What Does DALL-E 2 Know About Radiology?

TL;DR

This paper investigates DALL-E 2's ability to generate and manipulate radiology images, revealing promising zero-shot capabilities in X-ray image generation but limited performance in other medical imaging modalities.

Contribution

It demonstrates DALL-E 2's potential for radiology image generation and manipulation, highlighting the need for domain-specific fine-tuning for medical applications.

Findings

DALL-E 2 can generate X-ray images from text descriptions.

It can extend images and remove elements in radiology images.

Performance on CT, MRI, ultrasound, and pathology images is limited.

Abstract

Generative models such as DALL-E 2 could represent a promising future tool for image generation, augmentation, and manipulation for artificial intelligence research in radiology provided that these models have sufficient medical domain knowledge. Here we show that DALL-E 2 has learned relevant representations of X-ray images with promising capabilities in terms of zero-shot text-to-image generation of new images, continuation of an image beyond its original boundaries, or removal of elements, while pathology generation or CT, MRI, and ultrasound images are still limited. The use of generative models for augmenting and generating radiological data thus seems feasible, even if further fine-tuning and adaptation of these models to the respective domain is required beforehand.