Current medical image analysis systems are typically task-specific, requiring separate models for classification and segmentation, and lack the flexibility to support user-defined workflows. To address these challenges, we introduce MedPrompt, a unified framework that combines a few-shot prompted Large Language Model (Llama-4-17B) for high-level task planning with a modular Convolutional Neural Network (DeepFusionLab) for low-level image processing. The LLM interprets user instructions and generates structured output to dynamically route task-specific pretrained weights. This weight routing approach avoids retraining the entire framework when adding new tasks-only task-specific weights are required, enhancing scalability and deployment. We evaluated MedPrompt across 19 public datasets, covering 12 tasks spanning 5 imaging modalities. The system achieves a 97% end-to-end correctness in interpreting and executing prompt-driven instructions, with an average inference latency of 2.5 seconds, making it suitable for near real-time applications. DeepFusionLab achieves competitive segmentation accuracy (e.g., Dice 0.9856 on lungs) and strong classification performance (F1 0.9744 on tuberculosis). Overall, MedPrompt enables scalable, prompt-driven medical imaging by combining the interpretability of LLMs with the efficiency of modular CNNs.

Ultrasound Coherent Plane Wave Compounding (CPWC) enhances image contrast by combining echoes from multiple steered transmissions. While increasing the number of angles generally improves image quality, it drastically reduces the frame rate and can introduce blurring artifacts in fast-moving targets. Moreover, compounded images remain susceptible to noise, particularly when acquired with a limited number of transmissions. We propose a zero-shot denoising framework tailored for low-angle CPWC acquisitions, which enhances contrast without relying on a separate training dataset. The method divides the available transmission angles into two disjoint subsets, each used to form compound images that include higher noise levels. The new compounded images are then used to train a deep model via a self-supervised residual learning scheme, enabling it to suppress incoherent noise while preserving anatomical structures. Because angle-dependent artifacts vary between the subsets while the underlying tissue response is similar, this physics-informed pairing allows the network to learn to disentangle the inconsistent artifacts from the consistent tissue signal. Unlike supervised methods, our model requires no domain-specific fine-tuning or paired data, making it adaptable across anatomical regions and acquisition setups. The entire pipeline supports efficient training with low computational cost due to the use of a lightweight architecture, which comprises only two convolutional layers. Evaluations on simulation, phantom, and in vivo data demonstrate superior contrast enhancement and structure preservation compared to both classical and deep learning-based denoising methods.

Solving Electromagnetic Inverse Scattering Problems (EISP) is fundamental in applications such as medical imaging, where the goal is to reconstruct the relative permittivity from scattered electromagnetic field. This inverse process is inherently ill-posed and highly nonlinear, making it particularly challenging. A recent machine learning-based approach, Img-Interiors, shows promising results by leveraging continuous implicit functions. However, it requires case-specific optimization, lacks generalization to unseen data, and fails under sparse transmitter setups (e.g., with only one transmitter). To address these limitations, we revisit EISP from a physics-informed perspective, reformulating it as a two stage inverse transmission-scattering process. This formulation reveals the induced current as a generalizable intermediate representation, effectively decoupling the nonlinear scattering process from the ill-posed inverse problem. Built on this insight, we propose the first generalizable physics-driven framework for EISP, comprising a current estimator and a permittivity solver, working in an end-to-end manner. The current estimator explicitly learns the induced current as a physical bridge between the incident and scattered field, while the permittivity solver computes the relative permittivity directly from the estimated induced current. This design enables data-driven training and generalizable feed-forward prediction of relative permittivity on unseen data while maintaining strong robustness to transmitter sparsity. Extensive experiments show that our method outperforms state-of-the-art approaches in reconstruction accuracy, generalization, and robustness. This work offers a fundamentally new perspective on electromagnetic inverse scattering and represents a major step toward cost-effective practical solutions for electromagnetic imaging.

The accurate segmentation of myocardial scars from cardiac MRI is essential for clinical assessment and treatment planning. In this study, we propose a robust deep-learning pipeline for fully automated myocardial scar detection and segmentation by fine-tuning state-of-the-art models. The method explicitly addresses challenges of label noise from semi-automatic annotations, data heterogeneity, and class imbalance through the use of Kullback-Leibler loss and extensive data augmentation. We evaluate the model's performance on both acute and chronic cases and demonstrate its ability to produce accurate and smooth segmentations despite noisy labels. In particular, our approach outperforms state-of-the-art models like nnU-Net and shows strong generalizability in an out-of-distribution test set, highlighting its robustness across various imaging conditions and clinical tasks. These results establish a reliable foundation for automated myocardial scar quantification and support the broader clinical adoption of deep learning in cardiac imaging.

Spiculation, characterized by irregular, spike-like projections from nodule margins, serves as a crucial radiological biomarker for malignancy assessment and early cancer detection. These distinctive stellate patterns strongly correlate with tumor invasiveness and are vital for accurate diagnosis and treatment planning. Traditional computer-aided diagnosis (CAD) systems are limited in their capability to capture and use these patterns given their subtlety, difficulty in quantifying them, and small datasets available to learn these patterns. To address these challenges, we propose a novel framework leveraging variational autoencoders (VAE) to discover, extract, and vary disentangled latent representations of lung nodule images. By gradually varying the latent representations of non-spiculated nodule images, we generate augmented datasets containing spiculated nodule variations that, we hypothesize, can improve the diagnostic classification of lung nodules. Using the National Institutes of Health/National Cancer Institute Lung Image Database Consortium (LIDC) dataset, our results show that incorporating these spiculated image variations into the classification pipeline significantly improves spiculation detection performance up to 7.53%. Notably, this enhancement in spiculation detection is achieved while preserving the classification performance of non-spiculated cases. This approach effectively addresses class imbalance and enhances overall classification outcomes. The gradual attenuation of spiculation characteristics demonstrates our model's ability to both capture and generate clinically relevant semantic features in an algorithmic manner. These findings suggest that the integration of semantic-based latent representations into CAD models not only enhances diagnostic accuracy but also provides insights into the underlying morphological progression of spiculated nodules, enabling more informed and clinically meaningful AI-driven support systems.

To develop a high-performance machine learning model for predicting and interpreting features of pulmonary diseases. This retrospective study analyzed clinical and imaging data from patients with non-small cell lung cancer (NSCLC), granulomatous inflammation, and benign tumors, collected across multiple centers from January 2015 to October 2023. Data from two hospitals in Anhui Province were split into a development set (n = 1696) and a test set (n = 424) in an 8:2 ratio, with an external validation set (n = 909) from Zhejiang Province. Features with p < 0.05 from univariate analyses were selected using the Boruta algorithm for input into Random Forest (RF) and XGBoost models. Model efficacy was assessed using receiver operating characteristic (ROC) analysis. A total of 3030 patients were included: 2269 with NSCLC, 529 with granulomatous inflammation, and 232 with benign tumors. The Obuchowski indices for RF and XGBoost in the test set were 0.7193 (95% CI: 0.6567-0.7812) and 0.8282 (95% CI: 0.7883-0.8650), respectively. In the external validation set, indices were 0.7932 (95% CI: 0.7572-0.8250) for RF and 0.8074 (95% CI: 0.7740-0.8387) for XGBoost. XGBoost achieved better accuracy in both the test (0.81) and external validation (0.79) sets. Calibration Curve and Decision Curve Analysis (DCA) showed XGBoost offered higher net clinical benefit. The XGBoost model outperforms RF in the three-class classification of lung diseases. XGBoost surpasses Random Forest in accurately classifying NSCLC, granulomatous inflammation, and benign tumors, offering superior clinical utility via multicenter data. Lung cancer classification model has broad clinical applicability. XGBoost outperforms random forests using CT imaging data. XGBoost model can be deployed on a website for clinicians.

To investigate automated breast ultrasound (ABUS) features affecting the use of Multiview convolutional neural network (CNN) for breast lesions according to radiologists' experience. A total of 656 breast lesions (152 malignant and 504 benign lesions) were included and reviewed by six radiologists for background echotexture, glandular tissue component (GTC), and lesion type and size without as well as with Multiview CNN. The sensitivity, specificity, and the area under the receiver operating curve (AUC) for ABUS features were compared between two sessions according to radiologists' experience. Radiology residents showed significant AUC improvement with the Multiview CNN for mass (0.81 to 0.91, P=0.003) and non-mass lesions (0.56 to 0.90, P=0.007), all background echotextures (homogeneous-fat: 0.84 to 0.94, P=0.04; homogeneous-fibroglandular: 0.85 to 0.93, P=0.01; heterogeneous: 0.68 to 0.88, P=0.002), all GTC levels (minimal: 0.86 to 0.93, P=0.001; mild: 0.82 to 0.94, P=0.003; moderate: 0.75 to 0.88, P=0.01; marked: 0.68 to 0.89, P<0.001), and lesions ≤10mm (≤5 mm: 0.69 to 0.86, P<0.001; 6-10 mm: 0.83 to 0.92, P<0.001). Breast specialists showed significant AUC improvement with the Multiview CNN in heterogeneous echotexture (0.90 to 0.95, P=0.03), marked GTC (0.88 to 0.95, P<0.001), and lesions ≤10mm (≤5 mm: 0.89 to 0.93, P=0.02; 6-10 mm: 0.95 to 0.98, P=0.01). With the Multiview CNN, the performance of ABUS in radiology residents was improved regardless of lesion type, background echotexture, or GTC. For breast lesions smaller than 10 mm, both radiology residents and breast specialists showed better performance of ABUS.

Low-dose computed tomography (LDCT) denoising plays an important role in medical imaging for reducing the radiation dose to patients. Recently, various data-driven and diffusion-based deep learning (DL) methods have been developed and shown promising results in LDCT denoising. However, challenges remain in ensuring generalizability to different datasets and mitigating uncertainty from stochastic sampling. In this paper, we introduce a novel dose-aware diffusion model that effectively reduces CT image noise while maintaining structural fidelity and being generalizable to different dose levels.&#xD;Approach: Our approach employs a physics-based forward process with continuous timesteps, enabling flexible representation of diverse noise levels. We incorporate a computationally efficient noise calibration module in our diffusion framework that resolves misalignment between intermediate results and their corresponding timesteps. Furthermore, we present a simple yet effective method for estimating appropriate timesteps for unseen LDCT images, allowing generalization to an unknown, arbitrary dose levels.&#xD;Main Results: Both qualitative and quantitative evaluation results on Mayo Clinic datasets show that the proposed method outperforms existing denoising methods in preserving the noise texture and restoring anatomical structures. The proposed method also shows consistent results on different dose levels and an unseen dataset.&#xD;Significance: We propose a novel dose-aware diffusion model for LDCT denoising, aiming to address the generalization and uncertainty issues of existing diffusion-based DL methods. Our experimental results demonstrate the effectiveness of the proposed method across different dose levels. We expect that our approach can provide a clinically practical solution for LDCT denoising with its high structural fidelity and computational efficiency.

Post-surgical prediction of recurrence or metastasis for primary gastrointestinal stromal tumors (GISTs) remains challenging. We aim to develop individualized clinical follow-up strategies for primary GIST patients, such as shortening follow-up time or extending drug administration based on the clinical deep learning radiomics model (CDLRM). The clinical information on primary GISTs was collected from two independent centers. Postoperative recurrence or metastasis in GIST patients was defined as the endpoint of the study. A total of nine machine learning models were established based on the selected features. The performance of the models was assessed by calculating the area under the curve (AUC). The CDLRM with the best predictive performance was constructed. Decision curve analysis (DCA) and calibration curves were analyzed separately. Ultimately, our model was applied to the high-potential malignant group vs the low-malignant-potential group. The optimal clinical application scenarios of the model were further explored by comparing the DCA performance of the two subgroups. A total of 526 patients, 260 men and 266 women, with a mean age of 62 years, were enrolled in the study. CDLRM performed excellently with AUC values of 0.999, 0.963, and 0.995 for the training, external validation, and aggregated sets, respectively. The calibration curve indicated that CDLRM was in good agreement between predicted and observed probabilities in the validation cohort. The results of DCA's performance in different subgroups show that it was more clinically valuable in populations with high malignant potential. CDLRM could help the development of personalized treatment and improved follow-up of patients with a high probability of recurrence or metastasis in the future. This model utilizes imaging features extracted from CT scans (including radiomic features and deep features) and clinical data to accurately predict postoperative recurrence and metastasis in patients with primary GISTs, which has a certain auxiliary role in clinical decision-making. We developed and validated a model to predict recurrence or metastasis in patients taking oral imatinib after GIST. We demonstrate that CT image features were associated with recurrence or metastases. The model had good predictive performance and clinical benefit.

Fast and accurate organ-at-risk (OAR) and gross tumor volume (GTV) contour propagation methods are needed to improve the efficiency of magnetic resonance (MR) imaging-guided radiotherapy. We trained deformable image registration networks to accurately propagate contours from planning to fraction MR images.&#xD;Approach: Data from 140 stage 1-2 lung cancer patients treated at a 0.35T MR-Linac were split into 102/17/21 for training/validation/testing. Additionally, 18 central lung tumor patients, treated at a 0.35T MR-Linac externally, and 14 stage 3 lung cancer patients from a phase 1 clinical trial, treated at 0.35T or 1.5T MR-Linacs at three institutions, were used for external testing. Planning and fraction images were paired (490 pairs) for training. Two hybrid transformer-convolutional neural network TransMorph models with mean squared error (MSE), Dice similarity coefficient (DSC), and regularization losses (TM_{MSE+Dice}) or MSE and regularization losses (TM_{MSE}) were trained to deformably register planning to fraction images. The TransMorph models predicted diffeomorphic dense displacement fields. Multi-label images including seven thoracic OARs and the GTV were propagated to generate fraction segmentations. Model predictions were compared with contours obtained through B-spline, vendor registration and the auto-segmentation method nnUNet. Evaluation metrics included the DSC and Hausdorff distance percentiles (50th and 95th) against clinical contours.&#xD;Main results: TM_{MSE+Dice} and TM_{MSE} achieved mean OARs/GTV DSCs of 0.90/0.82 and 0.90/0.79 for the internal and 0.84/0.77 and 0.85/0.76 for the central lung tumor external test data. On stage 3 data, TM_{MSE+Dice} achieved mean OARs/GTV DSCs of 0.87/0.79 and 0.83/0.78 for the 0.35 T MR-Linac datasets, and 0.87/0.75 for the 1.5 T MR-Linac dataset. TM_{MSE+Dice} and TM_{MSE} had significantly higher geometric accuracy than other methods on external data. No significant difference between TM_{MSE+Dice} and TM_{MSE} was found.&#xD;Significance: TransMorph models achieved time-efficient segmentation of fraction MRIs with high geometrical accuracy and accurately segmented images obtained at different field strengths.