriptors are considered. Databases and descriptors based on experimental studies are further categorized. Finally, prospects for the future development of OSCs are proposed.

Robust chemical analysis with graphene chemosensors and machine learning.

Ion-sensitive field-effect transistors (ISFETs) have emerged as indispensable tools in chemosensing applications1-4. ISFETs operate by converting changes in the composition of chemical solutions into electrical signals, making them ideal for environmental monitoring5,6, healthcare diagnostics7 and industrial process control8. Recent advancements in ISFET technology, including functionalized multiplexed arrays and advanced data analytics, have improved their performance9,10. Here we illustrate the advantages of incorporating machine learning algorithms to construct predictive models using the extensive datasets generated by ISFET sensors for both classification and quantification tasks. This integration also sheds new light on the working of ISFETs beyond what can be derived solely from human expertise. Furthermore, it mitigates practical challenges associated with cycle-to-cycle, sensor-to-sensor and chip-to-chip variations, paving the way for the broader adoption of ISFETs in commercial applications. Specifically, we use data generated bynon-functionalized graphene-based ISFET arrays to train artificial neural networks that possess a remarkable ability to discern instances of food fraud, food spoilage and food safety concerns. We anticipate that the fusion of compact, energy-efficient and reusable graphene-based ISFET technology with robust machine learning algorithms holds the potential to revolutionize the detection of subtle chemical and environmental changes, offering swift, data-driven insights applicable across a wide spectrum of applications.

Decision tree learning for predicting chronic postpartum depression in the Japan Environment and Children's Study

BackgroundMany studies have used machine learning techniques to construct predictive models of postpartum depression, but few such models are simple enough to use in community maternal health settings with pen and paper. Here, we use a decision tree to construct a prediction model for chronic postpartum depression. MethodsParticipants were 84,091 mothers. Chronic postpartum depression was identified as an Edinburgh Postnatal Depression Scale score of ≥9 at both 1 and 6 months postpartum. The training dataset included 84 diverse variables measured during pregnancy, including health status and biomarkers. In learning, the branching depth was constrained to 3, the number of branches per branch to 4, and the minimum number of n in a branch was 100. The training to validation data ratio was set to 7:3. ResultsA decision tree with 35 branches and an area under the receiver operating characteristic of 0.84 was created. Ten of 84 variables were extracted, and the most effective in classification was “feeling worthless.” At training (n = 58,635), the most and least prevalent branches were 73.2 % and 0.84 % (mean = 6.29 %), respectively; at validation (n = 25,456), they were 60.4 % and 0.72 % (mean = 6.52 %), respectively. LimitationsChronic postpartum depression was identified using self-administered questionnaires. ConclusionsThis study created a simple and relatively high-performing prediction model. Because the model can be easily understood and used without expertise in machine learning, it is expected to be useful in maternal health settings, including grassroots community health.

Construction of frailty and risk prediction models in maintenance hemodialysis patients: a cross-sectional study.

As the prevalence of diabetic nephropathy and hypertensive nephropathy increases with age in mainland China, the number of patients with end-stage renal disease is increasing, leading to an increase in the number of patients receiving maintenance hemodialysis. Considering the harmful effects of frailty on the health of maintenance hemodialysis patients, this study aims to identify hemodialysis patients at risk for frailty at an early stage, in order to prevent or delay the progression of frailty in the early stage, so as to prevent the adverse consequences of frailty. A total of 479 patients admitted to the blood purification centers of two grade tertiary hospitals in Anhui Province, China, using convenient sampling. The Frailty Scale, the SARC-F questionnaire, the Simplified Food Appetite Questionnaire (SNAQ) and the mini nutritional assessment short-form (MNA-SF) were used in the study. Pearson correlation analysis was used to explore the correlation among the frailty influencing factors. The incidence of frailty was 24.0% among 479 Chinese hemodialysis patients. Gender (p < 0.05), Malnutrition (p < 0.001), sarcopenia (p < 0.001), and feel tired after dialysis (p < 0.001) were highly correlated with frailty in Chinese hemodialysis patients. Serum albumin concentration (p < 0.05) was a protective factor of frailty. This survey shows that frailty was highly prevalent among Chinese hemodialysis patients. Medical staff and their families should make early judgments and carry out interventions on the risk of frailty.

Low carbon energy industry and network economy prediction based on sensors and real-time data processing

The widespread use of sensors provides a large amount of real-time data for enterprises and decision-makers, providing more accurate information support for the prediction and decision-making of the network economy. With the help of Internet of Things technology, the data collected by sensors is transmitted in real time to data centers or cloud platforms. Real time data processing technology is used to clean, denoise, and analyze the data in real time, ensuring the accuracy and timeliness of the data. Perform pattern recognition and trend analysis on historical data, discover hidden patterns and correlations in the data, construct predictive and decision-making models to predict future economic trends and make reasonable decisions, continuously optimize and adjust the model to adapt to real-time data changes and dynamic changes in the economic environment, and improve the accuracy and efficiency of the model. The experimental results show that the network economy prediction and decision-making model based on sensor networks and Internet of Things technology can more accurately predict economic development trends, improve decision-making efficiency and accuracy. The large amount of data provided by sensor networks provides sufficient support for the construction and optimization of models.

Bulk integrated single-cell-spatial transcriptomics reveals the impact of preoperative chemotherapy on cancer-associated fibroblasts and tumor cells in colorectal cancer, and construction of related predictive models using machine learning

BackgroundPreoperative chemotherapy (PC) is an important component of Colorectal cancer (CRC) treatment, but its effects on the biological functions of fibroblasts and epithelial cells in CRC are unclear. MethodsThis study utilized bulk, single-cell, and spatial transcriptomic sequencing data from 22 independent cohorts of CRC. Through bioinformatics analysis and in vitro experiments, the research investigated the impact of PC on fibroblast and epithelial cells in CRC. Subpopulations associated with PC and CRC prognosis were identified, and a predictive model was constructed using machine learning. ResultsPC significantly attenuated the pathways related to tumor progression in fibroblasts and epithelial cells. NOTCH3 + Fibroblast (NOTCH3 + Fib), TNNT1 + Epithelial (TNNT1 + Epi), and HSPA1A + Epithelial (HSPA1A + Epi) subpopulations were identified in the adjacent spatial region and were associated with poor prognosis in CRC. PC effectively diminished the presence of these subpopulations, concurrently inhibiting pathway activity and intercellular crosstalk. A risk signature model, named the Preoperative Chemotherapy Risk Signature Model (PCRSM), was constructed using machine learning. PCRSM emerged as an independent prognostic indicator for CRC, impacting both overall survival (OS) and recurrence-free survival (RFS), surpassing the performance of 89 previously published CRC risk signatures. Additionally, patients with a high PCRSM risk score showed sensitivity to fluorouracil-based adjuvant chemotherapy (FOLFOX) but resistance to single chemotherapy drugs (such as Bevacizumab and Oxaliplatin). Furthermore, this study predicted that patients with high PCRSM were resistant to anti-PD1therapy. ConclusionIn conclusion, this study identified three cell subpopulations (NOTCH3 + Fib, TNNT1 + Epi, and HSPA1A + Epi) associated with PC, which can be targeted to improve the prognosis of CRC patients. The PCRSM model shows promise in enhancing the survival and treatment of CRC patients.

The risk factors determined by four machine learning methods for the change of difference of bone mineral density in post-menopausal women after three years follow-up

The prevalence of osteoporosis has drastically increased recently. It is not only the most frequent but is also a major global public health problem due to its high morbidity. There are many risk factors associated with osteoporosis were identified. However, most studies have used the traditional multiple linear regression (MLR) to explore their relationships. Recently, machine learning (Mach-L) has become a new modality for data analysis because it enables machine to learn from past data or experiences without being explicitly programmed and could capture nonlinear relationships better. These methods have the potential to outperform conventional MLR in disease prediction. In the present study, we enrolled a Chinese post-menopause cohort followed up for 4 years. The difference of T-score (δ-T score) was the dependent variable. Information such as demographic, biochemistry and life styles were the independent variables. Our goals were: (1) Compare the prediction accuracy between Mach-L and traditional MLR for δ-T score. (2) Rank the importance of risk factors (independent variables) for prediction of δ T-score. Totally, there were 1698 postmenopausal women were enrolled from MJ Health Database. Four different Mach-L methods namely, Random forest (RF), eXtreme Gradient Boosting (XGBoost), Naïve Bayes (NB), and stochastic gradient boosting (SGB), to construct predictive models for predicting δ-BMD after four years follow-up. The dataset was then randomly divided into an 80% training dataset for model building and a 20% testing dataset for model testing. A 10-fold cross-validation technique for hyperparameter tuning was used. The model with the lowest root mean square error for the validation dataset was viewed as the best model for each ML method. The averaged metrics of the RF, SGB, NB, and XGBoost models were used to compare the model performance of the benchmark MLR model that used the same training and testing dataset as the Mach-L methods. We defined that the priority demonstrated in each model ranked 1 as the most critical risk factor and 22 as the last selected risk factor. For Pearson correlation, age, education, BMI, HDL-C, and TSH were positively and plasma calcium level, and baseline T-score were negatively correlated with δ-T score. All four Mach-L methods yielded lower prediction errors than the MLR method and were all convincing Mach-L models. From our results, it could be noted that education level is the most important factor for δ-T Score, followed by DBP, smoking, SBP, UA, age, and LDL-C. All four Mach-L outperformed traditional MLR. By using Mach-L, the most important six risk factors were selected which are, from the most important to the least: DBP, SBP, UA, education level, TG and sleeping hour. δ T score was positively related to SBP, education level, UA and TG and negatively related to DBP and sleeping hour in postmenopausal Chinese women.

B-141 Analysis of risk factors and construction of prediction model for pneumonia in patients with pneumoconiosis based on the clinical laboratory indexes

Abstract Background Pneumoconiosis has been the leading industrial disease in China. Pneumonia is one of the common comorbidities for patients with pneumoconiosis, which may lead to the worse condition and poor prognosis. Timely and accurate diagnosis is necessary to retard the pneumonia progression in pneumoconiosis patients. Due to the similarity of clinical features between pneumoconiosis and pneumonia, clinical laboratory tests are needed for assessing the hazard of pneumonia occurrence in patients with pneumoconiosis. Therefore, we aim to explore the risk factors of pneumonia from laboratory test indexes for patients with pneumoconiosis, construct prediction model, and evaluate its feasibility. Methods A total of 243 patients with pneumoconiosis from the General Hospital of Xuzhou Mining Group from September 2021 to August 2023 were selected as the modeling set, which was further separated as the non-pneumonia group and the pneumonia group. Nineteen laboratory test indexes related to inflammation, tumor markers, renal function, and blood routine examination were compared between the two groups. The univariate and multivariate Logistic analysis was used to screen out risk factors and construct the prediction model. Two hundred and twenty-eight patients with pneumoconiosis from General Hospital of Xuzhou Mining Group (Yi'an Branch) were enrolled as the validation set. The predictive efficiency of the model on pneumonia for patients with pneumoconiosis was evaluated through Hosmer-Lemeshow(H-L)test and receiver operating characteristic (ROC) curve. Results C-reactive protein (CRP), interleukin-6 (IL-6), interleukin-8 (IL-8), interleukin-10 (IL-10), β2 microglobulin, creatinine, urea, neutrophil percentage, lymphocyte percentage, and lymphocyte count in pneumonia group were significant different from those in non-pneumonia group (&amp;lt;0.05). Logistic regression analysis suggested that CRP (OR=3.883, &amp;lt;0.001), IL-6 (OR=1.217, =0.003), IL-8 (OR=1.215, =0.002) and β2-microglobulin (OR=2.838, =0.002) were independent risk factors for pneumonia in patients with pneumoconiosis. The prediction model was constructed as Logit (P) =1.357×CRP+0.196×IL-6+0.195×IL-8+1.043×β2-microglobulin-9.772. The H-L test demonstrated that there was no significantly difference between predicted values and the true values. The areas under ROC curves of the model in the modeling set and the validation set were 0.909 and 0.848. The sensitivity was 86.8% and 75.6%, and accuracy 82.1% and 79.7%, respectively. Conclusions CRP, IL-6, IL-8, and β2-microglobulin were independent risk factors of pneumonia for pneumoconiosis patients. The risk model constructed by laboratory test indexes showed a good predictive performance, which might provide laboratory evidence for the pneumonia diagnosis in patients with pneumoconiosis.

Advanced hybrid empirical mode decomposition, convolutional neural network and long short-term memory neural network approach for predicting grain pile humidity based on meteorological inputs

Grain pile humidity prediction is beneficial to ensure food security, and establishing an effective humidity prediction model is of great significance to the field of grain storage. By taking meteorological and grain temperature data as inputs, we propose a prediction model that combines Empirical Mode Decomposition (EMD), Convolutional Neural Network (CNN), and Long Short-Term Memory Network (LSTM). The model was verified in experimental data of three different storage layers of grain piles. The prediction results show that the proposed EMD-CNN-LSTM model has better prediction accuracy than the other three comparison models: CNN-LSTM, CNN and LSTM. From the average results of the entire granary, the MAE, RMSE, and MAPE results are 0.14, 0.18, and 0.25%, respectively, and the MAE value is 44% higher than the previous research method that does not consider meteorological factors. The MAE, RMSE, and MAPE results of the CNN-LSTM method with EMD decomposition were improved by 58%, 53% and 58% respectively compared with the method without EMD decomposition. It can be concluded that taking meteorological factors as model input and integrating EMD methods can improve prediction accuracy. The constructed prediction model shows effective prediction results in different storage layers of grain pile, which provides new insights for ensuring food security and also provides valuable references for multivariate time series prediction in other fields.

Precision molecular insights for prostate cancer prognosis: tumor immune microenvironment and cell death analysis of senescence-related genes by machine learning and single-cell analysis

BackgroundProstate cancer (PCa) is a prevalent malignancy among men, primarily originating from the prostate epithelium. It ranks first in global cancer incidence and second in mortality rates, with a rising trend in China. PCa's subtle initial symptoms, such as urinary issues, necessitate diagnostic measures like digital rectal examination, prostate-specific antigen (PSA) testing, and tissue biopsy. Advanced PCa management typically involves a multifaceted approach encompassing surgery, radiation, chemotherapy, and hormonal therapy. The involvement of aging genes in PCa development and progression, particularly through the mTOR pathway, has garnered increasing attention.MethodsThis study aimed to explore the association between aging genes and biochemical PCa recurrence and construct predictive models. Utilizing public gene expression datasets (GSE70768, GSE116918, and TCGA), we conducted extensive analyses, including Cox regression, functional enrichment, immune cell infiltration estimation, and drug sensitivity assessments. The constructed risk score model, based on aging-related genes (ARGs), demonstrated superior predictive capability for PCa prognosis compared to conventional clinical features. High-risk genes positively correlated with risk, while low-risk genes displayed a negative correlation.ResultsAn ARGs-based risk score model was developed and validated for predicting prognosis in prostate adenocarcinoma (PRAD) patients. LASSO regression analysis and cross-validation plots were employed to select ARGs with prognostic significance. The risk score outperformed traditional clinicopathological features in predicting PRAD prognosis, as evidenced by its high AUC (0.787). The model demonstrated good sensitivity and specificity, with AUC values of 0.67, 0.675, 0.696, and 0.696 at 1, 3, 5, and 8 years, respectively, in the GEO cohort. Similar AUC values were observed in the TCGA cohort at 1, 3, and 5 years (0.67, 0.659, 0.667, and 0.743). The model included 12 genes, with high-risk genes positively correlated with risk and low-risk genes negatively correlated.ConclusionsThis study presents a robust ARGs-based risk score model for predicting biochemical recurrence in PCa patients, highlighting the potential significance of aging genes in PCa prognosis and offering enhanced predictive accuracy compared to traditional clinical parameters. These findings open new avenues for research on PCa recurrence prediction and therapeutic strategies.

Incidence trends, overall survival, and metastasis prediction using multiple machine learning and deep learning techniques in pediatric and adolescent population with osteosarcoma and Ewing's sarcoma: nomogram and webpage.

The objective of this study was to analyze the incidence and overall survival (OS) of osteosarcoma (OSC) and Ewing's sarcoma (EWS) in a pediatric and adolescent population, employing machine learning (ML) and deep learning (DL) models to predict the likelihood of metastasis. Involving 2465 OSC and 1373 EWS patients aged 0-19 years, from 2004 to 2020. ML techniques-Lasso, Ridge Regression, Elastic Net, and Random Forest-were used alongside a deep learning model based on TensorFlow and Keras, to construct predictive models for metastasis. These models were optimized using grid search with cross-validation and evaluated on their performance metrics, including AUC, sensitivity, and accuracy. The variables' importance in metastasis prediction was determined using SHAP values. Statistical analysis was performed using R software, and an online nomogram was developed for clinical use. The age-adjusted incidence of OSC and EWS from 2004 to 2020 showed a significant uptrend. The deep learning model, iterated 50 times, outperformed the Random Forest model in both loss and accuracy stabilization. The nomogram created demonstrated accurate survival predictions, as evidenced by its calibration curves and the distinction between high and low-risk groups. The increasing trend in age-adjusted incidence of OSC and EWS highlights the need for continued research and improved therapeutic strategies in this domain. The study employed ML and DL models to predict distant metastasis in pediatric and adolescent patients with OSC and EWS, providing a valuable tool for prognosis. The online nomogram developed as a part of this research enhances the models' clinical utility, offering an accessible means for clinicians to predict survival outcomes effectively.

Association of plasma proteomics with mortality in individuals with and without type 2 diabetes: Results from two population-based KORA cohort studies

BackgroundProtein biomarkers may contribute to the identification of vulnerable subgroups for premature mortality. This study aimed to investigate the association of plasma proteins with all-cause and cause-specific mortality among individuals with and without baseline type 2 diabetes (T2D) and evaluate their impact on the prediction of all-cause mortality in two prospective Cooperative Health Research in the Region of Augsburg (KORA) studies.MethodsThe discovery cohort comprised 1545 participants (median follow-up 15.6 years; 244 with T2D: 116 total, 62 cardiovascular, 31 cancer-related and 23 other-cause deaths; 1301 without T2D: 321 total, 114 cardiovascular, 120 cancer-related and 87 other-cause deaths). The validation cohort comprised 1031 participants (median follow-up 6.9 years; 203 with T2D: 76 total, 45 cardiovascular, 19 cancer-related and 12 other-cause deaths; 828 without T2D: 169 total, 74 cardiovascular, 39 cancer-related and 56 other-cause deaths). We used Cox regression to examine associations of 233 plasma proteins with all-cause and cause-specific mortality and Lasso regression to construct prediction models for all-cause mortality stratifying by baseline T2D. C-index, category-free net reclassification index (cfNRI), and integrated discrimination improvement (IDI) were conducted to evaluate the predictive performance of built prediction models.ResultsThirty-five and 62 proteins, with 29 overlapping, were positively associated with all-cause mortality in the group with and without T2D, respectively. Out of these, in the group with T2D, 35, eight, and 26 were positively associated with cardiovascular, cancer-related, and other-cause mortality, while in the group without T2D, 55, 41, and 47 were positively associated with respective cause-specific outcomes in the pooled analysis of both cohorts. Regulation of insulin-like growth factor (IGF) transport and uptake by IGF-binding proteins emerged as a unique pathway enriched for all-cause and cardiovascular mortality in individuals with T2D. The combined model containing the selected proteins (five and 12 proteins, with four overlapping, in the group with and without T2D, respectively) and clinical risk factors improved the prediction of all-cause mortality by C-index, cfNRI, and IDI.ConclusionsThis study uncovered shared and unique mortality-related proteins in persons with and without T2D and emphasized the role of proteins in improving the prediction of mortality in different T2D subgroups.

Radiomics prediction models of left atrial appendage hypercoagulability based on machine learning algorithms: an exploration about cardiac computed tomography angiography imaging.

Transesophageal echocardiography (TEE) is the standard method for diagnosing left atrial appendage (LAA) hypercoagulability in patients with atrial fibrillation (AF), which means LAA thrombus/sludge, dense spontaneous echo contrast and slow LAA blood flow velocity (< 0.25m/s). Based on machine learning algorithms, cardiac computed tomography angiography (CCTA) radiomics features were adopted to construct prediction models and explore a suitable approach for diagnosing LAA hypercoagulability and adjusting anticoagulation. This study included 652 patients with non-valvular AF. The univariate analysis were used to select meaningful clinical characteristics to predict LAA hypercoagulability. Then 3D Slicer software was adopted to extract radiomics features from CCTA imaging. The radiomics score was calculated using the least absolute shrinkage and selection operator logistic regression analysis to predict LAA hypercoagulability. We then combined clinical characteristics and radiomics scores to construct a nomogram model. Finally, we got prediction models based on machine learning algorithms and logistic regression separately. The area under the receiver operating characteristic curve of radiomics score was 0.8449 in the training set and 0.7998 in the validation set. The nomogram model had a concordance index of 0.838. The final machine-learning based prediction models had good performances (best f1 score = 0.85). Radiomics features of long maximum diameter and high uniformity of Hounsfield unit in left atrial were significant predictors of the hypercoagulable state in LAA, with better predictive efficacy than clinical characteristics. Our combined models based on machine learning were reliable for hypercoagulable state screening and anticoagulation adjustment.

Determination and Prediction of Amino Acid Digestibility in Rapeseed Cake for Growing-Finishing Pigs.

The experiment was conducted to determine the apparent or standardized ileal digestibility (AID or SID) of crude protein (CP) and amino acids (AA) in 10 rapeseed cake samples fed to pigs, and to construct predictive models for the SID of CP and AA based on the chemical composition of rapeseed cakes. Twenty-two cannulated pigs (initial body weight: 39.8 ± 1.2 kg) were assigned to two 11 × 3 incomplete Latin square designs, including an N-free diet and 10 diets containing rapeseed cake. Each experimental period included 5 days of adaptation and 2 days of ileal digesta collection. Titanium dioxide (TiO2) was added at 0.3% to all the diets as an indigestible marker for calculating the ileal CP and AA digestibility. The coefficients of variation (CV) of the content of crude fat (EE), crude fiber (CF), neutral detergent fiber (NDF), acid detergent fiber (ADF), and total glucosinolates (TGS) in 10 samples of rapeseed cake were greater than 10%. The standardized ileal digestibility (SID) of crude protein (CP), lysine (Lys), methionine (Met), threonine (Thr), and tryptophan (Trp) in rapeseed cake was 73.34% (61.49 to 81.12%), 63.01% (41.41 to 73.10%), 69.47% (50.55 to 88.16%), 79.61% (74.41 to 87.58%), and 94.43% (91.34 to 97.20%), respectively. The best prediction equations for SIDCP, SIDLys, and SIDVal were as follows: SIDCP = 90.124 - 0.54NDF (R2 = 0.58), SIDLys = 100.107 - 1.229NDF (R2 = 0.94), and SIDVal = 151.012 - 2.990TGS (R2 = 0.57). Overall, great variation exists among the 10 rapeseed cakes, and the NDF, TGS, and heating temperature can be used as the key predictors for the SID of CP and AA.