Measurement of liver blood Measurement of liver blood flow using [flow using [1515O]HO]H22O and PETO and PET

Literature reviewLiterature review

77thth Modelling Workshop in Turku PET Centre, 20 Modelling Workshop in Turku PET Centre, 20thth October 2005 October 2005

Turku PET CentreTurku PET Centre20052005

Why measuring hepatic blood flow ?Why measuring hepatic blood flow ? Hepatocellular carcinoma is the most frequent malignant tumor of Hepatocellular carcinoma is the most frequent malignant tumor of

the liver. Liver metastase play a role in the morbidity and the liver. Liver metastase play a role in the morbidity and mortality associated with this cancermortality associated with this cancer

Metastatic tumor of liver: Multiple Metastatic tumor of liver: Multiple liver nodules liver nodules

are consistent with metastasesare consistent with metastases

((www.pathology.vcu.edu/education/dental2/lab4.html).).

Non-invasive measurement of regional blood flow within the liver Non-invasive measurement of regional blood flow within the liver is of particular clinical interest, since changes in the tumor’s is of particular clinical interest, since changes in the tumor’s blood flow have been recognized as an important aspect for the blood flow have been recognized as an important aspect for the assessment of treatment response of hepatic cancersassessment of treatment response of hepatic cancers

Many methods exist : classified in direct and indirect methodsMany methods exist : classified in direct and indirect methods

1) Direct methods can measure hepatic blood flow but they are 1) Direct methods can measure hepatic blood flow but they are invasive (exposure, cannulation)invasive (exposure, cannulation)-Microsphere method-Microsphere method-Plethysmography methodPlethysmography method-Electromagnetic -Electromagnetic

2) Indirect methods can be invasive or non-invasive :2) Indirect methods can be invasive or non-invasive :-Doppler ultrasonography (operator dependant, reproducibility)Doppler ultrasonography (operator dependant, reproducibility)-Dynamic computed tomography-Dynamic computed tomography-Dilution techniques-Dilution techniques-Radiolabelled methods-Radiolabelled methods

Positron emission tomography with [Positron emission tomography with [1515O]O] HH22O tracerO tracer

PET using an intravenous bolus injection of [PET using an intravenous bolus injection of [1515O] HO] H22O tracer O tracer permit non invasive measurement of hepatic blood flow and permit non invasive measurement of hepatic blood flow and

separate measurement of arterial and portal blood flowseparate measurement of arterial and portal blood flow

Curves obtained after tracer Curves obtained after tracer injectioninjection

ROI for the liver segments ROI for the liver segments

(Taniguchi et al, 1999)(Taniguchi et al, 1999)

Kety-Schmidt model (Kety & Schmidt, Kety-Schmidt model (Kety & Schmidt, 1948)1948)

This model is applicable to the study of blood flow in various This model is applicable to the study of blood flow in various organs (brain, heart) but, also in liver with modifications to organs (brain, heart) but, also in liver with modifications to reflect unique delivery environment of liver : incorporation of reflect unique delivery environment of liver : incorporation of two separate input functions (dual-input Kety schmidt model) or two separate input functions (dual-input Kety schmidt model) or using a combined input function. using a combined input function.

But the dual-input is mostly used (Taniguchi But the dual-input is mostly used (Taniguchi et alet al, 1993, 1996, , 1993, 1996, Ziegler Ziegler et alet al 1996, Munk 1996, Munk et alet al 2003) 2003)

Liver has dual blood supply :Liver has dual blood supply :

Hepatic artery (arterial blood from heart)Hepatic artery (arterial blood from heart) Portal vein (venous blood)Portal vein (venous blood)

The compartmental model has to account to The compartmental model has to account to this two supply this two supply

Scheme for general Scheme for general modelmodel

Dual-input modelDual-input model

Liver(CL)

Hepatic artery (CA)

Portal vein (Cp)

Fa/VL

FP/VL

Fa+ FP /VL

(portal organ )

(Cp)

Liver(CL)

Hepatic artery (CA)

artery (CA)

Fa/VL

FP/VL

Fa+ FP /VL

Single-input modelSingle-input model

Model parameterModel parameter

Fa : arterial blood flowFa : arterial blood flow Fp : portal veinous blood flowFp : portal veinous blood flow Fa + Fp = total hepatic blood flow Fa + Fp = total hepatic blood flow CCAA: Tracer concentration in arterial blood: Tracer concentration in arterial blood

CCpp: Tracer concentration in portal vein: Tracer concentration in portal vein

CCLL: Tracer concentration in liver: Tracer concentration in liver

VVLL: tracer volume of distribution in liver : tracer volume of distribution in liver

The ordinary differential equation The ordinary differential equation describing the tracer dynamicsdescribing the tracer dynamics

( )( ) ( ) ( )

( ) ( )

p a paLA p L

L L L

p p pA p

L L

F F FFdCC t C t C t

dt V V V

dC F FC t C t

dt V V

_

_ _

( )( ) ( ) ( )

( ) ( )

p a paLA p L

L L L

p org p pA p

p org p org

F F FFdCC t C t C t

dt V V V

dC F FC t C t

dt V V

And CAnd CLL(t=0) = 0(t=0) = 0

Assuming transport of tracer from vessel into tissue compartment Assuming transport of tracer from vessel into tissue compartment is fast compared to delivery by arterial and portal-venous blood is fast compared to delivery by arterial and portal-venous blood flows : flows :

tracer dynamics tracer dynamics one-tissue compartment model with 2 one-tissue compartment model with 2 different input functions.different input functions.

Dual input model :Dual input model :

Single input model :Single input model :

Method of time delay correction (taniguchi et al, Method of time delay correction (taniguchi et al, 1999)1999)

Method of time delay Method of time delay correction. The time correction. The time delay (delay (t) between sample t) between sample site and portal organ is site and portal organ is calculated using calculated using nonlinear regression nonlinear regression analysis. And the hepatic analysis. And the hepatic blood flow is calculated blood flow is calculated using the new input using the new input function that function that incorporated incorporated t.t.

Data are also corrected for dispersionData are also corrected for dispersion

For solving model we have to had two delay parameter Ta and For solving model we have to had two delay parameter Ta and Tp which represent the transit time from the aorta and the Tp which represent the transit time from the aorta and the portal vein regions to the liver respectively :portal vein regions to the liver respectively :

12( ')

10

( ) ( ' ) ( ' ) 'a

t k t tL a a p p pC t k C t T k C t T e dt

kk1a1a = F = Faa/V/VLL

kk1p1p =F =Fpp/V/VLL

kk22 = F = Faa+F+Fpp/V/VLL

A prioriA priori identifiability (Becker identifiability (Becker et alet al, 2005), 2005)

Investigation on the a priori identifiability of the 3 paramater FInvestigation on the a priori identifiability of the 3 paramater Faa, F, Fpp

and Vand VLL

Results indicate that the full identifiability of the model depends on Results indicate that the full identifiability of the model depends on the form of the portal-venous input function (cthe form of the portal-venous input function (cpp(t)) which is assumed (t)) which is assumed

to be a sum of m exponentials convolved with the arterial input to be a sum of m exponentials convolved with the arterial input function (cfunction (caa(t)).(t)).

When m When m 2, all 3-model parameters are uniquely identifiable 2, all 3-model parameters are uniquely identifiable

For m = 1 identifiability of FFor m = 1 identifiability of Fpp fails if cfails if cpp(t) coincides with tissue (t) coincides with tissue

concentrationconcentration

Ideally vascular input functions based on ROI derived from aorta Ideally vascular input functions based on ROI derived from aorta and portal vein lead to truly measurement of blood flow with better and portal vein lead to truly measurement of blood flow with better identifiabilityidentifiability

Example of results (Taniguchi, 1999)Example of results (Taniguchi, 1999)

ConclusionConclusion

Liver perfusion can be studied with PET [Liver perfusion can be studied with PET [1515O]HO]H22O and O and

compartmental modelling. Many articles this method combined compartmental modelling. Many articles this method combined with a dual-input one compartment model is a reproducible with a dual-input one compartment model is a reproducible method that provides perfusion values. Arterial and portal-venous method that provides perfusion values. Arterial and portal-venous perfusion can be determined separately.perfusion can be determined separately.

Despite promising reports, this technique is limited by the low Despite promising reports, this technique is limited by the low spatial and temporal resolutionspatial and temporal resolution

ReferencesReferences : :1- Kety SS & Schmidt CF. 1948. The Nitrous oxide methode for the quantitative determination 1- Kety SS & Schmidt CF. 1948. The Nitrous oxide methode for the quantitative determination of cerebral blood flow in man: theory, procedure and normal values. J. Clin. Invest. 27 476-83.of cerebral blood flow in man: theory, procedure and normal values. J. Clin. Invest. 27 476-83.

2- Taniguchi 2- Taniguchi et alet al. . Difference in regional hepatic blood flow in liver segments--non-invasive Difference in regional hepatic blood flow in liver segments--non-invasive measurement of regional hepatic arterial and portal blood flow in human by positron emission measurement of regional hepatic arterial and portal blood flow in human by positron emission

tomography with [tomography with [1515O]HO]H22O.O. Ann Nucl Med. 1993 Aug;7(3):141-5. Ann Nucl Med. 1993 Aug;7(3):141-5.

3- Taniguchi 3- Taniguchi et alet al. . Analysis of models for quantification of arterial and portal blood flow in the Analysis of models for quantification of arterial and portal blood flow in the

human liver using PET.human liver using PET. J Comput Assist Tomogr. 1996 Jan-Feb;20(1):135-44.J Comput Assist Tomogr. 1996 Jan-Feb;20(1):135-44. 4- Taniguchi 4- Taniguchi et alet al. . Using the spleen for time-delay correction of the input function in measuring Using the spleen for time-delay correction of the input function in measuring hepatic blood flow with oxygen-15 water by dynamic PET.hepatic blood flow with oxygen-15 water by dynamic PET. Ann Nucl Med. 1999 Ann Nucl Med. 1999 Aug;13(4):215-21. Aug;13(4):215-21.

5- Ziegler 5- Ziegler et alet al. Measurement of liver blood flow using oxygen-15 labelled water and dynamic . Measurement of liver blood flow using oxygen-15 labelled water and dynamic positron emission tomography: limitations of model description. Eur J Nucl Med. 1996 23: 169-positron emission tomography: limitations of model description. Eur J Nucl Med. 1996 23: 169-177177

6- Munk 6- Munk et alet al. Determination of regional flow by use of intravascular PET tracers: . Determination of regional flow by use of intravascular PET tracers: Microvascular theory and experimental validation for pig livers. J Nucl Med 2003; 44: 1862-Microvascular theory and experimental validation for pig livers. J Nucl Med 2003; 44: 1862-1870.1870.

7- Becker 7- Becker et alet al. . A prioriA priori identifiability identifiability of a one-compartment model with two input functions for of a one-compartment model with two input functions for liver blood flow measurements. Phys Med Biol. 2005 Apr 7;50(7):1393-404liver blood flow measurements. Phys Med Biol. 2005 Apr 7;50(7):1393-404

Different modelsDifferent models