Quantitative Parametric Perfusion Images Using 15O-Labeled Water and a Clinical PET/CT Scanner: Test–Retest Variability in Lung Cancer
Open Access
- 18 October 2010
- journal article
- clinical trial
- Published by Society of Nuclear Medicine in Journal of Nuclear Medicine
- Vol. 51 (11) , 1684-1690
Abstract
Quantification of tumor perfusion using radioactive water (H215O) and PET is a promising method for monitoring treatment with antiangiogenic agents. However, use of dynamic H215O scans together with a fully 3-dimensional clinical PET/CT scanner needs to be validated. The purpose of the present study was to assess validity and reproducibility of dynamic H215O PET/CT scans for measuring tumor perfusion and validate the quantitative accuracy of parametric perfusion images. Methods: Eleven patients with non–small cell lung cancer were included in this study. Patients underwent 2 dynamic H215O (370 MBq) PET scans on the same day. During the first scan, arterial blood was withdrawn continuously. Input functions were derived from blood sampler data and the ascending aorta as seen in the images themselves (image-derived input function [IDIF]). Parametric perfusion images were computed using a basis function implementation of the standard single-tissue-compartment model. Volumes of interest (VOIs) were delineated on low-dose CT (LD-CT) and parametric perfusion images. Results: VOIs could be accurately delineated on both LD-CT and parametric perfusion images. These parametric perfusion images had excellent image quality and quantitative accuracy when compared with perfusion values determined by nonlinear regression. Good correlation between perfusion values derived from the blood sampler input function and IDIF was found (Pearson correlation coefficient, r = 0.964; P < 0.001). Test–retest variability of tumor perfusion was 16% and 20% when delineated on LD-CT and parametric perfusion images, respectively. Conclusion: The use of ascending aorta IDIFs is an accurate alternative to arterial blood sampling for quantification of tumor perfusion. Image quality obtained with a clinical PET/CT scanner enables generation of accurate parametric perfusion images. VOIs delineated on LD-CT have the highest reproducibility, and changes of more than 16% in tumor perfusion are likely to represent treatment effects.Keywords
This publication has 25 references indexed in Scilit:
- Choi response criteria for early prediction of clinical outcome in patients with metastatic renal cell cancer treated with sunitinibBritish Journal of Cancer, 2010
- Early MRI response monitoring of patients with advanced hepatocellular carcinoma under treatment with the multikinase inhibitor sorafenibBMC Cancer, 2009
- Sorafenib in Advanced Hepatocellular CarcinomaNew England Journal of Medicine, 2008
- Sunitinib for Treatment of Advanced Renal Cell Cancer: Primary Tumor ResponseClinical Cancer Research, 2008
- Paclitaxel plus Bevacizumab versus Paclitaxel Alone for Metastatic Breast CancerNew England Journal of Medicine, 2007
- Sorafenib in Advanced Clear-Cell Renal-Cell CarcinomaNew England Journal of Medicine, 2007
- Assessment of Pharmacodynamic Vascular Response in a Phase I Trial of Combretastatin A4 PhosphateJournal of Clinical Oncology, 2003
- STATISTICAL METHODS FOR ASSESSING AGREEMENT BETWEEN TWO METHODS OF CLINICAL MEASUREMENTThe Lancet, 1986
- Estimating the Dimension of a ModelThe Annals of Statistics, 1978
- A new look at the statistical model identificationIEEE Transactions on Automatic Control, 1974