Non-invasive Prediction Research Articles

The value of ultrasound (US)-based radiomics in predicting axillary lymph node metastasis in patients with T1 stage breast invasive ductal carcinoma with negative axillary US results.

To evaluate the potential application value of qualitative and quantitative contrast-enhanced ultrasound (CEUS) features for predicting axillary lymph node metastasis in early-stage breast cancer, with special emphasis on area ratio. 146 patients with 146 T1 stage breast cancers were subjected to conventional ultrasound (US) and CEUS before surgeries. Logistic regression analysis was used to identify the associated risk factors and a prediction model was created to predict T1 stage breast IDCs with positive of axillary lymph node metastasis (ALNM). The diagnostic performance of the prediction model was assessed by the analysis of the area under the receiver operating characteristic curve (AUC), sensitivity, specificity, and accuracy. The value of area ratio, a quantitative CEUS feature, was evaluated. Logistic regression analysis indicated that taller than wide shape on conventional US, coarse or twist penetrating vessels, and area ratio > 1.93 on CEUS were independent risk factors for IDCs with ALNM-positive (all P < 0.05). Among all risk factors, the area ratio > 1.93 showed the highest AUC (i.e. 0.818). The prediction equation was established as follows: P = 1/1 + Exp∑ [-2.665 + 1.750×(if taller than wide shape) + 1.791×(if coarse or twist penetrating vessels) + 4.372×(if area ratio > 1.93)]. In comparison with US BI-RADS alone, the AUCs of the prediction model for both readers increased significantly (AUC: 0. 919 vs. 0.677 in reader 1, 0.919 vs. 0.707 in reader 2, both P < 0.001). Conventional US and CEUS features, especially the area ratio > 1.93 on CEUS, may be useful in the noninvasive prediction of ALNM-positive in breast IDCs.

Respiratory Severity Score (RSS): A Predictive Tool for Successful Extubation in Preterm Neonates

Introduction: The coexistence of immature lungs, poor respiratory drive, overly compliant chest wall, and surfactant deficiency in preterm neonates, leads to dependency on mechanical ventilation during the first days of life. Respiratory Severity Score (RSS) is a readily available, simple and non-invasive extubation failure prediction tool. Objective This study aims to evaluate the utility of pre-extubation RSS as a tool for predicting successful extubation among preterms. Methods and Analysis: This was a retrospective study with a duration of two years started after obtaining clearance from the Institutional Ethics Committee. All preterm neonates ventilated in NICU were included in the study. The hospital unit number was noted and the respective patient records were procured from the hospital records department. Demographic data were recorded from the files. Pre-extubation RSS was calculated. Other findings related to clinical review such as duration of invasive and non-invasive ventilation, surfactant requirement, intraventricular hemorrhage (IVH), early and late-onset sepsis, length of hospital stay (LOS), retinopathy of prematurity (ROP), bronchopulmonary dysplasia (BPD) mortality data etc. were also recorded. The age, time of extubation and weight of those babies who underwent planned extubation were also recorded. Results: The RSS and RSS/kg were significantly higher among the neonates who failed extubation ( P &lt; .001). The Area under curve (AUC) for RSS and RSS/kg was 0.921 and 0.85, respectively. There was a significantly higher proportion of morbidities such as late-onset sepsis ( P = .046), grade ≥3 IVH ( P = .004) and ROP ( P = .043) observed among the neonates with extubation failure. The prevalence of in-hospital mortality was observed only among neonates with extubation failure which was statistically significant ( P &lt; .001). Conclusion: This study shows that pre-extubation RSS is an excellent tool for predicting successful extubation and is more accurate than RSS/kg. RSS could be a useful tool for clinicians to assess readiness of extubation. However, more prospective, multicentric studies are necessary to validate the generalizability of RSS.

Development and validation of a radiomics nomogram for preoperative prediction of BRAFV600E mutation status in adult patients with craniopharyngioma.

Although craniopharyngiomas are rare benign brain tumors primarily managed by surgery, they are often burdened by a poor prognosis due to tumor recurrence and long-term morbidity. In recent years, BRAFV600E-targeted therapy has been promising, showing potential as an adjuvant or neoadjuvant approach. Therefore, we aim to develop and validate a radiomics nomogram for preoperative prediction of BRAFV600E mutation in craniopharyngiomas. A total of 398 patients with craniopharyngioma (training cohort: n = 278; validation cohort: n = 120) were retrospectively reviewed. We extracted 851 radiomic features from MRI images and adopted a support vector machine (SVM) classifier to develop a radiomic model. Also, a clinical-radiomics nomogram was constructed based on a multivariable logistic regression analysis. The performance of the nomogram was evaluated by its discrimination, calibration, and clinical utility. The radiomic model using the SVM based on three selected features showed good discrimination in the training and validation cohorts (area under the curve [AUC], 0.941 and 0.945, respectively). A higher Rad-score, smaller tumor volume, and homogenous enhancement were demonstrated as independent predictors of BRAFV600E mutation in craniopharyngioma. The nomogram incorporating the Rad-score and clinical-radiological factors exhibited AUCs of 0.958 (95% CI, 0.936-0.980) and 0.956 (95% CI, 0.921-0.991) in the training and validation cohorts, respectively, showing good clinical benefit and calibration. The radiomics nomogram could provide an accurate, non-invasive preoperative prediction of BRAFV600E mutation in craniopharyngioma and may provide potential guidance for the preoperative administration of BRAF V600E mutation inhibitors and promote personalized treatment. Further prospective validation is still needed.

Plasma acute phase proteins as potential predictors of intra-amniotic inflammation and infection in preterm premature rupture of membranes.

We aimed to investigate the potential of altered levels of various acute phase proteins (APPs) in the plasma, either used alone or in combination with ultrasound-, clinical-, and conventional blood-based tests, for predicting the risk of intra-amniotic inflammation (IAI), microbial invasion of the amniotic cavity (MIAC), histologic chorioamnionitis (HCA), and funisitis in women with preterm premature rupture of membranes (PPROM). A total of 195 consecutive pregnancies involving singleton women with PPROM (at 23 + 0-34 + 0 weeks) who underwent amniocentesis and from whom plasma samples were obtained at amniocentesis were retrospectively included in this study. Amniotic fluid (AF) was cultured to assess the MIAC and analyzed for interleukin (IL)-6 levels to define IAI (AF IL-6 level of ≥2.6 ng/mL). The plasma concentrations of hepcidin, mannose-binding lectin (MBL), pentraxin-2, retinol-binding protein 4 (RBP4), serum amyloid A1 (SAA1), and serpin A1 were determined using ELISA. Ultrasonographic cervical length (CL), neutrophil-to-lymphocyte ratio (NLR), and C-reactive protein levels were measured. IAI/MIAC was defined as IAI, MIAC, or both. Multivariate logistic regression analyses showed the following: (1) elevated plasma levels of hepcidin and SAA1 and decreased levels of RBP4 in the plasma were independently associated with IAI/MIAC and (2) decreased plasma RBP4 levels were independently associated with funisitis; however, (3) none of the plasma APPs investigated were associated with acute HCA when adjusted for baseline covariates. Using stepwise regression analysis, noninvasive prediction models comprising plasma RBP4 levels, CL, NLR, and gestational age at sampling were proposed, which provided a good prediction of IAI/MIAC and funisitis (area under the curve: 0.80 and 0.72, respectively). Hepcidin, RBP4, and SAA1 were identified as potential APP biomarkers in the plasma predictive of IAI/MIAC or funisitis in patients with PPROM. In particular, combination of these APP biomarkers with ultrasound-, clinical-, and conventional blood-based markers can significantly support the diagnosis of IAI/MIAC and funisitis.

Development and validation of non-invasive prediction models for assessing kidney histopathological activity index in lupus nephritis.

To develop and validate prediction models for estimating the activity index (AI) of kidney histopathology in lupus nephritis (LN) using clinical and laboratory data. This study used single-center data from patients with kidney biopsy-confirmed LN between January 2012 and March 2022. The development and validation datasets were temporally cut. We discriminated AI > 10 and ≤ 10 as high and mild/moderate activity status, respectively. We constructed four models for AI: Model 1 included demographic information; Model 2 additionally incorporated data of systemic conditions; Model 3 further included kidney-specific conditions; and Model 4 included all the aforementioned predictors. Logistic regression was employed in Models 1 to 3, while Model 4 utilized least absolute shrinkage and selection operator for predictor selection and model building. Internal validation was performed using 1000 bootstrap resampling, while external validation was performed in the temporal validation dataset. Both calibration and discrimination metrics were evaluated. There were 160 patients in the development dataset and 70 patients in the validation dataset. In the temporal validation, all the models achieved acceptable calibration and excellent discrimination. Model 2 which contained relatively fewer predictors achieved the highest area under the receiver operator characteristic curve of 0.86 (95% confidence interval 0.76 to 0.94). Our Model 2 incorporating demographic and systemic indicators exhibited good performance in estimating the AI of LN. We thus provide a simple yet effective algorithm to predict AI in patients with LN, potentially aiding clinicians in non-invasively assessing disease activity and guiding treatment decisions. Key Points • We developed a prediction model (Model 2) incorporating demographic and systemic indicators to predict AI in patients with LN. • The prediction model can aid clinicians in noninvasively assessing disease activity and guiding treatment decisions.

Identification of D842V mutation in gastrointestinal stromal tumors based on CT radiomics: a multi-center study

BackgroundGastrointestinal stromal tumors (GISTs) are the most common mesenchymal tumors of the gastrointestinal tract. Recent advent of tyrosine kinase inhibitors (TKIs) has significantly improved the prognosis of GIST patients. However, responses to TKI therapy can vary depending on the specific gene mutation. D842V, which is the most common mutation in platelet-derived growth factor receptor alpha exon 18, shows no response to imatinib and sunitinib. Radiomics features based on venous-phase contrast-enhanced computed tomography (CECT) have shown potential in non-invasive prediction of GIST genotypes. This study sought to determine whether radiomics features could help distinguish GISTs with D842V mutations.MethodsA total of 872 pathologically confirmed GIST patients with CECT data available from three independent centers were included and divided into the training cohort (n=487) and the external validation cohort (n=385). Clinical features including age, sex, tumor size and location were collected. Radiomics features on the largest axial image of venous-phase CECT were analyzed and a total of two radiomics features were selected after feature selection. Random forest models trained on non-radiomics features only (the non-radiomics model) and on both non-radiomics and radiomics features (the combined model) were compared.ResultsThe combined model showed better average precision (0.250 vs. 0.102, p = 0.039) and F1 score (0.253 vs. 0.155, p = 0.012) than the non-radiomics model. There was no significant difference in ROC-AUC (0.728 vs. 0.737, p = 0.836) and geometric mean (0.737 vs. 0.681, p = 0.352).ConclusionsThis study demonstrated the potential of radiomics features based on venous-phase CECT images to identify D842V mutation in GISTs. Our model may provide an alternative approach to guide TKI therapy for patients inaccessible to sequence variant testing, potentially improving treatment outcomes for GIST patients especially in resource-limited settings.

A non-invasive model for diagnosis of primary Sjogren’s disease based on salivary biomarkers, serum autoantibodies, and Schirmer’s test

BackgroundMinor salivary gland (MSG) biopsy is a critical but invasive method for the classification of primary Sjögren’s disease (pSjD). Here we aimed to identify salivary proteins as potential biomarkers and to establish a non-invasive prediction model for pSjD.MethodsLiquid chromatography-tandem mass spectrometry was conducted on whole saliva samples from patients with pSjD and non-Sjögren control subjects (non-pSjD). Proteins involved in immune processes were upregulated in the pSjD group, such as complement C3 (C3), complement factor B (CFB), clusterin (CLU), calreticulin (CALR), and neutrophil elastase (NE), which were further confirmed by ELISA. Multivariate logistic regression analyses were performed to identify markers that differentiated pSjD from non-pSjD; receiver operating characteristic (ROC) curves were constructed. A diagnostic model based on the combination of salivary biomarkers (CFB, CLU, and NE), serum autoantibodies (anti-SSA /Ro60 and anti-SSA/Ro52), and Schirmer’s test was evaluated in 186 patients (derivation cohort) with replication in 72 patients (validation cohort).ResultsIn multivariate analyses, CFB, CLU, and NE were independent predictors of pSS. A model based on the combination of salivary biomarkers (CFB, CLU, and NE), serum autoantibodies (anti-SSA and anti-Ro52), and Schirmer’s test achieved significant discrimination of pSS. In the derivation cohort, the area under curve (AUC) of the ROC was 0.930 (95% CI 0.877–0.965, P < 0.001), with a sensitivity and specificity of 84.85% and 92.45%, respectively. Notably, similar results were obtained in a validation cohort.ConclusionThe 6-biomarker panel could provide a novel non-invasive tool for the classification of pSjD.

Non-invasive prediction of interstitial fibrosis and tubular atrophy (IFTA) in lupus nephritis.

Non-invasive prediction of interstitial fibrosis and tubular atrophy (IFTA) in lupus nephritis.

Noninvasive early identification of durable clinical benefit from immune checkpoint inhibition: a prospective multicenter study (NCT04566432)

Immune checkpoint inhibitors (ICIs) have changed the treatment landscape for patients with non-small cell lung cancer (NSCLC). In spite of durable responses in some patients, many patients develop early disease progression during the ICI treatment. Thus, early identification of patients with no durable benefit would facilitate the clinical decision for these patients. In this prospective, multicenter study, 101 non-EGFR/ALK patients who received ICI treatment were enrolled after screening 328 stage III-IV NSCLC patients. At the date of cutoff, 83 patients were eligible for ICI efficacy evaluation, with 56 patients having progress-free survival (PFS) over 6 months, which was defined as durable clinical benefit (DCB). A multimodal model was established by integrating normalized bTMB, early dynamic of ctDNA and the first RECIST response. This model could robustly predict DCB with area under the curve (AUC) of 0.878, sensitivity of 79.2% at 86.4% specificity (accuracy = 80.0%). This model was further validated in the independent cohort of the DIREct-On study with AUC of 0.887, sensitivity of 94.7% at 85.3% specificity (accuracy = 90.3%). Patients with higher predict scores had substantially longer PFS than those with lower scores (training cohort: median PFS 13.6 vs 4.2 months, P < 0.001, HR = 0.24; validation cohort: median PFS 11.0 vs 2.2 months, P < 0.001, HR = 0.17). Taken together, these results demonstrate that integrating early changes of ctDNA, normalized bTMB, and the first RECIST response can provide accurate, noninvasive, and early prediction of durable benefits for NSCLC patients treated with ICIs. Further prospective studies are warranted to validate these findings and guide clinical decision-making for optimal immunotherapy in NSCLC patients.

Improvement of an Edge-IoT Architecture Driven by Artificial Intelligence for Smart-Health Chronic Disease Management.

One of the health challenges in the 21st century is to rethink approaches to non-communicable disease prevention. A solution is a smart city that implements technology to make health smarter, enables healthcare access, and contributes to all residents' overall well-being. Thus, this paper proposes an architecture to deliver smart health. The architecture is anchored in the Internet of Things and edge computing, and it is driven by artificial intelligence to establish three foundational layers in smart care. Experimental results in a case study on glucose prediction noninvasively show that the architecture senses and acquires data that capture relevant characteristics. The study also establishes a baseline of twelve regression algorithms to assess the non-invasive glucose prediction performance regarding the mean squared error, root mean squared error, and r-squared score, and the catboost regressor outperforms the other models with 218.91 and 782.30 in MSE, 14.80 and 27.97 in RMSE, and 0.81 and 0.31 in R2, respectively, on training and test sets. Future research works involve extending the performance of the algorithms with new datasets, creating and optimizing embedded AI models, deploying edge-IoT with embedded AI for wearable devices, implementing an autonomous AI cloud engine, and implementing federated learning to deliver scalable smart health in a smart city context.

Hemoglobin value prediction with bayesian optimization assisted machine learning models

This study presents a framework for predicting hemoglobin (Hb) levels utilizing Bayesian optimization-assisted machine learning models, incorporating both time-domain and frequency-domain features derived from photoplethysmography (PPG) signals. Hemoglobin, a crucial protein for oxygen and carbon dioxide transport in the blood, has levels that indicate various health conditions, including anemia and diseases affecting red blood cell production. Traditional methods for measuring Hb levels are invasive, posing potential risks and discomfort. To address this, a dataset comprising PPG signals, along with demographic data (gender and age), was analyzed to predict Hb levels accurately. Our models employ support vector regression (SVR), artificial neural networks (ANNs), classification and regression trees (CART), and ensembles of trees (EoT) optimized through Bayesian optimization algorithm. The results demonstrated that incorporating age and gender as features significantly improved model performance, highlighting their importance in Hb level prediction. Among the tested models, ANN provided the best results, involving normalized raw signals, feature selection, and reduction methods. The model achieved a mean squared error (MSE) of 1.508, root mean squared error (RMSE) of 1.228, and R-squared (R²) of 0.226. This study's findings contribute to the growing body of research on non-invasive hemoglobin prediction, offering a potential tool for healthcare professionals and patients for convenient and risk-free Hb level monitoring.

Predicting lymph node metastasis of clinical T1 non-small cell lung cancer: a brief review of possible methodologies and controversies

Non-small cell lung cancer (NSCLC) is a major subtype of lung cancer and poses a serious threat to human health. Due to the advances in lung cancer screening, more and more clinical T1 NSCLC defined as a tumor with a maximum diameter of 3cm surrounded by lung tissue or visceral pleura have been detected and have achieved favorable treatment outcomes, greatly improving the prognosis of NSCLC patients. However, the preoperative lymph node staging and intraoperative lymph node dissection patterns of operable clinical T1 NSCLC are still subject to much disagreement, as well as the heterogeneity between primary tumors and metastatic lymph nodes poses a challenge in designing effective treatment strategies. This article comprehensively describes the clinical risk factors of clinical T1 NSCLC lymph node metastasis, and its invasive and non-invasive prediction, focusing on the genetic heterogeneity between the primary tumor and the metastatic lymph nodes, which is significant for a thoroughly understanding of the biological behavior of early-stage NSCLC.

Interleukin-6 and thyroid-stimulating hormone index predict plaque stability in carotid artery stenosis: analyses by lasso-logistic regression

ObjectiveTo develop and validate a new prediction model based on the Lass-logistic regression with inflammatory serologic markers for the assessment of carotid plaque stability, providing clinicians with a reliable tool for risk stratification and decision-making in the management of carotid artery disease.MethodsIn this study, we retrospectively collected the data of the patients who underwent carotid endarterectomy (CEA) from 2019 to 2023 in Nanjing Drum Tower Hospital. Demographic characteristics, vascular risk factors, and the results of preoperative serum biochemistry were measured and collected. The risk factors for vulnerable carotid plaque were analyzed. A Lasso-logistic regression prediction model was developed and compared with traditional logistic regression models. The Akaike information criterion (AIC) and Bayesian information criterion (BIC) were used to evaluate the performance of three models.ResultsA total of 131 patients were collected in this study, including 66 (50.4%) in the vulnerable plaque group and 65 (49.6%) in the stable plaque group. The final Lasso-logistic regression model included 4 features:IL-6, TSH, TSHI, and TT4RI; AIC = 161.6376, BIC = 176.0136, both lower than the all-variable logistic regression model (AIC = 181.0881, BIC = 261.5936), and the BIC was smaller than the stepwise logistic regression model (AIC = 154.024, BIC = 179.9007). Finally, the prediction model was constructed based on the variables screened by the Lasso regression, and the model had favorable discrimination and calibration.ConclusionsThe noninvasive prediction model based on IL-6 and TSHI is a quantitative tool for predicting vulnerable carotid plaques. It has high diagnostic efficacy and is worth popularizing and applying.

Predicting the diabetic foot in the population of type 2 diabetes mellitus from tongue images and clinical information using multi-modal deep learning.

Based on the quantitative and qualitative fusion data of traditional Chinese medicine (TCM) and Western medicine, a diabetic foot (DF) prediction model was established through combining the objectified parameters of TCM and Western medicine. The ResNet-50 deep neural network (DNN) was used to extract depth features of tongue demonstration, and then a fully connected layer (FCL) was used for feature extraction to obtain aggregate features. Finally, a non-invasive DF prediction model based on tongue features was realized. Among the 391 patients included, there were 267 DF patients, with their BMI (25.2 vs. 24.2) and waist-to-hip ratio (0.953 vs. 0.941) higher than those of type 2 diabetes mellitus (T2DM) group. The diabetes (15 years vs. 8 years) and hypertension durations (10 years vs. 7.5 years) in DF patients were significantly higher than those in T2DM group. Moreover, the plantar hardness in DF patients was higher than that in T2DM patients. The accuracy and sensitivity of the multi-mode DF prediction model reached 0.95 and 0.9286, respectively. We established a DF prediction model based on clinical features and objectified tongue color, which showed the unique advantages and important role of objectified tongue demonstration in the DF risk prediction, thus further proving the scientific nature of TCM tongue diagnosis. Based on the qualitative and quantitative fusion data, we combined tongue images with DF indicators to establish a multi-mode DF prediction model, in which tongue demonstration and objectified foot data can correct the subjectivity of prior knowledge. The successful establishment of the feature fusion diagnosis model can demonstrate the clinical practical value of objectified tongue demonstration. According to the results, the model had better performance to distinguish between T2DM and DF, and by comparing the performance of the model with and without tongue images, it was found that the model with tongue images performed better.

Assessment of four-dimensional flow MRI for prediction of varices risk in cirrhotic patients.

This study aims to validate the application of abdominal four-dimensional flow magnetic resonance imaging (MRI) for cirrhotic patients and quantify its effectiveness in assessing the hemodynamic impacts of cirrhosis to evaluate varices. All consecutive patients who underwent MRIs between September 2022 and June 2023 were enrolled. Groups were divided into varicose, non-varicose, and healthy groups. ANOVA and post hoc LSD-t tests were used for statistical analysis. The correlation between hemodynamic parameters and liver function grade was evaluated using Kendall's correlation coefficient. A total of 80 patients were included (53 cirrhotic, 27 healthy). Significant disparities were found in main portal vein flow (MPV-FR), splenic vein flow (SV-FR), and vessel diameters (MPV-VD, SV-VD) among the groups (p < 0.05). MPV-FR was higher in the varicose group (24.81 ± 8.52) compared to non-varicose (19.52 ± 5.07) and healthy groups (17.26 ± 5.48). The most robust assessment of variceal risk was achieved by combining the flow rates (FRs) and VDs of MPV and SV (AUC 0.83, 95% CI 0.72-0.94). The combined indices of FRs and VDs of MPV and SV effectively predict the occurrence of varicose veins in cirrhotic patients. Question Non-invasive prediction of variceal risk is essential for the clinical management of advanced chronic cirrhosis, yet existing clinical examinations are inadequate. Findings The effective assessment of variceal risk was achieved by combining the flow rates and vessel diameters of the main portal vein and splenic vein. Clinical relevance Four-dimensional flow MRI can reveal hemodynamic changes in cirrhotic patients and assist in identifying gastroesophageal varices, serving as a marker for varices risk prediction.

Contrast-Enhanced Computed Tomography Radiomics Predicts Colony-Stimulating Factor 3 Expression and Clinical Prognosis in Ovarian Cancer

To develop a radiomics model for non-invasive prediction of colony-stimulating factor 3 (CSF3) expression in ovarian cancer (OC) and evaluate its prognostic value. We acquired clinical data, genetic information, and corresponding computed tomography (CT) scans of OC from The Cancer Genome Atlasand The Cancer Imaging Archiverepositories. We assessed the prognostic significance of CSF3 and its association with clinical features through the utilization of Kaplan-Meier analysis, univariate and multivariate Cox regression analysis, along with subgroup analysis. To explore the potential molecular mechanisms associated with CSF3 expression, we utilized gene set enrichment analysisand conducted an analysis on immune-cell infiltration. The max-relevance and min-redundancyand recursive feature elimination (RFE) algorithms were used for feature screening. The CT-based radiomics prediction model was built using support vector machine (SVM) and logistic regression (LR). The expression of CSF3 was found to be decreased in OC, and high expression of CSF3 was associated with poor overall survival. Moreover, it was noted that the expression of CSF3 exhibited a positive correlation with programmed death ligand 1 (PD-L1) and sialic acid-binding Ig-like lectin 15(SIGLEC15). Patients with high CSF3 expression exhibited a decrease in tumor necrosis factor receptor superfamily member 7 (CD27) expression. The infiltration of neutrophils increased and CD8+T cells decreased in CSF3 high expression group. The radiomics model, which utilized both LR and SVM methods, demonstrated significant clinical applicability. The expression level of CSF3 was related to the prognosis of OC. Radiomics based on CT can serve as a novel tool for forecasting prognosis.

Comment: Non-invasive prediction of post-sustained virological response hepatocellular carcinoma in hepatitis C virus.

Comment: Non-invasive prediction of post-sustained virological response hepatocellular carcinoma in hepatitis C virus.

Radiomic features of dynamic contrast-enhanced MRI can predict Ki-67 status in head and neck squamous cell carcinoma

PurposeThis study aimed to investigate the potential of radiomic features derived from dynamic contrast-enhanced MRI (DCE-MRI) in predicting Ki-67 and p16 status in head and neck squamous cell carcinoma (HNSCC). Materials and methodsA cohort of 124 HNSCC patients who underwent pre-surgery DCE-MRI were included and divided into training and test set (7:3), further subgroup analysis was performed for 104 cases with oral squamous cell carcinoma (OSCC). Radiomics features were extracted from DCE images. The least absolute shrinkage and selection operator (LASSO) was used for radiomics features selection, and receiver operating characteristics analysis for predictive performance assessment. The nomogram's performance was evaluated using decision curve analysis (DCA). ResultsTen DCE-MRI features were identified to build the predictive model of HNSCC, demonstrating excellent predictive value for Ki-67 status in both the training set (AUC of 0.943) and test set (AUC of 0.801). The nomograms based on the predictive model showed good fit in the calibration curves (p > 0.05), and DCA indicated its high clinical usefulness. In subgroup analysis of OSCC, fourteen features were selected to build the predictive model for Ki-67 status with an AUC of 0.960 in training set and 0.817 in test set. No features could be included to establish a model to predict p16 status. ConclusionThe radiomics model utilizing DCE-MRI features could effectively predict Ki-67 status in HNSCC patients, offering potential for noninvasive preoperative prediction of Ki-67 status.

A non-invasive heart rate prediction method using a convolutional approach.

The research focuses on leveraging convolutional neural networks (CNNs) to enhance the analysis of physiological signals, specifically photoplethysmogram (PPG) data which is a valuable tool for non-invasive heart rate prediction. Recognizing the global challenge of heart failure, the study seeks to provide a rapid, accurate, and non-invasive alternative to traditional, uncomfortable blood pressure cuffs. To achieve more accurate and efficient heart rate estimates, a k-fold CNN model with an optimal number of convolutional layers is employed. While the findings show promising results, the study addresses potential sources of error in cuffless PPG-based heart rate measurement, including motion artifacts and skin color variations, emphasizing the need for validation against gold standard methods. The research optimizes a CNN model with optimal layers, operating on 1D arrays of 8-s data slices and employing k-fold cross-validation to mitigate probabilistic uncertainties. Finally, the model yields a remarkable minimum absolute error (MAE) rate of 6.98 beats per minute (bpm), marking a significant 10% improvement over recent studies. The study also advances medical diagnostics and data analysis, then lays a strong foundation for developing cost-effective, reliable devices that offer a more comfortable and efficient way of predicting heart rate.

Advancing EGFR mutation subtypes prediction in NSCLC by combining 3D pretrained ConvNeXt, radiomics, and clinical features.

The aim of this study was to develop a novel approach for predicting the expression status of Epidermal Growth Factor Receptor (EGFR) and its subtypes in patients with Non-Small Cell Lung Cancer (NSCLC) using a Three-Dimensional Convolutional Neural Network (3D-CNN) ConvNeXt, radiomics features and clinical features. A total of 732 NSCLC patients with available CT imaging and EGFR expression data were included in this retrospective study. The region of interest (ROI) was manually segmented, and clinicopathological features were collected. Radiomic and deep learning features were extracted. The instances were randomly divided into training, validation, and test sets. Feature selection was performed, and XGBoost was used to create solo models and combined models to predict the presence of EGFR and subtypes mutations. The effectiveness of the models was assessed using ROC and PRC curves. We established the following models: ModelCNN, Modelradiomic, Modelclinical, ModelCNN+radiomic, ModelCNN+clinical, Modelradiomic+clinical, and ModelCNN+radiomic+clinical, which were based on deep learning features, radiomic features, clinical data and combinations of these, respectively. In predicting EGFR mutations, ModelCNN+radiomic+clinical demonstrated superior performance compared to other prediction models, achieving an AUC of 0.801. For distinguishing between EGFR subtypes ex19del and L858R, ModelCNN+radiomic reached the highest AUC value of 0.775. Both deep learning models and radiomic signature-based models offer reasonably accurate non-invasive predictions of EGFR status and its subtypes. Fusion models hold the potential to enhance noninvasive methods for predicting EGFR mutations and subtypes, presenting a more reliable prediction approach.