Computerized Medical Imaging and Graphics最新文献

Semi-supervised learning is crucial for medical image segmentation due to the scarcity of labeled data. However, existing methods that combine consistency regularization and pseudo-labeling often suffer from inadequate feature representation, suboptimal subnetwork disagreement, and noisy pseudo-labels. To address these limitations, this paper proposed a novel Confidence-Calibrated Contrastive Mean Teacher (C3MT) framework. First, C3MT introduces a Contrastive Learning-based co-training strategy, where an adaptive disagreement adjustment mechanism dynamically regulates the divergence between student models. This not only preserves representation diversity but also stabilizes the training process. Second, C3MT introduces a Confidence-Calibrated and Category-Aligned uncertainty-guided region mixing strategy. The confidence-calibrated mechanism filters out unreliable pseudo-labels, whereas the category-aligned design restricts region swapping to patches of the same semantic category, preserving anatomical coherence and preventing semantic inconsistency in the mixed samples. Together, these components significantly enhance feature representation, training stability, and segmentation quality, especially in challenging low-annotation scenarios. Extensive experiments on ACDC, Synapse, and LA datasets show that C3MT consistently outperforms recent state-of-the-art methods. For example, on the ACDC dataset with 20% labeled data, C3MT achieves up to a 4.3% improvement in average Dice score and a reduction in HD95 of more than 1.0 mm compared with strong baselines. The implementation is publicly available at https://github.com/l1654485/C3MT.

Machine learning methods based on imaging and other clinical data have shown great potential for improving the early and accurate diagnosis of Alzheimer's disease (AD). However, for most deep learning models, especially those including high-dimensional imaging data, the decision-making process remains largely opaque which limits clinical applicability. Explainable Boosting Machines (EBMs) are inherently interpretable machine learning models, but are typically applied to low-dimensional data. In this study, we propose an interpretable machine learning framework that integrates data-driven feature extraction based on Convolutional Neural Networks (CNNs) with the intrinsic transparency of EBMs for AD diagnosis and prediction. The framework enables interpretation at both the group-level and individual-level by identifying imaging biomarkers contributing to predictions. We validated the framework on the Alzheimer's Disease Neuroimaging Initiative (ADNI) cohort, achieving an area-under-the-curve (AUC) of 0.969 for AD vs. control classification and 0.750 for MCI conversion prediction. External validation was performed on an independent cohort, yielding AUCs of 0.871 for AD vs. subjective cognitive decline (SCD) classification and 0.666 for MCI conversion prediction. The proposed framework achieves performance comparable to state-of-the-art black-box models while offering transparent decision-making, a critical requirement for clinical translation. Our code is available at: https://gitlab.com/radiology/neuro/interpretable_ad_classification.

A considerable increase in the incidence of heart failure (HF) has recently posed major challenges to the medical field, underscoring the urgent need for early detection and intervention. Medical vision-and-language pretraining learns general representation from medical images and texts and shows great prospects in multimodal data-based diagnosis. During model pretraining, patient images may contain multiple symptoms simultaneously and medical research faces a considerable challenge from noisy labels owing to factors such as differences among experts and machine-extracted labels. Furthermore, parameter-efficient fine-tuning (PEFT) is important for promoting model development. To address these challenges, we developed a multimodal vision-language pretrained model for HF, called HF-VLP. In particular, label calibration loss is adopted to solve the labeling noise problem of multisource pretraining data. During pretraining, the data labels are calibrated in real time using the correlation between the labels and prediction confidence. Considering the efficient migratability of the pretrained model, a PEFT method called decomposed singular value weight-decomposed low-rank adaptation is developed. It can learn the downstream data distribution quickly by fine-tuning <1% of the parameters to obtain a better diagnosis rate than zero-shot. Simultaneously, the developed model fuses chest X-ray image features and radiology report features through the dynamic fusion graph module, enhancing the interaction and expression ability of multimodal information. The validity of the model is verified on multiple medical datasets. The average AUC of multisymptom prediction in the Open-I dataset and the hospital dataset PPL-CXR reached 83.67% and 91.28%, respectively. The developed model can accurately classify the symptoms of patients, thereby assisting doctors in diagnosis.

As a fundamental task, medical image segmentation plays a crucial role in various clinical applications. In recent years, deep learning-based segmentation methods have achieved significant success. However, these methods typically represent the image and objects within it as grid-structural data, while insufficient attention is given to relationships between the objects to segment. To address this issue, we propose a novel model called MedSegViG, which consists of a hierarchical encoder based on Vision GNN (ViG) and a hybrid feature decoder. During the segmentation process, our model first represents the image as a graph and then utilizes the encoder to extract multi-level graph features and image features. Finally, our hybrid feature decoder fuses these features to generate the final segmentation map. To validate the effectiveness of the proposed model, we conducted extensive experiments on six datasets across three types of lesions: polyps, skin lesions, and retinal vessels. The results demonstrate that MedSegViG achieves superior segmentation accuracy, robustness, and generalizability.

The adoption of Artificial Intelligence in medical imaging holds great promise, yet it remains hindered by challenges such as data scarcity, privacy concerns, and the need for robust multimodal integration. While recent advances in generative modeling have enabled high-quality synthetic data generation, existing approaches are often limited to unimodal, unidirectional synthesis and therefore lack the ability to jointly synthesize multiple modalities while preserving clinical consistency. To address this challenge, we introduce XGeM, a 6.77-billion-parameter multimodal generative model designed to support flexible, any-to-any synthesis between medical data modalities. XGeM constructs a shared latent space via contrastive learning and introduces a novel Multi-Prompt Training strategy, enabling conditioning on arbitrary subsets of input modalities. This design allows the model to adapt to heterogeneous clinical inputs and generate multiple outputs jointly, preserving both semantic and structural coherence. We extensively validate XGeM by first benchmarking it against five competitors on the MIMIC-CXR dataset, a state-of-the-art dataset for multi-view Chest X-ray and radiological report generation. Secondly, we perform a Visual Turing Test with expert radiologists to assess the realism and clinical relevance of the generated data, ensuring alignment with real-world scenarios. Finally, we demonstrate how XGeM can support key medical data challenges such as anonymization, class imbalance, and data scarcity, underscoring its utility as a foundation model for medical data synthesis. Project page is at https://cosbidev.github.io/XGeM/.

Background: The pretreatment identification of post-radiation nasopharyngeal necrosis (PRNN) combined with recurrent nasopharyngeal carcinoma (referred to as cancer-infiltrative PRNN) is crucial for the diagnosis and treatment of PRNN. As the first study to identify recurrent nasopharyngeal carcinoma in patients with PRNN, we aimed to develop a deep learning (DL)-based predictive model using routine MRI to distinguish cancer-infiltrative PRNN from cancer-free PRNN.

Methods: MRIs of 437 patients with PRNN were manually labeled and randomly divided into training and validation cohorts. Video Swin Transformer and Multilayer Perceptron were employed to construct the DL model. The integrated DL and clinical model (DC_combined model) and the integrated radiomics and clinical model (RC_combined model) were constructed using linear weighted fusion of the prediction results from the two models. The predictive value of each model was evaluated using the area under the curve (AUC), accuracy, sensitivity, and specificity.

Results: The DC_combined model significantly outperformed the radiologists in terms of AUC (0.83 vs. 0.60, p < 0.001), accuracy (0.78 vs. 0.60, p = 0.002), and sensitivity (0.86 vs. 0.62, p = 0.002) in the validation cohort. The DC_combined model showed the highest validation sensitivity of 0.86 (95 % CI 0.77-0.94), whereas the RC_combined model demonstrated the highest specificity of 0.88 (95 % CI 0.81-0.96).

Conclusions: Our DC_combined model based on DL can noninvasively distinguish cancer-infiltrative PRNN from cancer-free PRNN with higher AUC, accuracy, and sensitivity than those of radiologists and better sensitivity than that of the RC_combined model based on radiomics.

Colorectal cancer (CRC) is the third most commonly diagnosed malignancy worldwide and a leading cause of cancer-related mortality. This study aims to investigate an automatic detection pipeline for identification and localization of the primary CRC in portal venous phase contrast-enhanced CT scans, which is a crucial first step for downstream CRC staging, prognostication, and treatment planning. We propose a deep learning-based automated detection pipeline using YOLOv11 as the baseline architecture. A ResNet50 module was incorporated into the YOLOv11 backbone to enhance image feature extraction. Additionally, a scale-adaptive loss function, which introduces an adaptive coefficient and a scaling factor to adaptively measure the Intersection over Union (IoU) and center point distance for improving box regression performance, was designed to further improve detection performance. The proposed pipeline achieved a recall of 0.8092, precision of 0.8187, and F-1 score of 0.8139 for CRC detection on our in-house dataset at the patient level (inter-patient evaluation) and a recall of 0.9949, precision of 0.9894, and F-1 score of 0.9921 at the slice level (intra-patient evaluation). Validation on an external public dataset demonstrated that our pipeline, when trained on a patient-level in-house dataset, obtained a recall of 0.8283, precision of 0.8414, and F-1 score of 0.8348 and, when trained on a slice-level in-house dataset, achieved a recall of 0.6897, precision of 0.7888, and F-1 score of 0.7358, outperforming existing representative detection methods. The superior CRC detection performance on the in-house CT dataset and state-of-the-art generalization performance on the public dataset (with a 31.97 %age point improvement in detection sensitivity (recall) over the next closest state-of-the-art method), highlight the potential translational value of our pipeline for CRC clinical decision support, conditional upon validation in larger cohorts.