The effect of modern PET technology and techniques on the EANM paediatric dosage card

The EANM dosage card is designed to achieve the same effective dose irrespective of the child’s size. This is achieved by using a weight-based scaling dependent on radiopharmaceutical, together with a minimum activity [4]. In adult PET, weight-based scaling is also used. However, in that case, the emphasis is more on achieving similar levels of image noise across different patient sizes. This concept was disregarded in the EANM paediatric dosage card because it can lead to excessive effective doses [17], although this approach has been questioned for PET [18]. Instead, to achieve an acceptable level of image noise in paediatric imaging, increasing the scanning duration from that used in adults can be useful.

In the EANM adult tumour PET imaging guidelines [14], in addition to linear (and non-linear) weight-based scaling of injected activity, the concept of including scan duration has been incorporated, together with differing factors depending on the applied bed overlap. The latter being suggested because of the influence of overlap on image noise in whole body PET imaging [19]. Given the history of the EANM dosage card project, and the generality of its approach to cover all nuclear medicine and PET investigations, the dosage card did not explicitly suggest a scan time and bed overlap. Instead, it assumed the same scan duration as for adults. Based on the literature at the time the dosage card was written, for PET imaging, a scanning duration of 3 min per bed position and manufacturer recommended bed overlaps were advised. While scanning times are relatively standard at around 3 min per bed position at many PET centres, any adjustments that reduce or increase scan duration should come with corresponding adjustments to injected activity. While the dosage card is applicable for standard bed overlaps, given the range of bed overlaps provided by the manufacturers (~ 20% to ~ 50%), further optimisation in this area is still possible, potentially using the recommendations of the EANM adult guidelines as a starting point. For sites that use continuous bed motion PET acquisitions [20], manufacturer recommended table speeds and/or noise equivalent count (NEC)-based conversions between time per bed position and overlap to table speed should be used [21].

In recent years, there has been a significant shift in PET hardware technology, from traditional PET/CT systems to the introduction of PET/MR, “digital” semiconductor readout of light from the scintillation crystals, and longer (≥ 25 cm) axial field of view detector rings (Table 1). While the current EANM dosage card is still relevant and applicable with these shifts in technology, the improved sensitivity that comes with these developments allows the potential reduction of injected activities.

While 15–18 cm axial field of view systems were traditionally standard, the extension of the PET detector ring beyond 20 cm and now 25 cm on PET/MR and PET/CT systems has improved performance. Extending the axial field of view allows the scanner to capture more lines of response from the positron emissions, which in turn improves the systems sensitivity almost by the square of the length. As shown in Table 1, whereas sensitivity measurements of around 6–9 kcps/kBq have been common in traditional systems with axial field of views less than 20 cm, the sensitivity of contemporary long field of view (≥ 25 cm) PET/CT and PET/MR scanners is 2–3 times this figure at 16–23 cps/kBq. Also of note is the relatively similar sensitivities of these long field of view PET/MR and PET/CT systems.

The transition from traditional photomultiplier systems to semiconductor SiPMs detectors has seen further gains in PET performance. New scintillation detectors using SiPMs found in both PET/CT and PET/MR systems have a slightly higher sensitivity but, more significantly, are capable of timing resolutions close to 200 ps. The latter translates into significant improvements to time-of-flight PET performance [38] and associated signal to noise benefits [39].

For the three PET/MR scanner systems (Siemens Biograph mMR, GE Signa, UI uPMR790) and long-field-of-view scanners (≥ 25 cm) with semiconductor readout (Siemens Vision 600, GE Discovery MI 5R) currently available, the sensitivities are higher than those for conventional systems (see Table 1).

At the time of writing, there were three publications, describing a potential reduction of the injected activity for [¹⁸F]FDG PET/MR in a paediatric population [11‐13]. Overall, in these three papers, 52 patients were scanned with two different PET/MR scanner types. Gatidis et al. [11] (24 patients) scanned at 4 min per bed position with an activity of 3.1 MBq/kg. They found that an administered activity of 1.5 MBq/kg was sufficient for adequate results. This value corresponds to 6 MBq.min bed⁻¹ kg⁻¹ (4 min × 1.5 MBq/kg). Zucchetta et al. [12] (17 patients) applied 5 min per bed position with an administered activity of 3 MBq/kg body weight. They also found that an administered activity of 1.5 MBq/kg is adequate for sufficient image quality (7.5 MBq.min bed⁻¹ kg⁻¹). Both groups used a SIEMENS Biograph mMR. Schmall et al. [13] (11 patients) injected 3.7 MBq/kg with a bed overlap of 33%, using a GE Signa scanner. They found adequate image quality for an administered activity of 2.5 MBq/kg, however for a scan duration of 3 min per bed position (7.5 MBq.min bed⁻¹ kg⁻¹). In each study, image quality was assessed based on its visual impression. In addition, the study of Zucchetta et al. [12] demonstrated by using several image quality metrics the potential of reducing the activity in clinical PET applications.

It can be concluded that an injected activity between 6 and 7.5 MBq.min bed⁻¹ kg⁻¹ is sufficient to provide adequate image quality for a standard bed overlap. As a result, we recommend using the equivalent of 2.5 MBq/kg for a 3 min/bed position for both PET/MR scanners, provided the bed overlap is similar to the values reported [13]. For different scanner types, bed overlaps, and scan duration per bed, the activity values should be adjusted accordingly. For example, an increase of the scan time to 5 min per bed would result in an activity that could be reduced to 60% of the calculated value. The activities are calculated for commonly used PET/MR systems and currently available long-field-of-view scanners (≥ 25 cm) with semiconductor readout. Systems with a long-field-of-view of more than 32 cm are not within the scope of this recommendation, although they may be addressed at a later time.

The dosage recommendations should be taken in context of “good practice” of nuclear medicine and do not substitute for national and international legal or regulatory provisions. Lower activities can be administered if the PET systems used suits the needs of the clinic with respect to their specific equipment, clinical preference, and the particular needs of their patients.

In order to be consistent with the previous versions of the EANM dosage card and the methodology described [4, 5, 7, 8], the following reduced values for [¹⁸F]FDG PET/MR and long-field-of-view PET/CT are proposed for a scan duration of 3 min per bed position: