Liver cancer ranks among the most lethal cancers. Hepatocellular carcinoma (HCC) is the most common type of primary liver cancer and better diagnostic tools are needed to diagnose patients at risk. The aim is to develop a machine learning algorithm that enhances the sensitivity and specificity of the Contrast-Enhanced Ultrasound Liver Imaging Reporting and Data System (CEUS-LIRADS) in classifying indeterminate at-risk liver nodules (LR-M, LR-3, LR-4) as HCC or non-HCC. Our study includes patients at risk for HCC with untreated indeterminate focal liver observations detected on US or contrast-enhanced CT or MRI performed as part of their clinical standard of care from January 2018 to November 2022. Recursive partitioning was used to improve HCC diagnosis in indeterminate at-risk nodules. Demographics, blood biomarkers, and CEUS imaging features were evaluated as potential predictors for the algorithm to classify nodules as HCC or non-HCC. We evaluated 244 indeterminate liver nodules from 224 patients (mean age 62.9 y). Of the nodules, 73.2% (164/224) were from males. The algorithm was trained on a random 2/3 partition of 163 liver nodules and correctly reclassified more than half of the HCC liver nodules previously categorized as indeterminate in the independent 1/3 test partition of 81 liver nodules, achieving a sensitivity of 56.3% (95% CI: 42.0%, 70.2%) and specificity of 93.9% (95% CI: 84.4%, 100.0%). Machine learning was applied to the multicenter, multinational study of CEUS LI-RADS indeterminate at-risk liver nodules and correctly diagnosed HCC in more than half of the HCC nodules.

Epilepsy affects over 50 million people worldwide, with anti-seizure medications (ASMs) as the primary treatment for seizure control. However, ASM selection remains a "trial and error" process due to the lack of reliable predictors of effectiveness and tolerability. While machine learning approaches have been explored, existing models are limited to predicting outcomes only for ASMs encountered during training and have not leveraged recent biomedical foundation models for this task. This work investigates ASM outcome prediction using only patient MRI scans and reports. Specifically, we leverage biomedical vision-language foundation models and introduce a novel contextualized instruction-tuning framework that integrates expert-built knowledge trees of MRI entities to enhance their performance. Additionally, by training only on the four most commonly prescribed ASMs, our framework enables generalization to predicting outcomes and effectiveness for unseen ASMs not present during training. We evaluate our instruction-tuning framework on two retrospective epilepsy patient datasets, achieving an average AUC of 71.39 and 63.03 in predicting outcomes for four primary ASMs and three completely unseen ASMs, respectively. Our approach improves the AUC by 5.53 and 3.51 compared to standard report-based instruction tuning for seen and unseen ASMs, respectively. Our code, MRI knowledge tree, prompting templates, and TREE-TUNE generated instruction-answer tuning dataset are available at the link.

BackgroundBreast microcalcification diagnostics are challenging due to their subtle presentation, overlapping with benign findings, and high inter-reader variability, often leading to unnecessary biopsies. While deep learning (DL) models - particularly deep convolutional neural networks (DCNNs) - have shown potential to improve diagnostic accuracy, their clinical application remains limited by the need for large annotated datasets and the "black box" nature of their decision-making. PurposeTo develop and validate a deep learning model (DCNN) using a double transfer learning (d-TL) strategy for classifying suspected mammographic microcalcifications, with explainable AI (XAI) techniques to support model interpretability. Material and methodsA retrospective dataset of 396 annotated regions of interest (ROIs) from full-field digital mammography (FFDM) images of 194 patients who underwent stereotactic vacuum-assisted biopsy at the Womens Hospital radiological department, Helsinki University Hospital, was collected. The dataset was randomly split into training and test sets (24% test set, balanced for benign and malignant cases). A ResNeXt-based DCNN was developed using a d-TL approach: first pretrained on ImageNet, then adapted using an intermediate mammography dataset before fine-tuning on the target microcalcification data. Saliency maps were generated using Gradient-weighted Class Activation Mapping (Grad-CAM) to evaluate the visual relevance of model predictions. Diagnostic performance was compared to a radiologists BI-RADS-based assessment, using final histopathology as the reference standard. ResultsThe ensemble DCNN achieved an area under the ROC curve (AUC) of 0.76, with 65% sensitivity, 83% specificity, 79% positive predictive value (PPV), and 70% accuracy. The radiologist achieved an AUC of 0.65 with 100% sensitivity but lower specificity (30%) and PPV (59%). Grad-CAM visualizations showed consistent activation of the correct ROIs, even in misclassified cases where confidence scores fell below the threshold. ConclusionThe DCNN model utilizing d-TL achieved performance comparable to radiologists, with higher specificity and PPV than BI-RADS. The approach addresses data limitation issues and may help reduce additional imaging and unnecessary biopsies.

Considerable socioeconomic disparities exist among pediatric traumatic brain injury (TBI) patients. This study aims to analyze the effects of social determinants of health on head injury outcomes and to create a novel machine-learning algorithm (MLA) that incorporates socioeconomic factors to predict the likelihood of a positive or negative trauma-related finding on head computed tomography (CT). A cohort of blunt trauma patients under age 15 who presented to the largest safety net hospital in New England between January 2006 and December 2013 (n=211) was included in this study. Patient socioeconomic data such as race, language, household income, and insurance type were collected alongside other parameters like Injury Severity Score (ISS), age, sex, and mechanism of injury. Multivariable analysis was performed to identify significant factors in predicting a positive head CT outcome. The cohort was split into 80% training (168 samples) and 20% testing (43 samples) datasets using stratified sampling. Twenty-two multi-parametric MLAs were trained with 5-fold cross-validation and hyperparameter tuning via GridSearchCV, and top-performing models were evaluated on the test dataset. Significant factors associated with pediatric head CT outcome included ISS, age, and insurance type (p<0.05). The age of the subjects with a clinically relevant trauma-related head CT finding (median= 1.8 years) was significantly different from the age of patients without such findings (median= 9.1 years). These predictors were utilized to train the machine learning models. With ISS, the Fine Gaussian SVM achieved the highest test AUC (0.923), with accuracy=0.837, sensitivity=0.647, and specificity=0.962. The Coarse Tree yielded accuracy=0.837, AUC=0.837, sensitivity=0.824, and specificity=0.846. Without ISS, the Narrow Neural Network performed best with accuracy=0.837, AUC=0.857, sensitivity=0.765, and specificity=0.885. Key predictors of clinically relevant head CT findings in pediatric TBI include ISS, age, and social determinants of health, with children under 5 at higher risk. A novel Fine Gaussian SVM model outperformed other MLA, offering high accuracy in predicting outcomes. This tool shows promise for improving clinical decisions while minimizing radiation exposure in children. TBI = Traumatic Brain Injury; ISS = Injury Severity Score; MLA = Machine Learning Algorithm; CT = Computed Tomography; AUC = Area Under the Curve.

The core role of medical images in disease diagnosis makes their quality directly affect the accuracy of clinical judgment. However, due to factors such as low-dose scanning, equipment limitations and imaging artifacts, medical images are often accompanied by non-uniform noise interference, which seriously affects structure recognition and lesion detection. This paper proposes a medical image adaptive denoising model (MI-ND) that integrates multi-scale convolutional and Transformer architecture, introduces a noise level estimator (NLE) and a noise adaptive attention module (NAAB), and realizes channel-spatial attention regulation and cross-modal feature fusion driven by noise perception. Systematic testing is carried out on multimodal public datasets. Experiments show that this method significantly outperforms the comparative methods in image quality indicators such as PSNR, SSIM, and LPIPS, and improves the F1 score and ROC-AUC in downstream diagnostic tasks, showing strong prac-tical value and promotional potential. The model has outstanding benefits in structural recovery, diagnostic sensitivity, and cross-modal robustness, and provides an effective solution for medical image enhancement and AI-assisted diagnosis and treatment.

Generative artificial intelligence models facilitate open-data sharing by proposing synthetic data as surrogates of real patient data. Despite the promise for healthcare, some of these models are susceptible to patient data memorization, where models generate patient data copies instead of novel synthetic samples, resulting in patient re-identification. Here we assess memorization in unconditional latent diffusion models by training them on a variety of datasets for synthetic data generation and detecting memorization with a self-supervised copy detection approach. We show a high degree of patient data memorization across all datasets, with approximately 37.2% of patient data detected as memorized and 68.7% of synthetic samples identified as patient data copies. Latent diffusion models are more susceptible to memorization than autoencoders and generative adversarial networks, and they outperform non-diffusion models in synthesis quality. Augmentation strategies during training, small architecture size and increasing datasets can reduce memorization, while overtraining the models can enhance it. These results emphasize the importance of carefully training generative models on private medical imaging datasets and examining the synthetic data to ensure patient privacy.

Microinvasive adenocarcinoma (MIA) and invasive adenocarcinoma (IAC) require distinct treatment strategies and are associated with different prognoses, underscoring the importance of accurate differentiation. This study aims to develop a predictive model that combines radiomics and deep learning to effectively distinguish between MIA and IAC. In this retrospective study, 252 pathologically confirmed cases of ground-glass nodules (GGNs) were included, with 177 allocated to the training set and 75 to the testing set. Radiomics, 2D deep learning, and 3D deep learning models were constructed based on CT images. In addition, two fusion strategies were employed to integrate these modalities: early fusion, which concatenates features from all modalities prior to classification, and late fusion, which ensembles the output probabilities of the individual models. The predictive performance of all five models was evaluated using the area under the receiver operating characteristic curve (AUC), and DeLong's test was performed to compare differences in AUC between models. The radiomics model achieved an AUC of 0.794 (95% CI: 0.684-0.898), while the 2D and 3D deep learning models achieved AUCs of 0.754 (95% CI: 0.594-0.882) and 0.847 (95% CI: 0.724-0.945), respectively, in the testing set. Among the fusion models, the late fusion strategy demonstrated the highest predictive performance, with an AUC of 0.898 (95% CI: 0.784-0.962), outperforming the early fusion model, which achieved an AUC of 0.857 (95% CI: 0.731-0.936). Although the differences were not statistically significant, the late fusion model yielded the highest numerical values for diagnostic accuracy, sensitivity, and specificity across all models. The fusion of radiomics and deep learning features shows potential in improving the differentiation of MIA and IAC in GGNs. The late fusion strategy demonstrated promising results, warranting further validation in larger, multicenter studies.

Alzheimer's disease is a progressive neurodegenerative disorder marked by cognitive decline, memory loss, and behavioral changes. Early diagnosis, particularly identifying Early Mild Cognitive Impairment (EMCI), is vital for managing the disease and improving patient outcomes. Detecting EMCI is challenging due to the subtle structural changes in the brain, making precise slice selection from MRI scans essential for accurate diagnosis. In this context, the careful selection of specific MRI slices that provide distinct anatomical details significantly enhances the ability to identify these early changes. The chief novelty of the study is that instead of selecting all slices, an approach for identifying the important slices is developed. The ADNI-3 dataset was used as the dataset when running the models for early detection of Alzheimer's disease. Satisfactory results have been obtained by classifying with deep learning models, vision transformers (ViT) and by adding new structures to them, together with the model proposal. In the results obtained, while an accuracy of 99.45% was achieved with EfficientNetB2 + FPN in AD vs. LMCI classification from the slices selected with SSIM, an accuracy of 99.19% was achieved in AD vs. EMCI classification, in fact, the study significantly advances early detection by demonstrating improved diagnostic accuracy of the disease at the EMCI stage. The results obtained with these methods emphasize the importance of developing deep learning models with slice selection integrated with the Vision Transformers architecture. Focusing on accurate slice selection enables early detection of Alzheimer's at the EMCI stage, allowing for timely interventions and preventive measures before the disease progresses to more advanced stages. This approach not only facilitates early and accurate diagnosis, but also lays the groundwork for timely intervention and treatment, offering hope for better patient outcomes in Alzheimer's disease. The study is finally evaluated by a statistical significance test.

Stress is a natural response of the body to perceived threats, and it can have both positive and negative effects on brain hemodynamics. Stress-induced changes in pupil and eyelid size/shape have been used as a biomarker in several fMRI studies. However, there were limited knowledges regarding changes in behavior of pupil and eyelid dynamics, particularly on animal models. In the present study, the pupil and eyelid dynamics were carefully investigated and characterized in a newly developed awake rodent fMRI protocol. Leveraging deep learning techniques, the mouse pupil and eyelid diameters were extracted and analyzed during different training and imaging phases in the present project. Our findings demonstrate a consistent downwards trend in pupil and eyelid dynamics under a meticulously designed training protocol, suggesting that the behaviors of the pupil and eyelid can be served as reliable indicators of stress levels and motion artifacts in awake fMRI studies. The current recording platform not only enables the facilitation of awake animal MRI studies but also highlights its potential applications to numerous other research areas, owing to the non-invasive nature and straightforward implementation.

Cervical cancer is a leading cause of death from malignant tumors in women, and accurate evaluation of occult lymph node metastasis (OLNM) is crucial for optimal treatment. This study aimed to develop several predictive models-including Clinical model, Radiomics models (RD), Deep Learning models (DL), Radiomics-Deep Learning fusion models (RD-DL), and a Clinical-RD-DL combined model-for assessing the risk of OLNM in cervical cancer patients.The study included 130 patients from Center 1 (training set) and 55 from Center 2 (test set). Clinical data and imaging sequences (T1, T2, and DWI) were used to extract features for model construction. Model performance was assessed using the DeLong test, and SHAP analysis was used to examine feature contributions. Results showed that both the RD-combined (AUC = 0.803) and DL-combined (AUC = 0.818) models outperformed single-sequence models as well as the standalone Clinical model (AUC = 0.702). The RD-DL model yielded the highest performance, achieving an AUC of 0.981 in the training set and 0.903 in the test set. Notably, integrating clinical variables did not further improve predictive performance; the Clinical-RD-DL model performed comparably to the RD-DL model. SHAP analysis showed that deep learning features had the greatest impact on model predictions. Both RD and DL models effectively predict OLNM, with the RD-DL model offering superior performance. These findings provide a rapid, non-invasive clinical prediction method.