Training Dataset Research Articles

Automated classification of online reviews of otolaryngologists.

The study aimed to extract online comments of otolaryngologists in the 20 most populated cities in the United States from healthgrades.com, develop and validate a natural language processing (NLP) logistic regression algorithm for automated text classification of reviews into 10 categories, and compare 1- and 5-star reviews in directly-physician-related and non-physician-related categories. 1977 1-star and 12,682 5-star reviews were collected. The primary investigator manually categorized a training dataset of 324 1-star and 909 5-star reviews, while a validation subset of 100 5-star and 50 1-star reviews underwent dual manual categorization. Using scikit-learn, an NLP algorithm was trained and validated on the subsets, with F1 scores evaluating text classification accuracy against manual categorization. The algorithm was then applied to the entire dataset with comparison of review categorization according to 1- and 5-star reviews. F1 scores for NLP validation ranged between 0.71 and 0.97. Significant associations emerged between 1-star reviews and treatment plan, accessibility, wait time, office scheduling, billing, and facilities. 5-star reviews were associated with surgery/procedure, bedside manner, and staff/mid-levels. The study successfully validated an NLP text classification system for categorizing online physician reviews. Positive reviews were found to have an association with directly-physician related context. 1-star reviews were related to treatment plan, accessibility, wait time, office scheduling, billing, and facilities. This method of text classification effectively discerned the nuances of human-written text, providing valuable insights into online healthcare feedback that is scalable.Level of evidence: Level 3.

Machine learning interatomic potential with DFT accuracy for general grain boundaries in α-Fe

AbstractTo advance the development of high-strength polycrystalline metallic materials towards achieving carbon neutrality, it is essential to design materials in which the atomic level control of general grain boundaries (GGBs), which govern the material properties, is achieved. However, owing to the complex and diverse structures of GGBs, there have been no reports on interatomic potentials capable of reproducing them. This accuracy is essential for conducting molecular dynamics analyses to derive material design guidelines. In this study, we constructed a machine learning interatomic potential (MLIP) with density functional theory (DFT) accuracy to model the energy, atomic structure, and dynamics of arbitrary grain boundaries (GBs), including GGBs, in α-Fe. Specifically, we employed a training dataset comprising diverse atomic structures generated based on crystal space groups. The GGB accuracy was evaluated by directly comparing with DFT calculations performed on cells cut near GBs from nano-polycrystals, and extrapolation grades of the local atomic environment based on active learning methods for the entire nano-polycrystal. Furthermore, we analyzed the GB energy and atomic structure in α-Fe polycrystals through large-scale molecular dynamics analysis using the constructed MLIP. The average GB energy of α-Fe polycrystals calculated by the constructed MLIP is 1.57 J/m2, exhibiting good agreement with experimental predictions. Our findings demonstrate the methodology for constructing an MLIP capable of representing GGBs with high accuracy, thereby paving the way for materials design based on computational materials science for polycrystalline materials.

Abstract 4125281: Detection of Acute Myocardial Infarction and Ischemia From Lead-I ECG Using Deep Learning

Introduction: While smartwatches that acquire Lead-I electrocardiogram (ECG) have shown promise in detecting arrhythmias, the detection of ischemic events using Lead-I alone remains understudied. We aim to build and evaluate a deep-learning model to determine whether Lead-I ECG can identify ischemia/infarction in a 12-lead ECG. Methods: We developed a deep-learning model that can identify MI/ischemia using only Lead-I ECG signal as input. For this task, we utilized PTB-XL, a large publicly available dataset of 12-Lead clinical ECG with associated diagnostic labels. Our experiment utilized 8781 ECGs whose labels were validated by at-least one cardiologist with 100% confidence scores. This dataset includes 1391 ECGs that are labeled as either ischemic ST/T changes or MI, and 7390 ECGs with either “normal” labels or non-ischemic ST/T abnormalities. We split the dataset in training, validation, and holdout sets (80-10-10) with equal stratification of the diagnosis classes. The holdout test set contains 909 non-ischemic ECGs and 233 ECGs with MI/ischemia diagnosis, 106 of which are labeled as acute MI or ischemic ST/T changes and 127 labeled as other MI chronicity. The discriminatory ability of the model is evaluated with the Area Under the receiver-operating Curve (AUC), sensitivity, specificity, and the predictive values. We further analyze the model’s capability in discriminating the cohort of acute MI and ischemia diagnoses from other MI as well as from the non-ischemic cohort. Results: The model has an AUC of 0.88 for discriminating an MI/ischemia vs normal. Using a threshold corresponding to 80% sensitivity, the corresponding specificity is 79% and the negative and positive predictive values are 94% and 50% (Table-1). Although the model was trained to identify the combined class of MI/ischemia, it can differentiate acute MI and ischemia cohorts from other MI as well as from non-ischemic diagnoses with p-values<10 -4 (Table-2). Lastly, with the same threshold, the sensitivity for detecting acute inferior myocardial infarctions or inferior ischemia – a class that is difficult to detect with Lead-I alone – is 82%. Conclusions: Our deep learning model, which was trained on Lead-I ECG alone, can effectively detect acute MI and ischemia that would have been identified using a 12-lead ECG. This method presents a proof-of-concept for applications in home settings. Further prospective validation using wearable ECG monitors that record lead-I is warranted.

Abstract 4136273: A Machine Learning Approach to Simplify Risk Stratification of Patients with Atherosclerotic Cardiovascular Disease

Introduction: The 2018 AHA/ACC/Multisociety Blood Cholesterol Guideline recommends that patients with atherosclerotic cardiovascular disease (ASCVD) be classified as very high-risk (VHR) or not very high-risk (NVHR). This designation is based on the presence of > 2 major ASCVD events (recent acute coronary syndrome [ACS], history of myocardial infarction [MI], history of ischemic stroke, symptomatic peripheral artery disease [PAD]) or 1 major ASCVD event and > 2 high-risk conditions (age ≥65 years, heterozygous familial hypercholesterolemia, history of coronary revascularization, diabetes, hypertension, chronic kidney disease, current smoking, persistently elevated LDL-cholesterol [LDL-C], history of heart failure). While this approach was initially instituted to better predict future ASCVD events alone, more intensive LDL-C lowering is now recommended for those at VHR (LDL-C treatment threshold of 55 mg/dL for VHR patients vs 70 mg/dL for NVHR patients). We aimed to simplify ASCVD risk stratification in patients with clinical ASCVD using machine learning models. Methods: We used electronic health record data (Providence Health system) from 2022 to identify patients with ASCVD based on ICD-10 codes. Patient demographics, comorbidities, and laboratory data stratified patients as VHR or NVHR. Then, a classification and regression tree analysis was performed, with the outcome being VHR classification. Patients were randomly assigned to either the testing or training dataset in equal proportion. Variables in the model included age, sex, race, ethnicity, and the Guideline criteria. Model performance was assessed using misclassification rate and area under the receiver operative curve (AUC). Results: A total of 491,514 patients with ASCVD were included, of which 56% were male, 77% were white, and the mean (SD) age was 72.5 (13.0) years. Over half of patients (59%) were VHR. Those with recent ACS, history of MI, history of ischemic stroke, or PAD were classified as VHR 93%, 92%, 84%, and 91%, respectively. Presence of hypertension in these patients increased the likelihood of being VHR to 97%, 97%, 94%, and 98%, respectively. The misclassification rate and AUC for the testing set were 6.8% and 96.2%, respectively. Conclusions: Most patients with a major ASCVD event and hypertension were found to be VHR. Relying on these two factors alone can help simplify risk assessment. This approach may also help increase use of guideline-recommended LDL-C lowering therapy.

Adrenal indeterminate nodules: CT-based radiomics analysis of different machine learning models for predicting adrenal metastases in lung cancer patients

ObjectiveThere is a paucity of research using different machine learning algorithms for distinguishing between adrenal metastases and benign tumors in lung cancer patients with adrenal indeterminate nodules based on plain and biphasic-enhanced CT radiomics.Materials and MethodsThis study retrospectively enrolled 292 lung cancer patients with adrenal indeterminate nodules (training dataset, 205 (benign, 96; metastases, 109); testing dataset, 87 (benign, 42; metastases, 45)). Radiomics features were extracted from the plain, arterial, and portal CT images, respectively. The independent risk radiomics features selected by least absolute shrinkage and selection operator (LASSO) and multivariate logistic regression (LR) were used to construct the single-phase and combined-phase radiomics models, respectively, by support vector machine (SVM), decision tree (DT), random forest (RF), and LR. The independent clinical-pathological and radiological risk factors for predicting adrenal metastases selected by using univariate and multivariate LR were used to develop the traditional model. The optimal model was selected by ROC curve, and the models’ clinical values were estimated by decision curve analysis (DCA).ResultsIn the testing dataset, all SVM radiomics models showed the best robustness and efficiency, and then RF, LR, and DT models. The combined radiomics model had the best ability in predicting adrenal metastases (AUC=0.938), and then the plain (AUC=0.935), arterial (AUC=0.870), and portal radiomics model (AUC=0.851). Besides, compared to clinical-pathological-radiological model (AUC=0.870), the discriminatory capability of the plain and combined radiomics model were further improved. All radiomics models had good calibration curves and DCA showed the plain and combined radiomics models had more optimal clinical efficacy compared to other models, with the combined radiomics model having the largest net benefit.ConclusionsThe combined SVM radiomics model can non-invasively and efficiently predict adrenal metastatic nodules in lung cancer patients. In addition, the plain radiomics model with high predictive performance provides a convenient and accurate new method for patients with contraindications in enhanced CT.

Abstract 4114276: Identification of the heterogeneity in the risk of heart failure in those with left bundle branch block pattern using artificial intelligence

Introduction: The association between left bundle branch block (LBBB) and heart failure is well-established. However, the lack of a practical risk assessment tool for patients with LBBB poses a clinical challenge to the management despite the heterogeneity of the LBBB population. Hypothesis: Artificial intelligence (AI) can extract features to stratify the risk of heart failure (HF) within the heterogeneity of LBBB. Methods: All 12-lead ECGs diagnosed with LBBB recorded from 1/1/2015 to 12/31/2019 were identified in our institution. ECG data were analyzed as 2D matrices with a shape of 12×2500. The Uniform Manifold Approximation and Projection (UMAP) was used to visualize the heterogeneity of LBBB ECG. Additionally, a 2-dimensional convolutional neural network model was trained to detect LBBB ECGs associated with past HF diagnosis. The model was then applied to ECGs from an external cohort of patients with LBBB but without a history of HF. Cumulative incidences of HF admission were compared by stratifying according to tertiles of the model output (low, intermediate and high AI score groups) Results: We identified 15,124 LBBB ECGs in our institution (training dataset) and 2,463 individuals with LBBB ECGs in the external cohort (external test dataset). The UMAP projection of the ECGs revealed distinct heterogeneity in the data ( Fig A ). This heterogeneity was not fully explained by patient demographics, ECG features, or institutions. When the external cohort was divided into 3 groups according to the tertile of AI prediction, patients with high AI scores had a higher risk of HF admission (high vs low AI score group: hazard ratio (HR), 2.10; 95% confidence interval (CI), 1.66-2.65; intermediate vs low AI score group: HR, 1.39; 95%CI, 1.08-1.78: log-rank p <0.0001; Fig 1B ). This finding was unchanged after adjusting for multiple clinical characteristics and ECG parameters, including QRS duration. Conclusion: We observed significant heterogeneity in ECGs presenting LBBB, and the AI model excellently stratified the risk of heart failure admission in patients with LBBB from a single recording of 12-lead ECG, suggesting great potential for clinical utility in managing patients with LBBB.

Abstract 4145826: Atrial Fibrillation Screening During Sinus Rhythm Periods by Interrelated Systems Dynamics Analysis

Background: Previous studies have shown that AF screening in at-risk populations can reduce stroke incidence. However, non-targeted screening approaches often result in high false positive rates, placing an unnecessary burden on the healthcare system. In contrast, artificial intelligence-guided screening has been demonstrated to increase diagnostic yield in large prospective clinical trials. This approach, however, requires recording an ECG and a large-scale dataset for model training. Heart rate variability (HRV) analysis has proven effective in deciphering key heart dynamics. By analyzing HRV as interrelated dynamic systems, it may be possible to facilitate targeted AF screening using wearable devices that measure heart rate. The Koopman operator, used for data-driven modeling of interrelated dynamic systems, has been shown to accurately predict complex phenomena in chaotic systems such as climate forecasting and drug adverse reaction prediction. This is achieved by utilizing common characteristics of the systems for most model parameters, with only a small fraction of the parameters being specific to a certain system. Methods: Long (>10 hour) records from 361 individuals (AFDB, LTAFDB) and healthy individuals' datasets from PhysioNet and THEW were analyzed for inter-beat intervals. The unified dataset was then split into 94 training, 17 validation, and 250 test set patients. Recordings from the training set were used to train both the common and specific parts of the interrelated dynamic systems model for each patient, along with a shared small neural network classifying patients into low and high risk for AF based on the unique (not shared between patients) singular values of the dynamic system model. Patient-specific dynamic system models were then fitted for the validation and test sets to calculate the patient dynamic singular values, which were used to classify patients into low and high risk for AF groups. Results: Atrial fibrillation occurred in 48 of 202 (23%) patients classified as low risk and 35 of 48 (72.9%) patients classified as high risk (odds ratio 8.63, 95% CI 4.23-17.64), yielding 72.9% sensitivity with 76.2% specificity. Conclusion: In this retrospective analysis, classification of the dynamic system model singular values identified patients at high risk for atrial fibrillation from sinus rhythm period.

Finite Element Modeling and Artificial Neural Network Analyses on the Flexural Capacity of Concrete T-Beams Reinforced with Prestressed Carbon Fiber Reinforced Polymer Strands and Non-Prestressed Steel Rebars

The use of carbon fiber reinforced polymer (CFRP) strands as prestressed reinforcement in prestressed concrete (PC) structures offers an effective solution to the corrosion issues associated with prestressed steel strands. In this study, the flexural behavior of PC beams reinforced with prestressed CFRP strands and non-prestressed steel rebars was investigated using finite element modeling (FEM) and artificial neural network (ANN) methods. First, three-dimensional nonlinear FE models were developed. The FE results indicated that the predicted failure mode, load-deflection curve, and ultimate load agreed well with the previous test results. Variations in prestress level, concrete strength, and steel reinforcement ratio shifted the failure mode from concrete crushing to CFRP strand fracture. While the ultimate load generally increased with a higher prestressed level, an excessively high prestress level reduced the ultimate load due to premature fracture of CFRP strands. An increase in concrete strength and steel reinforcement ratio also contributed to a rise in the ultimate load. Subsequently, the verified FE models were utilized to create a database for training the back propagation ANN (BP-ANN) model. The ultimate moments of the experimental specimens were predicted using the trained model. The results showed the correlation coefficients for both the training and test datasets were approximately 0.99, and the maximum error between the predicted and test ultimate moments was around 8%, demonstrating that the BP-ANN method is an effective tool for accurately predicting the ultimate capacity of this type of PC beam.

Diffusion MRI with machine learning

Abstract Diffusion-weighted magnetic resonance imaging (dMRI) of the brain offers unique capabilities including noninvasive probing of tissue microstructure and structural connectivity. It is widely used for clinical assessment of disease and injury, and for neuroscience research. Analyzing the dMRI data to extract useful information for medical and scientific purposes can be challenging. The dMRI measurements may suffer from strong noise and artifacts, and may exhibit high intersession and interscanner variability in the data, as well as intersubject heterogeneity in brain structure. Moreover, the relationship between measurements and the phenomena of interest can be highly complex. Recent years have witnessed increasing use of machine learning methods for dMRI analysis. This manuscript aims to assess these efforts, with a focus on methods that have addressed data preprocessing and harmonization, microstructure mapping, tractography, and white matter tract analysis. We study the main findings, strengths, and weaknesses of the existing methods and suggest topics for future research. We find that machine learning may be exceptionally suited to tackle some of the difficult tasks in dMRI analysis. However, for this to happen, several shortcomings of existing methods and critical unresolved issues need to be addressed. There is a pressing need to improve evaluation practices, to increase the availability of rich training datasets and validation benchmarks, as well as model generalizability, reliability, and explainability concerns.

An empirical model of carbon-ion relative biological effectiveness based on the linear correlation between radiosensitivity to photons and carbon ions.

To develop an empirical model to predict carbon ion (C-ion) relative biological effectiveness (RBE). 

Approach. We used published cell survival data comprising 360 cell line/energy combinations to characterize the linear energy transfer (LET) dependence of cell radiosensitivity parameters describing the dose required to achieve a given survival level, e.g. 5% (D5%), which are linearly correlated between photon and C-ion radiations. Based on the LET response of the metrics D5% and D37%, we constructed a model containing four free parameters that predicts cells' linear quadratic model (LQM) survival curve parameters for C-ions, αCand βC, from the reference LQM parameters for photons, αXand βX, for a given C-ion LET value. We fit our model's free parameters to the training dataset and assessed its accuracy via leave-one out cross-validation. We further compared our model to the local effect model (LEM) and the microdosimetric kinetic model (MKM) by comparing its predictions against published predictions made with those models for clinically relevant LET values in the range of 23-107 keV/μm. 

Main Results. Our model predicted C-ion RBE within ±7%-15% depending on cell line and dose which was comparable to LEM and MKM for the same conditions. 

Significance. Our model offers comparable accuracy to the LEM or MKM but requires fewer input parameters and is less computationally expensive and whose implementation is so simple we provide it coded into a spreadsheet. Thus, our model can serve as a pragmatic alternative to these mechanistic models in cases where cell-specific input parameters cannot be obtained, the models cannot be implemented, or for which their computational efficiency is paramount.

Abstract 4138272: CMR Quantification of Myocardial Oxygenation Extraction Fraction Maps Using a Model-based Self-supervised Learning Network: Initial Experience in Patients with Cardiomyopathies

Background: Measurements of myocardial oxygenation extraction fraction (mOEF) maps with cardiovascular magnetic resonance (CMR) has been recently developed to assess myocardial oxygen metabolism in healthy. Recent studies focus on developing supervised machine learning methods to improve the quantification of mOEF. However, these supervised learning methods may be limited by the lack of ground-truth mOEF maps because MRI cannot directly measure mOEF with the same precision as PET. Hypothesis: Model-based self-supervised learning neural networks maybe feasible to generate mOEF maps accurately. Aims: To develop a physical model-based self-supervised learning neural network to accurately estimate mOEF for the patients with a variety of cardiomyopathies. Methods: An ECG triggered, two-dimensional asymmetric spin-echo (ASE) prepared sequence was employed to scan 71 patients with dilated, hypertrophic, or ischemic cardiomyopathies. At least 3 slices were scanned for each patient and each ASE scan was performed free-breathing with a spatial resolution of 1.7 × 1.7 × 8 mm 3 and scan time of 18 RR-interval. Acquisition of late gadolinium enhancement (LGE) was performed in each patient after the administration of 0.15 mmol/kg gadolinium agent for the visualization of myocardial infarction regions. A fully convolutional neural network with U-Net architecture was optimized with a physical model-based signal loss function (see Figure 1) on the ASE training dataset from 67 patients to generate mOEF maps. To evaluate if the method can accurately quantify abnormal mOEF, four patients with myocardial infarction were selected as the test set. Results: A previous publication reported normal mOEF levels at approximately 0.6-0.7 in healthy subjects. Figure 2 shows our proposed method generated mOEF maps from two representative patients and their LGE images for comparisons. In comparison with the remote region (mOEF = 0.70±0.10), the mOEF of myocardial regions with infarction was significantly lower (mOEF = 0.36±0.12, p < 0.001). The infarction areas predicted by the network overlapped well with the LGE. Conclusion: Quantification of mOEF using the proposed model-based self-supervised U-Net is feasible to detect abnormalities in regional myocardial oxygenation. Current study is ongoing to evaluate the proposed technique in a large cardiac patient dataset.

AI bias in lung cancer radiotherapy

Aim: In lung cancer research, AI has been trained to read chest radiographs, which has led to improved health outcomes. However, the use of AI in healthcare settings is not without its own set of drawbacks, with bias being primary among them. This study seeks to investigate AI bias in diagnosing and treating lung cancer patients. The research objectives of this study are threefold: 1) To determine which features of patient datasets are most susceptible to AI bias; 2) to then measure the extent of such bias; and 3) from the findings generated, offer recommendations for overcoming the pitfalls of AI in lung cancer therapy for the delivery of more accurate and equitable healthcare. Methods: We created a synthetic database consisting of 50 lung cancer patients using a large language model (LLM). We then used a logistic regression model to detect bias in AI-informed treatment plans. Results: The empirical results from our synthetic patient data illustrate AI bias along the lines of (1) patient demographics (specifically, age) and (2) disease classification/histology. As it concerns patient age, the model exhibited an accuracy rate of 82.7% for patients < 60 years compared to 85.7% for patients ≥ 60 years. Regarding disease type, the model was less adept in identifying treatment categories for adenocarcinoma (accuracy rate: 83.7%) than it was in predicting treatment categories for squamous cell carcinoma (accuracy rate: 92.3%). Conclusions: We address the implications of such results in terms of how they may exacerbate existing health disparities for certain patient populations. We conclude by outlining several strategies for addressing AI bias, including generating a more robust training dataset, developing software tools to detect bias, making the model’s code open access and soliciting user feedback, inviting oversight from an ethics review board, and augmenting patient datasets by synthesizing the underrepresented data.

Prediction of bone mineral density based on computer tomography images using deep learning model.

Introduction The problem of population aging is intensifying worldwide. Osteoporosis has become an important cause affecting the health status of older populations. However, the diagnosis of osteoporosis and people's understanding of it are seriously insufficient. We aim to develop a deep learning model to automatically measure bone mineral density (BMD) and improve the diagnostic rate of osteoporosis. Methods The images of 801 subjects with 2080 vertebral bodies who underwent abdominal paired computer tomography (CT) and quantitative computer tomography (QCT) scanning was retrived from June 2020 to January 2022. The BMD of T11-L4 vertebral bodies was measured by QCT. Developing a multi-stage deep learning-based model to simulate the segmentation of the vertebral body and predict BMD. The subjects were randomly divided into training dataset, validation dataset and test dataset. Analyze the fitting effect between the BMD measured by the model and the standard BMD by QCT. Accuracy, precision, recall and f1- score were used to analyze the diagnostic performance according to categorization criterion measured by QCT. Results 410 males (51.2%) and 391 females (48.8%) were included in this study. Among them, there were 154 (19.2%) males and 118 (14.7%) females aged 23-44; 182 (22.7%) males and 205 (25.6%) females aged 45-64; 74 (9.2%) males and 68 (8.5%) females aged 65-84. The number of vertebral bodies in the training dataset, the validation dataset, and the test dataset was 1433, 243, 404, respectively. In each dataset, the BMD of males and females decreases with age. There was a significant correlation between the BMD measured by the model and QCT, with the coefficient of determination (r2) 0.95-0.97. The diagnostic accuracy based on the model in the three datasets was 0.88, 0.91, and 0.91, respectively. Conclusion The proposed multi-stage deep learning-based model can achieve automatic measurement of vertebral BMD and performed well in the prediction of osteoporosis.

Machine-learning-based prediction of cardiovascular events for hyperlipidemia population with lipid variability and remnant cholesterol as biomarkers.

Dyslipidemia poses a significant risk for the progression to cardiovascular diseases. Despite the identification of numerous risk factors and the proposal of various risk scales, there is still an urgent need for effective predictive models for the onset of cardiovascular diseases in the hyperlipidemic population, which are essential for the prevention of CVD. We carried out a retrospective cohort study with 23,548 hyperlipidemia patients in Shenzhen Health Information Big Data Platform, including 11,723 CVD onset cases in a 3-year follow-up. The population was randomly divided into 70% as an independent training dataset and remaining 30% as test set. Four distinct machine-learning algorithms were implemented on the training dataset with the aim of developing highly accurate predictive models, and their performance was subsequently benchmarked against conventional risk assessment scales. An ablation study was also carried out to analyze the impact of individual risk factors to model performance. The non-linear algorithm, LightGBM, excelled in forecasting the incidence of cardiovascular disease within 3years, achieving an area under the 'receiver operating characteristic curve' (AUROC) of 0.883. This performance surpassed that of the conventional logistic regression model, which had an AUROC of 0.725, on identical datasets. Concurrently, in direct comparative analyses, machine-learning approaches have notably outperformed the three traditional risk assessment methods within their respective applicable populations. These include the Framingham cardiovascular disease risk score, 2019 ESC/EAS guidelines for the management of dyslipidemia and the 2016 Chinese recommendations for the management of dyslipidemia in adults. Further analysis of risk factors showed that the variability of blood lipid levels and remnant cholesterol played an important role in indicating an increased risk of CVD. We have shown that the application of machine-learning techniques significantly enhances the precision of cardiovascular risk forecasting among hyperlipidemic patients, addressing the critical issue of disease prediction's heterogeneity and non-linearity. Furthermore, some recently-suggested biomarkers, including blood lipid variability and remnant cholesterol are also important predictors of cardiovascular events, suggesting the importance of continuous lipid monitoring and healthcare profiling through big data platforms.

Combining LiDAR, SAR, and DEM Data for Estimating Understory Terrain Using Machine Learning-Based Methods

Currently, precise estimation of understory terrain faces numerous technical obstacles and challenges that are difficult to overcome. To address this problem, this paper combines LiDAR, SAR, and DEM data to estimate understory terrain. The high multivariable-precision spaceborne LiDAR ICESat-2 data, validated by the NEON, are divided into training and validation sets. The training dataset is used as a dependent variable, the SRTM DEM and Sentinel-1 SAR data are regarded as independent variables, a total of 13 feature parameters with high contributions are extracted to construct a Multiple Linear Regression model (MLR), BAGGING model, Random Forest model (RF), and Long Short-Term Memory model (LSTM). The results indicate that the RF model exhibits the highest accuracy among the four models, with R2 = 0.999, RMSE = 0.701 m, and MAE = 0.249 m. Then, based on the RF model, the understory terrain at the regional scale is generated, and an accuracy assessment is performed using the validation dataset, yielding R2 = 0.999, RMSE = 0.847 m, and MAE = 0.517 m. Furthermore, this paper quantitatively analyzes the effects of slope, vegetation coverage, and canopy height on the estimation accuracy of understory terrain. The results show that as slope, and canopy height increase, the estimation accuracy of the RF model for understory terrain gradually decreases. The accuracy of the understory terrain estimated by the RF model is relatively stable and not easily affected by slope, vegetation coverage, and canopy height. The research on the estimation of understory terrain holds significant practical implications for forest resource management, ecological conservation, and biodiversity protection, as well as natural disaster prevention.

A nomogram model to predict recurrence of early-onset endometrial cancer after resection based on clinical parameters and immunohistochemical markers: a multi-institutional study

ObjectiveThis study aimed to investigate the prognosis value of the clinical parameters and immunohistochemical markers of patients with early-onset endometrial cancer (EC) and establish a nomogram to accurately predict recurrence-free survival (RFS) of early-onset EC after resection.MethodsA training dataset containing 458 patients and an independent testing dataset consisting of 170 patients were employed in this retrospective study. The independent risk factors related to RFS were confirmed using Cox regression models. A nomogram model was established to predict RFS at 3 and 5 years post-hysterectomy. The C-index, area under the curve (AUC) of the receiver operating characteristic (ROC) curve, and calibration curve were calculated to assess the predictive accuracy of the nomogram.ResultsIn all early-onset EC patients, more than half (368/628, 58.6%) were diagnosed in the age range of 45-49 years. Meanwhile, the recurrence rate of early-onset EC is approximately 10.8%. Multivariate Cox regression analyses showed that histological subtype, FIGO stage, myometrial invasion, lymphovascular space invasion (LVSI), P53 expression, and MMR status were independent prognostic factors related to RFS (all P < 0.05) and established the nomogram predicting 3- and 5-year RFS. The C-index and calibration curves of the nomogram demonstrated a close correlation between predicted and actual RFS. Patients were divided into high- and low-risk groups according to the model of RFS.ConclusionsCombining clinical parameters and immunohistochemical markers, we developed a robust nomogram to predict RFS after surgery for early-onset EC patients. This nomogram can predict prognosis well and guide treatment decisions.

NIMG-71. T2 RELAXIVITY AND DIFFUSION TENSOR METRICS IN THE PREDICTION OF PROGRESSION-FREE SURVIVAL IN NON-RECURRENT HIGH GRADE GLIOMA

Abstract High grade glioma (HGG) is the most common primary brain tumour; we aimed to evaluate T2 relaxivity and diffusion tensor imaging (DTI) metrics in the prediction of progression-free survival. Seventy-two non-recurrent HGG subjects (IDH: MUT=10, Unknown=20) followed at our institution after surgery and combined chemo-radiotherapy were included. All data used were prior to progression as per RANO criteria; only the first 2 years of follow up was included in the analyses. T2 relaxation rates (1/T2) were calculated voxel-wise assuming mono-exponential decay. Diffusion data were corrected from two opposite phase encode acquisitions and DTI metrics evaluated included FA, MD, p, q, and L1. Whole- and half-brain (ipsilateral/contralateral) masks were generated by brain extraction and registration; data extracted from these masks were analyzed using python 3.8 with the exception of numerous histogram features extracted using fsl tools. Radiomics features were also extracted using the pyradiomics package. The lofo-importance package was used for feature selection. A multi-level approach was used to reduce the initial 32,000 MRI features to a final number of 443; from these, the 50 highest-ranked features were used further. Within-subject feature engineering was then performed including 1–4 step forward and backward lag, lag difference, as well as area, maximum, mean and cumulative sum. Non-MRI features included available demographic, histological and treatment-related data. A LightGBM classification model was used for the prediction of progression within 90 days; the data were grouped by subject, such that each subject was present only in training or testing datasets. Hyperparameters were tuned using randomized search cross validation, and the final model using 5-fold cross validation reached a mean ROC-AUC of 0.89. Our results show the potential of quantitative MRI features derived from qualitatively unremarkable images within the progress-free interval to identify subjects at risk of imminent progression.

DeepMEns: an ensemble model for predicting sgRNA on-target activity based on multiple features.

The CRISPR/Cas9 system developed from Streptococcus pyogenes (SpCas9) has high potential in gene editing. However, its successful application is hindered by the considerable variability in target efficiencies across different single guide RNAs (sgRNAs). Although several deep learning models have been created to predict sgRNA on-target activity, the intrinsic mechanisms of these models are difficult to explain, and there is still scope for improvement in prediction performance. To overcome these issues, we propose an ensemble interpretable model termed DeepMEns based on deep learning to predict sgRNA on-target activity. By using five different training and validation datasets, we constructed five sub-regressors, each comprising three parts. The first part uses one-hot encoding, wherein 0-1 representation of the secondary structure is used as the input to the convolutional neural network (CNN) with Transformer encoder. The second part uses the DNA shape feature matrix as the input to the CNN with Transformer encoder. The third part uses positional encoding feature matrices as the proposed input into a long short-term memory network with an attention mechanism. These three parts are concatenated through the flattened layer, and the final prediction result is the average of the five sub-regressors. Extensive benchmarking experiments indicated that DeepMEns achieved the highest Spearman correlation coefficient for 6 of 10 independent test datasets as compared to previous predictors, this finding confirmed that DeepMEns can accomplish state-of-the-art performance. Moreover, the ablation analysis also indicated that the ensemble strategy may improve the performance of the prediction model.

SFNet: Spatial and Frequency Domain Networks for Wide-Field OCT Angiography Retinal Vessel Segmentation.

Automatic segmentation of blood vessels in fundus images is important to assist ophthalmologists in diagnosis. However, automatic segmentation for Optical Coherence Tomography Angiography (OCTA) blood vessels has not been fully investigated due to various difficulties, such as vessel complexity. In addition, there are only a few publicly available OCTA image data sets for training and validating segmentation algorithms. To address these issues, we constructed a wild-field retinal OCTA segmentation data set, the Retinal Vessels Images in OCTA (REVIO) dataset. Second, we propose a new retinal vessel segmentation network based on spatial and frequency domain networks (SFNet). The proposed model are tested on three benchmark data sets including REVIO, ROSE and OCTA-500. The experimental results show superior performance on segmentation tasks compared to the representative methods.

Automatic Classification of Focal Liver Lesions Based on Multi-Sequence MRI.

Accurate and automated diagnosis of focal liver lesions is critical for effective radiological practice and patient treatment planning. This study presents a deep learning model specifically developed for classifying focal liver lesions across eight different MRI sequences, categorizing them into seven distinct classes. The model includes a feature extraction module that derives multi-level representations of the lesions, a feature fusion attention module to integrate contextual information from the various sequences, and an attention-guided data augmentation module to enrich the training dataset. The proposed model achieved a patient-wise classification accuracy of 0.9302 and a lesion-wise accuracy of 0.8592, along with an F1-score of 0.8395, a recall of 0.8296, and a precision of 0.8551. These findings demonstrate the effectiveness of combining multi-sequence MRI with advanced deep learning methodologies, providing a robust tool to support radiologists in accurately classifying liver lesions in clinical settings.