Download Evaluating Splenic Blood Flow: A Comparison of Steady-State and Dynamic PET Methods and more Lecture notes Medical statistics in PDF only on Docsity!
with a steady-state method using 15C-wateras well as the reliabilityof the results in relation to those obtained with the dynamic-state method using ‘5O-water.
METhODS
MathemaOcal Mod& and Theory In the steadybloodflow state, ‘5C-wateris continuouslyin haledand‘@Ois rapidlyconvertedinthe lungsfrom‘5C-waterto ‘50-water(5). Therefore,the followingequationholdsbetween the radioactivityin the spleen(Cs(t)),the inputfunctionof the aorta (Ca(t)), the total splemc blood flow (F), the volume of the spleen(V),andthespleen-bloodpartitioncoefficientforwater(p):
Methods: Reg@nalsplenic bloodflow(SBF)was quantifiedby PEr usinga steady-state methodwith15Ø@e@ dioxide.SBFs were determined using 104 tOmOgraphICplanes obtained from 49 patients. Results: Whenthe spleen-bloodpartitioncoefficient forwater (p) was 0.85, algn@cant correlations (p < 0.005) were found between SBF values determined by the steady-state and dynamicmethods.The best correlationbetween SBFs deter minedbythe twomethods (r = 0.571)was foundwhen p = 0.93. The best regressionline,however,wasthoughtto be the line when p = 0.93. The regression linebetween SBF calculated by thesteady-statemethod(y)andSBFdeterminedbythedynamic method (x) was y = 0.57 x + 0.03 wIth an F ratio of 48. (d.f. = 103, p = 5.0 x 10°%, by ANOVA) when p = 0.92. Conclusion: A quickevaluationof SBF can be made by using the newlydefined regression line. Key Words: positron emission tomogmphy re@onalsplenic @@ fio@ @oncoefficient
J NuciMed1995;36:599-
Eq.
where @Li5 the decayconstantof ‘50-carbondioxide.Regional splenicbloodflowper100g of splenictissue(SBF)canbe solved as follows:
Eq.
ollowing Ter-Pogossian's pioneering studies with ‘@Oto measure cerebral blood flow (1), rapid progress has been made in the understandingof cerebral blood flow and ox ygen metabolism using the steady-state method (2). This method, however, has never been applied to other organs, especially intra-abdominal organs. PET studies enable the exact measurement of the distribution of positron-emitting radioisotopes within the human body and can be used to effectively quantify blood flow in many organs. We have studied splertichemodynamics, which change as liver damageprogresses (3,4), with PET usingwater and a dynamic-state method. Since splenic hemodynamics are
probably importantin liver disorders and portal hyperten
sion, PET is a powerful tool to study this relationship. In this study, we estimated splenic blood flow (SBF)
ReceivedApr.21, 1994;revisionacceptedAug.1, 1994. ForcorrespondenceorreprintscontactHWoidTaniguchi,MD,FirstDepartment ofSurgery,KyotoPrefectui@UnWers@yofMed@ne,Kawaiwnachi.Hirokchji,Ks migyo-ku.Kyob602,Japan.
p@ Forty-ninepatients[28malesand21 females,agesrangefrom 32 to 77 yr (mean:58.1 yr)]were investigated.No patienthad hepaticfunctionaldisorders. Materials ThePETsystemconsistedof awhole-bodyPETscanneranda cyclotronwith a gas purifyingsystem. The performancecharac teristics of the PET system were set as follows: an image resolu tionof8.2 mmFWHManda slicethicknessof 11—13mmFWHM. The matrixsize of the imagewas 128 x 128with a 2-mmpixel size.The sliceintervalofthe planeswas 15mm.The scanposition foreachpatientwasdeterminedusingx-rayCT.Emissiondatafor three slicesat intervalsof 15mmwere collectedsimultaneously. One hundred four sliceswere chosen as havingplanes which encompassedthe spleenandprovidedregionsof interest(ROIs) with full signalimaging.Under steady-stateconditions,5-min PETscanningwas performedduringcontinuousinhalationof 185 MBq—370MBq of 15C-water. Blood samples were collected from
Splenic Bicod Aow DetermInatiOnsUsing PET •Tar@guchiet al. 599
Determination of the Spleen-Blood Partition
Coefficient for Water with Oxygen-15-Water and
Oxygen-15-Carbon Dioxide Dynamic PET
Steady-State Methods
Hiroki Taniguchi, Atsushi Oguro, Hiroshi Koyama, Keigo Miyata, Kazumi Takeuchi and Toshio Takahashi
Fi@rt Department ofSu,ge,y, Kyoto Prefectural University ofMedicuze, Kyoto, Japan
dCs(t) F FCs(t) —@-—=@Ca(t)—Vp —@Cs(t)=0,
100F 100k SBF=@@@=(@) 1'
@Cs(t) p) whereD is the specificgravityof the spleen.
0.
0.
0.
8
-0.
U U
§ C 0 U
.@ -0. 8
0.95 1 1.05 1.
—0. spleen —blood partition coefficient for water
@t
@@ - @•
•-
@@T;_@-:-@@-
@ , Th@---
d r @@,
-.FiGURE
- Correlationcoefficient(r)betweenSBF determined by the dynamic and Steady-State methods when different spleen bk)od partition coefficients for water (p) were applied to the steady I Vstate
method.the rvaluewasthe h@hestamount(O.571)whenpfor the steady-statemethodwas0.92.correlations
during-@ were found between the SBFs obtained the dynamic and steady states. As p increased from 0.85,
. the correlation coefficient (r) between SBFs determined CT images of a patientwithnormal using the two methods increased and reached a maximum (6)]@‘@n@tk@i;. [r = 0.571, y = 0.76 x —16.5 after Brace's correction
PET and x-raywhen
pthe p was 0.92. Conversely, r gradually decreased as leftbrachialarteryinthe firstand last minutesof the scan^ increased beyond 0.92. In all cases where significant cor session,andTheCa(t).relation the averagecount in the two sampleswas takenas relations were found, the p value was less than 0.005. regressionDirectly of r to p is shown in Figure 2, and the after finishing the flingusingthedynamicmethod blood flow requiresthat p cannot be obtained with value of p was taken which steady-statemethodcorrelate fromthe dynamicmethod. methods pertainingto the scribed in a previous reportsteady-state
cxperiment,PET @. lines for each change in p are plotted in Figure 3. @ was performed. Determination of p was 0.93, the regiession line relating the SBFs is known, althoughthis parameter for the two methods passed on the origin. Figure 4 shows thesteady-state method. Consequently, a the correlation between SBFS for the two methods. The made the SBF derived from the... most @losclywiththe SBFdenved^.^.^ regression line relating SBF..^ estimated by the steady-state Detailsof the theoiy andthe actual method (y) to SBF detetinmed by the dynamic method (x) dynamic state method have been de- was y 0.57 x + 0.03 with an F ratio of 48.75 (df. = 103, betweenregionalcerebral (3), whichis basedon the methodfor P 5.0 X 108%, by ANOVA). The relationship flow(1).The blood whichclosely specificgravityofthe spleenwas takenas one, approximates the surgically removed spleens. Braces method(6) was the regressionlme between oris Because both the havemeasurementerrorsmean
specificgravityof 1.029for 15
usedto calculatethe correlationand the dynamicandsteadystatemeth dependentandthe independentvariables corrections were made by multiplying=i
@ o@ -i oo , =o. 95 p =0. 93 @@ 150 #.‘@:::.:.s : -:.-. -T - ... .@ @@ ,@-0 90 @ iooeach ,@
y standard
Aby
thestandarddeviation deviation of y F @ significance
typical PET imageofx
anddividingtheresultbythe test analysis (ANOVA) @asapplied to of linearrelationships
with a ROI and x-ray CF images-@
@ 50
@@@@@@@. : 85
@ 300areshown o so ioo 150 200 250 method100 in Figure 1.With the dynamic method, SBF persplenic^ blood flow by the dynamic to(mi/lOOg/min)347.9 gof splenic tissue(SBF = F/V) ranged from 46. ml/min/100g (mean: spleen-blood partition When p derived from163.9,
@@ s.e.: 6.6). The mean wi@UiBraC&Smethod
@ coefficient was 0.74. s@ d@fr@d bythesteady-stateanddynamicstatemethods @ arethe the dynamic method was used for ddte@ntspisen-bicodpartitioncoefficientsfor water (p) steady-state methodor p was 0.84, no significant shown.
600 TheJournalof NudearMedicine•Vol.36 •No.4 •Apdl
p values obtained with the dynamic method were used, although it soon became apparentthat p can have a wide range of values, which if low can yield a negative blood flow to the spleen. Therefore, p 0.85 was needed to maintainsignificantcorrelations between SBF determined using the two methods. In the current study, the highest correlation coefficient between the two methods (0.571) was found when p = 0.92, althoughif p was assumed to be 1.00, the correlation coefficient was only slightly less (0.539).Thus,assumingp = 1.00maybeadequatefor calculating SBF by the steady-state method. In another study, Lammertsma reported that if the partition coeffi cient increased above one, there would be only a small over- or underestimationofcerebral blood flow (10). Thus, many investigators of cerebralblood flow and oxygen me tabolism that use the steady-state method (11,12) assume that p = 1. Despite these arguments, however, it may not necessarily follow that setting p = 1 will provide reliable estimates of SBF. We found thatwith p = 1 the regression line for dynamic and steady-state evaluations of SBFs in tersected the axis for SBF by the steady state method on 66.0ml/100g/min (Fig. 3). By settingp = 0.93, the regres sion line relating SBFs determined by the two methods passed on the origin. This line was thought to be the best one. The spleen-blood partition coefficient for water, p, de nved from the steady-state method was higher when com pared to the value from the dynamic method. In the dy namic method, p can have a wide range of values as mentioned above because it is detennined using the non linear least squares method. In the steady-state method, however, p can have only a veiy narrow range of values and must be close to one, since p is a denominationof one in the second equation. This also explains why organblood flow derived from the steady-state method is inaccurate. Because the patients in our study have neither hepatic nor splenic disorders, these results are applicable to them. Data from a previous study (4) suggest significantcorrela tion between the spleen-blood partitioncoefficient for wa ter and hepatic function. This perspective on partitionco efficients in patients with liver disease requires further study.
In conclusion, we have established a relationship be tween dynamic and steady-state methods for measuring SBF. A quick evaluation of SBF can be made with the
followingregressionline:y (SBF by the dynamicmethod)
= 0.57 x (SBF by the steady-statemethod with p at 0.93) + 0.03.
ACKNOWLEDGMENTS
The authorsthankMr.HitoshiHorn,Mr.KazuoWakitaand Mr. Ryo Fujii, members of the Nishijin Hospital PET Center, for technicalassistance.
REFERENCES
1. Ter.PogosianMM, EichlingJO, DavisDO, Ctal. Thedeterminationof re@ cerebral blood flow by means of water labeledwith radioactive oxygen-15.Radiology1969;93:31—40. 2. Jones1, ChesterDA, Ter.PogossianMM. The continuousinhalationof oxygen-15for assessingregonal oxygenextractionin the brain of man. Br IRadiol 1976;49:339—343. 3. OguroA, TaniguchiH, KoyamaH, et al. Quantificationof humansplenic bloodflow(Quantitativemeasurementofsplenicblood flowwithH2150and dynamicstate method:I).Ann NuciMed 1993;7:245-250. 4. OguroA, TaniguchiH, KoyamaH, etal.Relationshipbetweenliverfunc tionand splenicbloodflow(Quantitativemeasurementofsplenicbloodflow **withH2150anddynamicstatemethod:II).AnnNuciMed1993;7:251-255.
- WestJB,DolleryCF.Uptakeofoxygen.15-IabeledCO2inthelung.JAppI** Physiol 1962;17:9—13. 6. BraceRA. Fittingstraightlinestoexperimentaldata.AmI Physiol1977; 233:R94-R99. 7. TaniguchiH, OgumA, TakeuchiK, et al. Differencein regionalhepatic bloodflowin liversegments—noninvasivemeasurementofregionalhepatic arterial and portal blood flowin humanby positron emissiontomography with H2150.Ann Nuci Med 1993;7:141—145. 8. TakeuchiK,TanigUChiH,OguroA,etal.Quantificationofhumanpancre atic blood flowby@tron emission tomographywith oxygen.l5water. Eur JNuclMed, in submission. 9. YamaguchiT,Kannol,IidaH,etal.Studyforthestabffityinthe15Ogas steady state method.JapJNucl Med 1986;23:9-15. 10. LammertsmaAA, JonesT, FrackowiakRSJ,etat.A theoreticalstudyof the steady-statemodelfor measuringregionalcerebralblood flowand ox. ygen utilization using oxygen-15.JConiputAssist Tomogr 1981,5:554-550. 11. FraCkOWiakRSJ, Lenzi GL, Teriy J, et aL Quantitativemeasurementof cerebralblood flowand oxygenmetabolismin man using“0and positron emission tomography: theoiy, procedure and normal values. I Comput Assist Tomogr 19804:727-736. 12. LammerismaAA, JonesT. Correctionfor the presenceof intravascular oxygen.15in the steady.statetechniquefor measuringregomi oxygen extraction ratio in the brain: I. Descriptionof the method.I CerebBlood FlowMetabol 1983;3:416—424.
602 The Journal of Nudear Medians •Vol.36 •No. 4 •ApilI
