Optimal image-derived input function models for multi-parameter analysis and acceptably reduced acquisition time in [18F]F-FAPI-42 PET total-body dynamic imaging for lung cancer.

PURPOSE: Lung tumors, which receive dual-blood-supply from the pulmonary and bronchial arteries, may exhibit distinct kinetic parameters compared to other malignancies. This study aimed to investigate the impact of various factors on the kinetic parameter quantification of [18F]F-FAPI-42 dynamic PET/CT and to establish an acceptable shortened acquisition time for lung cancer. METHODS: A total of 19 patients with lung tumors underwent 60-minute total-body dynamic [18F]F-FAPI-42 PET/CT imaging. Tumor kinetic metrics (K1 to K3 and Ki) were calculated using a two-tissue irreversible comparative (2TiC) model. The effects of different image-derived input function (IDIF) models (derived from the right ventricle [RV], left ventricle [LV], and descending aorta [DA]), as well as tumor location, pathohistological subtype and size on kinetic parameters were evaluated. Additionally, the mean standardized uptake value (SUVmean), tumor-to-background ratio (TBR), signal-to-noise ratio (SNR), and contrast-to-noise ratio (CNR) were assessed to determine an acceptable shortened acquisition time. RESULTS: The time-activity curve (TAC) of the RV exhibited the earliest and highest peak, followed by those of the LV and DA. Impact of IDIF model and tumor size on kinetic parameters of primary tumors was observed. Specifically, in the RVIF model, size of tumor > 3 cm exhibited higher K2 and K3 than those with size ≤ 3 cm (P < 0.05). Similar findings were also noted for K3 in the LVIF model (P < 0.05), but not in the DAIF model. Tumor location and pathohistological subtype had no significant impact on kinetic parameters quantification. Regarding acquisition time, the RVIF model achieved kinetic parameters equivalent to those at 60 min in 26 min, while the LVIF and DAIF models required 36 min. At 26 min, the tumors were clearly visualized, with SUVmean, SNR, CNR and TBR being equivalent or nearly approaching the values observed at 60 min. CONCLUSION: The RVIF model appears to be more suitable than the DAIF model for quantifying kinetic parameters in [18F]F-FAPI-42 PET dynamic imaging of lung cancer, with an acceptable shortened acquisition time of 26 min.