Advanced Search

Home > Journals > The Quarterly Journal of Nuclear Medicine and Molecular Imaging > Past Issues > The Quarterly Journal of Nuclear Medicine 2002 June;46(2) > The Quarterly Journal of Nuclear Medicine 2002 June;46(2):122-30



A Journal on Nuclear Medicine and Molecular Imaging

A Journal on Nuclear Medicine and Molecular Imaging
Affiliated to the Society of Radiopharmaceutical Sciences and to the International Research Group of Immunoscintigraphy
Indexed/Abstracted in: Current Contents/Clinical Medicine, EMBASE, PubMed/MEDLINE, Science Citation Index (SciSearch), Scopus
Impact Factor 2,413

Frequency: Quarterly

ISSN 1824-4785

Online ISSN 1827-1936


The Quarterly Journal of Nuclear Medicine 2002 June;46(2):122-30

Guest Editors: Freeman L. M., Bombardieri E. 


Qualitative and quantitative comparison between images obtained with filtered back projection and iterative reconstruction in prostate cancer lesions of 18F-FDG PET

Etchebehere E. C. S. C. 1,2, Macapinlac H. A. 1, Gonen M. 3, Humm K. 1, Yeung H. W. D. 1, Akhurst T. 1, Scher H. I. 4, Larson S. M. 1

1 Nuclear Med­i­cine Ser­vice, Depart­ment of ­Radiology, Memo­rial ­Sloan-Ket­tering ­Cancer ­Center, New ­York, USA
2 Divi­sion of ­Nuclear Med­i­cine, Depart­ment of ­Radiology School of Med­ical Sci­ences, Cam­pinas ­State Uni­ver­sity Uni­camp, Cam­pinas, ­Brazil
3 Depart­ment of Epi­dem­i­ology and Bio­sta­tis­tics, and
4 Genito-Uri­nary ­Oncology Ser­vice, Depart­ment of Med­i­cine Memo­rial ­Sloan Ket­tering ­Cancer ­Center, New ­York, USA

Back­ground. ­Recently, iter­a­tive recon­struc­tion ­with seg­mented atten­u­a­tion cor­rec­tions (­IRSAC) has ­been intro­duced for recon­struc­tion of 18F-FDG PET ­images. ­IRSAC pro­duces ­images ­that are ­more ­pleasing to the eye, but qual­ita­tive and quan­ti­ta­tive com­par­i­sons ­between ­IRSAC and fil­tered ­back pro­jec­tion (FBP) ­have not ­been ­reported for met­a­static ­cancer. ­Since quan­ti­ta­tive ­data has ­been ­widely ­used as an ­adjunct to inter­pre­ta­tion of PET ­scans, com­par­ison ­between ­IRSAC and FBP is ­needed. The pur­pose of ­this ­study was to com­pare ­image ­quality and the max­imum stan­dard­ized ­uptake value (SUVmax) ­obtained ­with FBP and ­with ­IRSAC in met­a­static ­lesions ­from pros­tate ­cancer.
­Methods. ­Twenty 18F-FDG PET ­scans (10 base­line and 10 ­follow-up) ­were per­formed in 10 ­patients ­with pros­tate ­cancer (­ages 66-85 yrs, ­mean 73.6 yrs). Acqui­si­tion ­began 45 min ­after injec­tion of 370 MBq of 18F-FDG. ­Images ­were recon­structed ­using FBP and ­IRSAC, and sub­mitted to ­visual and quan­ti­ta­tive anal­ysis. ­SUVmax was ­obtained for all metas­tases, on FBP and ­IRSAC. A ­Jaszczak ­phantom ­study was ­also per­formed.
­Results. ­IRSAC ­images ­showed ­better ­image ­quality ­than FBP espe­cially in ­regions of ­high ­activity con­cen­tra­tions. ­IRSAC ­detected 106 ­lesions on ­both base­line and ­follow-up ­scans, ­while FBP ­detected 100 and 95 ­lesions on base­line and ­follow-up ­scans, respec­tively. There­fore, 17 ­more ­lesions ­were ­seen on ­IRSAC. The ­mean ­SUVmax ­values on base­line ­scans for FBP and ­IRSAC ­were system­at­i­cally dif­ferent, at 4.46±1.99 and 5.13±2.67, respec­tively. On ­follow-up ­scans ­values ­were 3.89±1.72 for FBP and 4.29±1.93 for ­IRSAC. Com­par­ison of FBP ­with ­IRSAC on base­line and ­follow-up ­scans ­were sta­tis­ti­cally sig­nif­i­cant (base­line: ­paired “t”-­test p=0.0017; ­follow-up: ­paired “t”-­test p=0.0008). ­Phantom ­studies ­reveal ­that ­these dif­fer­ences can be ­explained by the ­type of recon­struc­tion fil­ters ­used, and ­IRSAC was ­more accu­rate ­than FBP.
Con­clu­sions. ­IRSAC ­detects ­smaller vol­umes in phan­toms, ­patient ­images are ­easier to inter­pret and ­more met­a­static ­lesions ­were ­detected. In addi­tion, ­IRSAC pro­vides repro­du­cible quan­ti­ta­tive ­data, com­par­able to ­data pro­vided by FBP. ­IRSAC SUV and FBP SUV are in ­close agree­ment but ­there was a sta­tis­ti­cally sig­nif­i­cant dif­fer­ence ­between the two, and there­fore ­threshold ­values of SUV ­will prob­ably ­need to be re-deter­mined ­with ­IRSAC, and are ­likely to be 10 to 19% ­higher ­than cur­rently ­reported.

language: English


top of page