Value of Volumetric and Morphological Parameters on Computed Tomography for Assessing Severity of Viral-Induced Liver Cirrhosis

Med. J. Cairo Univ., Vol. 80, No. 2, June: 189-194, 2012 www.medicaljournalofcairouniversity.com Value of Volumetric and Morphological Parameters on Computed Tomography for Assessing Severity of Viral-Induced Liver Cirrhosis MOHAMED SHERIF EL-SHARKAWY, M.D. The Department of Radiology and Medical Imaging, King Khalid University Hospital, King Saud University, Riyadh, Kingdom of Saudi Arabia. Abstract Aim of the Work : To assess the value of abdominal CT scoring system compared to Child-Pugh classification in determining the severity of viral-induced cirrhosis. Material and Method: Computed tomography examinations of the liver of 26 patients proven positive for hepatitis B (12 patients) and C (14 patients) were reviewed. All patients are examined using 64 slice multidetector CT triphasic liver protocol. Measurements are done for liver volume index, modified caudate lobe ratio (CRL-r), splenic volume index, ascites grading, presence or absence and grading of Porto systemic collaterals, grade of contour changes, confluent fibrosis if present. Patients were assessed by Child-Pugh classification in the same week of CT examination. Excluded from this study those who had end stage shrunken liver or those with focal hepatic lesions on CT. Results are statistically studied using Cohen's Kappa coefficient for measurement of agreement for quantitative items. Results: We found substantial agreement between clinical (Child-Pugh classification) and the measured CT values (except C/RL-r) as the measurement of agreement of Kappa were equal to 0.61 and p<0.0001. Based on CT data, patients were divided into three groups (A, B, and C) and they are assessed for statistical agreement for same groups in Child- Pugh's classification A,B and C groups. Highest agreement was in the Child B group which represents 50% of the patients of the study. We found no relation between clinical stage and modified caudate-right lobe ratio (CRL-r) as rs=0.293 and p>0.05. Also we found no relation between clinical and spleen volume index since rs=0.256 and p>0.05. Conclusion: CT scoring based on Liver volume index well-correlate with the clinical status in viral-induced cirrhotic patients in absence of global liver atrophy and focal hepatic lesions. It can be used with good accuracy rate in follow-up and assement of treatment response. Abbreviations: CT = Computed Tomography. IV = Intravenous. MRI = Magnetic resonance imaging. C/RL = Caudate-to-right lobe ratio. C/RL-r = Modified caudate-to-right lobe ratio. Correspondence to: Dr. Mohamd Sherif El-Sharkawy E-mail: sherif_elsharkawy@hotmail. com Key Words: Liver - Cirrhosis - Computed tomography - Abdomen - Organ volume. Introduction CIRRHOSIS is the final result of chronic damage to the liver from various etiologies, characterized by parenchymal injury leading to extensive fibrosis and nodular regeneration. The result is a diffuse disorganization of hepatic morphology with progressive loss of liver function. Cirrhosis is most commonly the result of hepatitis B and C virus infection or chronic alcoholism. Other causes are biliary, cryptogenic and metabolic [1]. Once cirrhosis develops, hepatic fibrosis proceeds slowly and gradually toward decompensated cirrhosis with risk of several complications [2]. In the stage of cirrhosis where the liver is still compensating, (particularly in cases secondary to hepatitis C virus infection), interferon treatment can help to prolong sufficient liver function. Thus, it is important to define the stage of cirrhosis in each patient and to follow the disease progression carefully [3]. After the diagnosis of cirrhosis has been established, it would be helpful if it were possible to determine the severity of cirrhosis clinically and by noninvasive imaging [2]. This is clinically determined using Child-Pugh classification. Radiologic evaluation is an important part of clinical follow-up in patients with chronic liver disease and its complications. The current gold standard, liver biopsy, cannot be used for population-based screening, has well known drawbacks if used for monitoring of the disease progression or treatment success. CT is more commonly used in these patients and in selected cases where needed magnetic resonance imaging (MRI) is performed. The development of helical CT allowed single breath-hold scanning and evaluation of the whole hepatic parenchyma without motion artifacts during the hepatic arterial 189

190 Value of Volumetric & Morphological Parameters phase and the portal venous phase [3,4]. Both CT and MRI provide valuable insights into the extent of hepatic injury from cirrhosis and complications including portal hypertension. CT has been used as the gold standard in most volume assessment studies [5]. Attention to scan technique is key, including the use of multiphasic imaging with a rapid IV bolus of contrast medium [1]. The physical properties associated with CT lead to highresolution image data sets that are spatially accurate. Although abdominal ultrasound is very useful, quantitative volume accuracy is compromised by the manual scanning procedure and speed of sound inaccuracies [5]. The most significant intra- and extra-hepatic changes related to this condition are as follows: The right liver lobe atrophy, hypertrophy of the caudate lobe and the lateral segment of the left lobe, liver nodularity and irregular contours, heterogeneous hepatic parenchyma on imaging studies, development of regenerative nodules, iron deposits, ascites, Splenomegaly, and Portosystemic collaterals [1,3]. Liver volume index and ascites on CT and MRI are good indicators of clinical severity of hepatic cirrhosis secondary to hepatitis. Ascites is also one of the main parameters used in Child-Pugh classification for clinical severity. On the other hand, although liver volume change is not one of the main parameters of Child-Pugh grading, it is correlated well with functional capacity of the liver in cirrhosis secondary to viral hepatitis [3]. A quantitative approach for assessing the caudate lobe/right lobe ratio has been reported. The diagnosis of cirrhosis can be made with a 96% confidence level if the caudate lobe/right lobe ratio exceeds 0.652 [6]. The ratio between the transverse width of the caudate lobe and the right lobe can be used for differentiating normal from advanced cirrhotic livers. However, this ratio does not help to identify the presence or absence of early cirrhosis [7]. The modified caudate lobe ratio (C/RL-r) appears to better represent the division between hypertrophied and atrophied liver and is more sensitive and accurate for diagnosing cirrhosis than is the previously used version of the C/RL [8]. Volume index of the spleen, volume indexes of the right posterior plus left medial plus left lateral liver segments, the presence of ascites, and the presence of varices and collaterals were important items in assessing the severity of cirrhosis [1]. The spleen commonly increases in size in response to conditions such as infection and hematologic or metabolic disorders. CT is considered the most accurate imaging technique for the in vivo evaluation of organ volume [3,9]. Splenic size can be reliably evaluated in vivo employing imaging methods [5]. The product of the width, the thickness and the length of the spleen (the so-called splenic index ) has also been proposed as an indicator for evaluating splenic size on CT [9]. The aim of this study: To assess the value of abdominal CT scoring system based on liver, splenic volume indexes, ascites, confluent fibrosis and Porto systemic collaterals in addition to modified caudate lobe ratio, compared with Child-Pugh classification for determining the severity of viralinduced cirrhosis based statistically on the agreement between similar grades of each CT assement and Child-Pugh classification. Material and Methods The subject of this study are 26 patients proven positive for hepatitis B (12 patients) and C (14 patients) assessed by both designed CT scoring system and Child-Pugh classification with both done in the same week. Their CT examinations of the liver (done routinely for follow-up and exclusion of hepatic neoplasm) were reviewed. All patients are examined by GE Multi-slice light speed VCT-XTE 64 slices (GE Medical Systems, Milwaukee WI, USA), with slice thickness=0.625mm interval 0.625mm pitch 1.351mm section overlapping every 0.7mm. 1.5mg/kg intravenous iodinated contrast was given injected at a rate of 2.5ml/sec. All patients are examined using triphasic CT protocol for exclusion of focal hepatic lesions. Measurements are done on the PACS station using electronic calibers. CT scoring criteria are illustrated in Table (1) [3]. To calculate liver volume index [3], we summed the volumes of the Posterior segment of the right lobe, the medial segment of the left lobe, and the lateral segment of the left lobe (Fig. 1). Measurements from the liver segments were achieved 1 cm below or above or at the bifurcation of main portal vein. To measure each of these volumes, from each segment we used the largest anterior-posterior and medial-lateral distances of the structure on axial CT images. The superior-inferior distance based on the slice positions at which each segment of the liver first appeared and ended on CT images were used as the third distance. The three distances obtained were then multiplied to calculate volume indices. The spleen volume index was determined measuring largest anteroposterior, mediolateral and superoinferior distances at the level of splenic

Mohamed S. El-Sharkawy 191 hilum. In addition to these parameters, each set of images was assessed for ascites, Portosystemic collateral vessels, contour irregularities and confluent fibrosis of the liver. For transverse diameters of the right and caudate lobes using right portal vein as the separating point for assessment of modified caudate lobe ratio (CRL-r) (Fig. 2). These indices are not true volumes of organs, rather they were values determined a standard way for comparison. For assessment of values of CRL-r and Splenomegaly in relation with different groups of Child-Pugh's classification we used non-parametric Kruskal-Wallis test to compare between different classes A,B and C. Patients were also assessed for ascites and it's grade, presence or absence and grading of portosystemic collaterals, splenic volume index and grade of contour changes, presence or absence of confluent liver fibrosis (wedge-shaped in appearance, radiated from the portal hilus, involving the medial segment of left lobe, the anterior segment of the right lobe, or both). In the same week of CT examination patients were assessed by Child-Pugh classification. If the gap is longer than one week or if the patients had shrunken liver and those with focal hepatic, patient lesions is excluded from this study. Patients were divided, based on CT data, into three scoring groups Table (1): Criteria for grading of each parameter evaluated on CT [3]. (A, B, and C) and they are assessed for statistical agreement for same groups in Child-Pugh's classification (A, B and C groups). Results Based on radiological data assessed by CT (Table 1), Patients are divided radiologically into three groups and correlated for groups of Child- Pugh's classification groups (Table 2). 9 patients are in radiological group A, 14 in group B and 3 in group C with a total number of 26 patients. Clinical matching (i.e. same group radiologically and on Child-Pugh's) was 7 in group A (n=7/9), 10 in group B (n=10/14) and 3 in group C with (n=3/3) a total matching of 20 patients out of the total number of patients which are 26 (77%) (Fig. 3 and Tables 3,4). We found substantial agreement between clinical (Child-Pugh classification) and the measured radiological values (except CRL-r) as the measurement of agreement of Kappa were equal to 0.61 and p<0.0001. Highest agreement was in the Child B group (Fig. 3). Meanwhile, we found no relation (no matching) between clinical and modified caudate lobe ratio (CRL-r) as rs=0.293 and p>0.05. Also we found no relation (no matching) between clinical and spleen volume since rs=0.256 and p>0.05 (Table 5). Parameter CT grade Grade 0 Grade 1 Grade 2 Grade 3 Liver volume index (cm 3 ) Spleen volume index (cm 3 ) Ascites Vascular collaterals Contour changes Fibrosis >1300 1300-1001 1000-651 <650 <400 401-800 801-1400 >1400 None Minimal around liver Intraperitoneal fluid Fluid causing significant and spleen with no distension distension None In one location In two locations In three or more locations None Mild irregularity Moderate irregularity Marked lobulation and irregularity Absent Present Table (2): Clinical and radiology matching and non-matching cases distribution. Group Radiology A n=7/26 (26.92%) B C A n=2=26 (7.69%) n=0/26 ( 0 %) * Total number of patients is 26. Clinical (child-pugh) B C n=4/26 n=0/26 (15.38%) ( 0 %) n=10 n=0/26 (38.46%) ( 0 %) n =0/26 n=3/26 ( 0 %) (11.54%) Table (3): Radiologic-clinical matching patients. Radiological Group Group A (n=9) Group B (n=14) Group C (n=3) Total (n=26) Clinical Matching Per Group Percentage Per Group Matching/total (n/26) N=7/9 78% 26.92% (n=7/26) N=10/14 71% 38.46% (n=10/26) N=3/3 100% 11.54% (n=3/26) N=20/26 77% Total Matching = 77% n (=20/26)

192 Value of Volumetric & Morphological Parameters Table (4): Radiology-clinical matching patients (n=20/26). Radiological Overall Clinical Radiological Group Percentage Matching Group Matching (n=20) (n=27) Group A (n=9) 7/20 7/9 (78%) Group B (n=14) 10/20 10/14 (71.1%) Group C (n=3) 3/20 3/3 (100%) Table (5): Comparison between Classes A,B and C regarding CRL-r and splenic volume. Clinical groups (Child-Pugh) A B C Radiology (n = 9) (n = 14) (n = 3) p-value Median (min.-max.) Median (min.-max.) Median (min.-max.) CRL-r 0.8 (0.5-1.5) 0.95 (0.6-2.3) 1.2 (0.8-1.5) 0.322 Spleen Volume 232.6 (101.7-544.3) 271.45 (100.4-634.7 ) 878.4 (216.9-1457.1) 0.192 * By non-parametric Kruskal - Wallis test to compare between different class A,B and C. 2 1 Caudate Right A B 2 1 Fig. (1): CT liver measurements: Hepatitis B patient with A class in radiology and Child-Pugh classification. Arterial phase of triphasic CT of the liver. Measurement 1 maximum AP diameter of the right lobe. Measurement 2 maximum transverse diameter of the left lobe. 40 38.46 35 30 26.92 25 20 15 11.54 10 5 0 A B C Different class Fig. (3): Distribution of matched patients between clinical and radiology in post-hepatitis cirrhosis patients (total number=26). Percentages Percentages of patients matched between clinical and radiology Fig. (2): Measurement of modified caudate lobe ratio (C/RLr): Hepatitis B patient with A class in radiology. Portovenous phase of triphasic CT liver examination. Line 1 at the most medial margin of caudate lobe. Line 2 through right lateral wall of the bifurcation of the right portal vein. Distance A transverse diameter of the caudate lobe. Distance B transverse diameter of the right lobe. C/RL-r = 0.94. Discussion Morphological changes due to viral-induced cirrhosis are well described on CT whether hepatic or extra hepatic. Those patients require sensitive follow-up particularly those treated conservatively for evaluation of response to treatment and for early detection of complication in the course of the disease. This requires accurate standard scoring system that can be reproduced and followed. Clinically for a long time Child-Pugh system is still holding for this purpose. However this system depends mainly on clinical and laboratory values. Invasive tests as liver biopsy is kept for first step

Mohamed S. El-Sharkawy 193 of diagnosis or in case of suspicion of neoplasm development. Researchers suggested different systems for scoring of severity of hepatic cirrhosis changes [10-12]. Recent studies have investigated the value of various imaging methods for diagnosing liver cirrhosis and evaluating related complications. The most significant intra-and extra-hepatic changes related to this condition are as follows: Atrophy of the right liver lobe, hypertrophy of the caudate lobe and the lateral segment of the left lobe, nodularity in the liver and irregular contours, heterogeneous appearance of the hepatic parenchyma on imaging studies, development of regenerative nodules, iron deposits, ascites, splenomegaly, and varicose veins [3]. There has been great interest in determining in vivo organ volumes with computed tomography (CT). The determination of abdominal organ volumes in particular has significant potential clinical value. For example, liver volumes are important not only in determining disease states and disease progression but also in estimating segmental liver volumes [8]. In their study Barutcu et al., [3] on liver volume index (measured by MRI) they found significant correlation between child-pugh grades and liver volume indices (measured on MRI). These indices are not true volumes of organs, rather they were values determined a standard way for comparison. With values of (p=0.0001), and ascites (p=0.009) they are strongly correlated with Child-Pugh grades. With CT, only ascites was correlated with Child- Pugh grades (p=0.002). This study indicated that the radiological scoring system described based on liver volume index, splenomegaly, ascites on CT are good indicator of clinical severity of hepatic cirrhosis secondary to hepatitis. The best values that are matching clinically and radiologically in their study are splenomegaly and ascites. In another study, Ito et al., [10] also found that in cirrhotic patients splenomegaly may develop in early stages of cirrhosis and is actually a better indicator of portal hypertension than cirrhosis. He also found out that there is positive correlation between spleen volume index and clinical severity of cirrhosis. CT is routinely used in the evaluating patients with chronic liver disease as it is fast and preferred in most patients with complications of liver disease. It also correlated well with functional liver capacity in cases of cirrhosis secondary to viral hepatitis [3]. We designed our scoring system based on Barutcu et al., [3] (Table 1). Using CT, In addition to this we measured separately the modified caudate lobe ratio in our study which was assessed by Awaya et al., [8] (Table 5). This was because we were hoping to use a practical single liver measurement that can be used daily with the large number of cases cirrhosis instead of complicated liver volume index calculations. As the caudate lobe ratio and the modified one were used by many researchers [8,12,14] and proven to be valuable ratios, we decided to study it. Unfortunately in our patients with viral-induced cirrhosis we found no relation between the modified caudate lobe ratio (CRL-r) and the clinical since rs=0.293 and p>0.05 (p=0.73) (Table 5). This was agreed the study of Okazaki et al., [13] who found that enlargement of the caudate lobe was more frequently seen in patients with alcoholic cirrhosis compared with viral-induced cirrhosis and it was also reported by Dodd et al., [15] that caudate lobe was the most frequently seen in the region of hypertrophy in patients with end-stage cirrhosis in primary sclerosing cholangitis. Based on our scoring system, we found that out of our 26 patients only 2 cases in class B radiologically fall in the Child A group and 4 cases of class A radiologically fall in Child B. The other 20 cases out of 26 cases (78%) matched well clinically to radiological stage (Tables 3,4). Using the Kappa method we found there was substantial agreement between clinical and radiology since the measure of agreement kappa is equal 0.61 and p<0.0001. Agreement was 78% in group A and 71% group B (which includes 53% of our patients and matching patients were 50% of the overall matching patients). Although the best matching was in group C however this is not strong as the number of these patients in this group were only 3 (100% but n=3/3) C which was the smallest number group. Better values are seen in group A (78% n=7/9) and less in group B (71% n=10/14 of this group) (Tables 2,3 and Fig. 3). The main limitation of our study was the small size of study sample (n=26), (although more than Barutcu et al., n=23), and the very low number of group C (n=3) but in the same time the largest group of our study was Child B which was with matching of 71% of the group (n=10/14) and group A with matched cases in 38% (n=10/26) of the total patients and in C group matched cases were 27% of the total number of patients overall radiological-clinical matching of 77% (n=20/26) (Table 3). The small sample size was because we wanted only to include patients which were evaluated by

194 Value of Volumetric & Morphological Parameters both CT and Child-Pugh classification done on the same week to avoid distortion of values by progression of morphological changes if the time gap between evaluations was longer and also we excluded all the patients that had shrunken liver or focal hepatic lesions on CT examinations as those can negative impact on the standardization of measurement. Conclusion: CT scoring based on Liver volume index wellcorrelates with the clinical status based on Child- Pugh classification in post-hepatitis cirrhotic patients in absence of global liver atrophy and focal hepatic lesions. It can be used with good accuracy rate in follow-up and assement of treatment response. There is no radiological-clinical matching as regard modified caudate lobe ratio and clinical stage of cirrhosis. Acknowledgment: We wish to thank Mr. Ameer Marzouk for his assistance in the statistical analysis, Mrs. Nezie Ocampo for helping in typing the manuscript and Dr. Abdelmajeed Alhazzani for his assistance in collecting clinical data. References 1- BRANCATELLI G., FEDERLE M.P., AMBROSINI R., LAGALLA R., CARRIERO A., MIDIRI M. and VILGRAIN V.: Cirrhosis: CT and MR imaging evaluation. European Journal of Radiology, 61: 57-69, 2007. 2- ITO K., MITCHELL D.G., HANN H.W., KIM Y., FUJITA T., OKAZAKI H, HONJO K. and MATSUNAGA N.: Viral-induced cirrhosis: Grading of severity using MR imaging. AJR: 173, September 1999. 3- SAYGILI O.B., TARHAN N.C., YILDIRIM T., SERIN E., OZER B. and AGILDERE A.M.: Value of computed tomography and magnetic resonance imaging for assessing severity of liver cirrhosis secondary to viral hepatitis. European Journal of Radiology, 54: 400-407, 2005. 4- BONEKAMP S., KAMEL I., SOLGA S. and CLARK J.: Can imaging modalities diagnose and stage hepatic fibrosis and cirrhosis accurately? Journal of Hepatology, 50: 17-35, 2009. 5- GERAGHTY E.M., BOONE J.M., MCGAHAN J.P. and JAIN K.: Normal organ volume assessment from abdominal CT Abdom Imaging, 29: 482-490, 2004. 6- ROFSKY N.M. and FLEISHAKER H.: CT and MRI of Diffuse Liver Disease Seminars in Ultrasound, CT, and MRI, Vo1. 16, No. 1 (February), pp 16-33, 1995. 7- ZHANG X.T., KANEMATS M., KATO H., ZHOU X., TAKESHI HARA T., YOKOYAMA R., FUJITA H. and HOSHI H.: Differentiation of cirrhosis by using 3D hepatic volume ratio of LTW in multi-detector row CT scans and MR imaging. International Congress Series, 1281: 1163-1168, 2005. 8- AWAYA H., MITCHELL D.G., KAMISHIMA T., HOL- LAND G., ITO K. and MATSUMOTO T.: Cirrhosis: Modified Caudate-Right Lobe Ratio1 Radiology, 224: 769-774, 2002. 9- PRASSOPOULOS P., DASKALOGIANNAKI M., RAIS- SAKI M., HATJIDAKIS A. and GOURTSOYIANNIS N.: Determination of normal splenic volume on computed tomography in relation to age, gender and body habitus. Eur. Radiol., 7: 246-248, 1997. 10- ITO K., MITCHELL D.G., HANN HWL, et al.: Progressive viral-induced cirrhosis: Serial MR imaging findings and clinical correlation. Radiology, 207: 729-35, 1998. 11- ITO K., MITCHELL D.G., HANN HWL, et al.: Viralinduced cirrhosis: Grading of severity using MR imaging. AJR, 173: 591-6, 1999. 12- ITO K., MITCHELL G.D., HANN H.W., et al.: Compensated cirrhosis due to viral hepatitis using MR imaging to predict clinical progression. AJR, 169: 801-5, 1997. 13- OKAZAKI H., ITO K., FUJITO T., KOIKE S. and MA- SUNAGA N.: Discrimination of alcoholic from virusinduced cirrhosis on MR. Imaging, AJR, 165: 1677-1681, 2001. 14- HARBIN W.P., ROBERT N.J. and FERRUCCI J.J.: Diagnosis of cirrhosis based on regional changes in hepatic morphology: A radiological and pathological analysis. Radiology, 135: 273-283, 1980. 15- DODD IIIGD, BARON R.L., OLIVER J.H. and FREDLE M.P.: Spectrum of imaging findings of liver end stage cirrhosis. Part II focal abnormalities. AJR, 173: 1185-92, 1999.