An interpretable deep learning model to map land subsidence hazard.

This study aims to develop and validate an interpretable deep learning (DL) model and a nomogram based on endoscopic ultrasound (EUS) images for the prediction of pathological grading in pancreatic neuroendocrine tumors (PNETs). This multicenter retrospective study included 108 patients with PNETs, who were divided into train (n = 81, internal center) and test cohorts (n = 27, external centers). Univariate and multivariate logistic regression were used for screening demographic characteristics and EUS semantic features. Deep transfer learning was employed using a pre-trained ResNet18 model to extract features from EUS images. Feature selection was conducted using the least absolute shrinkage and selection operator (LASSO), and various machine learning algorithms were utilized to construct DL models. The optimal model was then integrated with clinical features to develop a nomogram. The performance of the model was assessed using the area under the curve (AUC), calibration curves, decision curve analysis (DCA), and clinical impact curves (CIC). The nomogram, which integrates the optimal DL model (Naive Bayes) with clinical features, achieved AUC values of 0.928 (95% CI 0.849–0.981) in the train cohort and 0.882 (95% CI 0.778–0.954) in the test cohort. Calibration curves revealed minimal discrepancies between predicted and actual probabilities, with mean absolute errors of 4.5% and 6.6% in the train and test cohorts, respectively. DCA and CIC demonstrated substantial net benefit and clinical utility. The SHapley Additive exPlanations (SHAP) method provided insights into the contribution of each DL feature to the model’s predictions. This study developed and validated a novel interpretable DL model and nomogram using EUS images and machine learning, which holds promise for enhancing the clinical application of EUS in identifying PNETs’ pathological grading.

BackgroundThe subjective assessment of Breast Imaging Reporting and Data System (BI-RADS) 4A lesions leads to a high number of unnecessary biopsies, highlighting the need for more objective and accurate diagnostic tools. This study aimed to construct and validate a novel multimodal framework that integrates deep learning (DL), radiomics, and clinical and breast ultrasound (US) features to distinguish between benign and malignant breast lesions classified as BI-RADS 4A.MethodsA total of 935 patients with pathologically confirmed BI-RADS 4A lesions were retrospectively enrolled and randomly divided into a training cohort (n=654) and an internal validation cohort (n=281). An additional 488 patients were enrolled as the external validation cohort. DL and radiomics models were developed with the light gradient boosting machine (LightGBM) algorithm, a machine learning technique known for its efficiency in handling high-dimensional data. Univariate and multivariate logistic regression (LR) analyses identified significant clinical and US features. After handcrafted radiomics features, DL features, and clinical/breast US features were combined, dimensionality reduction and feature selection were performed. The integrated model was developed via LightGBM, and the SHapley Additive Explanations (SHAP) approach was applied to rank feature contributions.ResultsThe integrated model achieved robust discrimination between benign and malignant BI-RADS 4A lesions, with an area under the curve (AUC) of 0.938, 0.870, and 0.861 for the training cohort, internal validation cohort, and external validation cohort, respectively. Furthermore, in the external validation cohort, the integrated model significantly outperformed the handcrafted radiomics model (AUC =0.719; P<0.001) and the DL model (AUC =0.763; P=0.011). Decision curve analysis confirmed that the integrated model was the most clinically useful across the three cohorts.ConclusionsThe integrated model demonstrated high accuracy in differentiating between benign and malignant BI-RADS 4A lesions, exhibiting the ability to significantly reduce unnecessary biopsies while maintaining diagnostic accuracy.

Introduction: Arthritis is one of the most common chronic diseases. Early detection of arthritis and its progression can facilitate early intervention measures, lowering disease severity in patients. As electronic health records (EHR) become more accessible, this study assesses whether general health information and arthritis-related questionnaires can be used in arthritis diagnosis, without the involvement of costly imaging methods. Therefore, we created deep learning (DL) and machine learning (ML) models to explore the feasibility of combining EHR and modern computational tools to diagnose arthritis. Methods: A total of 782 arthritis patients and 4014 control patients were identified from the Osteoarthritis Initiative (OAI) – a ten-year-long observational study that included patient EHR in five time points. Six hundred variables were filtered by random forest classifier followed by manual filtering. Data were split properly to training, testing and validation set, and the training set was balanced. Sequential, nonsequential DL models, and five independent DL models for each time points were used. The accuracy, positive prevalence value (PPV), negative prevalence value (NPV), and area under curve (AUC), were assessed and compared with four classical ML models. SHAP (SHapley Additive exPlanations) summary analysis was also conducted. Results: Sequential and non-sequential deep learning models showed accuracies of ~ 0.97, and the four classical machine learning approaches showed accuracies of above 0.9. High positive and negative predicted values (&gt; 0.90) for all of the models suggested the potential clinical applicability of the model, while the SHAP analysis demonstrated its interpretability. Discussion: We tested various models and showed the ability to use machine learning methods for early diagnosis of arthritis with EHR. The models can be used as a screening tool to select susceptible patients for confirmatory tests such as X-ray and MRI. Identification of early disease states could facilitate protective measures that slow disease progression.

Explainable artificial intelligence (XAI) for interpreting the contributing factors feed into the wildfire susceptibility prediction model

Electrocardiogram signal analysis can be time-consuming, tedious, and error-prone. Therefore, automated analysis is need of time that will assist clinicians in detecting cardiac abnormalities accurately and efficiently. Recently, deep learning models have shown unprecedented progress and strong arrhythmia classification capabilities, but their deployment in the medical sector is constrained due to their “black-box” nature. This paper proposes a robust explainability method to assist in explaining the underlying decision-making process in deep neural networks (DNN) and help provide feedback on biases that would benefit in improving DNN models. To achieve these objectives, initially, a deep learning model is trained on the MIT-BIH Arrhythmia Database and their classification performance is evaluated. The classification findings are then interpreted using the post-hoc explanation methods such as the SHapley Additive exPlanations (SHAP), Local Interpretable Model-Agnostic Explanations (LIME) and Gradient-weighted Class Activation Mapping (Grad-CAM) that interpret the decision rationale. As these methods are initially proposed for image applications therefore a new masking approach is proposed to cater these post-hoc explainability methods for ECG time-series data. After evaluating these methods for ECG arrhythmia classification, several drawbacks are drawn such as they fail to locate a feature’s importance if there are multiple occurrences of the same feature in a signal and also, SHAP and LIME perform random perturbations that sometimes produce unreliable explanations. Therefore, to overcome the drawbacks associated with these post-hoc explainability methods on time-series data, a novel K-GradCam method is proposed. The proposed K-GradCam method ensemble the benefits of these gradient based and perturbation based approaches and has demonstrated advantages over SHAP, Grad-CAM and LIME in terms of interpreting the models’ decisions. To compare the the proposed technique with post-hoc explainability methods quantitatively, the confidence index of the proposed method is evaluated using dice loss. The proposed method shares 71% similarity with SHAP and 81% similarity with Grad-CAM methods; however, it shares the benefits of both methods and is computationally faster than SHAP and LIME.

PurposeTo conduct a head-to-head comparison between deep learning (DL) and radiomics models across institutions for predicting microvascular invasion (MVI) in hepatocellular carcinoma (HCC) and to investigate the model robustness and generalizability through rigorous internal and external validation.MethodsThis retrospective study included 2304 preoperative images of 576 HCC lesions from two centers, with MVI status determined by postoperative histopathology. We developed DL and radiomics models for predicting the presence of MVI using B-mode ultrasound, contrast-enhanced ultrasound (CEUS) at the arterial, portal, and delayed phases, and a combined modality (B + CEUS). For radiomics, we constructed models with enlarged vs. original regions of interest (ROIs). A cross-validation approach was performed by training models on one center’s dataset and validating the other, and vice versa. This allowed assessment of the validity of different ultrasound modalities and the cross-center robustness of the models. The optimal model combined with alpha-fetoprotein (AFP) was also validated. The head-to-head comparison was based on the area under the receiver operating characteristic curve (AUC).ResultsThirteen DL models and 25 radiomics models using different ultrasound modalities were constructed and compared. B + CEUS was the optimal modality for both DL and radiomics models. The DL model achieved AUCs of 0.802–0.818 internally and 0.667–0.688 externally across the two centers, whereas radiomics achieved AUCs of 0.749–0.869 internally and 0.646–0.697 externally. The radiomics models showed overall improvement with enlarged ROIs (P < 0.05 for both CEUS and B + CEUS modalities). The DL models showed good cross-institutional robustness (P > 0.05 for all modalities, 1.6–2.1% differences in AUC for the optimal modality), whereas the radiomics models had relatively limited robustness across the two centers (12% drop-off in AUC for the optimal modality). Adding AFP improved the DL models (P < 0.05 externally) and well maintained the robustness, but did not benefit the radiomics model (P > 0.05).ConclusionCross-institutional validation indicated that DL demonstrated better robustness than radiomics for preoperative MVI prediction in patients with HCC, representing a promising solution to non-standardized ultrasound examination procedures.

The neglect of occult lymph nodes metastasis (OLNM) is one of the pivotal causes of early non-small cell lung cancer (NSCLC) recurrence after local treatments such as stereotactic body radiotherapy (SBRT) or surgery. This study aimed to develop and validate a computed tomography (CT)-based radiomics and deep learning (DL) fusion model for predicting non-invasive OLNM. Patients with radiologically node-negative lung adenocarcinoma from two centers were retrospectively analyzed. We developed clinical, radiomics, and radiomics-clinical models using logistic regression. A DL model was established using a three-dimensional squeeze-and-excitation residual network-34 (3D SE-ResNet34) and a fusion model was created by integrating seleted clinical, radiomics features and DL features. Model performance was assessed using the area under the curve (AUC) of the receiver operating characteristic (ROC) curve, calibration curves, and decision curve analysis (DCA). Five predictive models were compared; SHapley Additive exPlanations (SHAP) and Gradient-weighted Class Activation Mapping (Grad-CAM) were employed for visualization and interpretation. Overall, 358 patients were included: 186 in the training cohort, 48 in the internal validation cohort, and 124 in the external testing cohort. The DL fusion model incorporating 3D SE-Resnet34 achieved the highest AUC of 0.947 in the training dataset, with strong performance in internal and external cohorts (AUCs of 0.903 and 0.907, respectively), outperforming single-modal DL models, clinical models, radiomics models, and radiomics-clinical combined models (DeLong test: P<0.05). DCA confirmed its clinical utility, and calibration curves demonstrated excellent agreement between predicted and observed OLNM probabilities. Features interpretation highlighted the importance of textural characteristics and the surrounding tumor regions in stratifying OLNM risk. The DL fusion model reliably and accurately predicts OLNM in early-stage lung adenocarcinoma, offering a non-invasive tool to refine staging and guide personalized treatment decisions. These results may aid clinicians in optimizing surgical and radiotherapy strategies.

Fast Deformable Image Registration for Real-Time Target Tracking During Radiation Therapy Using Cine MRI and Deep Learning

To develop and evaluate static ultrasound images of ovaries-based radiomics and deep learning models for predicting antral follicle count (AFC) and anti-Müllerian hormone (AMH). A retrospective analysis was performed on 395 infertility patients who completed AFC through transvaginal ultrasonography and AMH test in two branches of the hospital. The data set was divided into training cohort (n = 210, Branch A), internal test cohort (n = 91, Branch A), and external test cohort (n = 94, Branch B). The AFC results of static ultrasound images (AFCstatic) and dynamic scan (AFCdynamic) were analysed for correlation and consistency. The radiomics, deep learning and combined models (Rad-DL model) were developed to quantitatively predict AFC and AMH respectively, and the ordered 3-class classification (AFCdynamic≤5,6≤AFCdynamic≤19 and 20≤AFCdynamic) prediction was further carried out. There was a high degree of correlation and consistency between AFCstatic and AFCdynamic in both branches. In the quantitative prediction of AFCdynamic, the R2 of the radiomics, deep learning and Rad-DL models in the external test cohort were 0.572, 0.541, 0.583, respectively. In the quantitative prediction of AMH, the R2 of the radiomics, deep learning and Rad-DL models in the external test cohort were 0.420, 0.352, 0.509, respectively. For the ordered 3-class classification prediction, the radiomics, deep learning and Rad-DL models demonstrated a good discrimination performance with overall AUCs of 0.816 (95%CI, 0.747–0.887), 0.843 (95%CI, 0.772–0.901), and 0.881 (95%CI, 0.816–0.931) in the external test cohort, respectively. Radiomics and deep learning models developed by using static ultrasound images of ovaries may quantitatively predict AFCdynamic to a certain extent, and accurately screen infertility patients with different ovarian reserve levels. This study lays the foundation for further break through the rigid antral follicle count and explore an intelligent non-invasive ovarian reserve evaluation method.

COVID-19 is highly infectious and causes acute respiratory disease. Machine learning (ML) and deep learning (DL) models are vital in detecting disease from computerized chest tomography (CT) scans. The DL models outperformed the ML models. For COVID-19 detection from CT scan images, DL models are used as end-to-end models. Thus, the performance of the model is evaluated for the quality of the extracted feature and classification accuracy. There are four contributions included in this work. First, this research is motivated by studying the quality of the extracted feature from the DL by feeding these extracted to an ML model. In other words, we proposed comparing the end-to-end DL model performance against the approach of using DL for feature extraction and ML for the classification of COVID-19 CT scan images. Second, we proposed studying the effect of fusing extracted features from image descriptors, e.g., Scale-Invariant Feature Transform (SIFT), with extracted features from DL models. Third, we proposed a new Convolutional Neural Network (CNN) to be trained from scratch and then compared to the deep transfer learning on the same classification problem. Finally, we studied the performance gap between classic ML models against ensemble learning models. The proposed framework is evaluated using a CT dataset, where the obtained results are evaluated using five different metrics The obtained results revealed that using the proposed CNN model is better than using the well-known DL model for the purpose of feature extraction. Moreover, using a DL model for feature extraction and an ML model for the classification task achieved better results in comparison to using an end-to-end DL model for detecting COVID-19 CT scan images. Of note, the accuracy rate of the former method improved by using ensemble learning models instead of the classic ML models. The proposed method achieved the best accuracy rate of 99.39%.

Electricity load forecasting is a crucial undertaking within all the deregulated markets globally. Among the research challenges on a global scale, the investigation of deep transfer learning (DTL) in the field of electricity load forecasting represents a fundamental effort that can inform artificial intelligence applications in general. In this paper, a comprehensive study is reported regarding day-ahead electricity load forecasting. For this purpose, three sequence-to-sequence (Seq2seq) deep learning (DL) models are used, namely the multilayer perceptron (MLP), the convolutional neural network (CNN) and the ensemble learning model (ELM), which consists of the weighted combination of the outputs of MLP and CNN models. Also, the study focuses on the development of different forecasting strategies based on DTL, emphasizing the way the datasets are trained and fine-tuned for higher forecasting accuracy. In order to implement the forecasting strategies using deep learning models, load datasets from three Greek islands, Rhodes, Lesvos, and Chios, are used. The main purpose is to apply DTL for day-ahead predictions (1–24 h) for each month of the year for the Chios dataset after training and fine-tuning the models using the datasets of the three islands in various combinations. Four DTL strategies are illustrated. In the first strategy (DTL Case 1), each of the three DL models is trained using only the Lesvos dataset, while fine-tuning is performed on the dataset of Chios island, in order to create day-ahead predictions for the Chios load. In the second strategy (DTL Case 2), data from both Lesvos and Rhodes concurrently are used for the DL model training period, and fine-tuning is performed on the data from Chios. The third DTL strategy (DTL Case 3) involves the training of the DL models using the Lesvos dataset, and the testing period is performed directly on the Chios dataset without fine-tuning. The fourth strategy is a multi-task deep learning (MTDL) approach, which has been extensively studied in recent years. In MTDL, the three DL models are trained simultaneously on all three datasets and the final predictions are made on the unknown part of the dataset of Chios. The results obtained demonstrate that DTL can be applied with high efficiency for day-ahead load forecasting. Specifically, DTL Case 1 and 2 outperformed MTDL in terms of load prediction accuracy. Regarding the DL models, all three exhibit very high prediction accuracy, especially in the two cases with fine-tuning. The ELM excels compared to the single models. More specifically, for conducting day-ahead predictions, it is concluded that the MLP model presents the best monthly forecasts with MAPE values of 6.24% and 6.01% for the first two cases, the CNN model presents the best monthly forecasts with MAPE values of 5.57% and 5.60%, respectively, and the ELM model achieves the best monthly forecasts with MAPE values of 5.29% and 5.31%, respectively, indicating the very high accuracy it can achieve.

We developed a multimodal ultrasound (US) deep learning (DL) fusion model to automatically classify early fibrosis in patients with chronic kidney disease (CKD). This prospective study included patients with CKD who underwent continuous gray-scale US, superb microvascular imaging, and strain elastography from May to November 2022. According to the pathological tubular atrophy and interstitial fibrosis score, patients were divided into minimal and mild groups (affected area ≤10% and 11 − 25% of the total cortical volume, respectively). The dataset was divided into training (70%) and test (30%) sets. A DL model combining the features of the three US modes was developed to predict early fibrosis in patients with CKD. We compared these findings with the area under the receiver operating characteristic curve (AUC) of the clinical model by analyzing the receiver operating characteristic curve in the test set. The AUC of single-mode DL based on gray-scale US, superb microvascular imaging, and strain elastography was 0.682, 0.745, and 0.648, respectively, while that of the multimodal US DL model was 0.86. The accuracy, specificity, and sensitivity of the multimodal US DL model were 0.779, 0.767, and 0.796, respectively, and the negative and positive predictive values were 0.842 and 0.706, respectively. The AUC of the multimodal US DL model was significantly better than that of the single-mode DL and clinical models. The DL algorithm developed using multimodal US images can effectively predict early fibrosis in patients with CKD with significantly greater accuracy than single-mode DL or clinical models.

We aimed at comparing intratumoral and peritumoral deep learning, radiomics, and fusion models in predicting KRAS mutations in rectal cancer using endorectal ultrasound imaging. This study included 304 patients with rectal cancer from Fujian Medical University Union Hospital. The patients were randomly divided into a training group (213 patients) and a test group (91 patients) at a 7:3 ratio. Radiomics and deep learning models were established using primary tumor and peritumoral images. In the optimally performing regions-of-interest, two fusion strategies, a feature-based and a decision-based model, were employed to build the fusion models. The Shapley additive explanation (SHAP) method was used to evaluate the significance of features in the optimal radiomics, deep learning, and fusion models. The performance of each model was assessed using the area under the receiver operating characteristic curve (AUC) and decision curve analysis (DCA). In the test cohort, both the radiomics and deep learning models exhibited optimal performance with a 10-pixel patch extension, yielding AUC values of 0.824 and 0.856, respectively. The feature-based DLRexpand10_FB model attained the highest AUC (0.896) across all study sets. In addition, the DLRexpand10_FB model demonstrated excellent sensitivity, specificity, and DCA. SHAP analysis underscored the deep learning feature (DL_1) as the most significant factor in the hybrid model. The feature-based fusion model DLRexpand10_FB can be employed to predict KRAS gene mutations based on pretreatment endorectal ultrasound images of rectal cancer. The integration of peritumoral regions enhanced the predictive performance of both the radiomics and deep learning models.

Deep learning breast cancer risk models demonstrate improved accuracy compared with traditional risk models but have not been prospectively tested. We compared the accuracy of a deep learning risk score derived from the patient's prior mammogram to traditional risk scores to prospectively identify patients with cancer in a cohort due for screening. We collected data on 119 139 bilateral screening mammograms in 57 617 consecutive patients screened at 5 facilities between September 18, 2017, and February 1, 2021. Patient demographics were retrieved from electronic medical records, cancer outcomes determined through regional tumor registry linkage, and comparisons made across risk models using Wilcoxon and Pearson χ2 2-sided tests. Deep learning, Tyrer-Cuzick, and National Cancer Institute Breast Cancer Risk Assessment Tool (NCI BCRAT) risk models were compared with respect to performance metrics and area under the receiver operating characteristic curves. Cancers detected per thousand patients screened were higher in patients at increased risk by the deep learning model (8.6, 95% confidence interval [CI] = 7.9 to 9.4) compared with Tyrer-Cuzick (4.4, 95% CI = 3.9 to 4.9) and NCI BCRAT (3.8, 95% CI = 3.3 to 4.3) models (P &lt; .001). Area under the receiver operating characteristic curves of the deep learning model (0.68, 95% CI = 0.66 to 0.70) was higher compared with Tyrer-Cuzick (0.57, 95% CI = 0.54 to 0.60) and NCI BCRAT (0.57, 95% CI = 0.54 to 0.60) models. Simulated screening of the top 50th percentile risk by the deep learning model captured statistically significantly more patients with cancer compared with Tyrer-Cuzick and NCI BCRAT models (P &lt; .001). A deep learning model to assess breast cancer risk can support feasible and effective risk-based screening and is superior to traditional models to identify patients destined to develop cancer in large screening cohorts.

To develop and evaluate the performance of a deep learning model (DLM) that predicts eyes at high risk of surgical intervention for uncontrolled glaucoma based on multimodal data from an initial ophthalmology visit. Longitudinal, observational, retrospective study. 4898 unique eyes from 4038 adult glaucoma or glaucoma-suspect patients who underwent surgery for uncontrolled glaucoma (trabeculectomy, tube shunt, xen, or diode surgery) between 2013 and 2021, or did not undergo glaucoma surgery but had 3 or more ophthalmology visits. We constructed a DLM to predict the occurrence of glaucoma surgery within various time horizons from a baseline visit. Model inputs included spatially oriented visual field (VF) and optical coherence tomography (OCT) data as well as clinical and demographic features. Separate DLMs with the same architecture were trained to predict the occurrence of surgery within 3 months, within 3–6 months, within 6 months–1 year, within 1–2 years, within 2–3 years, within 3–4 years, and within 4–5 years from the baseline visit. Included eyes were randomly split into 60%, 20%, and 20% for training, validation, and testing. DLM performance was measured using area under the receiver operating characteristic curve (AUC) and precision-recall curve (PRC). Shapley additive explanations (SHAP) were utilized to assess the importance of different features. Model prediction of surgery for uncontrolled glaucoma within 3 months had the best AUC of 0.92 (95% CI 0.88, 0.96). DLMs achieved clinically useful AUC values (> 0.8) for all models that predicted the occurrence of surgery within 3 years. According to SHAP analysis, all 7 models placed intraocular pressure (IOP) within the five most important features in predicting the occurrence of glaucoma surgery. Mean deviation (MD) and average retinal nerve fiber layer (RNFL) thickness were listed among the top 5 most important features by 6 of the 7 models. DLMs can successfully identify eyes requiring surgery for uncontrolled glaucoma within specific time horizons. Predictive performance decreases as the time horizon for forecasting surgery increases. Implementing prediction models in a clinical setting may help identify patients that should be referred to a glaucoma specialist for surgical evaluation.