Alberto Villagran Asiares1, Teresa Vitadello2, Borjana Bogdanovic3, Esteban Lucas Solari3, Lachlan McIntosh4, Sylvia Schachoff3, Tareq Ibrahim2, Stephan G Nekolla3,5. 1. Nuklearmedizinische Klinik und Poliklinik, Klinikum Rechts der Isar, School of Medicine, Technical University of Munich, Munich, Germany. alberto.villagran@tum.de. 2. Klinik und Poliklinik für Innere Medizin I, Klinikum Rechts der Isar, School of Medicine, Technical University of Munich, Munich, Germany. 3. Nuklearmedizinische Klinik und Poliklinik, Klinikum Rechts der Isar, School of Medicine, Technical University of Munich, Munich, Germany. 4. Department of Physical Sciences, Peter MacCallum Cancer Centre, Melbourne, Australia. 5. DZHK (German Centre for Cardiovascular Research), Partner site Munich Heart Alliance, Munich, Germany.
Abstract
BACKGROUND: This work investigated the impact of different cardiac gating methods on the assessment of cardiac function by FDG-PET in a cross-validation PET/MR study. METHODS AND RESULTS: MR- and PET-based left ventricular end-diastolic, end-systolic volumes, and ejection fraction (EDV, ESV, and EF) were delineated in 30 patients with a PET/MR examination. Cardiac PET imaging was performed using three ECG gating methods: fixed number of gates per beat (STD), STD with a beat acceptance window (STD-BR), and fixed gate duration (FW). High MR-PET correlations were found in all the values. ESVs correlated better than EDVs and EFs: Pearson's r coefficient [0.92, 0.92, 0.92] in ESV vs [0.75, 0.81, 0.80] in EDV and [0.79, 0.91, 0.87] in EF, for each method [STD, STD-BR, FW]. Biases with respect to MRI for all the evaluated PET methods were less than 13% in EDV, 5% in ESV, and 14% in EF, but with wide limits of agreements, in the range (59-68)% in EDV, (65-70)% in ESV, and (49-71)% in EF. STD showed the strongest disagreement, while there were no marked differences between STD-BR and FW. CONCLUSION: Based on these findings, PET- and MR-based cardiac function parameters were highly correlated but in substantial disagreement with variabilities introduced by the selected PET ECG gating method. The most significant differences were associated with the ECG gating method susceptible to highly irregular beats, while similar performance was observed in the methods using uniform adjustment of gates width per beat with the beat acceptance window, and fixed gate width along all the beats. Thus, strict quality controls of R peak detection are needed to minimize its impact on the function assessment.
BACKGROUND: This work investigated the impact of different cardiac gating methods on the assessment of cardiac function by FDG-PET in a cross-validation PET/MR study. METHODS AND RESULTS: MR- and PET-based left ventricular end-diastolic, end-systolic volumes, and ejection fraction (EDV, ESV, and EF) were delineated in 30 patients with a PET/MR examination. Cardiac PET imaging was performed using three ECG gating methods: fixed number of gates per beat (STD), STD with a beat acceptance window (STD-BR), and fixed gate duration (FW). High MR-PET correlations were found in all the values. ESVs correlated better than EDVs and EFs: Pearson's r coefficient [0.92, 0.92, 0.92] in ESV vs [0.75, 0.81, 0.80] in EDV and [0.79, 0.91, 0.87] in EF, for each method [STD, STD-BR, FW]. Biases with respect to MRI for all the evaluated PET methods were less than 13% in EDV, 5% in ESV, and 14% in EF, but with wide limits of agreements, in the range (59-68)% in EDV, (65-70)% in ESV, and (49-71)% in EF. STD showed the strongest disagreement, while there were no marked differences between STD-BR and FW. CONCLUSION: Based on these findings, PET- and MR-based cardiac function parameters were highly correlated but in substantial disagreement with variabilities introduced by the selected PET ECG gating method. The most significant differences were associated with the ECG gating method susceptible to highly irregular beats, while similar performance was observed in the methods using uniform adjustment of gates width per beat with the beat acceptance window, and fixed gate width along all the beats. Thus, strict quality controls of R peak detection are needed to minimize its impact on the function assessment.
Authors: Riemer H J A Slart; Jeroen J Bax; Richard M de Jong; Jaep de Boer; Hildo J Lamb; Piet H Mook; Antoon T M Willemsen; Willem Vaalburg; Dirk J van Veldhuisen; Pieter L Jager Journal: J Nucl Med Date: 2004-02 Impact factor: 10.057
Authors: G Germano; H Kiat; P B Kavanagh; M Moriel; M Mazzanti; H T Su; K F Van Train; D S Berman Journal: J Nucl Med Date: 1995-11 Impact factor: 10.057
Authors: Wolfgang M Schaefer; Claudia S A Lipke; Bernd Nowak; Hans-Juergen Kaiser; Patrick Reinartz; Arno Buecker; Gabriele A Krombach; Udalrich Buell; Harald P Kühl Journal: J Nucl Med Date: 2004-01 Impact factor: 10.057
Authors: Aliasghar Khorsand; Senta Graf; Herbert Frank; Kurt Kletter; Heinz Sochor; Gerald Maurer; Ernst Schuster; Sebastian Globits; Robert Dudczak; Gerold Porenta Journal: J Nucl Med Date: 2003-11 Impact factor: 10.057