FM Lavender, RT Meades, A Al-Nahhas, KS Nijran Imperial College Healthcare NHS Trust, London, UK.

A Comparative Study of Software used for LVEF Quantification in Myocardial Perfusion Imaging

FM Lavender, RT Meades, A Al-Nahhas, KS Nijran

Imperial College Healthcare NHS Trust, London, UK

Quantitation software: uses thresholding, binarization, and clusterification to estimate position of myocardial surfaces.

Left ventricular ejection fraction :LVEF = (EDV-ESV )x100%

EDV The LVEF provides quantitative functional information about

the heart. Comparison of the LVEF at rest/stress gives a measure of

the level of reversible ischaemia. Comparison of the LVEF between successive scans gives an

indication of the level of deterioration or response to treatment.

A Comparative Study of Software used for LVEF Quantification in Myocardial Perfusion Imaging

The aims of this study were to compare the effects of:

1. Quantitation software : 101 patientsQuantitative Gated SPECT vs 4DMSPECT

2. Pre-reconstruction filtering : 32 patientsButterworth filter of varying order and cut-off frequency

on LVEF quantification.

Aims

101 patients

Acquisition method: Siemens ECAM dual-head gamma camera SPECT acquisition: 32 projections,

30second/projection. 400-750MBq Tc99m Tetrofosmin 64x64 matrix, 6.59mm pixel size ECG gated into 8 time bins

Quantitation software

Gated SPECT

AutoCardiac

4DMSPECT

University of

Michigan

QGSCedars-

Sinai Medical Centre

Data pathway

Reconstruction Quantitation Acquisition

Hermes workstation

esoft workstation

esoft workstation

Results: QGS vs 4DMSPECT

10 20 30 40 50 60 70 80 90 1000

10

20

30

40

50

60

70

80

90

100

f(x) = 0.77514013925 x + 16.3569864485792R² = 0.782482059277969

EF [QGS] (%)

EF

[4D

-MS

PE

CT

] (%

)

Results: QGS vs 4DMSPECT

0 10 20 30 40 50 60 70 80 90 100

-40

-30

-20

-10

0

10

20

30

f(x) = 2.14851485148515

f(x) = − 11.384810352954

f(x) = 15.6818400559243

Mean EF (%)

Dif

fere

nc

e in

EF

[4

DM

SP

EC

T-Q

GS

] (%

)

0 10 20 30 40 50 60 70 80 90 100

-40

-30

-20

-10

0

10

20

30

f(x) = 2.14851485148515

f(x) = − 11.384810352954

f(x) = 15.6818400559243

Mean EF (%)

Dif

fere

nc

e in

EF

[4

DM

SP

EC

T-Q

GS

] (%

)

Range +21% to -28%

Results: QGS vs 4DMSPECT

Factors affecting quantitation: Heart size

ESV≤30ml [1], 53 patients ESV>30ml, 48 patients

50 55 60 65 70 75 80 85 90 95 100

0

10

20

30

40

50

60

70

80

90

100

f(x) = 0.424530160654744 x + 41.9051682327978R² = 0.252662376020306

EF [QGS] (%)

EF

[4

DM

SP

EC

T]

(%)

10 20 30 40 50 60 70 80 900

10

20

30

40

50

60

70

80

90

f(x) = 0.915562989780487 x + 9.18942761559965R² = 0.77837318232447

EF [QGS] (%)

EF

[4

DM

SP

EC

T]

(%)

[1] Hambye et al. Eur J Nucl Med Mol Imaging. 2004; 31(12):1606-13

Small hearts: EF>50% for ALL patients

Partial volume effect ESV estimation of 5ml is

unlikely to be accurate. Khalil et al. reported that

LVEF of small hearts was overestimated by 12% compared to gated blood pool [2] .

Factors affecting quantitation: Heart size

[2] Khalil et al., Nucl Med Commun. 2006 Apr; 27(4):321-32.

0 20 40 60 80 100 120

-35

-25

-15

-5

5

15

f(x) = − 13.1108940712184

f(x) = 12.6203280334825

f(x) = − 0.245283018867925

Mean EF (%)

EF

[4

DM

]-E

F[Q

GS

] (%

)

Factors affecting quantitation: Heart size

ESV≤30mlMean difference ±SD

-0.2 ± 6.4%

0 20 40 60 80 100 120

-30

-20

-10

0

10

20

30

f(x) = − 7.70868695118679

f(x) = 17.1554954618251

f(x) = 4.72340425531915

Mean EF (%)

EF

[4D

M]-

EF

[QG

S]

(%)

ESV>30mlMean difference ±SD

4.7 ± 6.2%

Manual correction necessary in 5% QGS and 4% 4DMSPECT

Factors affecting quantitation: Manual correction

•High uptake of tracer in the gut

QGS

5/101

•High uptake of tracer in the gut•Inaccurate valve plane positioning [3]

4DMSPECT

4/101

Factors affecting quantitation: Manual correction

Surface that intersects most basal myocardial points

Surface perpendicular to the long axis of heart

[3] Nakajima et al. J Nucl Med. 2001 Oct;42(10):1571-8

Results: QGS vs 4DMSPECT

0 10 20 30 40 50 60 70 80 90 100

-40

-30

-20

-10

0

10

20

30

f(x) = 2.14851485148516

f(x) = − 11.384810352954

f(x) = 15.6818400559243

Mean EF (%)

Dif

fere

nc

e in

EF

[4

DM

SP

EC

T-Q

GS

] (%

)

Large perfusion defects

Manual contour correction to exclude activity in the gut

The effect of perfusion defects on LVEF is controversial.

Vanhove 2002: gated blood pool imaging Concluded that the accuracy of QGS was dependent

on the severity but not the extent of perfusion defects [4].

Valejo 2000: MRI Perfusion defects led to an increase in overestimation

of EF by QGS [5].

Factors affecting quantitation: Perfusion defects

[4] Vanhove et al. Eur J Nucl Med Mol Imag 2002 Jun;29(6):735-41[5] Valejo et al. J Nucl Med. 2000 May;41(5):874-82 

(a) (b) (c) (d)

Filtering

Short axis and vertical long axis slices from a patient exhibiting high tracer uptake in the gut. Order 5 Butterworth filter at cut-off frequencies of (a, b) 0.5cycles/cm, (c) 0.6cycles/cm, (d) 0.7cycles/cm.

0.5cycles/cm

0.6cyc/cm 0.7cyc/cm

Van Laere comments that, “Recommended choices for a SPECT filter in the scarce literature on this subject are almost as numerous as the number of authors.” [6]

As the cut-off frequency is increased, larger volumes and smaller LVEFs were estimated using QGS [7].

The volumes and LVEF reported to plateau above 0.50cyc/cm [8].

[6] Van Laere et al. Comput Med Imaging Graph. 2001 Mar-Apr;25(2):127-33. Review.

[7] Kakhki et al. Clin Nucl Med. 2007 May;32(5):404-6.

[8] Nakajima et al. Eur J Nucl Med. 2000 Sep;27(9):1372-9.

Filtering: Published literature

32 patients Reconstruction performed using Hermes

FBP SPET

Pre-reconstruction filter: Butterworth Orders 5 and 10 Cut-off frequencies 0.5 to 1.2 cycles/cm in

steps of 0.1cycles/cm. 512 reconstructions

Filtering

0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3

-10

0

10

20

30

40

50

Order 5

cut-off frequency (cycles/cm)

% d

iffer

ence

in E

F

Filtering: Results

No obvious plateau

Effect of cut-off frequency on EF is very patient specific

0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3-10

0

10

20

30

40

50Order 5

cut-off frequency (cycles/cm)

% d

iffer

ence

in E

F

Filtering: Results

At cut-off frequency of 0.5cyc/cm, maximum difference index was 42%.

At cut-off frequencies ≤ 0.8cyc/cm, the EF was overestimated compared to higher cut-off frequencies in 26/30 patients.

Difference index=

For z=0.5 to 1.2cyc/cm

Why do we get large variations? Reconstruction process is not fully

automated. Image is filtered, then axes have to be re-

aligned for each reconstruction.

Filtering: intraoperator reproducibility of reconstruction

Method: Repeated filtering for first ten patients

(320 reconstructions)

Results: Average difference between initial and repeated

results was <0.1% for both orders Therefore manual axes alignment does not

significantly affect LVEF estimation.

Filtering: intraoperator reproducibility of reconstruction

4DMSPECT and QGS estimate significantly different EFs (range 21% to -28%) and should not be used interchangeably.

The agreement between the packages is worse for small hearts (ESV≤30ml) and may also be affected by severe perfusion defects and manual contour correction.

LVEF is dependent on the cut-off frequency used during pre-reconstruction filtering but the dependency is very patient specific.

Changing the cut-off frequency by as little as 0.1cyc/cm sometimes resulted in changes to the EF of >10 absolute EF percentage points.

Conclusion

Thank you