Device-Less Data-Driven Cardiac and Respiratory Gating in PET Imaging

Project by Nanna Overbeck

Introduction

Respiratory and cardiac motion during Positron Emission Tomography (PET) scans can cause image blurring, reducing diagnostic accuracy and affecting treatment planning. This project introduces a device-less, data-driven approach utilizing high-sensitivity long axial field-of-view (LAFOV) PET histo images. By enabling gated image reconstruction, the method improves image clarity, enhances tumor detectability, and provides more reliable diagnostic information without the need for external devices.

Project Background

PET, combined with Computed Tomography (CT), is a valuable tool for diagnosing and treatment planning in oncology, neurology, and cardiology. However, motion during PET scans can lead to reduced image quality, which may lead to inaccurate assessments of tumor activity, size, and visibility.

Gated image reconstruction can address these issues by dividing scan data into phases of the motion cycle. Traditional gating methods rely on external devices, which increase preparation time and complexity.

This project explores a device-less alternative using LAFOV [18F]FDG PET histo images and Fourier Transform (FT) analysis to estimate cardiac and respiratory motion. Results from 18 patients show that this method reliably matches reference measurements. Gated reconstructions demonstrated increased tumor maximum standardized uptake values (SUVmax) and reduced tumor volume, improving tumor quantification and overall diagnostic precision.

Project Potential

This device-less, data-driven method has the potential to optimize PET imaging by reducing preparation time and eliminating the need for external devices. It enhances motion correction, leading to more accurate tumor measurements and improved diagnostic accuracy. LAFOV PET histo images have demonstrated consistent and reliable motion estimation, with gated reconstructions effectively isolating heart and lung motion. This approach enhances image clarity, making it particularly valuable for motion-sensitive cases such as lung and cardiac imaging, with potential for clinical integration.

Gated PET Videos of Cardiac and Respiratory Motion
Cardiac Motion
Respiratory Motion
This example shows two gated PET videos. On the left, the video illustrates a beating heart, created by combining individual gates into a continuous sequence. Similarly, the video on the right demonstrates respiratory motion, constructed from a series of respiratory-gated images. Both videos highlight how gating techniques capture and visualize motion during PET scans.
Approach: Data-Driven gated PET reconstruction
Figure 1: Approach
Figure 1: Device-less, data-driven cardiac and respiratory gating using LAFOV PET histogram images.
PET scan data is divided into time-based segments called "gates." The dynamic signals for cardiac and respiratory motion are detected and analyzed, allowing each motion phase to be assigned to a gate. Over the 5-minute scan duration, these gates represent different points in the motion cycle. Images from each gate are reconstructed separately to create clearer, motion-corrected images. This method effectively reduces motion artifacts without the need for external devices.
Results: Data-Driven gated PET reconstruction
Figure 2: Results
Figure 2 displays eight respiratory-gated PET images (numbered 1 to 8), each representing different phases of the respiratory cycle. The image in the black box shows a standard 5-minute non-gated reconstruction. The black horizontal line indicates the top position of a lung tumor. In the gated series, the tumor position shifts across the different phases due to respiratory motion. In contrast, the non-gated image shows a more blurred and less defined tumor structure, highlighting the impact of motion correction in improving tumor localization and clarity.

Contact Information

Name: Nanna Overbeck
Email: nanna.overbeck.petersen.01@regionh.dk
Location: Department of Clinical Physiology and Nuclear Medicine, Rigshospitalet, Denmark
Position: PhD Student

Publications

Device-Less Data-Driven Cardiac and Respiratory Gating Using LAFOV PET Histo Images