Impact of time-of-flight PET on quantification accuracy and lesion detection in simultaneous 18F-choline PET/MRI for prostate cancer

• Published in: EJNMMI research 2018-05, Vol.8 (1), p.1-12, Article 41
• Main Authors: Muehlematter, Urs J., Nagel, Hannes W., Becker, Anton, Mueller, Julian, Vokinger, Kerstin N., de Galiza Barbosa, Felipe, ter Voert, Edwin E. G. T., Veit-Haibach, Patrick, Burger, Irene A.
• Format: Article
• Language: English
• Subjects: Attenuation; Attenuation correction; Cardiac Imaging; Choline; Confidence intervals; Emission analysis; Fluorine isotopes; Imaging; Lesions; Lymph; Lymphatic system; Magnetic resonance imaging; Medicine; Medicine & Public Health; Metastasis; NMR; Nuclear magnetic resonance; Nuclear Medicine; Oncology; Original Research; Orthopedics; Patients; PET/MRI; Positron emission; Prostate cancer; Radiology; Readers; Time-of-flight; Tomography
• ISSN: 2191-219X; 2191-219X
• DOI: 10.1186/s13550-018-0390-8

Background Accurate attenuation correction (AC) is an inherent problem of positron emission tomography magnetic resonance imaging (PET/MRI) systems. Simulation studies showed that time-of-flight (TOF) detectors can reduce PET quantification errors in MRI-based AC. However, its impact on lesion detection in a clinical setting with 18 F-choline has not yet been evaluated. Therefore, we compared TOF and non-TOF 18 F-choline PET for absolute and relative difference in standard uptake values (SUV) and investigated the detection rate of metastases in prostate cancer patients. Results Non-TOF SUV was significantly lower compared to TOF in all osseous structures, except the skull, in primary lesions of the prostate, and in pelvic nodal and osseous metastasis. Concerning lymph node metastases, both experienced readers detected 16/19 (84%) on TOF PET, whereas on non-TOF PET readers 1 and 2 detected 11 (58%), and 14 (73%), respectively. With TOF PET readers 1 and 2 detected 14/15 (93%) and 11/15 (73%) bone metastases, respectively, whereas detection rate with non-TOF PET was 73% (11/15) for reader 1 and 53% (8/15) for reader 2. The interreader agreement was good for osseous metastasis detection on TOF (kappa 0.636, 95% confidence interval [CI] 0.453–0.810) and moderate on non-TOF (kappa = 0.600, CI 0.438–0.780). Conclusion TOF reconstruction for 18 F-choline PET/MRI shows higher SUV measurements compared to non-TOF reconstructions in physiological osseous structures as well as pelvic malignancies. Our results suggest that addition of TOF information has a positive impact on lesion detection rate for lymph node and bone metastasis in prostate cancer patients.