Ezgi Ilan1,2, Mattias Sandström3,2, Irina Velikyan3,4, Anders Sundin3,4, Barbro Eriksson5, Mark Lubberink3,2. 1. Section of Nuclear Medicine and PET, Department of Surgical Sciences, Uppsala University, Uppsala, Sweden ezgi.ilan@akademiska.se. 2. Medical Physics, Uppsala University Hospital, Uppsala, Sweden. 3. Section of Nuclear Medicine and PET, Department of Surgical Sciences, Uppsala University, Uppsala, Sweden. 4. PET-Centre, Medical Imaging Centre, Uppsala University Hospital, Uppsala, Sweden; and. 5. Section of Endocrine Oncology, Department of Medical Science, Uppsala University Hospital, Uppsala, Sweden.
Abstract
68Ga-DOTATOC and 68Ga-DOTATATE are radiolabeled somatostatin analogs used for the diagnosis of somatostatin receptor-expressing neuroendocrine tumors (NETs), and SUV measurements are suggested for treatment monitoring. However, changes in net influx rate (Ki) may better reflect treatment effects than those of the SUV, and accordingly there is a need to compute parametric images showing Ki at the voxel level. The aim of this study was to evaluate parametric methods for computation of parametric Ki images by comparison to volume of interest (VOI)-based methods and to assess image contrast in terms of tumor-to-liver ratio. Methods:Ten patients with metastatic NETs underwent a 45-min dynamic PET examination followed by whole-body PET/CT at 1 h after injection of 68Ga-DOTATOC and 68Ga-DOTATATE on consecutive days. Parametric Ki images were computed using a basis function method (BFM) implementation of the 2-tissue-irreversible-compartment model and the Patlak method using a descending aorta image-derived input function, and mean tumor Ki values were determined for 50% isocontour VOIs and compared with Ki values based on nonlinear regression (NLR) of the whole-VOI time-activity curve. A subsample of healthy liver was delineated in the whole-body and Ki images, and tumor-to-liver ratios were calculated to evaluate image contrast. Correlation (R2) and agreement between VOI-based and parametric Ki values were assessed using regression and Bland-Altman analysis. Results: The R2 between NLR-based and parametric image-based (BFM) tumor Ki values was 0.98 (slope, 0.81) and 0.97 (slope, 0.88) for 68Ga-DOTATOC and 68Ga-DOTATATE, respectively. For Patlak analysis, the R2 between NLR-based and parametric-based (Patlak) tumor Ki was 0.95 (slope, 0.71) and 0.92 (slope, 0.74) for 68Ga-DOTATOC and 68Ga-DOTATATE, respectively. There was no bias between NLR and parametric-based Ki values. Tumor-to-liver contrast was 1.6 and 2.0 times higher in the parametric BFM Ki images and 2.3 and 3.0 times in the Patlak images than in the whole-body images for 68Ga-DOTATOC and 68Ga-DOTATATE, respectively. Conclusion: A high R2 and agreement between NLR- and parametric-based Ki values was found, showing that Ki images are quantitatively accurate. In addition, tumor-to-liver contrast was superior in the parametric Ki images compared with whole-body images for both 68Ga-DOTATOC and 68Ga DOTATATE.
RCT Entities:
68Ga-DOTATOC and 68Ga-DOTATATE are radiolabeled somatostatin analogs used for the diagnosis of somatostatin receptor-expressing neuroendocrine tumors (NETs), and SUV measurements are suggested for treatment monitoring. However, changes in net influx rate (Ki) may better reflect treatment effects than those of the SUV, and accordingly there is a need to compute parametric images showing Ki at the voxel level. The aim of this study was to evaluate parametric methods for computation of parametric Ki images by comparison to volume of interest (VOI)-based methods and to assess image contrast in terms of tumor-to-liver ratio. Methods: Ten patients with metastatic NETs underwent a 45-min dynamic PET examination followed by whole-body PET/CT at 1 h after injection of 68Ga-DOTATOC and 68Ga-DOTATATE on consecutive days. Parametric Ki images were computed using a basis function method (BFM) implementation of the 2-tissue-irreversible-compartment model and the Patlak method using a descending aorta image-derived input function, and mean tumor Ki values were determined for 50% isocontour VOIs and compared with Ki values based on nonlinear regression (NLR) of the whole-VOI time-activity curve. A subsample of healthy liver was delineated in the whole-body and Ki images, and tumor-to-liver ratios were calculated to evaluate image contrast. Correlation (R2) and agreement between VOI-based and parametric Ki values were assessed using regression and Bland-Altman analysis. Results: The R2 between NLR-based and parametric image-based (BFM) tumor Ki values was 0.98 (slope, 0.81) and 0.97 (slope, 0.88) for 68Ga-DOTATOC and 68Ga-DOTATATE, respectively. For Patlak analysis, the R2 between NLR-based and parametric-based (Patlak) tumor Ki was 0.95 (slope, 0.71) and 0.92 (slope, 0.74) for 68Ga-DOTATOC and 68Ga-DOTATATE, respectively. There was no bias between NLR and parametric-based Ki values. Tumor-to-liver contrast was 1.6 and 2.0 times higher in the parametric BFM Ki images and 2.3 and 3.0 times in the Patlak images than in the whole-body images for 68Ga-DOTATOC and 68Ga-DOTATATE, respectively. Conclusion: A high R2 and agreement between NLR- and parametric-based Ki values was found, showing that Ki images are quantitatively accurate. In addition, tumor-to-liver contrast was superior in the parametric Ki images compared with whole-body images for both 68Ga-DOTATOC and 68Ga DOTATATE.
Authors: Christos Sachpekidis; Leyun Pan; Boris A Hadaschik; Klaus Kopka; Uwe Haberkorn; Antonia Dimitrakopoulou-Strauss Journal: Am J Nucl Med Mol Imaging Date: 2018-10-20
Authors: Julian M M Rogasch; Frank Hofheinz; Lutz van Heek; Conrad-Amadeus Voltin; Ronald Boellaard; Carsten Kobe Journal: Diagnostics (Basel) Date: 2022-02-10
Authors: Asma Bashir; Mark Bitsch Vestergaard; Tina Binderup; Helle Broholm; Lisbeth Marner; Morten Ziebell; Kåre Fugleholm; Tiit Mathiesen; Andreas Kjær; Ian Law Journal: Eur J Nucl Med Mol Imaging Date: 2020-03-13 Impact factor: 9.236
Authors: Philippe Thuillier; David Bourhis; Jean Philippe Metges; Romain Le Pennec; Karim Amrane; Ulrike Schick; Frédérique Blanc-Beguin; Simon Hennebicq; Pierre-Yves Salaun; Véronique Kerlan; Nicolas Karakatsanis; Ronan Abgral Journal: Sci Rep Date: 2021-03-01 Impact factor: 4.379