 Recent advancements in computer vision have enabled the creation of realistic images from textual inputs using diffusion probabilistic models. These models have been successfully applied to generate images from text in the field of image generation, such as DOLU2, Imogen, and Stable Diffusion. However, these models have yet to be explored in the context of medical imaging, which often involves three-dimensional volumes. In this study, we demonstrate that diffusion probabilistic models can be used to generate high-quality medical images from MI and CT scans. Additionally, we show that synthetic images can be used to augment small datasets and improve the performance of breast segmentation models. Finally, our results suggest that diffusion probabilistic models are capable of generating realistic images with anatomically consistent and accurate reconstructions. This article was authored by Feroz Cader, Gustav Muller-Franzies, Sarush Tairi Biarasta, and others.