Diffusion models were initially developed for text-to-image generation and are now being utilized to generate high quality synthetic images. Preceded by GANs, diffusion models have shown impressive results using various evaluation metrics. However, commonly used metrics such as FID and IS are not suitable for determining whether diffusion models are simply reproducing the training images. Here we train StyleGAN and a diffusion model, using  BRATS20, BRATS21 and a chest x-ray pneumonia dataset, to synthesize brain MRI and chest x-ray images, and measure the correlation between the synthetic images and all training images. Our results show that diffusion models are more likely to memorize the training images, compared to StyleGAN, especially for small datasets and when using 2D slices from 3D volumes. Researchers should be careful when using diffusion models (and to some extent GANs) for medical imaging, if the final goal is to share the synthetic images. 

Spine Computed Tomography (SCT) is essential for identifying fractures, tumors and degenerative spine diseases, assisting medical practitioners in formulating an accurate diagnosis and treatment. One of the core element of SCT is reporting. The effectiveness of spine reporting is often limited by challenges such as an inadequate infrastructure and lack of experts. Automated SCT analysis has the potential to revolutionize spinal healthcare and improve patient outcomes. To achieve this objective, we proposed a framework for spine report generation that utilizes transformer architecture, trained on textual reports alongside the visual features extracted from the sagittal slices of the SCT volume. A foundation model is used to perform Knowledge Distillation (KD) alongside an encoder to ensure an optimal performance. The proposed framework is evaluated on the public dataset (VerSe20). The incorporation of KD results improved both the BERT and BLEU1 score on the dataset, from 0.7486 to 0.7522 and 0.6361 to 0.7291. Additionally, the proposed framework is evaluated using four different types of reports: original radiologist reports, reports without spine-level annotations, rephrased reports, and reports generated by ChatGPT-4o (ChatGPT). The evaluation without spine-level annotations demonstrates superior performance across most metrics, achieving the highest BLEU-1 and ROUGE-L scores, with a BLEU-1 of 0.9293 and a ROUGE-L score of 0.9297. In contrast, the other techniques achieved moderate scores across all metrics. Finally, experienced radiologists assessed the spine report and have given high rating to the original reports across all three criteria (completeness, conciseness and correctness), in comparison to the generated reports. This study’s findings suggest that omitting spine-level annotations can improve the quality of text generation.

Diffusion magnetic resonance imaging (diffusion MRI) is widely employed to probe the diffusive motion of water molecules within the tissue. Numerous diseases and processes affecting the central nervous system can be detected and monitored via diffusion MRI thanks to its sensitivity to microstructural alterations in tissue. The latter has prompted interest in quantitative mapping of the microstructural parameters, such as the fiber orientation distribution function (fODF), which is instrumental for noninvasively mapping the underlying axonal fiber tracts in white matter through a procedure known as tractography. However, such applications demand repeated acquisitions of MRI volumes with varied experimental parameters demanding long acquisition times and/or limited spatial resolution. In this work, we present a deep-learning-based approach for increasing the spatial resolution of diffusion MRI data in the form of fODFs obtained through constrained spherical deconvolution. The proposed approach is evaluated on high quality data from the Human Connectome Project, and is shown to generate upsampled results with a greater correspondence to ground truth high-resolution data than can be achieved with ordinary spline interpolation methods. Furthermore, we employ a measure based on the earth mover’s distance to assess the accuracy of the upsampled fODFs. At low signal-to-noise ratios, our super-resolution method provides more accurate estimates of the fODF compared to data collected with 8 times smaller voxel volume.

Four-dimensional flow (4D Flow) CMR affords comprehensive 3D maps of advanced hemodynamics parameters such as wall shear stress (WSS). However, the evaluation of these data is often restricted to spatial averages in large regions of interests, such as the ascending aorta. Recent studies have explored ways of analyzing local intercohort WSS differences by using basic statistical tests with a p-value of 0.05 for determining significance, thus not accounting for the large number of comparisons made when exploring differences for multiple locations across the ascending aorta surface.

Permutation tests, frequently used in brain MRI, permit statistical analysis on a local level while controlling for the family-wise error rate by constructing the null hypothesis distribution based on the maximum statistic over the voxels at each permutation. We sought to use permutation tests to identify local regions of abnormal WSS in the ascending aorta in patients with aortic dilation.

 Despite well-defined criteria for radiographic diagnosis of atypical femur fractures (AFFs), missed and delayed diagnosis is common. An AFF diagnostic software could provide timely AFF detection to prevent progression of incomplete or development of contralateral AFFs. In this study, we investigated the ability for an artificial intelligence (AI)-based application, using deep learning models (DLMs), particularly convolutional neural networks (CNNs), to detect AFFs from femoral radiographs. A labelled Australian dataset of pre-operative complete AFF (cAFF), incomplete AFF (iAFF), typical femoral shaft fracture (TFF), and non-fractured femoral (NFF) X-ray images in anterior-posterior view were used for training (N = 213, 49, 394, 1359, respectively). An AFFnet model was developed using a pretrained (ImageNet dataset) ResNet-50 backbone, and a novel Box Attention Guide (BAG) module to guide the model's scanning patterns to enhance its learning. All images were used to train and internally test the model using a 5-fold cross validation approach, and further validated by an external dataset. External validation of the model's performance was conducted on a Sweden dataset comprising 733 TFF and 290 AFF images. Precision, sensitivity, specificity, F1-score and AUC were measured and compared between AFFnet and a global approach with ResNet-50. Excellent diagnostic performance was recorded in both models (all AUC >0.97), however AFFnet recorded lower number of prediction errors, and improved sensitivity, F1-score and precision compared to ResNet-50 in both internal and external testing. Sensitivity in the detection of iAFF was higher for AFFnet than ResNet-50 (82 % vs 56 %). In conclusion, AFFnet achieved excellent diagnostic performance on internal and external validation, which was superior to a pre-existing model. Accurate AI-based AFF diagnostic software has the potential to improve AFF diagnosis, reduce radiologist error, and allow urgent intervention, thus improving patient outcomes.

Large annotated datasets are required for training deep learning models, but in medical imaging data sharing is often complicated due to ethics, anonymization and data protection legislation. Generative AI models, such as generative adversarial networks (GANs) and diffusion models, can today produce very realistic synthetic images, and can potentially facilitate data sharing. However, in order to share synthetic medical images it must first be demonstrated that they can be used for training different networks with acceptable performance. Here, we therefore comprehensively evaluate four GANs (progressive GAN, StyleGAN 1–3) and a diffusion model for the task of brain tumor segmentation (using two segmentation networks, U-Net and a Swin transformer). Our results show that segmentation networks trained on synthetic images reach Dice scores that are 80%–90% of Dice scores when training with real images, but that memorization of the training images can be a problem for diffusion models if the original dataset is too small. Our conclusion is that sharing synthetic medical images is a viable option to sharing real images, but that further work is required. The trained generative models and the generated synthetic images are shared on AIDA data hub.

This study explores the application of computational pathology to analyze pediatric brain tumors utilizing hematoxylin and eosin (H&E) and immunohistochemistry (IHC) whole slide images (WSIs). Experiments were conducted on H&E images for predicting tumor diagnosis and fusing them with unregistered IHC images to investigate potential improvements. Patch features were extracted using UNI, a vision transformer (ViT) model trained on H&E data, and whole slide classification was achieved using the attention-based multiple instance learning CLAM framework. In the astrocytoma tumor classification, early fusion of the H&E and IHC significantly improved the differentiation between tumor grades (balanced accuracy: 0.82±0.05vs 0.84 ± 0.05). In the multiclass classification, H&E images alone had a balanced accuracy of 0.79 ± 0.03 without any improvement obtained when fused with IHC. The findings highlight the potential of using multi-stain fusion to advance the diagnosis of pediatric brain tumors, however, further fusion methods should be investigated.

Radiotherapy treatment planning takes substantial time, several hours per patient, as it involves manual segmentation of tumor and risk organs. Segmentation networks can be trained to automatically perform the segmentations, but typically require large annotated datasets for training. Sharing of sensitive data between hospitals, to create a larger dataset, is often difficult due to ethics and GDPR. Here we therefore demonstrate that federated learning is a solution to this problem, as then only the segmentation model is sent between each hospital and a global server. We export and preprocess brain tumor images from the oncology departments in Linköping and Lund, and use federated learning to train a global segmentation model using two different frameworks.

Atypical femur fractures (AFF) represent a very rare type of fracture that can be difficult to discriminate radiologically from normal femur fractures (NFF). AFFs are associated with drugs that are administered to prevent osteoporosis-related fragility fractures, which are highly prevalent in the elderly population. Given that these fractures are rare and the radiologic changes are subtle currently only 7% of AFFs are correctly identified, which hinders adequate treatment for most patients with AFF. Deep learning models could be trained to classify automatically a fracture as AFF or NFF, thereby assisting radiologists in detecting these rare fractures. Historically, for this classification task, only imaging data have been used, using convolutional neural networks (CNN) or vision transformers applied to radiographs. However, to mimic situations in which all available data are used to arrive at a diagnosis, we adopted an approach of deep learning that is based on the integration of image data and tabular data (from electronic health records) for 159 patients with AFF and 914 patients with NFF. We hypothesized that the combinatorial data, compiled from all the radiology departments of 72 hospitals in Sweden and the Swedish National Patient Register, would improve classification accuracy, as compared to using only one modality. At the patient level, the area under the ROC curve (AUC) increased from 0.966 to 0.987 when using the integrated set of imaging data and seven pre-selected variables, as compared to only using imaging data. More importantly, the sensitivity increased from 0.796 to 0.903. We found a greater impact of data fusion when only a randomly selected subset of available images was used to make the image and tabular data more balanced for each patient. The AUC then increased from 0.949 to 0.984, and the sensitivity increased from 0.727 to 0.849.

These AUC improvements are not large, mainly because of the already excellent performance of the CNN (AUC of 0.966) when only images are used. However, the improvement is clinically highly relevant considering the importance of accuracy in medical diagnostics. We expect an even greater effect when imaging data from a clinical workflow, comprising a more diverse set of diagnostic images, are used.