Quantitative and visual analyses of the effect of activity reduction on image metrics and quality in 18F-FDG PET/MRI in pediatric oncology.

BACKGROUND

: The aim of our study was to evaluate the effect of reduced injected tracer activities on the quantitative image metrics and the visual image quality in whole-body 18F-FDG PET/MRI with TOF capability in pediatric oncology.

METHODS

Seventy-seven PET/MRI examinations of 54 patients were analyzed (standard injected activity: 1.9 MBq/kg, standard PET scan duration: 5 min per bed position). Lower activity PET images (1.2 MBq/kg and 0.9 MBq/kg) were retrospectively simulated from the originally acquired list-mode data sets. Quantitative parameters were assessed by measuring the SUV metrics, signal-to-noise ratio (SNR), contrast-to-noise ratios (CNR), and textural features in each PET data set. PET images were also evaluated visually for image quality by using a scoring system.

RESULTS

SNRs were found as significantly different among PET data sets (p < 0.001) and showed increasing image noise with decreasing activities. CNR values did not show significant differences among PET data sets. The mean relative percentage changes in SUV metrics were found to be lower in 1.2 MBq/kg data set compared to 0.9 MBq/kg data set. Lesion SUVmax, SUVmean, SULpeak, and textural features were significantly different in 0.9 MBq/kg data set compared to the original data set (p < 0.05 for all). However, SUV metrics and textural features did not show a significant difference between the original and 1.2 MBq/kg data sets. While, the mean visual scores in 0.9 MBq/kg data set were significantly different compared to the original data set (p < 0.001), there was no significant difference between the original and 1.2 MBq/kg data sets in terms of general image quality and image sharpness.

DISCUSSION

Our analyses showed that the reduction of injected activity to 1.2 MBq/kg may be feasible in pediatric oncological PET/ MRI, with a smaller relative percentage change in quantitative parameters and with similar image quality to the original data set.