Anatomically constrained CT image translation for heterogeneous blood vessel segmentation

TL;DR

This paper introduces an anatomically constrained CycleGAN model that improves the synthesis of CT images, enhancing blood vessel segmentation accuracy while reducing radiation exposure by generating missing modalities with high fidelity.

Contribution

It extends CycleGAN with anatomical constraints and automatic ROI selection, enabling better 3D image translation between ceCT and CT modalities from unpaired datasets.

Findings

Improved structural consistency in generated images.

Enhanced segmentation performance compared to existing methods.

Qualitative and quantitative validation on real datasets.

Abstract

Anatomical structures such as blood vessels in contrast-enhanced CT (ceCT) images can be challenging to segment due to the variability in contrast medium diffusion. The combined use of ceCT and contrast-free (CT) CT images can improve the segmentation performances, but at the cost of a double radiation exposure. To limit the radiation dose, generative models could be used to synthesize one modality, instead of acquiring it. The CycleGAN approach has recently attracted particular attention because it alleviates the need for paired data that are difficult to obtain. Despite the great performances demonstrated in the literature, limitations still remain when dealing with 3D volumes generated slice by slice from unpaired datasets with different fields of view. We present an extension of CycleGAN to generate high fidelity images, with good structural consistency, in this context. We leverage anatomical constraints and automatic region of interest selection by adapting the Self-Supervised Body Regressor. These constraints enforce anatomical consistency and allow feeding anatomically-paired input images to the algorithm. Results show qualitative and quantitative improvements, compared to stateof-the-art methods, on the translation task between ceCT and CT images (and vice versa).