Analyzing spatial distribution between 18F-fluorodeoxyglucose and 18F-boronophenylalanine positron emission tomography to investigate selection indicators for boron neutron capture therapy

Abstract

Abstract Background 18F-FDG PET is often utilized to determine BNCT selection due to the limited availability of 18F-BPA PET, which is performed by synthesizing 18F into the boron drug used for BNCT, although the uptake mechanisms between those are different. Additionally, only a few non-spatial point parameters, such as maximum SUV (SUVmax), have reported a correlation between those in previous studies. This study aimed to investigate the spatial accumulation pattern between those PET images in tumors, which would be expected to either show higher uptake on 18F-BPA PET or be utilized in clinical, to verify whether 18F-FDG PET could be used as a selection indicator for BNCT. Methods A total of 27 patients with 30 lesions (11 squamous cell carcinoma, 9 melanoma, and 10 rhabdomyosarcoma) who received 18F-FDG and 18F-BPA PET within 2 weeks were enrolled in this study. The ratio of metabolic tumor volumes (MTVs) to GTV, histogram indices (skewness/kurtosis), and the correlation of total lesion activity (TLA) and non-spatial point parameters (SUVmax, SUVpeak, SUVmin, maximum tumor-to-normal tissue ratio (Tmax/N), and Tmin/N) were evaluated. After local rigid registration between those images, distances of locations at SUVmax and the center of mass with MTVs on each image and similarity indices were also assessed along its coordinate. Results In addition to SUVmax, SUVpeak, and Tmax/N, significant correlations were found in TLA. The mean distance in SUVmax was $$25.2 \pm 24.4\; {\text{mm}}$$ 25.2 ± 24.4 mm and significantly longer than that in the center of mass with MTVs. The ratio of MTVs to GTV, skewness, and kurtosis were not significantly different. However, the similarities of MTVs were considerably low. The similarity indices of Dice similarity coefficient, Jaccard coefficient, and mean distance to agreement for MTV40 were $$0.65 \pm 0.21$$ 0.65 ± 0.21 , $$0.51 \pm 0.21$$ 0.51 ± 0.21 , and $$0.27 \pm 0.30$$ 0.27 ± 0.30  cm, respectively. Furthermore, it was worse in MTV50. In addition, spatial accumulation patterns varied in cancer types. Conclusions Spatial accumulation patterns in tumors showed low similarity between 18F-FDG and 18F-BPA PET, although the various non-spatial point parameters were correlated. In addition, the spatial accumulation patterns were considerably different in cancer types. Therefore, the selection for BNCT using 18F-FDG PET should be compared carefully with using 18F-FBPA PET.