IEEE TRANSACTIONS ON MEDICAL IMAGING, VOL. 31, NO. 12, DECEMBER 2012 2213
Direct 4-D PET List Mode Parametric Reconstruction
With a Novel EM Algorithm
Jianhua Yan, Beata Planeta-Wilson, and Richard E. Carson*, Member, IEEE
Abstract—The production of images of kinetic parameters is
often the ultimate goal of positron emission tomography (PET)
imaging. The indirect method of PET parametric imaging, also
called the frame-based method (FM), is performed by fitting the
time-activity curve (TAC) for each voxel with an appropriate
compartment model after image reconstruction. The indirect
method is simple and easily implemented, however, it usually
leads to some loss of accuracy or precision, due to the use of
two separate s teps. This paper presents a direct 4-D method for
producing 3-D images of kinetic parameters from list mode PET
data. In this application, the TAC for each voxel is described
by a one-tissue compartment model (1T). Extending previous
EM algorithms, a new spatiotemporal complete data space was
introduced to optimize the maximum likelihood function. This
leads to a straightforward close d-form parametric image update
equation. This method was implemented by extending the current
list mode platform MOLAR to produce a parametric algorithm
PMOLAR-1T. Using an ordered subset approach, qualitative
and quantitative evaluations were performed using 2-D (x, t) and
4-D (x, y, z, t) simulated list mode data based on brain receptor
tracers and also with a human brain study. Comparisons with
the indirect method showed that the proposed direct method can
lead to accurate estimation of the parametric image values with
reduced variance, especially at low count levels.
In the 2-D test,
the direct method showed similar bias to the frame-based method
but with variance reduction of 23%–60%. I n the 4-D test, bias
values of both methods were no more than 4% and th
edirect
method had lower variability (coefficient of variation reduction
of 0%–64% compared to the frame-based method) at the normal
count level. The direct method had a larg
er reduction in variability
(27%–81%) and lower bias (1%–5% for 4-D and 1%–19% for
FM) at low count levels. The results in the human brain study are
similar with PMOLAR-1T showing lowe
r noise than FM.
Index Terms—Expectation maximizati
on, image reconstruction,
one-tissue compartment model, parametric imaging.
I. INTRODUCTION
D
YNAMIC positron emission to mo graphy (PET) permits
qu
antification of tracer dynamics. An important goal of
much of quantitative brain PET is to p roduce parametric images
Manuscript received May 17, 2012; revised July 30, 2012; a ccepted July 31,
2012. Date of publication August 23, 2012; date of current version November
27, 2012. This work was supported b y the N atio nal In stitute of Neu r ological
Disorders and Stroke under Grant R01NS058360, Clinical and Translational
Science Awards under Grant UL1 RR024139 from the National Center for Re-
search Resources (N CRR), National Institutes of Health (NIH) Roadmap for
Medical Research and Siem ens Medical System s. Its contents are solely the re-
sponsibility of the authors and do not necessarily represen t the official review
of NCRR or NIH. Asterisk indicates c o r responding author.
J. Yan was with PET Center, Department of Diagnostic Radiology, Yale Uni-
versity, New Haven, CT 06520 USA. He is now with the A*STAR-NUS, Clin-
ical Imaging Research Center, Center for Translational Medicine, Singapore
117599 (e-mail: yan_jianhua@circ.a-star.edu.sg).
B. P. Wilson is with the PET Center, D e par tm en t o f D iagnostic Radiology,
Yale University, New Haven, CT 06520 USA.
*R. E. Carson is with the PET Center, Department of Diagnostic Radiolog y,
Yale University, New Haven, CT 06520 USA (e-mail: richard.e.carson@yale.
edu).
D
igital Object Identifier 10.1109/TMI.2012.2212451
of kinetic parameters, such as t he uptake rate and t he total
volume of distribution
[1]. The current data processing
path for parametric imaging is to reconstruct a time series of
images from measured projection data independently, and then
estimate each voxel’s kinetic parameters from the time-activity
curve (TA C), typically with a compartmental model. This indi-
rect or frame-based approach requires selection of the duration
of each frame, involving a choice between using longer scans
with better counting statistics but poor tempo ral resolution, and
shorter scans that are noisy but preserve temporal resolution.
Optimal use of the dynamic data requires accurate noise esti-
mates for data weighting; this estimation is challenging fo r non-
linear iterative reconstruction methods because noise is spatially
variant and object dependent [2], [3].
Direct approaches for creation of parametric images have
been in the literature for nearly 30 years. In 1984, Snyder [4] de-
veloped a list mode expectation-maximization ( EM) maxim
um
likelihood (ML) algorithm for estimatio n of parametric images
using an inhomogeneous spatial-temporal Poisson processes
and a kinetic compartmental model. Carson and L
ange [5]
also proposed an EM framework for direct parametric image
reconstruction algorithm for a one-tissue model. Subsequ ently,
many direct kinetic e stimation methods [6
]–[10] for sinogram
data have been produced . However, for a high-resolu tio n
scanner such as the HRRT [11] with
potential l ines of
response, list mode data is prefer
red, since it can reduce the data
storage requirements w hile maintaining the highest resolution
by storing the measured attributes of each event in the list.
Other direct methods [12]–[2
0] were developed from linear
basis functions, w h ose temporal m odel is lin ear with respect to
the parameters. Although such basis functions can describe the
TAC, they have no direct
physiological meaning and kinetic
parameters must still be calculated from the dynamic images,
leading again to a two-step process.
The use of kinetic com
partment mode ls has a firmer bio-
logical basis, and is thus the primary goal of dynamic PET.
Tsoumpas et al. [9] and Wang et al. [10] proposed direct
parametric ima
ge reconstruction methods, whose compartment
model is linear ( P atlak plot), and is thus lim ited to ir rev ersible
tracers such as FDG. Furthermore, Rahmim et al. [21] proposed
aclosed-fo
rm direct 4-D parametric image estimation m ethod
for reversible tracers employing a recently published linear
graphical analysis m odel. Recently, Wang and Qi [22] proposed
a genera
lized direct parametric image reconstruction approach
applicable to any kinetic model, which was comprised of tw o
steps for each parameter update (reconstruction of intermediate
ima
ges and kinetic parameter estimation with accurate weights
determined from the derived algorithm) .
Here, we present a new EM-based direct 4-D parametric
i
mage reconstruction al gor ithm for list mode data by combining
0278-0062/$31.00 © 2012 IEEE