Risk Prediction Research Articles

Visceral adiposity index and cardiorespiratory fitness: Unmasking risk of impaired fasting glucose among adolescents

Visceral adiposity index and cardiorespiratory fitness: Unmasking risk of impaired fasting glucose among adolescents

Predictive Value of Pre-Transplant Monocyte-to-Lymphocyte Ratio for Delayed Graft Function in Kidney Transplant Recipients

Background: Delayed graft function (DGF) after kidney transplantation (KT) negatively impacts long-term allograft survival. Inflammatory and immune response markers in transplant recipients have been linked to allograft outcomes. However, the association between the pre-transplant monocyte-to-lymphocyte ratio (MLR) and DGF following KT has not been previously investigated.Methods: This study included 162 patients who underwent KT between January 1989 and December 2023. The optimal pre-transplant MLR cutoff for predicting DGF was identified using receiver operating characteristic (ROC) curve analysis. Univariate and multivariate logistic regression analyses were performed to identify factors associated with DGF.Results: DGF occurred in 58 patients (35.8%). The optimal MLR cut-off for predicting DGF was 0.255 (Area under the curve (95% confidence interval) = 0.686 (0.603–0.769), P < 0.001), with a sensitivity of 81.0% and specificity of 55.8%. In multivariate analysis, MLR ≥ 0.255 was independently associated with DGF (Odds ratio (95% confidence interval) = 3.74 (1.55–9.02), P = 0.003). Higher MLR values were also correlated with longer hospital stays.Conclusions: An elevated pre-transplant MLR was a significant predictor of DGF following KT. MLR may serve as a useful, non-invasive biomarker for risk stratification and prediction of post-transplant outcomes.

Prognostic Value of Matrix Metalloproteinase 9 (MMP9) in Patients Following Off-Pump Coronary Artery Bypass Grafting

Background: Matrix metalloproteinase 9 (MMP9) has recently emerged as a risk predictor in patients with cardiovascular diseases (CVD). However, little is known regarding the significance of elevated plasma MMP9 levels in patients during the long-term period following myocardial revascularisation. We aimed to investigate the role of MMP9 in relation to myocardial status before and after myocardial revascularisation and to assess its long-term prognostic value. Methods: This prospective observational study included 200 male patients with ischaemic heart disease. All patients underwent direct myocardial revascularisation on a beating heart (off-pump surgery). Plasma MMP9 levels were analysed preoperatively, at 48 h postoperatively, and during the long-term follow-up period (one year postoperatively). Key echocardiographic parameters, specifically left ventricular ejection fraction (LVEF) and Left Ventricular End-Diastolic Volume (LVEDV), were also assessed. Results: MMP9 levels decreased significantly at 48 h postoperatively (p < 0.0001). During the long-term postoperative period, a clear relationship was demonstrated: higher 1-year MMP9 levels were associated with lower 1-year LVEF, whilst lower 1-year MMP9 levels were associated with higher 1-year LVEF. No significant correlation was observed between preoperative MMP9 levels and age or most other baseline laboratory parameters. Conclusions: Our study established an association between 1-year postoperative MMP9 levels and key parameters of left ventricular function during the long-term follow-up period. This suggests that MMP9 may serve as a novel biomarker for predicting outcomes following myocardial revascularisation.

Research on Financial Risk Prediction and Management Models Based on Big Data Analysis

The increasing complexity and volatility of financial markets necessitate more advanced risk prediction and management techniques. Conventional financial risk models typically depend on linear assumptions and fixed statistical distributions, which constrain their capability to accurately reflect complex market behaviors. Recent advancements in big data analytics and deep learning provide new opportunities for more precise and adaptive risk assessment. This research introduces a novel framework for financial risk prediction that combines deep learning, probabilistic modeling, and reinforcement learning-driven risk management. Unlike conventional econometric models, our approach employs a risk-aware deep learning model (RDLM) to capture nonlinear dependencies among financial indicators while leveraging probabilistic estimation to quantify uncertainty in risk predictions. We introduce an adaptive risk mitigation strategy (ARMS), which dynamically adjusts risk exposure through reinforcement learning and market-responsive portfolio optimization. RDLM integrates deep neural networks with probabilistic risk estimation to enhance prediction accuracy and interpretability. By focusing on financial risk distributions instead of point estimates, this method effectively measures uncertainty, leading to more reliable risk evaluations. To enhance transparency and address critical regulatory issues, explainable AI methods like SHAP and LIME are utilized. ARMS leverages reinforcement learning and real-time data processing to dynamically refine investment strategies. The model includes market-regime detection. This allows it to adjust portfolio allocations as risk conditions change, ensuring adaptability in volatile environments. Experimental evaluations on real-world financial datasets demonstrate the effectiveness of our approach in enhancing risk prediction accuracy, minimizing financial losses, and optimizing risk-adjusted returns. The proposed framework combines big data analytics, deep learning, and adaptive risk management, providing a scalable and interpretable solution for financial stability and decision-making.

Blood Type as a Potential Predictor of Hemorrhagic Risk in Patients Undergoing Partial Hepatectomy for Colorectal Liver Metastasis.

Background: Hepatic resection is performed for liver lesions and requires careful preoperative planning to minimize bleeding. Blood type O, associated with lower von Willebrand factor (vWF) levels, may increase bleeding risk. This study investigates the relationship between the ABO blood type and perioperative bleeding in partial hepatectomy for colorectal liver metastases (CRLMs). Methods: Out of 563 patients who underwent hepatectomy, 135 cases were analyzed for CRLM at Carmel Medical Center (2013-2023). Patients were categorized into blood type O (61 patients) and non-O (74 patients) groups. Data on perioperative hemoglobin levels, blood loss, coagulation parameters, transfusion needs, and complications were assessed using χ2, t-tests, and ANOVA (p < 0.05). Results: No significant differences were observed for estimated blood loss (474.3 ± 696 mL for O vs. 527.8 ± 599 mL for non-O; p = 0.29), intraoperative hemoglobin drop (p = 0.613), or transfusion rates (24.59% for O vs. 28.37% for non-O; p = 0.698). Although non-O patients had a higher postoperative INR (p = 0.035), this did not correlate with increased bleeding or transfusion needs. Conclusions: Blood type O does not significantly affect perioperative bleeding or transfusion requirements in partial hepatectomy for CRLM. Further research is needed to better understand the significance of the ABO blood type.

Polygenic risk score for breast cancer in the Thai population: Addressing genetic disparities in underrepresented populations.

10572 Background: Effective breast cancer prevention and management require accurate risk prediction tools. Polygenic risk scores (PRS) have shown promise but are often less effective in non-European populations due to differences in genetic architecture. This study evaluates PRS performance and adaptation for breast cancer in the Thai population, addressing disparities in underrepresented groups. Methods: We retrospectively analyzed breast cancer cases from the Genomics Thailand project at Siriraj Hospital and general population controls from the National Health Examination Survey (NHES) in Thailand. Whole-genome sequencing was performed for cases, and genotyping with imputation was done for controls using the TOPMed r2 reference panel. Clinical data were extracted from electronic medical records. PRS were constructed using SBayesRC, incorporating variants from publicly available genome-wide association study (GWAS) summary statistics and variant functional annotations. Logistic regression and area under the receiver operating characteristic curve (AUC) analyses were conducted using R. Results: The discovery cohort included 975 cases and 1,502 controls, with 230 cases and 265 controls in the validation cohort. Of the 330 previously reported GWAS loci, only 231 lead variants were identified in our dataset. We further analyzed variants near these lead variants within the 330 loci, identifying nominal associations with breast cancer for 329 loci (p&lt;0.05). Four PRS models were tested: (1) 231 variants, (2) ~7 million functional variants based on European (EUR) data, (3) East Asian (EAS) models, and (4) combined EUR and EAS models. The EUR-based model (AUC 0.66) outperformed the 231-variant model (AUC 0.59) and the population-specific EAS model (AUC 0.58) at p&lt;0.05. The combined EUR and EAS models showed no significant improvement over the EUR model alone (AUC 0.66 for both, p=0.69). Individuals in the highest PRS risk group (above the 90 th percentile) had an odds ratio (OR) of 3.34 for breast cancer compared to the rest of the population (95% confidence interval: 2.54–4.42, p&lt;0.05). Among 249 patients with pathology data, PRS was not associated with tumor size, estrogen receptor status, or nodal metastasis. Conclusions: In the Thai population, PRS derived from large-scale European GWAS provided the highest prediction accuracy for breast cancer risk. The limited transferability of a top-variant PRS (e.g., 330-variant model) underscores the challenge posed by variant availability in this population. Validation in prospective studies is essential to optimize PRS utility and address disparities in genetic risk prediction.

The Role of Biomarkers in Acute Pain: A Narrative Review.

Acute pain, a critical aspect of patient care, presents a challenge due to its subjective nature and complex biological underpinnings. Biomarkers for acute pain promise a paradigm shift in how pain is perceived, diagnosed, and managed. The study of genetic, inflammatory, and neurotransmission markers associated with pain experience may hold the key for the development of personalized and effective pain management strategies. This narrative review explores the neurobiological pathways of acute pain, encompassing inflammatory responses and neurotransmission mechanisms. It synthesizes current research on the identification and clinical application of biomarkers, emphasizing their potential to enhance diagnostic precision, treatment effectiveness, and risk prediction. We underscore the promising role of acute pain biomarkers in identifying patients at risk for developing acute and potentially chronic pain, predicting patients' response to pharmacological interventions, and aiding in the development of novel therapeutic and pain preventive strategies. The evolving landscape of biomarker research not only deepens our understanding of pain mechanisms but also lays the foundation for more tailored and patient-specific healthcare interventions.

Analysis of risk factors and prediction model construction for poor healing of perineal wounds after vaginal delivery: A retrospective case-control study.

Analysis of risk factors and prediction model construction for poor healing of perineal wounds after vaginal delivery: A retrospective case-control study.

Establishment and validation of a dynamic prognostic model using serial circulating tumor DNA (ctDNA) for endemic EBV-related nasopharyngeal carcinoma (NPC): A secondary analysis of EP-SEASON.

6049 Background: Published risk prediction tools have focused on pretreatment factors, whereas the accuracy remains challenging in cancer care. Emerging evidences emphasize the dynamic rather than static recurrence risks during treatment course, and non-invasive diagnostics tools have advanced opportunities for serial tumor assessments. Here, we present an effective dynamic risk individualized prediction model (NPC-DRIM) incorporating serial ctDNA data, using the endemic EBV-related NPC as a model. Methods: This study included 1000 patients (pts) enrolled from a prospective biomarker study EP-SEASON, with complete longitudinal ctDNA data at 11 timepoints across treatment: after each neoadjuvant chemotherapy (NAC) circle (T1-3), every week during radiotherapy (T4-T9), within 1 week after radiotherapy (T10), and 1-3 months after radiotherapy (T11). Pts were divided into subcohort NAC (n=752) and subcohort no-NAC (n=248) according to receiving NAC or not, and randomly 70/30% split into training and validation cohort. Time-series and statistical features characterizing the dynamic change of ctDNA at each timepoint were extracted. The NPC-DRIM at T3-T11 were developed using the features selected via Cox univariate analysis in training cohort and then validated. The performance of NPC-DRIM was determined by C-index, time-dependent AUC, calibration curves, and decision curves, and compared with existing models. Results: The NPC-DRIM incorporated 4 clinical variables, 8 time-series features and 10 statistical features of ctDNA data. The C-index for predicting recurrence increased with time: 0.64 at T2, 0.69 at T3-T4, 0.70 at T5, 0.71 at T6, 0.73 at T7-T9, 0.77 at T10, and 0.76 at T11 in subcohort NAC ; 0.70 at T5, 0.68 at T6, 0.82 at T7, 0.78 at T8, 0.73 at T9, 0.74 at T10, and 0.83 at T11 in subcohort no-NAC . The NPC-DRIM at T11 had statistically improved outcome prediction compared to other dynamic models (Landmark Cox and Joint Model), and static models (AHR_Chen, RPA_Guo, RPA_Lee, and AJCC_8th staging system) (Table). For individualized dynamic risk prediction, we developed a web-based calculator to visualized the estimated changing recurrence risks. In addition, we showed that the high-risk pts identified by NPC-DRIM benefit from immune checkpoint inhibitors (ICI), while the low-risk pts did not. Conclusions: We introduce for the first time that the dynamic risk prediction model NPC-DRIM outperformed the conventional models, facilitating personalized therapeutic paradigms. Clinical trial information: NCT03855020 . Subcohort NAC Subcohort no-NAC C-index p value C-index p value NPC-DRIM 0.76 0.83 Landmark Cox 0.65 0.01 0.63 &lt;0.01 Joint Model 0.63 &lt;0.01 0.61 0.01 AHR Model (Chen et al. 2021) 0.61 &lt;0.01 0.57 &lt;0.01 RPA Model (Guo et al. 2019) 0.59 &lt;0.01 0.59 &lt;0.01 RPA Model (Lee et al. 2019) 0.59 &lt;0.01 0.66 0.02 AJCC_8th 0.56 &lt;0.01 0.57 &lt;0.01

Predicting 10-year risk of chronic kidney disease in lithium-treated patients with bipolar disorder: A risk model development and internal cross-validation study.

Predicting 10-year risk of chronic kidney disease in lithium-treated patients with bipolar disorder: A risk model development and internal cross-validation study.

Benchmarking machine learning algorithm for stunting risk prediction in Indonesia

Stunting is a condition caused by poor nutrition that results in below-average height development, potentially leading to long-term effects such as intellectual disability, low learning abilities, and an increased risk of developing chronic diseases. One effort to reduce stunting is to apply a machine learning algorithm with a data science approach to develop risk prediction models based on factors in stunting. The study used the current cross industry standard process for data mining (CRISP-DM) framework to gain insight and analyzed 1561 records of data collected from the Indonesia family life survey (IFLS) for the prediction models. Two sampling methods, random undersampling, and oversampling synthetic minority oversampling technique (SMOTE), were employed and compared to overcome the data imbalance problem. Four machine learning classifier algorithms were trained and tested to determine the best-performing model. The experiment results showed that the algorithms yielded an average accuracy of more than 75%. Using the undersampling technique, the accuracy obtained by logistic regression, k-nearest neighbor (KNN), support vector classifier (SVC), and decision tree classifier were 95.21%, 78.91%, 92.97%, and 86.26% respectively. Meanwhile, the oversampling technique reached 96.17%, 88.50%, 93.29%, and 95.21%, respectively. Logistic regression emerges as the best classification, with oversampling yielding superior performance.

GDM-BC: Non-invasive body composition dataset for intelligent prediction of Gestational Diabetes Mellitus.

GDM-BC: Non-invasive body composition dataset for intelligent prediction of Gestational Diabetes Mellitus.

Machine learning and Fuzzy logic fusion approach for osteoporosis risk prediction.

Machine learning and Fuzzy logic fusion approach for osteoporosis risk prediction.

MRI-Based Multimodal AI Model Enables Prediction of Recurrence Risk and Adjuvant Therapy in Breast Cancer.

Timely intervention and improved prognosis for breast cancer patients rely on early metastasis risk detection and accurate treatment predictions. This study introduces an advanced multimodal MRI and AI-driven 3D deep learning model, termed the 3D-MMR-model, designed to predict recurrence risk in non-metastatic breast cancer patients. We conducted a multicenter study involving 1,199 non-metastatic breast cancer patients from four institutions in China, with comprehensive MRI and clinical data retrospectively collected. Our model employed multimodal-data fusion, utilizing contrast-enhanced T1-weighted imaging (T1+C) and T2-weighted imaging (T2WI) volumes, processed through a modified 3D-UNet for tumor segmentation and a DenseNet121-based architecture for disease-free survival (DFS) prediction. Additionally, we performed RNA-seq analysis to delve further into the relationship between concentrated hotspots within the tumor region and the tumor microenvironment. The 3D-MR-model demonstrated superior predictive performance, with time-dependent ROC analysis yielding AUC values of 0.90, 0.89, and 0.88 for 2-, 3-, and 4-year DFS predictions, respectively, in the training cohort. External validation cohorts corroborated these findings, highlighting the model's robustness across diverse clinical settings. Integration of clinicopathological features further enhanced the model's accuracy, with a multimodal approach significantly improving risk stratification and decision-making in clinical practice. Visualization techniques provided insights into the decision-making process, correlating predictions with tumor microenvironment characteristics. In summary, the 3D-MMR-model represents a significant advancement in breast cancer prognosis, combining cutting-edge AI technology with multimodal imaging to deliver precise and clinically relevant predictions of recurrence risk. This innovative approach holds promise for enhancing patient outcomes and guiding individualized treatment plans in breast cancer care.

Altered vertebral biomechanical properties in prostate cancer patients following androgen deprivation therapy.

Altered vertebral biomechanical properties in prostate cancer patients following androgen deprivation therapy.

The predictive power of baseline metabolic and volumetric [18F]FDG PET parameters with different thresholds for early therapy failure and mortality risk in DLBCL patients undergoing CAR-T-cell therapy.

The predictive power of baseline metabolic and volumetric [18F]FDG PET parameters with different thresholds for early therapy failure and mortality risk in DLBCL patients undergoing CAR-T-cell therapy.

Psychological distress and suicidal ideation in patients with depressive disorders: the chain mediation of psychological resilience and neuroticism.

Psychological distress and suicidal ideation in patients with depressive disorders: the chain mediation of psychological resilience and neuroticism.

PRESCIENTai, an AI-based digital histopathological image signature for risk of late distant recurrence and extended endocrine therapy (EET) benefit in hormone receptor–positive breast cancer.

1556 Background: A subset of patients (pts) with hormone receptor-positive (HR+) breast cancer (BC) experiences late distant recurrence (DR) and is more likely to benefit from EET. Clinical practice guidelines recommend use of genomic assays such as Breast Cancer Index (BCI) to identify these pts. We developed an updated AI-based digital histopathological risk score model to predict risk of late DR and extended letrozole therapy (ELT) benefit in this population. Methods: The AI model, PRESCIENTai, was trained on eligible samples (N = 2,271) from the National Surgical Adjuvant Breast and Bowel Project (NSABP) B-42 cohort, which randomized postmenopausal women with HR+ BC who were disease-free after 5 yrs of endocrine therapy (aromatase inhibitor (AI) or tamoxifen followed by AI) to either 5 yrs of letrozole or placebo. A transformer-based end-to-end deep learning model predicted risk score from H&amp;E whole-slide images (WSI) in conjunction with clinical information (age at randomization, surgery type, node status, prior use of tamoxifen, race, lowest bone mineral density T-score, HER2 status). CTransPath was used for feature extraction from WSI tiles. 5-fold cross validation was performed with data split into training, validation, and test sets (60:20:20). The risk score threshold was defined by the 50% quantile of the training set for each fold. Cox regression and Kaplan-Meier analysis evaluated late DR and ELT benefit in high- and low-risk pts. Results: Hazard ratio (HR) was computed for DR in low- vs. high-risk pts [HR = 0.198 (95% CI: 0.124, 0.317); p &lt; 0.001], with absolute difference of 7.61% in 10-yr DR (1.84% vs. 9.46%). High-risk pts experienced greater ELT benefit over placebo (HR = 0.622; 95% CI: 0.416–0.929; p = 0.02) than low-risk pts (HR = 0.727; 95% CI: 0.305–1.733; p = 0.471), with 10-yr absolute benefit of 3.74% vs. 0.66%. Even among node(+) pts, PRESCIENTai identified greater ELT benefit for high-risk pts (HR = 0.521; 95% CI: 0.329–0.827; p = 0.006) than low-risk pts (HR = 0.53; 95% CI: 0.048-5.905, p = 0.606), with 10-yr absolute benefit of 6.66% vs. 1.89%. ELT benefit was also observed for high-risk pts in other clinical subgroups such as age ≤60 years and prior tamoxifen. However, p-interaction for ELT benefit in high- vs. low-risk groups was not significant for all pts ( p = 0.791) or node(+) pts ( p = 0.889). Conclusions: This novel digital signature predicts risk of late DR in pts with HR+ BC. Although absolute ELT benefit was greater in high- vs. low-risk pts, the treatment by risk score interaction was not statistically significant. This is, to our knowledge, the first AI model to predict long-term outcomes in pts with HR+, early BC using a single slide image and clinical information. Successful validation in additional pt cohorts will confirm the clinical utility of PRESCIENTai for prediction of late DR risk and EET benefit.

Development of an explainable machine learning model for predicting depression in adolescent girls with non-suicidal self-injury: A cross-sectional multicenter study.

Development of an explainable machine learning model for predicting depression in adolescent girls with non-suicidal self-injury: A cross-sectional multicenter study.

Evaluating the Khorana score in advanced EGFR-mutated NSCLC patients receiving first-line osimertinib.

e20634 Background: Venous thromboembolism (VTE) poses a significant risk in patients with non-small cell lung cancer (NSCLC), particularly those with driver mutations, such as in the Epidermal Growth Factor Receptor (EGFR). While the Khorana Score is commonly used to predict VTE risk in general oncology patients, its relevance in genetically defined subgroups, such as EGFR-mutated NSCLC patients, remains unclear. Methods: This retrospective study evaluated the correlation between the Khorana Score and VTE incidence in 47 advanced EGFR-mutated NSCLC patients receiving first-line osimertinib at the Inova Schar Cancer Institute. Baseline demographics and Khorana Score components, including platelet count, white blood cell count, body mass index, hemoglobin, and erythropoietin stimulating agent use, were recorded, with VTE incidence as the primary outcome. Results: VTE was observed in 19.1% of patients, but the Khorana Score did not significantly correlate with VTE development. Median progression-free survival (PFS) was 13.87 months in the VTE group compared to 9.9 months in those without VTE. Conclusions: Our findings suggest that the Khorana Score may not be a reliable predictor of VTE risk in EGFR-mutated NSCLC patients treated with osimertinib, likely due to its design for a more heterogeneous cancer population. Further research is warranted to develop tailored VTE risk models for this distinct patient subgroup.