Automatic evaluation of myocardial perfusion on SPECT: Need for “Normality”
Automatic evaluation of myocardial perfusion on SPECT: Need for ''Normality''
Riccardo Liga 0 2
Alessia Gimelli 0
0 Reprint requests: Alessia Gimelli, MD, Fondazione Toscana/CNR G. Monasterio , Pisa , Italy
1 Fondazione Toscana/CNR G. Monasterio , Pisa , Italy
2 University of Pisa , Pisa , Italy
Myocardial perfusion imaging (MPI) with
singlephoton emission computer tomography (SPECT) is still
the most performed and, probably, the most trusted
noninvasive imaging modality for the evaluation of patients
with suspected or known ischemic heart disease (IHD).
Despite the great advancement of all the other imaging
modalities that has taken place in the last two decades,
the ability of SPECT MPI in terms of in-depth
characterization and risk stratification of patients with IHD
remains unbeaten, with solid evidence proving its
additive value in a wide range of patients subsets (i.e.,
symptomatic vs asymptomatic subjects; normal weight
vs obese; sinus vs paced rhythm)1-4 and clinical
scenarios (i.e., chronic IHD vs acute coronary
syndromes).5,6 In particular, among the major
advantages of this technology, the intrinsic versatility of
SPECT MPI is probably the most relevant. In fact, it can
be easily coupled with any of the currently available
stress tests (i.e., exercise, vasodilators, dobutamine),
allowing obtaining reproducible measures of myocardial
ischemic susceptibility in practically every category of
patients.1,7 Moreover, SPECT MPI study can provide
information on LV structure and function that may both
contribute to the diagnostic process, and provide
independent measures of adverse patients prognosis.8,9
Nevertheless, even with the evident advantages of
nuclear cardiac imaging, SPECT MPI is also
characterized by some relevant drawbacks that should be
always taken into account, particularly if traditional
Anger cameras are employed. Above all, the relatively
high radiation burden of a stress/rest imaging protocol
performed with a classical SPECT camera equipped
with Na/I scintillating crystals—especially if 201Tl was
used as radiotracer instead of the more favorable
99mTcbased compounds—has always represented a theoretical
limitation of this imaging modality,10 making the
clinician often favor other non-ionizing non-invasive
myocardial imaging techniques. Secondly, the relatively
lengthy acquisition protocol of traditional myocardial
SPECT imaging, still not too infrequently performed on
a two-day basis, has made the other non-invasive
imaging techniques look more appealing.1,10 In the last
decade, both those limitations have been practically
solved by several improvements consistently both the
architecture and software settings of traditional Anger
Regarding innovation in technology, a substantial
increase of photon sensitivity and spatial resolution can
be also obtained by the implementation of confocal
collimators mounted on conventional dual-headed
gamma cameras (i.e., IQ SPECT).12 These collimators
are characterized by an intrinsically more convergent
field of view at the center, while the convergence is
reduced approaching the edges of the field of view. The
increased sensitivity of those devices allows reducing
the acquisition times to about 4 minutes per scan, with a
similar quality of traditional systems.12,13
Finally, the introduction of dedicated cardiac
cameras built around solid-state Cadmium-Zinc-Telluride
(CZT) detectors has represented a real revolution in the
field of nuclear cardiac imaging, allowing reducing
consistently both the radiation burden and acquisition
time of MPI, while maintaining if not improving the
overall diagnostic capability.7,14,15
Because of the absence of the photomultipliers
tubes, CZT cameras are consistently less bulky than
traditional SPECT devices, increasing patients comfort
while reducing overall logistic needs. Nevertheless,
while performing consistently better than Anger
cameras, CZT devices still constitute a limited fraction of
the working nuclear cardiac cameras, mainly because of
the higher initial costs of this technology.
On the other hand, nuclear cardiac imaging on
conventional SPECT devices has been also interested by
a consistent software evolution, mainly based on
iterative reconstruction algorithms that outperform the
‘‘old’’ filtered back projection (FBP) in terms of higher
image quality, with less background noise, and smoother
myocardial borders.16 Image quality is further improved
by the implementation of resolution recovery
algorithms, additively reducing image noise with respect to
conventional reconstruction software. While the additive
value of novel iterative resolution recovery (IRR)
reconstruction algorithm may be limited in high-count
SPECT studies, it becomes relevant in the case of
lowdose SPECT scans, where the improvement with respect
to FBP becomes more relevant.17 In fact, low-dose
cardiac images acquired with IRR software are superior
to those acquired with FBP, allowing reducing
acquisition times and/or injected radiotracers dose. It is not
surprising that in the last years the feasibility of
lowdose SPECT imaging protocols with traditional Anger
cameras has been demonstrated, allowing maintaining
excellent image quality and diagnostic accuracy while
reducing consistently the overall radiation burden of the
scan.18 However, the technical performance of
conventional SPECT devices in the new realm of low-dose
scans has to be still characterized, since the application
the old rules that were appropriate in the case of
fulldose acquisitions might not be appropriate in this novel
In this number of the Journal, Scabbio et al.
evaluate the possible clinical impact of the implementation
of low-dose SPECT studies on the automatic
quantification of myocardial perfusion information.19 The main
objective of the study was the analysis of the impact of
non-study-specific software database—derived from a
reference population imaged with a different acquisition
protocol and injected dose—on the accurate
quantification of regional myocardial perfusion abnormalities in
patients imaged with a traditional dual-headed SPECT
camera that used a dedicated reconstruction algorithm
with IRR technology.
In this setting, the introduction of reliable
quantitative software for the automatic evaluation of
myocardial perfusion heterogeneity has been a great
advancement in modern nuclear cardiology, with
different algorithms currently available from a number of
vendors.20 While the different software differ in terms of
the specific algorithm used for the quantification
process, the basic steps of LV perfusion analysis are the 3D
segmentation of the myocardium, the generation of polar
maps of relative myocardial count density, and the
comparison of the obtained polar map with average
values obtained in a reference population.
Unfortunately, while site- or protocol-specific normal limits
should be created using generally a number of SPECT
scans of patients with a low (\5%) pre-test likelihood of
CAD, most of the times the original vendor-provided
normality databases are employed, possibly introducing
relevant bias in the evaluation of myocardial perfusion
However, to date, the impact of the use of
nonspecific normality database for the analysis of cardiac
SPECT with novel IRR reconstruction algorithms and
dose-sparing protocols had not been evaluated.
In order to accomplish this task, the study by
Scabbio et al. includes both a pre-clinical assessment
with an anthropometric phantom and a clinical
validation on a real-life population of patients submitted to
clinically indicated stress/rest myocardial SPECT
In the phantom study, the impact of different
radiotracer dose regimens and of specific reconstruction
algorithms (FBP, OSEM, and IRR) on automatic
myocardial perfusion analysis was evaluated, showing
that even significant reduction of the injected dose (i.e.,
25% of the traditional radiotracer dose) did not translate
into significant errors in the regional evaluation of
myocardial perfusion on polar maps, with only minor
regional variations when IRR algorithms were applied.
For the patient-based study, 40 consecutive normal
weight patients and an identical number of overweight
subjects were selected, and the perfusion scans were
analyzed with either IRR and non-IRR reconstruction
software using different normality databases (i.e., full
counts and half counts). The main results of the study
were that the selection of the specific normality database
of myocardial perfusion had a significant impact on the
quantification of regional myocardial perfusion
heterogeneity, both at rest and after stress. This finding was
more evident in the case of low-dose studies as well as
in the overweight subjects, with a significantly adverse
impact on the accurate evaluation of stress-induced
myocardial perfusion abnormalities. Moreover,
significant difference in the outputs obtained with IRR and
non-IRR reconstruction software was reported,
confirming the protocol-specificity of automatic SPECT
perfusion analysis. Interestingly, the impact of dose
reduction on the accuracy of myocardial perfusion
analysis was particularly evident only in the case of
‘‘extremely’’ low-dose protocols (25% of the traditional
injected dose) and in overweight patients, while it
remained non-significant when normal weight subjects
were imaged. In the former setting, even the use of a
half-counts normality database for automatic perfusion
analysis could not eliminate the bias in the quantification
of myocardial perfusion heterogeneity.
The authors of the study should be commended for
having conducted a technically solid study that
addressed a topic of high clinical interest.
The peculiar analysis that was performed—the use
of three different image reconstructions with
progressively reduced acquisition times for SPECT
projection—allowed obtaining for the same population
of patients three sets of images that could mimic
accurately normal dose, half dose, and quarter dose scans.21
Accordingly, the obtained results indicated that the
radical reduction of the injected radiotracer dose, as
currently feasible with high-end SPECT systems, must
be also paralleled by an improvement of software
settings, particularly if automatic perfusions analysis is
In this respect, automatic quantification software
offer the chance to obtain accurate and highly
reproducible assessment of myocardial perfusion
heterogeneity that compete well, and generally
outperform, the performance of visual segmental scoring, even
when performed by expert readers.20 These algorithms
ideally offer the chance to unmask subtle changes of
regional LV ischemic burden on sequential myocardial
perfusion scans, as a frequent condition in many cardiac
patients, potentially guiding targeted therapeutic
interventions. However, while automatic quantitative
perfusion analysis is the rule in conditions where an
absolute evaluation of frequently minimal changes of
myocardial perfusion heterogeneity is mandatory—
randomized controlled trial settings with sequential
SPECT scans (i.e., before and after a therapeutic
intervention) 22,23—they are less frequently employed in
daily clinical practice, where visual scoring of
myocardial perfusion is generally performed. Nevertheless,
comparative analyses have shown how automatic
perfusion analysis of SPECT scans may be more accurate
than visual scoring in diagnosing the presence of
significant CAD, obviating the need of frequently
cumbersome image reconstructions.20
Beside the many advantages, the wide
implementation of automatic quantification perfusion software is
limited by some relevant pitfalls of these technology,
namely the need of a reference database of normal
patients scanned with the same imaging protocol and the
great susceptibility from image artifacts. As a matter of
fact, a significant difference in terms of protocol
characteristics between the normality database and the study
patients may impact profoundly the outputs of automatic
perfusion analysis, and, as demonstrated by Scabbio
et al., may be more relevant in overweight patients or
when IRR algorithms are employed.18,24 On the other
hand, the presence of any image artifact, while generally
ignored by expert visual readers, may be misdiagnosed
by automatic quantification software, especially when
attenuation correction is not available.
While adding some interesting information on these
topics, the study by Scabbio et al. has also some
limitations that prevent the wide generalizability of its findings.
First of all, the occurrence of attenuation artifacts not
corrected by any form of attenuation correction
algorithms might have diminished the reproducibility of
automatic perfusion analysis, reducing the comparability
of the different acquisition protocols of SPECT images.20
Second, the absence of any clinically meaningful
reference standard for the evaluation of myocardial perfusion
analysis (i.e., coronary angiography) does not allow
deriving any definitive conclusion on the results, whose
clinical impact is still not completely defined. In
conclusion, present and previous reports further underline the
great advancements that nuclear cardiac imaging has had
in the last decade, bringing this technique into an era
characterized by reduced patient dosimetry and ultra-fast
acquisition protocols. However, further improvements in
software settings are needed to allow the full exploitation
of the diagnostic potential of myocardial perfusion
imaging. Above all, the need of different reference
perfusion databases of normal patients studied with variable
protocols and under various study conditions has been
strongly remarked in order to allow a proper definition of
subtle alteration of myocardial perfusion heterogeneity
and definition of patient risk also in patients studied with
low-dose protocols and/or with innovative IRR
Conflict of Interest
The authors declare that no potential conflict of interest
exists and that all them have read and approved the
The authors have nothing to declare.
1. Underwood SR , Anagnostopoulos C , Cerqueira M , Ell PJ , Flint EJ , Harbinson M , et al. Myocardial perfusion scintigraphy: The evidence . Eur J Nucl Med Mol Imaging 2004 ; 31 : 261 - 91 .
2. Acampa W , Gaemperli O , Gimelli A , Knaapen P , Schindler TH , Verberne H , et al. Role of risk stratification by SPECT, PET, and hybrid imaging in guiding management of stable patients with ischaemic heart disease: Expert panel of the EANM cardiovascular committee and EACVI . Eur Heart J Cardiovasc Imaging 2015 ; 16 : 1289 - 98 .
3. Gimelli A , Bottai M , Giorgetti A , Genovesi D , Filidei E , Marzullo P . Evaluation of ischaemia in obese patients: Feasibility and accuracy of a low-dose protocol with a cadmium-zinc telluride camera . Eur J Nucl Med Mol Imaging 2012 ; 39 : 1254 - 61 .
4. Lapeyre AC 3rd, Poornima IG , Miller TD , Hodge DO , Christian TF , Gibbons RJ . The prognostic value of pharmacologic stress myocardial perfusion imaging in patients with permanent pacemakers . J Nucl Cardiol 2005 ; 12 : 37 - 42 .
5. Douglas PS , Hoffmann U , Patel MR , Mark DB , Al-Khalidi HR , Cavanaugh B , et al. Outcomes of anatomical versus functional testing for coronary artery disease . N Engl J Med 2015 ; 372 : 1291 - 300 .
6. Nabi F , Kassi M , Muhyieddeen K , Chang SM , Xu J , Peterson LE , et al. Optimizing evaluation of patients with low-to-intermediaterisk acute chest pain: A randomized study comparing stress myocardial perfusion tomography incorporating stress-only imaging versus cardiac CT . J Nucl Med 2016 ; 57 : 378 - 84 .
7. Gimelli A , Liga R , Pasanisi EM , Casagranda M , Coceani M , Marzullo P . Influence of cardiac stress protocol on myocardial perfusion imaging accuracy: The role of exercise level on the evaluation of ischemic burden . J Nucl Cardiol 2016 ; 23 : 1114 - 22 .
8. Gimelli A , Liga R , Giorgetti A , Kusch A , Pasanisi EM , Marzullo P . Relationships between myocardial perfusion abnormalities and poststress left ventricular functional impairment on cadmium-zinctelluride imaging . Eur J Nucl Med Mol Imaging 2015 ; 42 : 994 - 1003 .
9. Gimelli A , Liga R , Clemente A , Marras G , Kusch A , Marzullo P . Left ventricular eccentricity index measured with SPECT myocardial perfusion imaging: An additional parameter of adverse cardiac remodeling . J Nucl Cardiol 2017 . doi: 10 .1007/s12350-017-0777-3.
10. Einstein AJ , Pascual TN , Mercuri M , Karthikeyan G , Vitola JV , Mahmarian JJ , et al. Current worldwide nuclear cardiology practices and radiation exposure: Results from the 65 country IAEA nuclear cardiology protocols cross-sectional study (INCAPS) . Eur Heart J 2015 ; 36 : 1689 - 96 .
11. Slomka PJ , Dey D , Duvall WL , Henzlova MJ , Berman DS , Germano G. Advances in nuclear cardiac instrumentation with a view towards reduced radiation exposure . Curr Cardiol Rep 2012 ; 14 : 208 - 16 .
12. Gremillet E , Agostini D . How to use cardiac IQ•SPECT routinely? An overview of tips and tricks from practical experience to the literature . Eur J Nucl Med Mol Imaging 2016 ; 43 : 707 - 10 .
13. Caobelli F , Ren Kaiser S , Thackeray JT , Bengel FM , Chieregato M , Soffientini A , et al. IQ SPECT allows a significant reduction in administered dose and acquisition time for myocardial perfusion imaging: Evidence from a phantom study . J Nucl Med 2014 ; 55 : 2064 - 70 .
14. Nakazato R , Berman DS , Hayes SW , Fish M , Padgett R , Xu Y , et al. Myocardial perfusion imaging with a solid-state camera: Simulation of a very low dose imaging protocol . J Nucl Med 2013 ; 54 : 373 - 9 .
15. Nkoulou R , Fuchs TA , Pazhenkottil AP , Kuest SM , Ghadri JR , Stehli J , et al. Absolute myocardial blood flow and flow reserve assessed by gated SPECT with cadmium-zinc-telluride detectors using 99mTc-tetrofosmin: Head-to-head comparison with 13Nammonia PET . J Nucl Med 2016 ; 57 : 1887 - 92 .
16. Zoccarato O , Scabbio C , De Ponti E , Matheoud R , Leva L , Morzenti S , et al. Comparative analysis of iterative reconstruction algorithms with resolution recovery for cardiac SPECT studies. A multi-center phantom study . J Nucl Cardiol 2014 ; 21 : 135 - 48 .
17. DePuey EG. Advances in SPECT camera software and hardware: Currently available and new on the horizon . J Nucl Cardiol 2012 ; 19 : 551 - 81 .
18. Marcassa C , Campini R , Zoccarato O , Calza P . Wide beam reconstruction for half-dose or half-time cardiac gated SPECT acquisitions: Optimization of resources and reduction in radiation exposure . Eur J Nucl Med Mol Imaging 2011 ; 38 : 499 - 508 .
19. Scabbio C , Zoccarato O , Malaspina S , Lucignani G , Del Solec A , Lecchia M. Impact of non-specific normal databases on perfusion quantification of lowdose myocardial SPECT studies . J Nucl Cardiol 2017 . doi: 10 .1007/s12350-017-1079-5.
20. Slomka P , Xu Y. Berman D and Germano G Quantitative analysis of perfusion studies: Strengths and pitfalls . J Nucl Cardiol 2012 ; 19 : 338 - 46 .
21. Lecchi M , Martinelli I , Zoccarato O , Maioli C , Lucignani G , Del Sole A . Comparative analysis of full-time, half-time, and quartertime myocardial ECG-gated SPECT quantification in normalweight and overweight patients . J Nucl Cardiol 2017 ; 24 : 876 - 87 .
22. Shaw LJ , Berman DS , Maron DJ , Mancini GB , Hayes SW , Hartigan PM , et al. Optimal medical therapy with or without percutaneous coronary intervention to reduce ischemic burden: Results from the Clinical Outcomes Utilizing Revascularization and Aggressive Drug Evaluation (COURAGE) Trial Nuclear Substudy . Circulation 2008 ; 117 : 1283 .
23. Mahmarian JJ , Cerqueira MD , Iskandrian AE , Bateman TM , Thomas GS , Hendel RC , et al. Regadenoson induces comparable left ventricular perfusion defects as adenosine: A quantitative analysis from the ADVANCE MPI 2 trial . JACC Cardiovasc Imaging 2009 ; 2 : 959 - 68 .
24. Zoccarato O , Marcassa C , Lizio D , Leva L , Lucignani G , Savi A , et al. Differences in polar-map patterns using the novel technologies for myocardial perfusion imaging . J Nucl Cardiol 2016 . doi: 10 .1007/s12350-016-0500-9.