Background

Ischemic heart disease is the leading cause of death worldwide, and hence, diagnosis and guidance of medical treatment and interventional revascularization are of great importance. In nuclear medicine, regional myocardial ischemia can be assessed by single-photon emission computed tomography (SPECT) or positron emission tomography (PET) with comparison of a rest and a stress scan (exercise or pharmacologically induced). Recently, myocardial perfusion flow imaging (MPI) using PET has proven superior to SPECT MPI, partly due to improved image spatial resolution and partly due to the possibility to determine absolute myocardial blood flow (MBF) and myocardial flow reserve (MFR). MBF and MFR enable the diagnosis of balanced three-vessel- and small-vessel disease.1,2,3,4 The generator-produced isotope Rubidium-82-Chloride (82Rb) is the most widely used PET tracer for MPI, and offers the possibility of retention imaging, but is limited by its incomplete extraction from the blood.5 Nitrogen-13-Ammonia (13N-ammonia) is a cyclotron-produced tracer with more optimal extraction than 82Rb, but impractical in clinical routine as the longer half-life necessitates a waiting time of 30-40 minutes between rest and stress imaging.1 The tracer O-15-H2O is freely diffusible over the cell membrane and therefore the perfect flow tracer, reflected in its status as the gold standard of non-invasive blood flow imaging. However, contrary to previously used measurements of MPI relying on retention of tracer as a measure of perfusion, O-15-H2O PET MPI requires the calculation of MBF based on tracer kinetics. No retention images are therefore produced as part of a clinical evaluation. Also, with an isotope half-life of 122 seconds, O-15-H2O MPI requires continuous on-site cyclotron production of the positron-emitting 15O-isotope.6 A bedside O-15-H2O generator with an automatic infusion system, fed directly with 15O from the cyclotron, was recently developed to enable repeated scans. We implemented O-15-H2O PET MPI in the daily clinical routine in May 2020.

Multiple extracardiac findings with increased perfusion have previously been reported as coincidental findings on SPECT and PET MPI retention images.7,8 As one of the first PET-Centres to implement O-15-H2O PET MPI in daily clinical routine, we expected similar findings in our patient cohort but were uncertain as to whether these would be readily visible using the non-retained radiotracer O-15-H2O. Hence, the aim of this paper is to share our initial experience regarding extracardiac findings on O-15-H2O PET MPI.

Methods

O-15-H2O PET MPI was implemented in daily clinical routine in May 2020. All scans were interpreted as routine clinical reads by a nuclear medicine specialist, and in most cases also by a resident doctor. Cases were collected consecutively during daily clinical work, especially since June 2021. Additionally, the reports of all scans were retrospectively reviewed, assessing the rate of coincidental findings.

All patients were scanned according to the standard clinical MPI protocol for O-15-H2O at Aarhus University Hospital. Patients were scanned during rest and during pharmacologically induced hyperemia on a GE Discovery MI Digital Ready PET/CT (GE Healthcare, Waukesha, Wisconsin, USA). All patients were instructed to abstain from caffeine intake for 24 hours before the PET scan. The scans were 4 minutes dynamic acquisitions (frame structure: 10 × 5 seconds, 4 × 10 seconds, 2 × 15 seconds, 3 × 20 seconds and 2 × 30 seconds) initiated at the same time as the infusion of an intravenous bolus of 400 MBq O-15-H2O using an automated production and infusion system (MedTrace Pharma AS, Hørsholm, Denmark). To induce hyperemia, a 6-minute intravenous infusion of adenosine (0.14 mg⋅kg⋅min) was initiated 2 minutes prior to the stress scan start. The 4-minute-long stress scan was performed under continuous ECG surveillance. Blood pressure and pulse were measured during the procedure.

An ultra-low-dose CT scan (ACCT) was performed prior to the PET scans to correct for attenuation (100 kVp, rotation time 0.7 second, pitch 1, Noise Index 125, slice thickness 3.75 mm, slice spacing 3.27 mm, Field-of-view 50 cm). PET images were reconstructed using the VuePointFX algorithm in isotropic voxels (3.27 × 3.27 × 3.27 mm3 in a 41.9 cm Field-of-view).

Clinical MPI images were analysed using carPET (cardiac research software package developed at Aarhus and Uppsala Universities based on the algorithms described in Harms et al.9). The software displays parametric polarmaps of the left ventricle (MBF and MFR) as well as parametric images in standard cardiac orientations with a restricted field of view and the assessment of relative and absolute perfusion is based solely on parametric images. By contrast, extracardiac perfusion was assessed by simple unweighted summing of all dynamic frames. This was done using the commercially available software Hermes Hybrid Viewer v5.0 (Hermes Medical Solutions, Stockholm, Sweden). The qualitative agreement between summed images and perfusion was verified by comparison with full field of view parametric images (not produced routinely). Biopsy histopathology and other imaging modalities served as reference standard.

Results

In total, 2963 patients have undergone clinical O-15-H2O PET MPI at our department. 371 (13%) of these included reports of coincidental findings, generally findings on low-dose CT. 135 (36%) of the coincidental findings were unchanged lung nodules and 142 (38%) were various benign conditions. Benign lung conditions constitute the majority of benign findings, such as atelectasis, pleural effusion, pneumonia, pleural plaques, bullae, etc., but also cysts, fractures, hernia and benign mamma findings were reported. 94 (25%) reports were potential malignancies with 81 being new/unknown lung nodules or ground glass opacities, 10 being expressed as highly suspicious of malignancy and 3 being enlarged lymph nodes. In total 106 (4%) patients had downstream testing.

Various malignant lesions as well as benign findings with increased perfusion appeared during the scans. The malignant extracardiac findings with increased perfusion are listed in Table 1, whereas the benign findings are listed in Table 2. The extracardiac findings from this study using O-15-H2O are compared to those reported using 82Rb and 13N-ammonia in Table 3.

Table 1 Malignant extracardiac findings with increased perfusion
Full size table
Figure 1
figure 1

A tumour in the right breast of this 71-year old woman revealed focally increased perfusion on rest O-15-H2O PET/CT Maximum Intensity Projection (MIP), axial PET and axial PET/CT fusion image (A, broad arrows) and an axillary lymph node was even visible as well (A, thin arrow). However, the tumour was not detected until subsequent mammography and later on staged using F-18-FDG PET/CT with highly increased FDG uptake in both tumour (B, broad arrow) and lymph node (B, thin arrow)

Full size image
Figure 2
figure 2

Rest scan images of malignant findings. A a lung cancer with increased perfusion and FDG uptake in the periphery and without uptake in the necrotic centre. B a lung tumour with increased perfusion and FDG uptake turned out to be large-celled neuroendocrine tumour. C Mediastinal and hilar lymph node masses with increased perfusion and FDG uptake represented lymphoma. D Lung metastases from renal cell carcinoma with focally increased perfusion

Full size image
Table 2 Benign extracardiac findings with increased perfusion
Full size table
Table 3 Summary table comparing extracardiac findings from the present study with others reported using Rb82-chloride and N13-ammonia
Full size table
Figure 3
figure 3

Rest scan images of benign findings. A Highly increased perfusion on O-15-H2O PET/CT in pneumonic infiltrate, which completely regressed on subsequent F-18-FDG PET/CT (arrow). B Patchy, ground-glass like, peripheral lung infiltrates spread in both lungs 12 days after positive COVID-19 PCR test (arrows). C Lung infiltrate with highly increased perfusion on O-15-H2O PET/CT, and moderate FDG uptake turned out to be chronical inflammation and fibrosis on biopsy (arrow). D Pleural plaques where a part of the non-calcified plaque had increased perfusion (arrow), whereas other plaques had low perfusion. E Rib fractures a few weeks after cardiopulmonary resuscitation, where bilateral rib fractures and the sternum revealed focally increased perfusion (arrows). F Gynecomastia with increased perfusion (arrow). G Increased perfusion in atelectasis (broad arrows) in close relation to pleural effusions (thin arrows), which naturally has very low perfusion. H A lung nodule with slightly increased perfusion and no pathological FDG uptake, which turned out benign on biopsy (arrow)

Full size image
Figure 4
figure 4

Summed and parametric rest O-15-H2O PET/CT images with highly increased perfusion in known pulmonary sarcoidosis lesions (arrows) diagnosed on F-18-FDG PET/CT

Full size image

Discussion

The present case series demonstrates various malignant and benign coincidental findings with increased perfusion, detected during clinical O-15-H2O PET MPI. PET imaging of tumour perfusion has been proposed as a potential cancer biomarker10 and has been studied with O-15-H2O PET in various malignancies including brain,11,12 head and neck,13,14 lung,15 breast,16,17 pancreas,18 cervix uteri,19 abdominal,20 and prostate cancer.21,22 Furthermore, increased perfusion has been described in studies using 82Rb PET in primary brain tumours,23,24,25,26,27,28 brain metastases from breast cancer29 and prostate cancer.30 And finally, both lymphoma, multiple myeloma, lung- and breast cancer,,7,31 kidney cancer,32 pheochromocytoma,33,34 and metastases from neuroendocrine tumour35 were described as coincidental findings during 82Rb PET MPI.

Hence, our coincidental findings of primary cancers and metastases were highly expected. In all cases but one, the extracardiac findings were visible on low-dose CT. In a single case however, a breast cancer was located partly outside the low-dose CT cardiac-optimized field of view but was actually visible as focally highly increased perfusion on the PET images (Figure 1). The cancer was detected by mammography and staged using F-18-Fluorodeoxyglucose (FDG) PET/CT, and was retrospectively visible on O-15-H2O PET. In our view, this underscores the importance of viewing the O-15-H2O PET images in addition to the low-dose CT images. Nevertheless, malignant lesions were more clearly defined on F-18-FDG PET/CT images compared to summed O-15-H2O PET/CT images and the cardiac PET do not replace F-18-FDG PET/CT imaging.

Some of the benign coincidental findings reported were caused by inflammation or infection such as pneumonia, COVID-19 infection, and chronical inflammatory lung infiltrate. Others are probably a mixture of inflammation and increased metabolism such as healing rib fractures, atelectasis36,37 and gynecomastia. Both inflammation and hyperplasia are conditions characterized by increased perfusion.

With few exceptions, the perfusion in the extracardiac findings were identical. One of the malignant lesions was slightly more intense on stress and two were slightly more intense on rest examination. Of the benign findings, one gynecomastia case was intense on stress and without increased perfusion on rest scan. Besides, the fractures were more visible on rest examination, possibly due to lower adjacent muscular perfusion.

Splenic switch-off is proposed as a marker of adenosine response using 13N-ammonia PET and 82Rubidium PET, and this phenomenon was also seen in most cases using O-15-H2O PET (Figure 5A). It has been shown that the presence of splenic switch-off indicates sufficient adenosine-induced coronary vasodilatation, whereas absence of splenic switch-off is not a reliable marker of failed adenosine response.38,39 This information can be extremely helpful in distinguishing patients with balanced hypoperfusion from non-responders.

Figure 5
figure 5

A Patient with pronounced splenic switch-off after adenosine stress. B Patient with increased muscular perfusion in the serratus anterior muscle during adenosine stress. The same phenomenon is seen in the erector spinae muscle in A. C An apparent segmental lung perfusion defect in the right lower lobe on the O-15-H2O PET/CT early dynamic images. However, no pulmonary embolism was seen on neither conventional ventilation/perfusion scintigraphy nor CT-angiography

Full size image

Muscular perfusion generally increases during adenosine response (Figure 5B). Since this phenomenon is not limited to respiratory muscles, it may be explained by muscular vasodilation rather than by increased respiratory work and might also serve as a marker of adenosine response.

Using the early dynamic O-15-H2O images, an evaluation of regional lung perfusion should theoretically be possible. However, a well-defined segmental pulmonary defect in the O-15-H2O PET frame with maximum lung signal (Figure 5C) could not subsequently be verified as a lung embolus on CT lung angiography or ventilation-perfusion SPECT/CT. Thus, the clinical usefulness of examining the early pulmonary passage of O-15-H2O is therefore currently questionable and detection of regional pulmonary perfusion with O-15-H2O may require a more sophisticated analysis.40

With few exceptions, findings from the summed O-15-H2O images could also be seen in full field of view parametric images of blood flow suggesting that the interpretation of these as increased perfusion is correct. However, since the cardiac software used to optimize the view of the myocardium masks out the lung, only large lesions in the lungs with high density on the CT could be visualized in the parametric images. Consequently, the lung metastases from renal cell carcinoma were not visible and only around half of the COVID/19 changes were visible, whereas the primary lung adenocarcinoma and the lung sarcoidosis (Figure 4) were visible in the parametric images. The summed images are thus a convenient alternative to parametric images for extracardiac lessons.

New knowledge gained

Benign and malignant coincidental extracardiac findings with increased perfusion are frequently encountered on O-15-H2O PET/CT. These coincidental findings and a discussion of indirect measures of adenosine response has not previously been described using O-15-H2O, but fits in line with the clinical experience from perfusion imaging with other PET tracers. The paper provides clinically relevant information for colleagues at PET-centers who perform cardiac PET, and especially those who consider or plan to perform cardiac PET with O-15-H2O.

Conclusions

Both malignant and benign extracardiac coincidental findings with increased perfusion are readily visible and frequently encountered on summed O-15-H2O PET images. In most cases, the coincidental findings can be verified by lesions on the accompanying low-dose CT-scan. We therefore recommend evaluating summed O-15-H2O images in addition to the cardiac software’s display of parametric MBF images when doing an O-15-H2O MPI report.