Mark Geurts1, Bruce Thomadsen, Ph.D.2, Reed Selwyn3

1Departments of Medical Physics and Engineering Physics, University of Wisconsin, Madison, WI2Departments of Medical Physics, Human Oncology, Engineering Physics, and Biomedical Engineering, University of Wisconsin, Madison, WI3Department of Medical Physics, University of Wisconsin, Madison, WI

NCCAAPM 2007 Spring Meeting, Burnsville, MN

Introduction

Most clinical settings do not have methods for validating the activity of Yttrium-90

90Y is commonly labeled to microspheres for selective internal radiation therapy (SIRT)

90Y exhibits a small internal pair production branching ratio of 31.86±0.47 x10-6 (Selwyn et al, 2007)

This project investigated whether a PET scanner could be calibrated for accurate assay validation

Coincidence Spectrum of 90Y

Spectrum was obtained using coincident NaI-HPGedetectors (Nickles et al, 2004)

0.000

0.010

0.020

0.030

0 200 400 600 800 1000

Se

nsi

tiv

ity

(cp

s/m

Ci)

Energy (keV)

Methods

PET scans performed on a GE Discovery LS CT/PET scanner

Each 90Y sample vial was assayed using a single HPGedetector

Each vial was then placed in a plastic sleeve to absorb emitted electrons & positrons

Sinogram data was collected with and without each reconstruction correction

2D and 3D scans were compared

Sinogram Correction DetailSinogram Correction Purpose

RandomCorrects for random coincidences detected within the coincidence resolving time

WellCorrects for axial sensitivity variation within the field of view

GeometricCorrects for sensitivity variation caused by gaps in the detector configuration

Decay Corrects for isotope decay during the scan

Dead time Corrects for electronic dead time losses

Normalization Corrects for variations in detector pair gains

Attenuation & ScatterCorrects for sensitivity variation caused by photon attenuation and scatter within the scan object

2D vs 3D Acquisition

3D 2D

Sensitivity Linearity Results Twelve 2D PET scans were taken over a period of several

months

Poisson 2σ uncertainty associated with the uncorrected sinogram count rate was less than 0.9% for each measurement

0

200

400

600

800

1000

0 50 100

Co

rrec

ted

Co

un

t R

ate

(cp

s)

Activity (mCi)

0

0.4

0.8

1.2

0 10 20 30

2D

Un

corr

ect

ed

Se

nsi

tivi

ty

(cp

s/m

Ci)

PET Slice Number

Sinogram Correction ResultsType of Sinogram Sensitivity (2σ uncert.) Adjusted R-squared

Uncorrected 4.000 cps/mCi ± 1.754% 99.91%

Random Corrected 4.003 cps/mCi ± 1.756% 99.91%

Well Corrected 3.510 cps/mCi ± 1.174% 99.96%

Geometric Corrected 4.440 cps/mCi ± 1.700% 99.91%

Decay Corrected 4.006 cps/mCi ± 1.752% 99.91%

Dead time Corrected 4.137 cps/mCi ± 2.448% 99.82%

Normalization Corrected 3.826 cps/mCi ± 1.604% 99.92%

All Corrections Applied 8.227 cps/mCi ± 1.012% 99.97%

Discussion All sinogram count rates correlated strongly with 90Y activity (R2 > 98%, P <

0.001) Random and decay corrections do not significantly change the observed

sensitivity (P = 0.821 and P = 0.682) Well and normalization corrections improve correlation (P < 0.001) Dead time correction degrades the sensitivity correlation (P < 0.001),

suggesting that the dead time model may be inadequate The two highest count rates contain more than 75% of the total absolute

residual when the dead time correction is applied

1.00

1.02

1.04

1.06

1.08

0 200 400

Me

an

De

ad

Tim

e

Co

rre

ctio

n

Uncorrected Count Rate (cps)

0

10

20

30

40

110 109 88 85 85 85 66 54 41 30 18 14Ab

solu

te R

esid

ual

(cp

s)

Sample Activity (mCi)

Discussion

The count rate for the 2D scan was higher than the 3D scan for 90Y activities greater than 100 mCi (P < 0.001)

The increased sensitivity of the 3D modality is offset by the increased number of third-gamma coincidences associated with the bremsstrahlung spectrum from the dominant beta-minus decay of 90Y

This comparison is consistent with the results of other research on 2D and 3D PET modalities (Schueller, 2003; Lubberink, 1999)

Conclusions

The provided PET scanner responds well under conditions of high photon background with a sensitivity 2σ uncertainty of 1.75%

Most of the built in sinogram corrections of the GE Discovery LS improve performance, reducing the uncertainty to 1.01%

Adequate results can be obtained in several minutes

PET is a quick and accurate method for assay validation can be created for 90Y and used to perform quality assurance on SIRT treatments

Work in Progress Monte Carlo model was developed to simulate the PET

scanner response using MCNP5 (9x107 source decays)Read in

parameters, initialize storage

arrays

Import and combine MCNP5

data

Advance time to next source

emission

Combine electron collisions into event pulses

Determine event location within

detector

Convolve pulse with energy resolution

Check block dead time

Apply energy discrimination

Apply module electronic dead

time

Trigger coincidence

window

Reject multiple coincidences

Process & store coincidence in

sinogram

Bin coincident event energy

resolution

Bin event separation time

Apply efficiency corrections

Export results to Excel

Simulation Results Simulations fit measured PET scanner sensitivity to 3.6% Applying new dead time model reduces sensitivity 2σ

uncertainty from 1.01% to 0.92%

90Y activities greater than 100 mCi should use a separate dead time model

1.00

1.05

1.10

1.15

1.20

1.25

1.30

0 400 800 1200

Dea

d t

ime

corr

ecti

on

Uncorrected Count Rate (cps)

0

400

800

1200

1600

0 100 200

Co

rrec

ted

Co

un

t R

ate

(cp

s)

Y-90 Activity (mCi)

Acknowledgements

The research group would like to recognize Dr. Robert J. Nickles and Dr. Robert Pyzalski at the University of Wisconsin for their ideas regarding 90Y PET imaging and expertise in positron emission tomography and coincidence detection systems

No funding was required for this research

References Schueller, M.J., Mulnix, T.L., Christian, B.T., Jensen, M., Holm, S., Oakes, T.R. (2003). Addressing

the third gamma problem in PET. IEEE Transactions on Nuclear Science, 50(1), 50-52. Selwyn, R.G., Nickles, R.J., Thomadsen, B.R., DeWerd, L.A., Micka, J.A. (2007). A new internal

pair production branching ratio of 90Y: the development of a non-destructive assay for 90Y and 90Sr. Appl Radiat Isot., 65(3), 318-327.

Wuosmaa, A.H., Conner, C.M., Ahmad, I., Back, B.B., Betts, R.R., Dunford, R.W. (1998). Positron-electron angular correlations in internal pair conversion. Physics Rev. C, 57(6), 2794-2798.

Lederer, M., Shirley, V. (Eds.). (1978). Table of Isotopes (7th Ed.). New York: Wiley. Nickles, R.J., Roberts, A.D., Nye, J.A., Converse, A.K., Barnhart, T.E., Avila-Rodirguez, M.A.

(2004). Assaying and PET imaging of yttrium-90. IEEE Nuclear Science Symposium Record, 6, 3412-3414.

Phelps, M.E. (2004). PET: molecular imaging and its biological applications. New York: Springer-Verlag.

Fahey, Frederic. (2002). Data acquisition in PET imaging. Journal of Nuclear Medicine Technology, 30(2), 39-49.

Ahn, S., Fessler, J. (2004). Emission image reconstruction for randoms-precorrected PET allowing negative sinogram values. IEEE Transactions on Medical Imaging, 23 (5), 591-601.

Martin, C., Christian, B., Satter, M., Nickerson, L., Nickles, R. (1995). Quantitative PET with positron emitters that emit prompt gamma rays. IEEE Transactions on Medical Imaging, 14 (4), 681-687.

Lubberink, M., Lundqvist, H., Westlin, J., Tolmachev, V., Schneider, E., Lövqvist, A. (1999). Positron emission tomography and radioimmunotargeting. Acta Oncologica, 38(3), 343-349.