Non-invasive quantification of tumor blood flow in prostate cancer using ¹⁵O-H₂O PET/CT

Tumor blood flow (TBF) measurements in prostate cancer (PCa) provide an integrative index of tumor growth, which could be important for primary diagnosis and therapy response evaluation. ¹⁵O-water PET is the non-invasive gold standard but is technically demanding. The aim of this study was to compare the accuracy of three different non-invasive strategies with an invasively measured arterial input function (BSIF): Using image-derived input functions (IDIF) from either 1) a separate heart scan or 2) the pelvic scan or 3) a populations-based input function (PBIF). Nine patients with biopsy-verified PCa scheduled for prostatectomy were included. All patients were characterized with serum levels of PSA (s-PSA), multiparametric magnetic resonance imaging (mpMRI) and post-surgical histopathology Gleason Grade. Dynamic ¹⁵O-water was performed of the heart and the pelvic area 15 minutes apart. TBF estimated from both wash-in (K₁) and wash-out (k₂) constants was calculated using a one-compartmental model. Results: Mean (range) s-PSA was 12 (3-27) ng/mL, Gleason Grade Group was 2.9 (1-5), k₂ was 0.44 (0.007-1.2), and K₁ was 0.24 (0.07-0.55) mL/mL/min. k₂ (BSIF) correlated with s-PSA (r=0.86, P<0.01) and Gleason Grade Group (rho=0.78, P=0.01). BSIF, heart-IDIF and PBIF provided near-identical k₂ and K₁ (r>0.95, P<0.001) with slopes near unity. The correlations of BSIF and pelvic-IDIF rate constants were good (r>0.95, P<0.001), but individual errors high. In conclusion, non-invasive protocols for ¹⁵O-water PET with IDIF or PBIF accurately measures perfusion in prostate cancer and might be useful for evaluation of tumor aggressiveness and treatment response.