PET/CT provides both the anatomic and metabolic features of malignant tumors for clinical evaluation, which was also demonstrated to be more sensitive particularly in detecting lymph node, bone marrow involvement and distant recurrent metastases than CT alone [26, 27]. The cumulative ionizing radiation dose by repeated scans continues to be concern, although the relationship between exposure and radiation-related illness remains debatable . Protocols for low-dose 18F-FDG PET/CT by the influential imaging associations have been posted, involving the reduction of radiotracer activity and shortening of the acquisition time. Meanwhile, several studies have focused on investigation of the effectivity of the ultra-low-dose PET imaging. Work from Liu et al.  has shown the total-body dynamic 18F-FDG PET imaging 10× reduction of injected activity (0.37 MBq/kg) allows equal performance to full-activity PET imaging in healthy volunteers. Shi et al . found that the total-body PET/CT with half-dose 18F-FDG activity can achieve the comparable image quality to conventional PET/CT.
However, pediatric patients are at a higher risk of radiation-induced cancers due to their developing bodies and greater life expectancies [28, 29]. In addition, the reported sensitivities and specificities of 18F-FDG PET/CT for tumor staging in children are over 90% [30, 31]. Few studies were concentrated on PET/CT low-dose injection activity in pediatric patients, while most of the them were only theoretical [15, 16]. Our study is the first to evaluate the impact of radiotracer dose regimen in realistic low-dose pediatric PET imaging data in the larger sample size.
The image quality of PET/CT is affected by multiple factors including the various instruments, imaging agents, waiting time, reconstruction parameters, individual subject factors (including age, BMI, blood glucose level, disease history, etc.) [25, 32, 33]. In order to eliminate the deviations as little as possible, we set a relatively narrow subject inclusion criteria, excluding children with excessive waiting time, true injected dose and BMI. Our results from the realistic half-dose and low-dose images covered an age range from 1 to 13 years old.
The disease referred to the specific common pediatric malignancies, mainly lymphomas, sarcomas, and neuroblastoma. Pediatric patients in various clinical stages were all involved (including newly diagnosed cancer for an initial staging, restaging and therapy response assessment). Different backgrounds (liver, spleen and blood pool) were selected for thoroughly evaluation considering the variability in lesions and organs’ uptake pattern, where the mediastinal blood pool, liver, and spleen serve as the cutoffs for PET positivity in the evaluation of the response assessment and post-treatment surveillance [34-37]. However, VOI determination could be still difficult due to the large difference in children’s age, height and weight.
Our results have shown that the total-body PET/CT with half-dose (1.85 MBq/kg, estimated effective dose: 1.76–2.57 mSv) of 18F-FDG was feasible for clinical application in pediatric patients, which was much lower than previous initiatives. The mean and SD of the image quality scores in G600s were 4.9 ± 0.2 for overall quality, 5.0 ± 0.0 for lesion conspicuity and 4.9 ± 0.3 for image noise. The image quality and lesion detectability were performed well as the dose of radiation reduced. In terms of the acquisition time, a longer time always results in higher image quality. The image quality showed a slight degradation as the dosing regimen reduced in low-dose images from G600s (1/2-dose) to G20s (1/60-dose) based on the Likert scoring. While the degradation in noise level was more obvious. Taking the G600s image as a reference, image quality scores in G20s were significantly lower. Based on the clinical data, we revised our original theoretical estimations  and proposed updated recommended injection-dose and acquisition time (Supplementary Table 2). Safe and sufficient overall image quality and lesion conspicuity available for clinical use could be maintained down to G60s (0.185 MBq/kg, estimated effective dose: 0.18–0.26 mSv), none of half-dose images was rated as undiagnostic, which was even beyond our previous estimation. For patients needs multiple PET/CT scan for therapy response assessment, lower-dose PET was recommended to manage the over-all exposure, while ultra-fast-scan with a full or half dose was recommended for patients that intolerant of long scan duration.
The change in lesion detectability with dose level was comparatively minimal. Fig 5 showed a representative view of a PET examination reconstructed using OSEM. As expected, lower-dose images appeared to be nosier because of the reduced number of events in the PET list file. The overall image quality scores were typically evaluated as 5, 5, 4, 3, 3 and 2 points, respectively, for G600s to G20s. The lesion was even clearly identifiable reduced down to 1/60-dose. However, our results showed that even some small lesions with low SUV uptake became difficult to detect in low-dose images due to the increased noise of background. There were 8% and 44% of false-negative lesions which could not be clearly demonstrated in G40s (1/30-dose) and G20s (1/60-dose) images, respectively, indicating they may be not feasible for clinical diagnose and therapy response assessment.
Consistent with previous observation [19, 38], the objective analyses showed that there was a pronounced increase in SUVmax and SD of backgrounds as the simulated dose was reduced, which was consistent with the visible increase noise in PET images at low doses and previous research [38, 39]. The same tendency was demonstrated in all three backgrounds, which means that a shortened acquisition time might diminish the accuracy of the SUV measurement. This increase may be caused by noise amplification therefore causing the maximum pixel value to be higher in the background measurement. There was no statistically significant deviation in TBR values as the dose level systematically reduced from half-dose, which further indicated a good maintain in lesion conspicuity.
One of major challenges in the pediatrics low-dose PET/CT images is the large variation in patient’s age, range from infants, children and adolescents with different size and metabolism, which impacts the dosimetry measurement [40, 41]. The increase in 18F-FDG uptake may be caused by age related changes in body mass, blood volume, organ volume and function [34, 42]. Separate evaluations are warranted for multiple pediatric age groups. Our studies are based on PET imaging data from patients with age ranging from 1 to 13 years old, especially involving infants, which were more distinct from adults, therefore directly including the age-dependent pharmacokinetics of 18F-FDG. SNR has almost no correlation with age, which means that, the currently half-dose regimen already provided a constant image quality throughout our patient population. Smaller children showed high SNRnorm values, which was probably caused by less attenuation and scatter due to smaller body mass. The results indicated that for the younger patients, the less activity is needed to obtain optimal image quality, which means the dosage regimen can be further reduced in younger patients, somehow alleviating the concern of repeated PET/CT.
As shown in Fig 4, G300s (1/4-dose) showed that high consistency in both lesion SUVmax and SUVmean was observed compared to half-dose images. The variability in lesion SUVmax gradually increased as the dose reduced, generally lower than 20%, which was similar to previous studies [43, 44]. The average bias on lesion SUVmax was negative from almost all other age groups except G20s (1/60-dose), which was probably due to the higher noise in G20s. This observation also indicated unstable values of lesion SUVmax at low-dose images. Comparatively, lesion SUVmean showed a smaller bias and variability as dose levels were reduced, which suggested that SUVmean was a much more accurate and stable metric for image quality at lower dose.
Our study had several potential limitations. First, it was a single-centre study, and 18F-FDG PET/CT images were obtained relevant to only a specific scanner, therefore it may not be generalizable to other centers, other PET equipment or tracers beyond FDG. Secondly, pediatric patients in our center were mainly restricted to several specific common malignancies such as lymphomas, sarcomas and neuroblastoma. The small variety of tumors might have resulted in selection bias. Furthermore, we did not used the disease as a subgroup because of the relatively small population of each tumor and different disease stages.