/STRU^NI RAD UDK 616.36-073.75:616.149-089.819

Estimation of the relative liver perfusion using two methods of radionuclide in the patients with...... hemodynamic disorders in the portal system VM Artiko1, DP [obi}-[aranovi}1, SV Pavlovi}1, MS Peri{i}-Savi}2, MV Stojkovi}2, IB Radoman3, SJ. Kne‘evi}4, M@ Vlajkovi}1,VB Obradovi}1 1Institute for Nuclear Medicine, Clinical Center of Serbia 2Institute for Digestive Diseases, Clinical Center of Serbia 3Gastroenterology Department, Clinical Center of Montenegro 4Gastroenterology Department, Clinical Center of Serbia, Ni{

The aim of this study is the assessment of the rela- Noninvasive methods, such as echoangiography of the tive arterial and venous contribution to the total large hepatic blood vessels, duplex-Doppler ultrasonogra- liver blood flow (hepatic perfusion index-HPI), phy of the portal venous system1 and methods of nuclear with two methods (S1 and S2), and estimation of medicine2-4 have been applied in hepatic perfusion stud- their value. With this correction, HPI nonsignifi- ies. Angiography and other invasive methods follow5. cantly increases (p>0.05) in all the groups of pa- Computerized tomography and nuclear magnetic reso-

rezime tients, with a very high correlation between the nance, although used mainly for the morphological ex- HPI(S1) and HPI(S2) values (p<0.01). In comparison amination can be also performed in hepatic blood flow to the portal perfusion in controls, values were signifi- studies6,7,8. Nuclear medicine methods are basicly foun- cantly (p<0,01) lower in chronic active hepatitis and ded on the analysis of hepatic radionuclide angiogram liver and differed between themselves (HRA), obtained after the "first pass" of the radioactive (p<0.01). In the groups of cirrhotic patients with bolus and registered activity over liver and other abdomi- , sclerosated esophageal varices, re- nal organs after intravenous injection of 99mTc. The tech- canalized umbilical , portal thrombosis and caver- nique of and determination of nous portal vein, portal perfusion was lower (p<0.01) the hepatic perfusion index (HPI) proposed by Sarper et than in controls, chronic active hepatitis and liver cirr- al.9,10 is noninvasive method supplying well reproducible hosis without collaterals. information on portal blood flow. Both angioscintigraphic methods are useful for the es- Performing the different mathematical formulas, it is timation of the disturbances in the portal system. Be- possible to estimate selectively arterial and venous liver cause of the more exact estimation of the liver perfu- perfusion. Sarper et al.9 proposed estimation of the he- sion, S2 is recommended. patic perfusion index (HPI) using slopes of the arterial Key words: hepatic radionuclide angiography, "first and venous phase of HRA (S1). Later, the same author, pass", arterial "washout" phase modificated his method introducing the correction of HPI for the amount of the contribution of the arterial "wash- INTRODUCTION out" phase in the portal blood supply (S2)10. The aim of the study is the estimation of various porto- nder physiological condition, liver receives approxi- hepatic circulation parameters using hepatic radionuclide Umately 70% of the portal and 30% of the arterial angiography (HRA) in patients with portal vascular disor- blood supply, but various pathological changes can ders and the hypertension syndrome. The other aim is the disturb this ratio. Thus, studies of the liver blood flow estimation of the validity of the correction of HPI ob- are important in the evaluation of hemodynamic disor- tained by Sarper’s method (S1), using the correction for ders in various hepatic diseases. The estimation of portal the arterial hepatic "washout" phase (S2). blood flow can provide the valuable information on the degree of portal flow restriction, the progress of patho- PATIENTS AND METHODS physiological changes and assessment of their extent, as well as for the appropriate choice of therapy (shunt sur- Investigation was performed in 10 controls (C), in 8 pa- gery and its type). tients with chronic active hepatitis (CAH), 13 patients with liver cirrhosis without esophageal varices (LC), 18 with liver cirrhosis and esophageal varices (LCEV), in 9 12 V. Artiko et al. ACI Vol. LV

FIGURE 1. CONTROL PATIENT. HEPATIC RADIONUCLIDE ANGIOGRAPHY SHOWS PHYSIOLOGICAL SHAPE OF THE LIVER CURVE (HPIS1=0.67, HPIS2=0.72). FIGURE 2. with liver cirrhosis and sclerosated esophageal varices PATIENT WITH LIVER CIRRHOSIS AND RECANALIZED (LCSEV), in 12 with portal thrombosis, or the thrombosis UMBILICAL VEIN. HEPATIC RADIONUCLIDE ANGIOG- of the portal venous branches (PVT), in 5 with recanal- RAPHY SHOWED INCREASED ARTERIAL AND DIMIN- ized umbilical vein (RUV) and in 8 with cavernous portal ISHED PORTAL LIVER INFLOW (HPIS1=0.36, vein (CPV). HPIS2=0.39). HRA was performed with bolus injection of 740 MBq 99mTc-pertechnetate, (1 min, 1f/s), using ROTA scintilla- tion camera. From the ROI over liver parenchyma, TA ar- terial-hepatic and portal-venous phase of the HRA, were separated at the moment when maximal activity over the left kidney ROI was registered. According to S1 method, on HRA, arterial slope (Ba) was estimated from the begi- nning of the activity in the liver region (T0), during 7s. Similarly, the portal slope (Bp) was estimated on the liver curve from the time point which reflects the maximal ac- tivity over the left kidney region (Tm), during 7s. HPI was calculated using formula: HPI = Bp/(Ba+Bp),9. Thus, HPI reflects the value of the relative portal contribution to the liver blood flow. Improved (S2) method, demands cal- culation of the arterial (As- from T0 during 7 s) and ve- nous (descendent - Ds, from kidney Tm during 7 s) slopes on the splenic curve. True portal slope (Pt) is obtained FIGURE 3. with the correction of Bp for the value of the arterial PATIENT WITH INCOMPLETE PORTAL VENOUS "washout" phase through hepatic (P0), using formu- THROMBOSIS. HEPATIC RADIONUCLIDE ANGIOGRA- las: a) Pt = Bp – P0; b) P0 = (DsxBa)/As and, finally, e) 10 PHY SHOWED INCREASED ARTERIAL AND VERY DIS- HPI = Pt/(Ba+Pt), . CRETE PORTAL LIVER INFLOW (HPIS1=0.11, The final diagnoses were based according to the clinical HPIS2=0.13). findings, results of the functional and laboratory analysis, ultrasound and Doppler ultrasound, angiography, biopsy ric and nonparametric correlation tests were used to estab- with histopathology and other clinical examinations. Con- lish eventual corelation between the HPI values obtained sent was obtained from each patient, and the study proto- using S1 and S2 method. col conforms to the established ethical guidelines. RESULTS Besides the mean HPI values (X) and standard deviation (SD), statistical analysis included T test, U test, one and With this correction, HPI increases. However, signifi- two way analysis of variance as well as Multiple range cant differences between HPI values obtained with these test. two methods in the different groups of patients were not For the estimation of the significances of differences of observed (Table 1). On the other hand, there was very the HPI values obtained using two different methods in- high parametric corelation of the results obtained using side the groups of the patients, U test was used. Paramet- these two methods in the C group (r = 0.9139, p< 0.01). In Br. 1 Estimation of the relative liver perfusion using two 13 methods of radionuclide angiography

TABLE 1

VALUES AND SIGNIFICANCES OF THE DIFFERENCES OF HEPATIC PERFUSION INDEX OBTAINED BY SARPER’S BASIC (S1) AND SARPER’S MODIFIED (S2) METHOD IN THE DIFFERENT GROUPS OF PATIENTS STUDIED

Group n HPIS1 HPIS2 SIGNIFICANCES C 10 0.66 +0.06 0.71+ 0.05 p>0.05 CAH 8 0.59 +0.03 0.61+0.08 p>0.05 LC 13 0.49+0.13 0.53+0.12 p>0.05 LCEV 18 0.32+0.19 0.36+0.19 p>0.05 LCSEV 90.28+0.13 0.30+0.15 p>0.05 PVT 12 0.13+0.08 0.185+0.09 p>0.05 RUV 50.21+0.11 0.22+0.11 p>0.05 CPV 80.23+0.11 0.25+0.10 p>0.05 C=control; CAH=chronic active hepatitis; LC=liver cirrhosis; LCEV=liver cirrhosis esophageal varices; LCSEV=liver cirrhosis sclerosated esophageal varices; RUV=recanalized umbilical vein; PVT=portl vein thrombosis; CPV=cavernous portal vein the groups of patients with higher standard deviation of the degree of portal hypertension in the patients with liver the values, existed very high (p< 0.01) nonparametaric cirrhosis, especially in the most severe form10. correlation between the HPI results as follows: in CAH r According to our results, using Sarper’s modified =0.9683, in LC r=0.9545, in LCEV r=0.9183, in LCSEV r method with correction of HPI, using correction of HPI =0.9494, in PVT r=0.9193, in RUV r=0.9256 and in CPV for the value of the arterial "washout" component, ob- r=0.9173. tained values are higher, pointing out higher participation With both methods, HPI values were decreased in all of the portal blood in the total liver vascularization, in the groups of portal hypertensive patients in comparison comparison to the results obtained by basic (Sl) method. to normal values (Tables 1 and 2). However, significant difference between HPI values ob- tained with these two methods (p>0.05) was not proved DISCUSSION neither in the controls, nor in the groups of patients with portal hypertension. There is a high correlation between Initially, Sarper, using the first (S1) method, Sarper et the HPI values obtained with this methods (p<0.01). Any- al.9 estimated the arterial phase slope during 7 s after the way, the absence of the significant differences between beginning of the liver activity, (because its shortest dura- the HPI results obtained by Sarper’s basic and modified tion, obtained experimentally, lasted 7 s). Similarly, the method, does not exclude the need for the correction of initial slope of the portal phase was determined. The re- the slope of the portal component HRA because of the sults, obtained with this method, showed very high corela- more exact estimation of the liver perfusion with S2 tion with angiographically determined degree of portal method. hypertension9. However, at the same moment when portal In the patients with chronic active hepatitis, HPI was sli- component arrives in the liver, being late after arterial one ghtly, but significantly decreased (p<0,01) in comparison because of the time needed to pass through the abdominal to C (Figure 1) and still significantly higher in comparison vessels, arterial blood, which previously arrived in the to other groups of portal hypertensive patients (LC, liver and passed through liver parenchyma outflows thro- LCEV, LCSEV, RUV, PVT and CPV). ugh hepatic veins, so, the portal phase slope of the liver Venous liver perfusion is significantly (p) decreased in curve is decreased for the value of the slope of the arterial liver cirrhosis without and with esophageal varices, reca- washout phase. Because of that, Sarper, modified his nalisation of the umbilical vein (Figure 2) and portal oc- method and suggested the correction of the portal slope clusion (Figure 3). However, portal perfusion in the pa- for the slppe of the arterial washout phase, supposing that tients with liver cirrhosis without esophageal varices was the ratio of the ascendent, arterial phase of the spleen and still significantly (p<0,01) higher than in the groups of pa- its descendnent outflow” phase is the same as the ratio be- tients with collateral circulation, whole values didn’t dif- tween the arterial inflow of the spleen and splenic outflow fer between themselves (p>0.05). phase. Using this, modified method, he obtained better The literature data are similar to ours. The portal com- corelation of the results with angiographic findings and ponent of liver perfusion determined by scintigraphy was reduced in patients with liver cirrhosis and correlated to 14 V. Artiko et al. ACI Vol. LV

TABLE 2

SIGNIFICANCES OF THE DIFFERENCES OF HEPATIC PERFUSION INDEX VALUES OBTAINED BY BOTH SARPER’S METHODS BETWEEN THE DIFFERENT GROUPS OF PATIENTS STUDIED

Group CAH LC LCEV LCSEV RUV PVT CPV C p<0,01 p<0,01 p<0,01 p<0,01 p<0,01 p<0,01 p<0,01 CAH p<0,01 p<0,01 p<0,01 p<0,01 p<0,01 p<0,01 LC p<0,01 p<0,01 p<0,01 p<0,01 p<0,01 C=control; CAH=chronic active hepatitis; LC=liver cirrhosis; LCEV=liver cirrhosis esophageal varices; LCSEV=liver cirrhosis sclerosated esophageal varices; RUV=recanalized umbilical vein; PVT=portl vein thrombosis; CPV=cavernous portal vein the development of cirrhosis and to porto-hepatic gradient ing angiographic perfusion of grades 1 to 4 the index was pressure. Angioscintigraphy and Duplex Doppler appear 54+4.6, 37+2.6, 17+4.7% respectively which also proved to be complementary in the study and follow-up of portal high correlation with the angiographic grade of portal hypertension11. Gon et al.13 recommends the application flow (p<0.001). Hamato et al.24 used the method of of the assessment of the total liver blood flow, but relative Fleming et al.25 and obtained significantly lower mesen- portal inflow in the CAH and LC correspond to our val- teric and the relative portal flow in cirrhosis (42.0+16.4% ues. and 70.9+8.9%) than in the controls (68.6+8.5% and The results of Takegoshi et al.14 correspond to our val- 78.6+5.9%). ues.The ratio was proportionally decreased to the progres- Similar to our findings, the results of Sauerbruch et al26 sion of the diseases (normal 74.5+7.3%, chronic hepatitis show that while the portal venous fraction of hepatic 58.8+9.2%, compensated liver cirrhosis 29.3+19.3%) and blood flow is significantly reduced in patients with liver correlated well with the patohistological findings. He rec- cirrhosis and esophageal varices (17+10%), it is not influ- ommended it as a noninvasive and simple method valu- enced by (20+11%) but the values were sig- able in diagnosing the chronic liver disease, and in esti- nificantly lower (56+9%, p<0.001) in comparison to the mating its clinical course. Kralik et al.15 obtained a decre- controls. The results of Gianpaolo et al.2, although not ase in the HPI which corresponded to the hemodynamical quite in accordance to our and the results of other authors, efficiency of collateral circulation. According to Drag- found increased arterial supply in liver cirrhosis and in oteanu et al.16, liver angioscintigraphy and per-rectal shunted patients and in spite of some methodological lim- portal scintigraphy are related to the degree of portal hy- itations proved some clinical utility. pertension, thus supporting disease staging and follow-up In the patients with portal thrombosis, HPI was signifi- of evolution to cirrhosis and portal thromboses. Similarly, cantly decreased (p<0,01) in comparison to C, CAH and Miguel et al.17, claimed that the ratio of hepatic ar- LC. In a few patients with complete portal vein throm- tery/portal flow in cirrhotic and alcoholic hepatitis was bosis, hepatic perfusion was exclusively arterial, which significantly different from normal controls (p<0,001) and was similar to our findings27. Also, MacMathuna et al.28 recommended radionuclide methods as a non invasive test emphasized the importance of HRA as a safe and accurate in the study of chronic liver diseases. Santambrogio et screening method for evaluating portal vein patency or al.18 obtained the mean HPI 0.42+0.24 in the LC group occlusion, particularly at a portal venous inflow of 20%, and 0.66+0.19 in the non-LC group (p<0.02), and claimed where the specificity was 100% and sensitivity 90%. that HPI is useful factor in estimating hepatic blood flow By performing echo-angiography and Duplex-Doppler and liver function19. The results of other authors20 are echosonography, one can successfully determine the di- also in accordance. Koranda et al.21 using HPI, confirmed rection and rate of blood flow, the presence of the collat- that the net hepatic arterial blood flow is also increased in eral circulation, the existence of thrombosis in blood ves- patients with cirrhosis. Fanfani et al.22 proved that HPI sels and detected other portal system vascular disorders, was reduced only in cirrhotic patients but not in those while hepatic radionuclide angiography enables the evalu- with chronic hepatitis, which is only partly in accordance ation of relative liver perfusion. The results show that por- to our findings. According to Sarper et al.9, abnormal val- tal inflow was normal in control group and decreased in ues of HPI in spite of physiological values of liver func- liver cirrhosis and portal hypertenson, proportionally to tion tests, proved this method to be a sensitive, noninva- the degree of portal hypertension. The lowest values for sive detector of early pathophysiological changes in the the participation of portal blood in the total liver vascular- splanchnic organs of alcohol and drug abusers. The same isation, were obtained in those with complete portal ve- author10 proved it to be correlated with the degree of por- nous thrombosis. Angioscintigraphy is reasonably accu- tal hypertension established by angiography. The observa- rate, reproducible, safe and can be considered suitable for tions of Rypins et al.23 were similar: HPI in normal sub- routine use in the assessment of liver hemodynamics. jects was 66.0+3.4 and in cirrhotic patients with decreas- Br. 1 Estimation of the relative liver perfusion using two 15 methods of radionuclide angiography

CONCLUSION 5. Nelson RC, Lovet KE, Chezma JL. Comparison of pulsed Doppler sonography and angiography in patients Hepatic radionuclide angiography can be used, among with portal hypertension. AJA 1987;149:77-81. other methods for the detection, evaluation, and follow-up 6. Ganeshan B, Miles KA, Young RC, Chatwin CR. In of hemodynamic disturbances in the portal venous sys- search of biologic correlates for liver texture on portal- tem. The HPI values obtained by both methods are pro- phase CT. 2007 Sep;14(9):1058-68. portional to the degree of portal hypertension. 7. Rhee TK, Omary RA, Gates V, Monajjed T, Larson However, because of the more exact estimation of the AC, Barakat O, Sato KT, Mulcahy M, Gordon S, Lewan- relative liver blood flow, S2 method is recommended. dowski RJ, Salem R. The effect of catheter-directed CT angiography on yttrium-90 radioembolization treatment of REZIME hepatocellular carcinoma. 2005 Aug;16(8):1085-91. PROCENA RELATIVNE PERFUZIJE JETRE 8. Annet L, Materne R, Danse E, et al. Hepatic flow pa- PRIMENOM DVE METODE RADIONUKLIDNE rameters measured with MR imaging and Doppler US: ANGIOGRAFIJE U BOLESNIKA SA POREME]AJIMA correlations with degree of cirrhosis and portal hyperten- PROTOKA KRVI U PORTNOM SISTEMU sion. Radiology 2003; 229:409-414. 9. Sarper R, Fajman WA, Rypins EB, et al. A noninva- Cilj rada je odredjivanje relativnog doprinosa arterijske sie method for measuring portal venous/total hepatic i portne komponente u vaskularizaciji jetre (hepati~ki per- blood flow by hepatosplenic radionuclide angiography. fuzioni indeks - HPI) primenom dve metode (Sl) i (S2), i Radiology 1981;141:179-184. procena njihove vrednosti. 10. Sarper R and Tarcan YA. An improved method of U odnosu na S1, primenom S2 metode, HPI se pove}a- estimating teh portal venous fraction of total hepatic blood va, ali ne zna~ajno (p>0.05) u svim grupama ispitanih flow from computerized radionuclide angiography. Radi- bolesnika, uz visoku korelaciju u vrednostima (p<0.01). U ology 1983;147:559-562. poredjenju sa portnom perfuzijom u kontrolnoj grupi, vre- 11. Bouvard G, Bouvard N, Dao T, et al. Value of he- dnosti HPI su zna~ajno ni‘e (p<0,01) u grupama boles- patic angioscintigraphy combined to pulsed Echo-Dop- nika sa hroni~nim aktivnim hepatitisom i cirozom jetre, u pler: application to the study of portal hypertension of cir- kojima se i medjusobno razlikuju (p<0,01). U grupama rhotic patients. Apropos of 148 cases. Ann Radiol (Paris). bolesnika sa variksima jednjaka, skleroziranim variksima 1991;34:362-368. jednjaka, rekanalizovanom umbilikalnom venom, trombo- 12. Lafortune M, Constantin A, Baeton G, et al. The re- zom vene porte i kavernoznom portnom venom, portna canalised umbilical vein in portal hypertension: a myth. perfuzija je bila zna~ajno ni‘a (p<0,01) od vrednosti u AJA 1985:144:549-553. kontrolnoj grupi, hroni~nom aktivnom hepatitisu i cirozi 13. Gon J, Nakagawa T, Tagami T, et al. Measurement jetre bez razvijene kolateralne cirkulacije. of intrahepatic shunt using Rutland method. Kaku Igaku Obe metode angioscintigrafije su korisne za procenu 1991;28:1-5. hemodinamskih poreme}aja u portnom sistemu. Zbog eg- 14. Takegoshi K, Seto H, Tohyama T, Okuda K, et zaktnije procene relativne perfuzije jetre, preporu~uje se al.The clinical usefulness of portal venous flow ratio by S2 metoda. hepatic angiography with 99mTc-Sn colloid in chronic Klju~ne re~i: hepati~ka radionuklidna angiografija, liver diseases: a comparison with histological finding. "prvi prolaz", arterijska "washout "faza Kaku Igaku 1989;26:499-505. 15. Kralik J, Koranda P, Neoral C, et al.Experimental BIBLIOGRAPHY verification of the hepatic perfusion index for monitoring direct liver perfusion by portal blood Rozhl Chir. 1. Grant GG, Tessler F, Perrella R. Color Doppler in- 1993;72:212-215. valuable in imaging liver vessels. Diagnostic Imaging In- 16. Dragoteanu M, Cotul SO, Tamas S, et al. Nuclear ternational 1989;1:50-54. medicine dynamic nvestigations of diffuse chronic liver 2. Gianpaolo M, Massimo B, Enzo S, et al. Assessment diseases and portal hypertension. Rom J Gastroenterol. of liver circulation by quantitative scintiangiography: 2004;13: 351-357. evaluation of the relative contribution of hepatic arterial 17. Miguel F, Boveda J, Solano D, et al.Isotopic ra- and portal venous blood flows to the liver perfusion. Eur J dioangiography in hepatic cirrhosis. Changes induced by Nucl Med 1989;15:211-216. somatostatin. Rev Clin Esp 1990;187:61-64. 3. Kosti} K, Artiko V, Peri{i}-Savi} M, et al. Evaluation 18. Santambrogio R, Opocher E, Bruno S, et al. The of hepatic and portal perfusion in live diseases and vascu- noninvasive assessment of the effects of penbutolol on lar disorders. Period Biol 1989; 91:497-498. liver hemodynamics in cirrhotic patients using angioscin- 4. Kosti} K, Artiko V, Peri{i}-Savi} M, et al. Ulta- tigraphy. A randomized controlled double-blind study. sonography (Doppler) and hepatic radionuclide angiogra- Recenti Prog Med 1990;81:705-709. phy in various disturbances of portal venous system. In: 19. Ono K, Kato H, Nishibe T, et al. Clinical evaluation Schmidt HA, Van der Schoot JB(ed.) The State of the Art of hepatic blood flow and liver function with 99mTc- of Nuclear Medicine in Europe, Nuclear Medicine. Edi- DTPA-HSA. Nippon Geka Gakkai Zasshi. 1993;94:816- tions Schattauer 1990:265-267. 823. 16 V. Artiko et al. ACI Vol. LV

20. Seidlova V, Hobza J, Pumprla J, et al.Radionuclide hepatic perfusion index and ultrasonography: assessment of portal hypertension in clinical practice. Acta Univ Palacki Olomuc Fac Med. 1989;122:213-221. 21. Koranda P, Myslivecek M, Erban J, et al.Hepatic perfusion changes in patients with cirrhosis indices of he- patic arterial blood flow. Clin Nucl Med 1999 ;24: 507- 510. 22. Fanfani G, Fratello A, Mele M, et al.Renal perfusion in chronic liver diseases: evaluation by radiotechnetium renography. Eur J Nucl Med 1985;10:317-320. 23. Rypins EB, Fajman W, Sarper R, et al. Radionuclide angiography of the liver and spleen. Noninvasive method for assessing the ratio of portal venous to total hepatic blood flow and portasystemic shunt patency. Am J Surg. 1981; 142:574-579. 24. Hamato N, Moriyasu F, Someda H, et al.Distribu- tion of mesenteric and splenic blood flow in cirrhosis of the liver as estimated by radionuclide angiography. Intern Med. 1993;32:445-448. 25. Fleming JS, Ackery DM, Walmsley BH, et al. Scin- tigraphic estimation of arterial and portal blood supplies to the liver. J Nucl Med 1983;24:1108-1113. 26. Sauerbruch T, Moser E.Influence of long-term injec- tion sclerotherapy on portal venous component of total liver perfusion measured by hepatosplenic radionuclide angiography. Hepatogastroenterology 1986;33:17-19. 27. Fakhay J, Gosink BE, Leopold GA. Recanalized Umbilical Vein Due to Portal Vein Occlusion: Documen- tation by Sonography. AJA 1981;137:410-412. 28. MacMathuna P, O’Connor MK, Weir DG, Keeling PW. Non-invasive diagnosis of portal vein occlusion by radionuclide angiography. Gut 1992;33:1671-1674.