PET IN ONCOLOGY
Guest Editor: I. Carrio 

The Quarterly Journal of Nuclear Medicine 2001 September;45(3):235-44

Copyright © 2009 EDIZIONI MINERVA MEDICA

language: English

Local detection of prostate cancer by positron emission tomography with 2-fluorodeoxyglucose: comparison of filtered back projection and iterative reconstruction with segmented attenuation correction

Turlakow A. ¹, Larson S. M. ¹, Coakley F. ¹, Akhurst T. ¹, Gonen M. ², Macapinlac H. A. ¹, Kelly W. ³, Leibel S. ⁴, Humm J. ⁵, Scardino P. ⁶, Scher H. ³, Hricak H. ¹

From the 1 Department of Radiology 2 Department of Biostatistics 3 Department of Medical Oncology 4 Department of Radiation Oncology 5 Department of Medical Physics 6 Department of Urology Memorial Sloab-Kettering Cancer Center, New York, USA

Back­ground. To com­pare fil­tered ­back pro­jec­tion (FBP) and iter­a­tive recon­struc­tion ­with seg­mented atten­u­a­tion cor­rec­tion (­IRSAC) in the ­local ­imaging of pros­tate ­cancer by posi­tron emis­sion tomog­raphy ­with 2-flu­o­rod­e­ox­y­glu­cose (FDG-PET).
­Methods. We ret­ro­spec­tively iden­ti­fied 13 ­patients ­with pri­mary (n=7) or recur­rent (n=6) pros­tate ­cancer who had ­increased ­uptake in the pros­tate on FDG-PET per­formed ­without uri­nary cath­et­er­iza­tion, con­tem­po­ra­neous ­biopsy con­firming the pres­ence of ­active ­tumor in the pros­tate, and cor­rel­a­tive ­cross-sec­tional ­imaging by MRI (n=8) or CT (n=5). FDG-PET ­images ­were recon­structed by FBP and ­IRSAC. Two inde­pen­dent ­nuclear med­i­cine phy­si­cians sep­ar­ately ­rated FBP and ­IRSAC ­images for vis­u­al­iza­tion of pros­tatic ­activity on a ­4-­point ­scale. ­Results ­were com­pared ­using ­biopsy and ­cross-sec­tional ­imaging find­ings as the stan­dard of ref­er­ence.
­Results. ­IRSAC ­images ­were sig­nif­i­cantly ­better ­that FBP in ­terms of vis­u­al­iza­tion of pros­tatic ­activity in 12 of 13 ­patients, and ­were equiv­a­lent in 1 ­patient (p<0.001, Wil­coxon ­signed ­ranks ­test). In par­tic­ular, 2 ­foci of ­tumor ­activity in 2 dif­ferent ­patients ­seen on ­IRSAC ­images ­were not vis­ible on FBP ­images. In 11 ­patients who had a ­gross ­tumor ­mass evi­dent on ­cross-sec­tional ­imaging, ­there was ­good agree­ment ­between PET and ­cross-sec­tional ana­tomic ­imaging ­with ­respect to ­tumor local­iza­tion.
Con­clu­sions. In ­selected ­patients, ­cancer can be ­imaged ­within the pros­tate ­using FDG-PET, and ­IRSAC is ­superior to FBP in ­image recon­struc­tion for ­local ­tumor vis­u­al­iza­tion.