18F-FDG:18F-NaF PET/MR multi-parametric imaging with kinetics-based bone segmentation for enhanced dual-tracer PET quantification

Dual-tracer Positron Emission Tomography (PET) imaging involves the administration of two tracers for their co-registered evaluation within the same imaging subject and session. Previously, the concept of mixing 18F-Fluorodeoxyglucose (18F-FDG) and 18F-sodium fluoride (18F-NaF or 18F-Fuoride) tracers had been proposed in oncology to enable characterization of a wider spectrum of tumor types from a single PET/CT exam. In this study, we are proposing an optimized F18-FDG:F18-NaF cocktail PET/MR imaging framework to enable multi-parametric PET/MR assessments of enhanced quantification. Our aims are: (i) to benefit from the superior soft tissue characterization from MR and, thus, enable more accurate PET lesion localization, (ii) significantly reduce radiation exposure, therefore allowing for multiple low-dose exams to closely monitor treatment response in oncology, (iii) facilitate multi-parametric evaluations for cardiovascular diseases too, and iv) exploit the high bone uptake rate of the cocktail NaF compound to improve PET/MR attenuation correction. For that purpose, we introduced a clinically feasible single-bed dynamic 18F-FDG:18F-NaF PET/MR protocol employing the same recommended dosage as for single-tracer PET studies. Subsequently, the dynamic PET images were robustly analyzed with the Patlak method to produce tracer uptake rate Ki parametric images. Then, the bone regions were segmented from the Ki and PET images and included as a fifth tissue class in the co-registered MR Dixon-based 4-class attenuation map. Our series of rabbit dual-tracer PET/CT and PET/MR exams, as well as the subsequent human carotid PET/MR studies, suggested a maximum FDG:NaF ratio of 4:1. This cocktail synthesis yielded sufficient signal for the robust segmentation of the major bone tissues, while FDG count statistics remained high. Finally, an average 15% improvement was observed for PET FDG:NaF bone quantification with the proposed 5-tissue class attenuation correction method.