The winding road towards respiratory motion correction: is this just another dead-end or do we finally get breathing under control?

Abstract

Cardiac positron emission tomography (PET) imaging steadily continues to evolve into a pivotal imaging modality for the management of patients with cardiovascular diseases. While there is accumulating evidence that PET myocardial perfusion imaging may be the first choice in the evaluation of ischemic heart disease, other indications such as prosthetic heart valve endocarditis, atherosclerotic plaque imaging, cardiac amyloidosis, 5 or cardiac sarcoidosis 6 are either already established or about to be implemented into clinical routine in the near future. There are, however, several important limitations still to be addressed. Despite the substantial refinements in various aspects of hardware and software technology, exploiting the potential spatial resolution of PET systems remains challenging in the field of cardiac imaging as image quality is inherently and simultaneously affected by cardiac contractions, respiratory motion, and patient movements. As a consequence, the high theoretical spatial resolution of state-of-the-art PET scanners does not translate into the effective global spatial resolution of PET images obtained in the clinical setting. Naturally, several techniques have been developed over the years to compensate for this loss in spatial resolution. However, while the road towards correction of cardiac contractions basically starts and ends with ECG gating, the road towards getting respiratory motion under control has proved to be much less straightforward. Initial studies tested the feasibility and impact of scanning at deep inspiration on myocardial perfusion imaging. Although the technique significantly reduces image artifacts and improves image quality, patient cooperation is crucial. Another approach is to monitor and measure respiratory motion by an external device such as a pneumatic belt around the patient’s abdomen or by an infrared tracking camera with a reflective marker placed on the patient’s chest (real-time position management systems). By correlating the breathing cycle and the acquisition of the imaging data (i.e., respiratory gating), image acquisition or reconstruction is limited to certain respiratory phases. The tradeoff for this image truncation and reduction in counts, however, is either increased image noise levels or longer acquisition time which in turn may even provoke more patient motion. Furthermore, the external devices may produce significant patient discomfort. Fortunately, more recent developments allow the derivation of the respiratory signal directly from the scan data (i.e., data-driven). The need for any external devices, additional staff action, or patient cooperation can then be omitted. Moreover, respiratory motion can be compensated without losing any image data. Although the potential of these data-driven motion correction algorithms to improve spatial resolution is auspicious, the techniques have not yet found widespread adoption into clinical routine. The reason for this may very well be due to complex set-up procedures and vendor-specific algorithms (Table 1). In the current issue of the Journal, Lassen et al. investigate a data-driven, projection-based respiratory motion compensation (DPR-MoCo) approach for cardiac PET data by modeling each motion in projection space and performing correction in the listmode data Funding The University Hospital Zurich holds a research agreement with GE Healthcare. Dr. Benz reports a research grant from the Theodor und Ida Herzog-Egli-Foundation. Reprint requests: Ronny R. Buechel, MD, Department of Nuclear Medicine, Cardiac Imaging, University Hospital Zurich, Ramistrasse 100, 8091 Zurich, Switzerland; [email protected] J Nucl Cardiol 2020;27:2231–3. 1071-3581/$34.00 Copyright 2019 American Society of Nuclear Cardiology.