We thank Panagi et al. [1] for their thoughtful approach to constructing a liver phantom for myocardial perfusion imaging (MPI). There are a number of sources within the gastrointestinal system which may be responsible for artefacts in MPI (in terms of motion, scatter and attenuation). The stomach and bowel are others [2]. Would they require similar principles of design?

Stomach and bowel are also highly variable in position [3]. This would make sense particularly in viscera which are not anatomically fixed within peritoneum. Specifically, the study concentrated on cranio-caudal motion. Are other directions clinically significant in the authors’ experience? How can the phantom be further enhanced to account adequately for intra-abdominal motion?

Beyond that, how could we account for artefactual gastric radiotracer uptake exacerbated by medications such as proton pump inhibitors? This is an increasingly common problem as this therapy becomes more common. We have found it to increase the frequency of image re-acquisition [4].