Premium
Impact of wavelet based denoising of [ C 11 ] ( R ) ‐PK11195 time activity curves on accuracy and precision of kinetic analysis
Author(s) -
Yaqub Maqsood,
Boellaard Ronald,
Schuitemaker Alie,
van Berckel Bart N. M.,
Lammertsma Adriaan A.
Publication year - 2008
Publication title -
medical physics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.473
H-Index - 180
eISSN - 2473-4209
pISSN - 0094-2405
DOI - 10.1118/1.2995834
Subject(s) - wavelet , noise reduction , mathematics , parametric statistics , noise (video) , filter (signal processing) , residual , pattern recognition (psychology) , artificial intelligence , algorithm , computer science , statistics , computer vision , image (mathematics)
The purpose of the present study was to investigate the use of various wavelets based techniques for denoising of[ C11 ] ( R ) ‐PK11195 time activity curves (TACs) in order to improve accuracy and precision of PET kinetic parameters, such as volume of distribution ( V T ) and distribution volume ratio with reference region ( DVR ). Simulated and clinical TACs were filtered using two different categories of wavelet filters: (1) wave shrinking thresholds using a constant or a newly developed time varying threshold and (2) “statistical” filters, which filter extreme wavelet coefficients using a set of “calibration” TACs. PET pharmacokinetic parameters were estimated using linear models (plasma Logan and reference Logan analyses). For simulated noisy TACs, optimized wavelet based filters improved the residual sum of squared errors with the original noise free TACs. Furthermore, both clinical results and simulations were in agreement. Plasma Logan V T values increased after filtering, but no differences were seen in reference Logan DVR values. This increase in plasma Logan V T suggests a reduction of noise induced bias by wavelet based denoising, as was seen in the simulations. Wavelet denoising of TACs for[ C11 ] ( R ) ‐PK11195 PET studies is therefore useful when parametric Logan based V T is the parameter of interest.