We present Causal-Adapter, a modular framework that adapts frozen text-to-image diffusion backbones for counterfactual image generation. Our method enables causal interventions on target attributes, consistently propagating their effects to causal dependents without altering the core identity of the image. In contrast to prior approaches that rely on prompt engineering without explicit causal structure, Causal-Adapter leverages structural causal modeling augmented with two attribute regularization strategies: prompt-aligned injection, which aligns causal attributes with textual embeddings for precise semantic control, and a conditioned token contrastive loss to disentangle attribute factors and reduce spurious correlations. Causal-Adapter achieves state-of-the-art performance on both synthetic and real-world datasets, with up to 91\% MAE reduction on Pendulum for accurate attribute control and 87\% FID reduction on ADNI for high-fidelity MRI image generation. These results show that our approach enables robust, generalizable counterfactual editing with faithful attribute modification and strong identity preservation.

Deep generative models have demonstrated remarkable success in medical image synthesis. However, ensuring conditioning faithfulness and high-quality synthetic images for direct or counterfactual generation remains a challenge. In this work, we introduce a cycle training framework to fine-tune diffusion models for improved conditioning adherence and enhanced synthetic image realism. Our approach, Cycle Diffusion Model (CDM), enforces consistency between generated and original images by incorporating cycle constraints, enabling more reliable direct and counterfactual generation. Experiments on a combined 3D brain MRI dataset (from ABCD, HCP aging & young adults, ADNI, and PPMI) show that our method improves conditioning accuracy and enhances image quality as measured by FID and SSIM. The results suggest that the cycle strategy used in CDM can be an effective method for refining diffusion-based medical image generation, with applications in data augmentation, counterfactual, and disease progression modeling.

In the clinical treatment of bone defects that exceed the critical size threshold, traditional methods using metal fixation devices, autografts, and allografts exhibit significant limitations. Meanwhile, bone scaffolds with minimal risks of secondary injury, low immune rejection are emerging as a promising alternative. The effective design of porosity, pore size, and trabecular thickness in bone scaffolds is critical; however, current strategies often struggle to optimally balance these parameters. Here, we propose a bionic bone scaffold design method that mimics multiple properties of natural cancellous bone using a diffusion model. First, we develop a classifier-free conditional diffusion model and train it on a Micro-CT (μCT) image dataset of porcine vertebral cancellous bone. The training model can produce personalized 2-dimensional images of natural-like bone with tunable microstructures. Subsequently, we stack images layer by layer to form 3-dimensional scaffolds, mimicking the CT/μCT image reconstruction process. Finally, computational fluid dynamics analysis is conducted to validate the scaffold models' fluid properties, while bioresin bone scaffold samples are 3D-printed for mechanical testing and biocompatibility assessment. The three key morphological parameters of the generated images-porosity (50-70%), pore size (468-936 μm), and trabecular thickness (156-312 μm)-can be precisely and independently controlled. Fluid simulation and mechanical testing confirm scaffolds' robust performance in permeability (10⁻⁹ to 10⁻⁸ m²), average fluid shear stress (0.1-0.3 Pa), Young's modulus (14-fold adjustable range), compressive strength (9-fold adjustable range), and viscoelastic properties. The scaffolds also exhibit good biocompatibility, meeting the basic requirements for clinical implantation. These promising results highlight the potential of our method for the personalized design of scaffolds to effectively repair large bone defects.

Neoadjuvant chemotherapy (NAC) is a common therapy option before the main surgery for breast cancer. Response to NAC is monitored using follow-up dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI). Accurate prediction of NAC response helps with treatment planning. Here, we adopt maximum intensity projection images from DCE-MRI to generate post-treatment images (i.e., 3 or 12 weeks after NAC) from pre-treatment images leveraging the emerging diffusion model. We introduce prompt tuning to account for the known clinical factors affecting response to NAC. Our model performed better than other generative models in image quality metrics. Our model was better at generating images that reflected changes in tumor size according to pCR compared to other models. Ablation study confirmed the design choices of our method. Our study has the potential to help with precision medicine.

Objective: Gadolinium-based contrast agents (GBCAs) are commonly employed with T1w MRI to enhance lesion visualization but are restricted in patients at risk of nephrogenic systemic fibrosis and variations in GBCA administration can introduce imaging inconsistencies. This study develops an efficient 3D deep-learning framework to generate T1-contrast enhanced images (T1C) from pre-contrast multiparametric MRI. Approach: We propose the 3D latent rectified flow (T1C-RFlow) model for generating high-quality T1C images. First, T1w and T2-FLAIR images are input into a pretrained autoencoder to acquire an efficient latent space representation. A rectified flow diffusion model is then trained in this latent space representation. The T1C-RFlow model was trained on a curated dataset comprised of the BraTS 2024 glioma (GLI; 1480 patients), meningioma (MEN; 1141 patients), and metastases (MET; 1475 patients) datasets. Selected patients were split into train (N=2860), validation (N=612), and test (N=614) sets. Results: Both qualitative and quantitative results demonstrate that the T1C-RFlow model outperforms benchmark 3D models (pix2pix, DDPM, Diffusion Transformers (DiT-3D)) trained in the same latent space. T1C-RFlow achieved the following metrics - GLI: NMSE 0.044 +/- 0.047, SSIM 0.935 +/- 0.025; MEN: NMSE 0.046 +/- 0.029, SSIM 0.937 +/- 0.021; MET: NMSE 0.098 +/- 0.088, SSIM 0.905 +/- 0.082. T1C-RFlow had the best tumor reconstruction performance and significantly faster denoising times (6.9 s/volume, 200 steps) than conventional DDPM models in both latent space (37.7s, 1000 steps) and patch-based in image space (4.3 hr/volume). Significance: Our proposed method generates synthetic T1C images that closely resemble ground truth T1C in much less time than previous diffusion models. Further development may permit a practical method for contrast-agent-free MRI for brain tumors.

The validation of multimodal deep learning models for medical image translation is limited by the lack of high-quality, paired datasets. We propose a novel framework that leverages computational phantoms to generate realistic CT and MRI images, enabling reliable ground-truth datasets for robust validation of artificial intelligence (AI) methods that generate synthetic CT (sCT) from MRI, specifically for radiotherapy applications. Two CycleGANs (cycle-consistent generative adversarial networks) were trained to transfer the imaging style of real patients onto CT and MRI phantoms, producing synthetic data with realistic textures and continuous intensity distributions. These data were evaluated through paired assessments with original phantoms, unpaired comparisons with patient scans, and dosimetric analysis using patient-specific radiotherapy treatment plans. Additional external validation was performed on public CT datasets to assess the generalizability to unseen data. The resulting, paired CT/MRI phantoms were used to validate a GAN-based model for sCT generation from abdominal MRI in particle therapy, available in the literature. Results showed strong anatomical consistency with original phantoms, high histogram correlation with patient images (HistCC = 0.998 ± 0.001 for MRI, HistCC = 0.97 ± 0.04 for CT), and dosimetric accuracy comparable to real data. The novelty of this work lies in using generated phantoms as validation data for deep learning-based cross-modality synthesis techniques.

To develop a new method to generate synthetic MR spectroscopic imaging (MRSI) data for training machine learning models. This study targeted routine MRI examination protocols with single voxel spectroscopy (SVS). A novel model derived from pix2pix generative adversarial networks was proposed to generate synthetic MRSI data using MRI and SVS data as inputs. T1- and T2-weighted, SVS, and reference MRSI data were acquired from healthy brains with clinically available sequences. The proposed model was trained to generate synthetic MRSI data. Quantitative evaluation involved the calculation of the mean squared error (MSE) against the reference and metabolite ratio value. The effect of the location of and the number of the SVS data on the quality of the synthetic MRSI data was investigated using the MSE. The synthetic MRSI data generated from the proposed model were visually closer to the reference. The 95% confidence interval (CI) of the metabolite ratio value of synthetic MRSI data overlapped with the reference for seven of eight metabolite ratios. The MSEs tended to be lower in the same location than in different locations. The MSEs among groups of numbers of SVS data were not significantly different. A new method was developed to generate MRSI data by integrating MRI and SVS data. Our method can potentially increase the volume of MRSI data training for other machine learning models by adding SVS acquisition to routine MRI examinations.

Deep learning necessitates a substantial amount of data, yet obtaining sufficient medical images is difficult due to concerns about patient privacy and high collection costs. To address this issue, we propose a conditional latent diffusion model-based medical image enhancement network, referred to as the Medical Image Enhancement Network (MedIENet). To meet the rigorous standards required for image generation in the medical imaging field, a multi-attention module is incorporated in the encoder of the denoising U-Net backbone. Additionally Rotary Position Embedding (RoPE) is integrated into the self-attention module to effectively capture positional information, while cross-attention is utilised to embed integrate class information into the diffusion process. MedIENet is evaluated on three datasets: Chest CT-Scan images, Chest X-Ray Images (Pneumonia), and Tongue dataset. Compared to existing methods, MedIENet demonstrates superior performance in both fidelity and diversity of the generated images. Experimental results indicate that for downstream classification tasks using ResNet50, the Area Under the Receiver Operating Characteristic curve (AUROC) achieved with real data alone is 0.76 for the Chest CT-Scan images dataset, 0.87 for the Chest X-Ray Images (Pneumonia) dataset, and 0.78 for the Tongue Dataset. When using mixed data consisting of real data and generated data, the AUROC improves to 0.82, 0.94, and 0.82, respectively, reflecting increases of approximately 6%, 7%, and 4%. These findings indicate that the images generated by MedIENet can enhance the performance of downstream classification tasks, providing an effective solution to the scarcity of medical image training data.

We present a patch-based 3D nnUNet adaptation for MR to CT and CBCT to CT image translation using the multicenter SynthRAD2025 dataset, covering head and neck (HN), thorax (TH), and abdomen (AB) regions. Our approach leverages two main network configurations: a standard UNet and a residual UNet, both adapted from nnUNet for image synthesis. The Anatomical Feature-Prioritized (AFP) loss was introduced, which compares multilayer features extracted from a compact segmentation network trained on TotalSegmentator labels, enhancing reconstruction of clinically relevant structures. Input volumes were normalized per-case using zscore normalization for MRIs, and clipping plus dataset level zscore normalization for CBCT and CT. Training used 3D patches tailored to each anatomical region without additional data augmentation. Models were trained for 1000 and 1500 epochs, with AFP fine-tuning performed for 500 epochs using a combined L1+AFP objective. During inference, overlapping patches were aggregated via mean averaging with step size of 0.3, and postprocessing included reverse zscore normalization. Both network configurations were applied across all regions, allowing consistent model design while capturing local adaptations through residual learning and AFP loss. Qualitative and quantitative evaluation revealed that residual networks combined with AFP yielded sharper reconstructions and improved anatomical fidelity, particularly for bone structures in MR to CT and lesions in CBCT to CT, while L1only networks achieved slightly better intensity-based metrics. This methodology provides a stable solution for cross modality medical image synthesis, demonstrating the effectiveness of combining the automatic nnUNet pipeline with residual learning and anatomically guided feature losses.

Cone-beam computed tomography (CBCT) is widely used for image-guided radiotherapy (IGRT). It provides real time visualization at low cost and dose. However, photon scattering and beam hindrance cause artifacts in CBCT. These include inaccurate Hounsfield Units (HU), reducing reliability for dose calculation, and adaptive planning. By contrast, computed tomography (CT) offers better image quality and accurate HU calibration but is usually acquired offline and fails to capture intra-treatment anatomical changes. Thus, accurate CBCT-to-CT synthesis is needed to close the imaging-quality gap in adaptive radiotherapy workflows. To cater to this, we propose a novel diffusion-based conditional generative model, coined EqDiff-CT, to synthesize high-quality CT images from CBCT. EqDiff-CT employs a denoising diffusion probabilistic model (DDPM) to iteratively inject noise and learn latent representations that enable reconstruction of anatomically consistent CT images. A group-equivariant conditional U-Net backbone, implemented with e2cnn steerable layers, enforces rotational equivariance (cyclic C4 symmetry), helping preserve fine structural details while minimizing noise and artifacts. The system was trained and validated on the SynthRAD2025 dataset, comprising CBCT-CT scans across multiple head-and-neck anatomical sites, and we compared it with advanced methods such as CycleGAN and DDPM. EqDiff-CT provided substantial gains in structural fidelity, HU accuracy and quantitative metrics. Visual findings further confirm the improved recovery, sharper soft tissue boundaries, and realistic bone reconstructions. The findings suggest that the diffusion model has offered a robust and generalizable framework for CBCT improvements. The proposed solution helps in improving the image quality as well as the clinical confidence in the CBCT-guided treatment planning and dose calculations.