Computerized Tomography Pulmonary Angiography Image Simulation using Cycle Generative Adversarial Network from Chest CT imaging in Pulmonary Embolism Patients

TL;DR

This paper presents a deep learning approach using Cycle GANs to simulate pulmonary angiography images from chest CT scans, aiming to assist in pulmonary embolism diagnosis and reduce the need for harmful, expensive procedures.

Contribution

It introduces a novel application of Cycle GANs for generating simulated CTPA images from chest CT scans to aid clinical diagnosis of pulmonary embolism.

Findings

Generated images enhance pulmonary vessel features

System can assist in PE screening process

Potential to reduce unnecessary CTPA procedures

Abstract

The purpose of this research is to develop a system that generates simulated computed tomography pulmonary angiography (CTPA) images clinically for pulmonary embolism diagnoses. Nowadays, CTPA images are the gold standard computerized detection method to determine and identify the symptoms of pulmonary embolism (PE), although performing CTPA is harmful for patients and also expensive. Therefore, we aim to detect possible PE patients through CT images. The system will simulate CTPA images with deep learning models for the identification of PE patients' symptoms, providing physicians with another reference for determining PE patients. In this study, the simulated CTPA image generation system uses a generative antagonistic network to enhance the features of pulmonary vessels in the CT images to strengthen the reference value of the images and provide a basis for hospitals to judge PE patients. We used the CT images of 22 patients from National Cheng Kung University Hospital and the corresponding CTPA images as the training data for the task of simulating CTPA images and generated them using two sets of generative countermeasure networks. This study is expected to propose a new approach to the clinical diagnosis of pulmonary embolism, in which a deep learning network is used to assist in the complex screening process and to review the generated simulated CTPA images, allowing physicians to assess whether a patient needs to undergo detailed testing for CTPA, improving the speed of detection of pulmonary embolism and significantly reducing the number of undetected patients.