# Imaging small dynamic lesions using positron emission tomography and computed tomography: an 18F-sodium fluoride valvular phantom study

## TL;DR

Researchers created a phantom to improve imaging of small, moving heart valve lesions using PET and CT scans.

## Contribution

A cardiac phantom was developed to optimize 18F-NaF PET/CT imaging for small dynamic valvular structures.

## Key findings

- Phantom imaging showed higher signal-to-noise and target-to-background ratios than human cases initially.
- Reducing scan duration to 15 seconds with motion correction improved phantom image quality to match human results.
- The phantom mimics clinical 18F-NaF valve imaging and allows testing of acquisition and reconstruction methods.

## Abstract

18F-sodium fluoride (18F-NaF) positron emission tomography (PET) detects active microcalcification and predicts adverse outcomes including bioprosthetic valve deterioration. However, measuring small areas of 18F-NaF uptake within moving structures remains challenging, requiring further optimization. We developed a representative cardiac phantom to optimize 18F-NaF imaging of bioprosthetic valves.

We placed a bioprosthetic valve with two pockets sutured to the leaflets mimicking valvular lesions and a subvalvular ring mimicking the valve remnant into the phantom and injected each with 18F-radionuclide (1 μCi pockets, 4 μCi ring). We injected the cardiac chambers with iohexol and 18F-radionuclide (0.176 mCi) for background activity. PET and computed tomography (CT) images were acquired using a Siemens Biograph Vision high-resolution digital PET/CT scanner. We analysed target-to-background ratio (TBR) and signal-to-noise ratio (SNR) and subjective measures of image quality. We compared results with a human case of transcatheter aortic valve replacement. Initially the SNR and TBR in the phantom greatly exceeded those from human imaging. We reduced the scan duration used for reconstruction to 30 and 15 s, achieving comparable results (30 s vs. 15 s vs. patient: SNR 45.6 vs. 13.9 vs. 44.3, TBRmax 6.5 vs. 5.4 vs. 4.1, noise 10.2% vs. 8.8% vs. 12.0%). With motion correction, SNR and image quality improved in the phantom (30 s 135.8 vs. 45.6, 15 s 32.9 vs. 13.9) but remained similar in the human case (47.3 vs. 44.3).

A cardiac phantom can mimic clinical 18F-NaF valve bioprosthesis imaging, providing an opportunity to explore acquisition, reconstruction, and post-processing of 18F-NaF PET/CT for small mobile cardiac structures.

Graphical Abstract