3D PET shows efficacy for quantitative myocardial blood flow analysis
Two-dimensional PET using 13N-ammonia as a tracer has established itself for the assessment of quantitative regional myocardial blood flow (MBF) and coronary flow reserve. A multinational team of researchers from Switzerland and Japan, in a recent study published in the Journal of Nuclear Medicine, sought to determine if 3D PET (with septa retracted) could provide similar accuracy.
“Although the use of 3D PET has demonstrated significant advantages over the use of 2D PET for brain imaging, the relative benefits of 3D PET for whole-body oncology and cardiac applications are less clear,” the authors wrote.
The researchers, from the University Hospital of Zurich in Zurich, Switzerland, and Osaka Medical College, in Takatsuki, Japan, compared 2D and 3D PET acquisition modes with analytic and iterative reconstruction algorithms for the absolute quantification of MBF with 13N-ammonia.
The study group consisted of 21 patients with a mean age of 63 years admitted for the assessment of myocardial perfusion with dynamic 13N-ammonia PET. Of the cohort, seven were referred for suspected coronary artery disease, nine had known coronary artery disease, and five patients with a normal coronary angiogram were evaluated for microvascular disease.
The 2D and 3D studies were conducted on an integrated PET/CT system, the 16-slice Discovery ST RX (GE Healthcare). The MBF measurements were acquired in both 2D and 3D modes at rest and during adenosine stress testing. The researchers noted that the 3D data acquisition was performed with the interplane septa removed from the field of view.
The image data was corrected for random coincidences, geometry, normalization, dead-time losses, scatter, and attenuation, according to the scientists. The 2D emission data was reconstructed with a filtered backplane projection (FBP), while the 3D data was converted into sets of contiguous transaxial 2D sinograms by use of Fourier rebinning (FORE). The 3D images were reconstructed using three techniques: FORE with FBP; FORE with ordered-subsets expectation maximization (OSEM); and a reprojection algorithm.
The researchers reported that there were no significant differences between the heart rate and the systolic blood pressure during 2D and 3D MBF measurements at rest and during peak pharmacologic stress.
“Visual inspection of the 2D static PET images revealed fixed perfusion defects in nine patients and mixed [fixed and reversible defects] in three patients,” they wrote. “Three-dimensional PET with any of the three reconstruction algorithms provided findings identical to those obtained by 2D PET, as determined by visual analysis.”
Of note, in a subset of the patients imaged in 3D mode, the researchers used a lower dose (30%) of 13N-ammonia and obtained similar results.
“These parameters may allow reductions in the radiation dose to the patient and potentially shorten the time delay between consecutive studies, as the activity reaches the background more rapidly,” they noted.
Although iterative reconstruction algorithms that use FORE-OSEM for 3D images may be an alternative to the standard 2D FDP algorithm for 13N-ammonia cardiac PET studies, the researchers acknowledged that the results are dependent on the specific parameters of the PET scanner, acquisition protocols, and reconstruction methods.
“Although the use of 3D PET has demonstrated significant advantages over the use of 2D PET for brain imaging, the relative benefits of 3D PET for whole-body oncology and cardiac applications are less clear,” the authors wrote.
The researchers, from the University Hospital of Zurich in Zurich, Switzerland, and Osaka Medical College, in Takatsuki, Japan, compared 2D and 3D PET acquisition modes with analytic and iterative reconstruction algorithms for the absolute quantification of MBF with 13N-ammonia.
The study group consisted of 21 patients with a mean age of 63 years admitted for the assessment of myocardial perfusion with dynamic 13N-ammonia PET. Of the cohort, seven were referred for suspected coronary artery disease, nine had known coronary artery disease, and five patients with a normal coronary angiogram were evaluated for microvascular disease.
The 2D and 3D studies were conducted on an integrated PET/CT system, the 16-slice Discovery ST RX (GE Healthcare). The MBF measurements were acquired in both 2D and 3D modes at rest and during adenosine stress testing. The researchers noted that the 3D data acquisition was performed with the interplane septa removed from the field of view.
The image data was corrected for random coincidences, geometry, normalization, dead-time losses, scatter, and attenuation, according to the scientists. The 2D emission data was reconstructed with a filtered backplane projection (FBP), while the 3D data was converted into sets of contiguous transaxial 2D sinograms by use of Fourier rebinning (FORE). The 3D images were reconstructed using three techniques: FORE with FBP; FORE with ordered-subsets expectation maximization (OSEM); and a reprojection algorithm.
 | |
| Static transaxial images obtained for representative patient (weight, 74 kg; height, 178 cm) with 2D FBP after injection of 900 MBq of 13NNH3 (A) and images obtained with low-dose protocol (500 MBq of 13N-NH) and 3D FORE–FBP (B), 3D FORE–OSEM (C), and 3D RP (D). Image and caption copyright © by the Society of Nuclear Medicine Inc. from "Absolute Quantification of Myocardial Blood Flow with 13N-Ammonia and 3-Dimensional PET," by Tiziano Schepis, Oliver Gaemperli, Valerie Treyer, Ines Valenta, Cyrill Burger, Pascal Koepfli, Mehdi Namdar, Itaru Adachi, Hatem Alkadhi, and Philipp Kaufmann, Journal of Nuclear Medicine. | |
The researchers reported that there were no significant differences between the heart rate and the systolic blood pressure during 2D and 3D MBF measurements at rest and during peak pharmacologic stress.
“Visual inspection of the 2D static PET images revealed fixed perfusion defects in nine patients and mixed [fixed and reversible defects] in three patients,” they wrote. “Three-dimensional PET with any of the three reconstruction algorithms provided findings identical to those obtained by 2D PET, as determined by visual analysis.”
Of note, in a subset of the patients imaged in 3D mode, the researchers used a lower dose (30%) of 13N-ammonia and obtained similar results.
“These parameters may allow reductions in the radiation dose to the patient and potentially shorten the time delay between consecutive studies, as the activity reaches the background more rapidly,” they noted.
Although iterative reconstruction algorithms that use FORE-OSEM for 3D images may be an alternative to the standard 2D FDP algorithm for 13N-ammonia cardiac PET studies, the researchers acknowledged that the results are dependent on the specific parameters of the PET scanner, acquisition protocols, and reconstruction methods.