Predictive Performance Research Articles

Nomogram prediction of vessels encapsulating tumor clusters in small hepatocellular carcinoma ≤ 3 cm based on enhanced magnetic resonance imaging

BACKGROUND Vessels encapsulating tumor clusters (VETC) represent a recently discovered vascular pattern associated with novel metastasis mechanisms in hepatocellular carcinoma (HCC). However, it seems that no one have focused on predicting VETC status in small HCC (sHCC). This study aimed to develop a new nomogram for predicting VETC positivity using preoperative clinical data and image features in sHCC (≤ 3 cm) patients. AIM To construct a nomogram that combines preoperative clinical parameters and image features to predict patterns of VETC and evaluate the prognosis of sHCC patients. METHODS A total of 309 patients with sHCC, who underwent segmental resection and had their VETC status confirmed, were included in the study. These patients were recruited from three different hospitals: Hospital 1 contributed 177 patients for the training set, Hospital 2 provided 78 patients for the test set, and Hospital 3 provided 54 patients for the validation set. Independent predictors of VETC were identified through univariate and multivariate logistic analyses. These independent predictors were then used to construct a VETC prediction model for sHCC. The model’s performance was evaluated using the area under the curve (AUC), calibration curve, and clinical decision curve. Additionally, Kaplan-Meier survival analysis was performed to confirm whether the predicted VETC status by the model is associated with early recurrence, just as it is with the actual VETC status and early recurrence. RESULTS Alpha-fetoprotein_lg10, carbohydrate antigen 199, irregular shape, non-smooth margin, and arterial peritumoral enhancement were identified as independent predictors of VETC. The model incorporating these predictors demonstrated strong predictive performance. The AUC was 0.811 for the training set, 0.800 for the test set, and 0.791 for the validation set. The calibration curve indicated that the predicted probability was consistent with the actual VETC status in all three sets. Furthermore, the decision curve analysis demonstrated the clinical benefits of our model for patients with sHCC. Finally, early recurrence was more likely to occur in the VETC-positive group compared to the VETC-negative group, regardless of whether considering the actual or predicted VETC status. CONCLUSION Our novel prediction model demonstrates strong performance in predicting VETC positivity in sHCC (≤ 3 cm) patients, and it holds potential for predicting early recurrence. This model equips clinicians with valuable information to make informed clinical treatment decisions.

Prognostic Value of Complete Resection of the High-Frequency Oscillation Area in Intracranial EEG: A Systematic Review and Meta-Analysis.

High-frequency oscillations (HFOs; ripples 80-250 Hz; fast ripples [FRs] 250-500 Hz) recorded with intracranial electrodes generated excitement and debate about their potential to localize epileptogenic foci. We performed a systematic review and meta-analysis on the prognostic value of complete resection of the HFOs-area (crHFOs-area) for epilepsy surgical outcome in intracranial EEG (iEEG) accessing multiple subgroups. We searched PubMed, Embase, and Web of Science for original research from inception to October 27, 2022. We defined favorable surgical outcome (FSO) as Engel class I, International League Against Epilepsy class 1, or seizure-free status. The prognostic value of crHFOs-area for FSO was assessed by (1) the pooled FSO proportion after crHFOs-area; (2) FSO for crHFOs-area vs without crHFOs-area; and (3) the predictive performance. We defined high combined prognostic value as FSO proportion >80% + FSO crHFOs-area >without crHFOs-area + area under the curve (AUC) >0.75 and examined this for the clinical subgroups (study design, age, diagnostic type, HFOs-identification method, HFOs-rate thresholding, and iEEG state). Temporal lobe epilepsy (TLE) was compared with extra-TLE through dichotomous variable analysis. Individual patient analysis was performed for sex, affected hemisphere, MRI findings, surgery location, and pathology. Of 1,387 studies screened, 31 studies (703 patients) met our eligibility criteria. Twenty-seven studies (602 patients) analyzed FRs and 20 studies (424 patients) ripples. Pooled FSO proportion after crHFOs-area was 81% (95% CI 76%-86%) for FRs and 82% (73%-89%) for ripples. Patients with crHFOs-area achieved more often FSO than those without crHFOs-area (FRs odds ratio [OR] 6.38, 4.03-10.09, p < 0.001; ripples 4.04, 2.32-7.04, p < 0.001). The pooled AUCs were 0.81 (0.77-0.84) for FRs and 0.76 (0.72-0.79) for ripples. Combined prognostic value was high in 10 subgroups: retrospective, children, long-term iEEG, threshold (FRs and ripples) and automated detection and interictal (FRs). FSO after complete resection of FRs-area (crFRs-area) was achieved less often in people with TLE than extra-TLE (OR 0.37, 0.15-0.89, p = 0.006). Individual patient analyses showed that crFRs-area was seen more in patients with FSO with than without MRI lesions (p = 0.02 after multiple correction). Complete resection of the brain area with HFOs is associated with good postsurgical outcome. Its prognostic value holds, especially for FRs, for various subgroups. The use of HFOs for extra-TLE patients requires further evidence.

External Validation of the PREdiction of Survival on Extracorporeal Membrane Oxygenation Therapy (PRESET) Score: A Single Center Cohort Experience.

Acute respiratory distress syndrome (ARDS) is a life-threatening condition affecting >10% of intensive care unit (ICU) patients worldwide with a mortality of up to 59% depending on severity. Extracorporeal membrane oxygenation (ECMO) is a potentially life-saving procedure in severe ARDS but is technically and financially challenging. In recent years, various scoring systems have been proposed to select patients most likely to benefit from ECMO, with the PREdiction of Survival on ECMO Therapy (PRESET) score being one of the most used. We collected data from 283 patients with ARDS of various etiology who underwent veno-venous (V-V) ECMO therapy at a German tertiary care ICU from January 2012 to December 2022. Median age in the cohort was 56 years, and 64.31% were males. The in-hospital mortality rate was 50.88% (n = 144). The median (25%; 75% quartile) severity scores were 38 (31; 49) for Simplified Acute Physiology Score (SAPS) II, 12 (10; 13) for Sequential Organ Failure Assessment (SOFA) and 7 (5; 8) for PRESET. Simplified Acute Physiology Score-II displayed the best prognostic value (area under the receiver operating characteristic [AUROC]: 0.665 [confidence interval (CI): 0.574-0.756; p = 0.046]). Prediction performance was weak in all analyzed scores despite good calibration. Simplified Acute Physiology Score-II had the best discrimination after adjustment of our original cohort. The use of scores explored in this study for patient selection for eligibility for V-V ECMO is not recommendable.

Externally validated treatment algorithm acceptably predicts nonoperative treatment success in patients with anterior cruciate ligament rupture.

This study aims to develop and externally validate a treatment algorithm to predict nonoperative treatment success or failure in patients with anterior cruciate ligament (ACL) rupture. Data were used from two completed studies of adult patients with ACL ruptures: the Conservative versus Operative Methods for Patients with ACL Rupture Evaluation study (development cohort) and the KNee osteoArthritis anterior cruciate Ligament Lesion study (validation cohort). The primary outcome variable is nonoperative treatment success or failure. Potential predictor variables were collected, entered into the univariable logistic regression modeland then incorporated into the multivariable logistic regression model for constructing the treatment algorithm. Finally, predictive performance and goodness-of-fit were assessed and externally validated by discrimination and calibration measures. In the univariable logistic regression model, a stable knee measured with the pivot shift test and a posttrauma International Knee Documentation Committee (IKDC) score <50 were predictive of needing an ACL reconstruction. Age >30 years and a body mass index > 30 kg/m2 were predictive for not needing an ACL reconstruction. Age, pretrauma Tegner score, the outcome of the pivot shift testand the posttrauma IKDC score are entered into the treatment algorithm. The predictability of needing an ACL reconstruction after nonoperative treatment (discrimination) is acceptable in both the development and the validation cohort: area under the curve = resp. 0.69 (95% confidence interval [CI]: 0.58-0.81) and 0.68 (95% CI: 0.58-0.78). This study shows that the treatment algorithm can acceptably predict whether an ACL injury patient will have a(n) (un)successful nonoperative treatment (discrimination). Calibration of the treatment algorithm suggests a systematical underestimation of the need for ACL reconstruction. Given the limitations regarding the sample size of this study, larger data sets must be constructed to improve the treatment algorithm further. Level II.

Extracellular volume-based scoring system for tracking tumor progression in pancreatic cancer patients receiving intraoperative radiotherapy

ObjectivesTo investigate the value of extracellular volume (ECV) derived from portal-venous phase (PVP) in predicting prognosis in locally advanced pancreatic cancer (LAPC) patients receiving intraoperative radiotherapy (IORT) with initial stable disease (SD) and to construct a risk-scoring system based on ECV and clinical-radiological features.Materials and methodsOne hundred and three patients with LAPC who received IORT demonstrating SD were enrolled and underwent multiphasic contrast-enhanced CT (CECT) before and after IORT. ECV maps were generated from unenhanced and PVP CT images. Clinical and CT imaging features were analyzed. The independent predictors of progression-free survival (PFS) determined by multivariate Cox regression model were used to construct the risk-scoring system. Time-dependent receiver operating characteristic (ROC) curve analysis and the Kaplan–Meier method were used to evaluate the predictive performance of the scoring system.ResultsMultivariable analysis revealed that ECV, rim-enhancement, peripancreatic fat infiltration, and carbohydrate antigen 19-9 (CA19-9) response were significant predictors of PFS (all p < 0.05). Time-dependent ROC of the risk-scoring system showed a satisfactory predictive performance for disease progression with area under the curve (AUC) all above 0.70. High-risk patients (risk score ≥ 2) progress significantly faster than low-risk patients (risk score < 2) (p < 0.001).ConclusionECV derived from PVP of conventional CECT was an independent predictor for progression in LAPC patients assessed as SD after IORT. The scoring system integrating ECV, radiological features, and CA19-9 response can be used as a practical tool for stratifying prognosis in these patients, assisting clinicians in developing an appropriate treatment approach.Critical relevance statementThe scoring system integrating ECV fraction, radiological features, and CA19-9 response can track tumor progression in patients with LAPC receiving IORT, aiding clinicians in choosing individual treatment strategies and improving their prognosis.Key PointsPredicting the progression of LAPC in patients receiving IORT is important.Our ECV-based scoring system can risk stratifying patients with initial SD.Appropriate prognostication can assist clinicians in developing appropriate treatment approaches.Graphical

Predicting early mortality and severe intraventricular hemorrhage in very-low birth weight preterm infants: a nationwide, multicenter study using machine learning

Our aim was to develop a machine learning-based predictor for early mortality and severe intraventricular hemorrhage (IVH) in very-low birth weight (VLBW) preterm infants in Taiwan. We collected retrospective data from VLBW infants, dividing them into two cohorts: one for model development and internal validation (Cohort 1, 2016–2021), and another for external validation (Cohort 2, 2022). Primary outcomes included early mortality, severe IVH, and early poor outcomes (a combination of both). Data preprocessing involved 23 variables, with the top four predictors identified as gestational age, birth body weight, 5-min Apgar score, and endotracheal tube ventilation. Six machine learning algorithms were employed. Among 7471 infants analyzed, the selected predictors consistently performed well across all outcomes. Logistic regression and neural network models showed the highest predictive performance (AUC 0.81–0.90 in both internal and external validation) and were well-calibrated, confirmed by calibration plots and the lowest two mean Brier scores (0.0685 and 0.0691). We developed a robust machine learning-based outcome predictor using only four accessible variables, offering valuable prognostic information for parents and aiding healthcare providers in decision-making.

Multi Task Learning Based Shallow Parsing for Indian Languages

Shallow Parsing is an important step for many Natural Language Processing tasks. Although shallow parsing has a rich history for resource rich languages, it is not the case for most Indian languages. Shallow Parsing consists of POS Tagging and Chunking. Our study focuses on developing shallow parsers for Indian languages. As part of shallow parsing we included morph analysis as well. For the study, we first consolidated available shallow parsing corpora for 7 Indian Languages (Hindi, Kannada, Bangla, Malayalam, Marathi, Urdu, Telugu) for which treebanks are publicly available. We then trained models to achieve state of the art performance for shallow parsing in these languages for multiple domains. Since analyzing the performance of model predictions at sentence level is more realistic, we report the performance of these shallow parsers not only at the token level, but also at the sentence level. We also present machine learning techniques for multitask shallow parsing. Our experiments show that fine-tuned contextual embedding with multi-task learning improves the performance of multiple as well as individual shallow parsing tasks across different domains. We show the transfer learning capability of these models by creating shallow parsers (only with POS and Chunk) for Gujarati, Odia, and Punjabi for which no treebanks are available. As a part of this work, we will be releasing the Indian Languages Shallow Linguistic (ILSL) benchmarks for 10 Indian languages including both the major language families Indo-Aryan and Dravidian as common building blocks that can be used to evaluate and understand various linguistic phenomena found in Indian languages and how well newer approaches can tackle them.

Personalized venlafaxine dose prediction using artificial intelligence technology: a retrospective analysis based on real-world data.

Venlafaxine dose regimens vary considerably between individuals, requiring personalized dosing. This study aimed to identify dose-related influencing factors of venlafaxine through real-world data analysis and to construct a personalized dose model using advanced artificial intelligence techniques. We conducted a retrospective study on patients with depression treated with venlafaxine. Significant variables were selected through a univariate analysis. Subsequently, the predictive performance of seven models (XGBoost, LightGBM, CatBoost, GBDT, ANN, TabNet, and DT) was compared. The algorithm that demonstrated optimal performance was chosen to establish the dose prediction model. Model validation used confusion matrices and ROC analysis. Additionally, a dose subgroup analysis was conducted. A total of 298 patients were included. TabNet was selected to establish the venlafaxine dose prediction model, which exhibited the highest performance with an accuracy of 0.80. The analysis identified seven crucial variables correlated with venlafaxine daily dose, including blood venlafaxine concentration, total protein, lymphocytes, age, globulin, cholinesterase, and blood platelet count. The area under the curve (AUC) for predicting venlafaxine doses of 75mg, 150mg, and 225mg were 0.90, 0.85, and 0.90, respectively. We successfully developed a TabNet model to predict venlafaxine doses using real-world data. This model demonstrated substantial predictive accuracy, offering a personalized dosing regimen for venlafaxine. These findings provide valuable guidance for the clinical use of the drug.

Advanced tree-based machine learning methods for predicting the seismic response of regular and irregular RC frames

In this study, predicting the maximum drift ratio (MDR) for two types of structures: 8-story regular (R8) and vertically irregular (IR8) reinforced concrete (RC) frames are focused. To accomplish this, advanced tree-based machine learning (ML) methods, namely Random Forest (RF), eXtreme Gradient Boosting (XGBoost), and Stochastic Gradient Boosting (SGB) are utilized. 2300 ground motion (GM) records, includes twenty-one input parameters and their nonlinear time history analyses results are considered as a dataset. The performance of the RF, XGBoost, and SGB models was examined considering the impact of the data pre-processing issue to test whether the use of data pre-processing approaches in model development could improve the prediction performance of the advanced ML models. Various data pre-processing techniques, including outlier detection using boxplot, missing data imputation using RF, data transformation using log transformation, and feature selection using Hill Climbing are employed as part of ML pipeline. The models' performance is evaluated using six different well-established statistical evaluation metrics. Subsequently, the results are systematically compared to identify the most effective prediction model. Our findings highlight the substantial impact of data pre-processing on the performance of ML models when predicting seismic responses, specifically the MDR of RC frames. Notably, RF consistently outperformed other ML models for both R8 and IR8 structures in the case of the simple ML pipeline strategy and data pre-processing approach. The RF model achieved R2 scores of R2 = 0.87 and R2 = 0.91 on the test data before implementing data pre-processing for R8 and IR8, respectively. After data pre-processing, R2 scores showed an improvement of 12% and 6.22% in the cases of R8 and IR8. In contrast, the performance scores of SGB consistently fell behind those of RF and XGBoost models.

GEnDDn: An lncRNA-Disease Association Identification Framework Based on Dual-Net Neural Architecture and Deep Neural Network.

Accumulating studies have demonstrated close relationships between long non-coding RNAs (lncRNAs) and diseases. Identification of new lncRNA-disease associations (LDAs) enables us to better understand disease mechanisms and further provides promising insights into cancer targeted therapy and anti-cancer drug design. Here, we present an LDA prediction framework called GEnDDn based on deep learning. GEnDDn mainly comprises two steps: First, features of both lncRNAs and diseases are extracted by combining similarity computation, non-negative matrix factorization, and graph attention auto-encoder, respectively. And each lncRNA-disease pair (LDP) is depicted as a vector based on concatenation operation on the extracted features. Subsequently, unknown LDPs are classified by aggregating dual-net neural architecture and deep neural network. Using six different evaluation metrics, we found that GEnDDn surpassed four competing LDA identification methods (SDLDA, LDNFSGB, IPCARF, LDASR) on the lncRNADisease and MNDR databases under fivefold cross-validation experiments on lncRNAs, diseases, LDPs, and independent lncRNAs and independent diseases, respectively. Ablation experiments further validated the powerful LDA prediction performance of GEnDDn. Furthermore, we utilized GEnDDn to find underlying lncRNAs for lung cancer and breast cancer. The results elucidated that there may be dense linkages between IFNG-AS1 and lung cancer as well as between HIF1A-AS1 and breast cancer. The results require further biomedical experimental verification. GEnDDn is publicly available at https://github.com/plhhnu/GEnDDn.

External validation and updating of the infusion rate individualization of soybean oil-based intravenous lipid emulsion: A descriptive cohort study.

Safe and efficient provision of intravenous lipid emulsion (ILE) requires a strategy to individualize infusion rates. Estimating the maximum acceptable infusion rate (MaxInfRate) of soybean oil-based ILE (SO-ILE) in individuals by using a triglyceride (TG) kinetic model was reported to be feasible. In this study, we aimed to externally validate and, if needed, update the MaxInfRate estimation. The maximum TG concentration (TGmax) in patients receiving SO-ILE at MaxInfRate was evaluated to determine if it met the definition of being <400 mg/dl for 90th percentile of patients. The TG kinetic model was evaluated through prediction performance checks and was subsequently updated using the data set of both the previous model development and present validation studies. Out of 83 patients, 74 had TGmax <400 mg/dl, corresponding to a probability of 89.2% (95% CI, 81.9%-95.2%), and the 90th percentile of TGmax was 400 mg/dl (95% CI, 328-490 mg/dl), closely aligned with the theoretical values. However, the individual TGmax values were biased by the infusion rate because the covariate effects were overestimated in the TG kinetic model, requiring a minor revision. The updated MaxInfRate with the combined data set showed unbiased and more accurate predictions. The MaxInfRate was validated in external inpatients and updated with all available data. MaxInfRate estimation for individuals could be an option for the safe and efficient provision of SO-ILE.

A hybrid machine learning forecasting model for photovoltaic power

The increasing use of photovoltaic (PV) power generation presents a significant opportunity for global energy transformation. However, accurately forecasting PV power remains a challenge. This study proposes a hybrid approach that combines variational mode decomposition (VMD), whale optimization algorithm (WOA), and long short-term memory neural network (LSTM) to forecast photovoltaic (PV) power accurately. The model decomposes the time series of PV power using VMD to address the non-stationary nature of the time series. The VMD parameters are optimized using WOA. Subsequently, and the PV power time series data is then decomposed using the optimized VMD parameters to obtain multiple intrinsic mode functions (IMFs) components and a residual component. These components are then reconstructed with meteorological parameters to obtain the reconstructed IMF components and residuals. Finally, multiple LSTM sub-models are built, with each of them taking the IMF components and residual from the previous reconstruction as inputs. The sub-models are optimized using the WOA method to determine their hyperparameters and then constructed with these optimized hyperparameters. The predicted values of each IMF and residual are output and added for sequence reconstruction to derive the final predicted value of PV power. The model's effectiveness was verified for one-hour-ahead forecasting at the 1.8 MW solar system in Yulara, central Australia. The test results show that the proposed model outperforms other benchmark models, with mean absolute error (MAE), root mean square error (RMSE), mean absolute percentage error (MAPE), and R-squared value (R2) of 15.247 kW, 19.753 kW, 4.405 % and 0.997, respectively. Compared to the LSTM, MAE, RMSE, and MAPE decreased by 84.942 %, 86.746 %, and 82.611 %, respectively, while R2 increased by 23.438 %. The proposed model has better predictive performance for both stable power changes and large fluctuations, essential for effectively integrating renewable energy sources into the power grid.

Functional hemodynamic imaging markers for the prediction of pathological outcomes in breast cancer patients treated with neoadjuvant chemotherapy.

Achieving pathologic complete response (pCR) after neoadjuvant chemotherapy (NACT) is a significant predictor of increased likelihood of survival in breast cancer patients. Early prediction of pCR is of high clinical value as it could allow personalized adjustment of treatment regimens in non-responding patients for improved outcomes. We aim to assess the association between hemoglobin-based functional imaging biomarkers derived from diffuse optical tomography (DOT) and the pathological outcome represented by pCR at different timepoints along the course of NACT. Twenty-two breast cancer patients undergoing NACT were enrolled in a multimodal DOT and X-ray digital breast tomosynthesis (DBT) imaging study in which their breasts were imaged at different compression levels. Logistic regressions were used to study the associations between DOT-derived imaging markers evaluated after the first and second cycles of chemotherapy, respectively, with pCR status determined after the conclusion of NACT at the time of surgery. Receiver operating characteristic curve analysis was also used to explore the predictive performance of selected DOT-derived markers. Normalized tumor HbT under half compression was significantly lower in the pCR group compared to the non-pCR group after two chemotherapy cycles (). In addition, the change in normalized tumor upon reducing compression from full to half mammographic force was identified as another potential indicator of pCR at an earlier time point, i.e., after the first chemo cycle (). Exploratory predictive assessments showed that AUCs using DOT-derived functional imaging markers as predictors reach as high as 0.75 and 0.71, respectively, after the first and second chemo cycle, compared to AUCs of 0.50 and 0.53 using changes in tumor size measured on DBT and MRI. These findings suggest that breast DOT could be used to assist response assessment in women undergoing NACT, a critical but unmet clinical need, and potentially enable personalized adjustments of treatment regimens.

Investigating a hybrid extreme learning machine coupled with Dingo Optimization Algorithm for modeling liquefaction triggering in sand-silt mixtures

Liquefaction is a devastating consequence of earthquakes that occurs in loose, saturated soil deposits, resulting in catastrophic ground failure. Accurate prediction of such geotechnical parameter is crucial for mitigating hazards, assessing risks, and advancing geotechnical engineering. This study introduces a novel predictive model that combines Extreme Learning Machine (ELM) with Dingo Optimization Algorithm (DOA) to estimate strain energy-based liquefaction resistance. The hybrid model (ELM-DOA) is compared with the classical ELM, Adaptive Neuro-Fuzzy Inference System with Fuzzy C-Means (ANFIS-FCM model), and Sub-clustering (ANFIS-Sub model). Also, two data pre-processing scenarios are employed, namely traditional linear and non-linear normalization. The results demonstrate that non-linear normalization significantly enhances the prediction performance of all models by approximately 25% compared to linear normalization. Furthermore, the ELM-DOA model achieves the most accurate predictions, exhibiting the lowest root mean square error (484.286 J/m3), mean absolute percentage error (24.900%), mean absolute error (404.416 J/m3), and the highest correlation of determination (0.935). Additionally, a Graphical User Interface (GUI) has been developed, specifically tailored for the ELM-DOA model, to assist engineers and researchers in maximizing the utilization of this predictive model. The GUI provides a user-friendly platform for easy input of data and accessing the model's predictions, enhancing its practical applicability. Overall, the results strongly support the proposed hybrid model with GUI serving as an effective tool for assessing soil liquefaction resistance in geotechnical engineering, aiding in predicting and mitigating liquefaction hazards.

Lysosome-related genes predict acute myeloid leukemia prognosis and response to immunotherapy

BackgroundAcute myeloid leukemia (AML) is a highly aggressive and pathogenic hematologic malignancy with consistently high mortality. Lysosomes are organelles involved in cell growth and metabolism that fuse to form specialized Auer rods in AML, and their role in AML has not been elucidated. This study aimed to identify AML subtypes centered on lysosome-related genes and to construct a prognostic model to guide individualized treatment of AML.MethodsGene expression data and clinical data from AML patients were downloaded from two high-throughput sequencing platforms. The 191 lysosomal signature genes were obtained from the database MsigDB. Lysosomal clusters were identified by unsupervised consensus clustering. The differences in molecular expression, biological processes, and the immune microenvironment among lysosomal clusters were subsequently analyzed. Based on the molecular expression differences between lysosomal clusters, lysosomal-related genes affecting AML prognosis were screened by univariate cox regression and multivariate cox regression analyses. Algorithms for LASSO regression analyses were employed to construct prognostic models. The risk factor distribution, KM survival curve, was applied to evaluate the survival distribution of the model. Time-dependent ROC curves, nomograms and calibration curves were used to evaluate the predictive performance of the prognostic models. TIDE scores and drug sensitivity analyses were used to explore the implication of the model for AML treatment.ResultsOur study identified two lysosomal clusters, cluster1 has longer survival time and stronger immune infiltration compared to cluster2. The differences in biological processes between the two lysosomal clusters are mainly manifested in the lysosomes, vesicles, immune cell function, and apoptosis. The prognostic model consisting of six prognosis-related genes was constructed. The prognostic model showed good predictive performance in all three data sets. Patients in the low-risk group survived significantly longer than those in the high-risk group and had higher immune infiltration and stronger response to immunotherapy. Patients in the high-risk group showed greater sensitivity to cytarabine, imatinib, and bortezomib, but lower sensitivity to ATRA compared to low -risk patients.ConclusionOur prognostic model based on lysosome-related genes can effectively predict the prognosis of AML patients and provide reference evidence for individualized immunotherapy and pharmacological chemotherapy for AML.

Orthodontic treatment outcome predictive performance differences between artificial intelligence and conventional methods

ABSTRACT Objectives To evaluate an artificial intelligence (AI) model in predicting soft tissue and alveolar bone changes following orthodontic treatment and compare the predictive performance of the AI model with conventional prediction models. Materials and Methods A total of 1774 lateral cephalograms of 887 adult patients who had undergone orthodontic treatment were collected. Patients who had orthognathic surgery were excluded. On each cephalogram, 78 landmarks were detected using PIPNet-based AI. Prediction models consisted of 132 predictor variables and 88 outcome variables. Predictor variables were demographics (age, sex), clinical (treatment time, premolar extraction), and Cartesian coordinates of the 64 anatomic landmarks. Outcome variables were Cartesian coordinates of the 22 soft tissue and 22 hard tissue landmarks after orthodontic treatment. The AI prediction model was based on the TabNet deep neural network. Two conventional statistical methods, multivariate multiple linear regression (MMLR) and partial least squares regression (PLSR), were each implemented for comparison. Prediction accuracy among the methods was compared. Results Overall, MMLR demonstrated the most accurate results, while AI was least accurate. AI showed superior predictions in only 5 of the 44 anatomic landmarks, all of which were soft tissue landmarks inferior to menton to the terminal point of the neck. Conclusions When predicting changes following orthodontic treatment, AI was not as effective as conventional statistical methods. However, AI had an outstanding advantage in predicting soft tissue landmarks with substantial variability. Overall, results may indicate the need for a hybrid prediction model that combines conventional and AI methods.

Association of inflammatory markers based on routine blood with prognosis in patients underwent percutaneous coronary intervention.

Inflammation contributes to the pathophysiological processes of coronary artery disease. We evaluated the association between inflammatory biomarkers, neutrophil-to-lymphocyte ratio (NLR), red cell distribution width (RDW), systemic inflammatory index, platelet-lymphocyte ratio, and 1-year all-cause mortality in patients underwent percutaneous coronary intervention (PCI). In this retrospective cohort, we consecutively enrolled 4651 patients who underwent PCI. Baseline demographic details, clinical data, and laboratory parameters on admission were analyzed. The primary outcome was 1-year all-cause mortality after PCI. We performed Cox regression and restricted cubic spline analysis to assessed the association between the inflammatory biomarkers and the clinical outcome. The area under the curve from receiver operating characteristic analysis was determined for the ability to classify mortality outcomes. A total of 4651 patients were included. Of these, 198 (4.26%) died on follow-up. Univariate Cox regression showed that NLR (heart rate [HR]: 1.070, 95% confidence interval [CI]: 1.060-1.082, P < .001), RDW (HR: 1.441, 95% CI 1.368-1.518, P < .001), systemic inflammatory index (HR: 1.000, 95% CI 1.000-3.180, P < .001), platelet-lymphocyte ratio (HR: 3.812, 95% CI 1.901-3.364, P < .001) were significant predictors of 1-year all-cause mortality. After adjusting for other confounders in multivariate analysis, NLR (HR: 01.038, 95% CI 1.022-1.054, P < .001) and RDW (HR: 1.437, 95% CI 1.346-1.535, P < .001) remained significant predictors. Restricted cubic spline analysis showed the relationship between RDW, NLR, and 1-year all-cause mortality was linear after adjusting for the covariables (P for non-linearity < 0.001). The multivariable adjusted model led to improvement in the area under the curve to 0.83 (P < .05). Nomogram was created to predict the probability of 1 year mortality. Among the laboratory indices, RDW and NLR showed the best performance for mortality risk prediction. Multivariate predictive models significantly improved risk stratification.

Damping accumulative NDAGM(1,N, α) power model and its applications

PurposeIn this essay, a new NDAGM(1,N,α) power model is recommended to resolve the hassle of the distinction between old and new information, and the complicated nonlinear traits between sequences in real behavior systems.Design/methodology/approachFirstly, the correlation aspect sequence is screened via a grey integrated correlation degree, and the damped cumulative generating operator and power index are introduced to define the new model. Then the non-structural parameters are optimized through the genetic algorithm. Finally, the pattern is utilized for the prediction of China’s natural gas consumption, and in contrast with other models.FindingsBy altering the unknown parameters of the model, theoretical deduction has been carried out on the newly constructed model. It has been discovered that the new model can be interchanged with the traditional grey model, indicating that the model proposed in this article possesses strong compatibility. In the case study, the NDAGM(1,N,α) power model demonstrates superior integrated performance compared to the benchmark models, which indirectly reflects the model’s heightened sensitivity to disparities between new and old information, as well as its ability to handle complex linear issues.Practical implicationsThis paper provides a scientifically valid forecast model for predicting natural gas consumption. The forecast results can offer a theoretical foundation for the formulation of national strategies and related policies regarding natural gas import and export.Originality/valueThe primary contribution of this article is the proposition of a grey multivariate prediction model, which accommodates both new and historical information and is applicable to complex nonlinear scenarios. In addition, the predictive performance of the model has been enhanced by employing a genetic algorithm to search for the optimal power exponent.

Machine learning approach for the development of a crucial tool in suicide prevention: The Suicide Crisis Inventory-2 (SCI-2) Short Form.

The Suicide Crisis Syndrome (SCS) describes a suicidal mental state marked by entrapment, affective disturbance, loss of cognitive control, hyperarousal, and social withdrawal that has predictive capacity for near-term suicidal behavior. The Suicide Crisis Inventory-2 (SCI-2), a reliable clinical tool that assesses SCS, lacks a short form for use in clinical settings which we sought to address with statistical analysis. To address this need, a community sample of 10,357 participants responded to an anonymous survey after which predictive performance for suicidal ideation (SI) and SI with preparatory behavior (SI-P) was measured using logistic regression, random forest, and gradient boosting algorithms. Four-fold cross-validation was used to split the dataset in 1,000 iterations. We compared rankings to the SCI-Short Form to inform the short form of the SCI-2. Logistic regression performed best in every analysis. The SI results were used to build the SCI-2-Short Form (SCI-2-SF) utilizing the two top ranking items from each SCS criterion. SHAP analysis of the SCI-2 resulted in meaningful rankings of its items. The SCI-2-SF, derived from these rankings, will be tested for predictive validity and utility in future studies.

Nomogram for predicting post-therapy recurrence in BCLC A/B hepatocellular carcinoma with Child-Pugh B cirrhosis

IntroductionThis study conducts a retrospective analysis on patients with BCLC stage A/B hepatocellular carcinoma (HCC) accompanied by Child-Pugh B cirrhosis, who underwent transarterial chemoembolization (TACE) in combination with local ablation therapy. Our goal was to uncover risk factors contributing to post-treatment recurrence and to develop and validate an innovative 1-, 3-, and 5-year recurrence free survival (RFS) nomogram.MethodsData from 255 BCLC A/B HCC patients with Child-Pugh B cirrhosis treated at Beijing You’an Hospital (January 2014 - January 2020) were analyzed using random survival forest (RSF), LASSO regression, and multivariate Cox regression to identify independent risk factors for RFS. The prognostic nomogram was then constructed and validated, categorizing patients into low, intermediate, and high-risk groups, with RFS assessed using Kaplan-Meier curves.ResultsThe nomogram, integrating the albumin/globulin ratio, gender, tumor number, and size, showcased robust predictive performance. Harrell’s concordance index (C-index) values for the training and validation cohorts were 0.744 (95% CI: 0.703–0.785) and 0.724 (95% CI: 0.644–0.804), respectively. The area under the curve (AUC) values for 1-, 3-, and 5-year RFS in the two cohorts were also promising. Calibration curves highlighted the nomogram’s reliability and decision curve analysis (DCA) confirmed its practical clinical benefits. Through meticulous patient stratification, we also revealed the nomogram’s efficacy in distinguishing varying recurrence risks.ConclusionThis study advances recurrence prediction in BCLC A/B HCC patients with Child-Pugh B cirrhosis following TACE combined with ablation. The established nomogram accurately predicts 1-, 3-, and 5-year RFS, facilitating timely identification of high-risk populations.