Comparison of Sonograms and Liver Histologic Findings in Patients with Chronic Hepatitis C Virus Infection

Transcription

1 Comparison of Sonograms and Liver Histologic Findings in Patients with Chronic Hepatitis C Virus Infection Rosalyn Kutcher, MD, Gail S. Smith, MD, Filiz Sen, MD, Scott F. Gelman, MD, Sumi Mitsudo, MD, Swan N. Thung, MD, John F. Reinus, MD Gray scale ultrasonographic images of the liver were correlated with histologic findings in patients with chronic hepatitis C virus infection. The gray scale patterns of 64 livers with chronic hepatitis C virus infection were categorized as normal, fatty, fibrofatty, fibrotic, or inflammatory and were graded as mild, moderate, or severe. Liver biopsy specimens also were analyzed for the presence of fat, inflammation, and fibrosis and graded similarly. No correlation was found between fatty and fibrofatty sonographic findings with any of the three histologic patterns. Correlations were found between fibrotic sonographic findings and both fibrotic and inflammatory histologic findings (r = 0.27; P = 0.03). lthough some pathologic features of liver disease were detected by ultrasonography, no useful correlation was noted between results of sonography and histologic examination. KEY WORDS: Hepatitis C virus; Liver, ultrasonography; Infection, hepatitic C. H epatitis C virus is the most common cause of chronic hepatitis in the United States. 1 In patients with chronic HCV infection, REVITIONS HCV, Hepatitis C virus; LT, lanine aminotransferase; SD, Standard deviation; ULN, Upper limit of normal Received July 31, 1997, from the Departments of Radiology (R.K., G.S.S.), Pathology (F.S., S.M.), and Medicine, Division of Gastroenterology (S.F.G., J.F.R.) The lbert Einstein College of Medicine and Montefiore Medical Center, ronx, New York; and The Lillian and Henry Stratton Hans Popper Department of Pathology, The Mount Sinai Medical Center, New York, New York. Revised manuscript accepted for publication February 9, ddress correspondence and reprint requests to Rosalyn Kutcher, MD, Department of Radiology, The lbert Einstein College of Medicine and Montefiore Medical Center, 111 East 210th Street, ronx, New York assessments of the severity of inflammatory liver injury (grade) and the degree of hepatic scarring (stage) are used to guide clinical management. 2 Gray scale ultrasonography has been reported to detect histologic features of chronic liver disease We compared retrospectively the gray scale liver images and histologic results in hepatitis C patients evaluated with sonographically guided liver biopsy to see if any correlation existed between sonographic and biopsy findings. If a positive correlation did exist, ultrasonography could be an important costeffective means of monitoring the disease in these patients. MTERILS ND METHODS Over a four year period ( ), 77 patients with chronic HCV infection and a positive anti-hcv antibody test (ELIS-2, Ortho-Diagnostics, Raritan, NJ) underwent percutaneous ultrasonographically 1998 by the merican Institute of Ultrasound in Medicine J Ultrasound Med 17: , /98/$3.50

2 322 CHRONIC HEPTITIS C VIRUS INFECTION J Ultrasound Med 17: , 1998 guided liver biopsy at our institution. Sixty-four of these patients formed our study population. Ultrasonograms, liver biopsy results, and clinical information were available for this group. ll images of the liver were acquired with cuson XP 10 and 128 units (cuson, Mountain View, C) using V4, V328, V3.5, or C366 transducers. Two or three core liver biopsies were obtained from the anterior segment of the right lobe in each patient using an 18 gauge needle (ard iopty or Magnum assembly, C.R. ard, Covington, G). Patient age, sex, serum LT level, sonographic size and texture of the liver, and liver histologic findings were reviewed. Size was determined by standard measurement of cranial to caudal length in the right midaxillary line. Sonographic images were classified simultaneously by two blinded readers (R.K., G.S.) as having normal, fatty, fibrofatty, fibrotic, or inflammatory patterns. liver was considered fatty if it demonstrated increased echogenicity, loss of portal venule walls, closely packed echoes, and decreased through-transmission. liver was considered fibrotic if it demonstrated a coarsened echotexture with a pinhead pattern but with preservation of the diaphragm. fibrofatty liver had both fibrotic and fatty features. liver was considered inflammatory if echogenic portal triads were observed in a hypoechoic liver or preservation of portal triads was noted in a background of fatty infiltration. 5,8 If any features of cirrhosis were present, including an irregular contour or nodularity, the liver disease was graded as severe (3) fibrosis. Grading was otherwise done subjectively for degree of severity of the various ultrasonographic patterns. fter the images of the liver were categorized appropriately on the basis of the sonographic pattern, the images within the given category were then graded subjectively on a spectrum as mild (1), moderate (2), or severe (3). iopsy specimens were examined blindly for the presence or absence of steatosis, inflammation, and fibrosis and were graded using the method of Desmet and coworkers. 11 Statistical comparison of sonographic patterns with corresponding histologic findings was performed using a Spearman rank correlation coefficient. Results from a subset of 35 patients with endstage renal disease or a history of end-stage renal disease and a functioning renal allograft were analyzed separately by a Wilcoxon rank sum test for differences in sonographic and histologic findings. ecause biopsy specimens from 12 of these patients were examined by two pathologists (F.S., S.N.T.) separately, the histologic results from these specimens were analyzed for correlation by an intraclass correlation coefficient. RESULTS Clinical evaluations, sonograms, and liver biopsy specimens from 64 patients (39 male, 25 female) aged 17 to 67 years (mean, 43.9 ± 11.6 SD; median, 44 years) were analyzed (Figs. 1 to 3). Liver size was 12 to 20.2 cm (mean, 15.2 cm ± 2 SD). LT values were one to 68 times the ULN (which equaled 40 units) with a mean of 2.9 times the ULN and a median of one times the ULN (54.8%). Thirty-five patients had end-stage renal disease, four had a history of alcohol abuse, two had biopsy-proved cirrhosis, and one had congestive heart failure. Two patients had ascites; Figure 1, Normal liver. Longitudinal ultrasonogram of the right lobe of the liver. Pathologic findings normal., Ultrasonogram: normal; pathologic findings: mild inflammation, mild fibrosis.

3 J Ultrasound Med 17: , 1998 KUTCHER ET L 323 none had hepatoma. No significant clinical history was present in the remaining 22 patients. Correlation was found between abnormal serum levels and histologic inflammation (r = 0.3, P < 0.01) in study patients. Ultrasonographic and histologic findings are presented in Table 1. Note that no sonographic categories were considered inflammatory. No correlation was found between the fatty or fibrofatty patterns on ultrasonography and the histologic findings of steatosis, fibrosis, or inflammation. However, a correlation was found between the fibrotic pattern on ultrasonography and the presence of fibrosis and inflammation at histologic examination (r = 0.27, P = 0.03). In addition, no correlation was noted between grade of fibrosis on ultrasonography and grade of fibrosis on histologic examination. n inverse correlation was observed between increasing liver size and fibrotic pattern on ultrasonography (r = 0.33, P < 0.01). This is expected since the liver shrinks as it becomes increasingly fibrotic or cirrhotic. similar correlation could not be made between liver size and the presence of fibrosis at histologic examination. The reason for this lack of correlation remains unclear, but it may be due to sampling error. In the subset of 35 patients with endstage renal disease, no additional correlation between sonographic findings and histologic findings was present. In the group of 12 patients in whom the biopsy specimens were examined by two pathologists separately, poor interobserver correlation was noted in grading of inflammation and fibrosis (Table 2). DISCUSSION The degree and type of liver injury in patients with acute or chronic liver disease has been difficult to evaluate using ultrasonography, although gray scale criteria for the diagnosis of hepatic inflammation, scarring, and fatty infiltration have been established. We used these criteria in our analysis of sonograms in patients with chronic HCV infection, 3 10 as described in the Materials and Methods section. The difference between fatty, fibrofatty, and fibrotic patterns often was difficult to perceive sonographically. Similarly, the severity of the patterns occurred over a continuous spectrum and was difficult to grade. The ultrasonongraphic pattern of inflammation is the most difficult to perceive, and none of the gray scale images in our series demonstrated either a hypoechoic liver or preserved portal triads. 4,9 lthough these variable gray scale patterns have been described to be typical of fat, fibrosis, and inflammation, this study suggests that these diagnoses are not entirely accurate in patients with chronic HCV infection. When a gray scale sonographic diagnosis of a fibrofatty or fatty liver was made, for example, fat or fibrosis was not necessarily present histologically, and often associated inflammatory changes were present that were not detected at all by ultrasonography. No patient had an inflammatory component on ultrasonography, although 56 patients had this clinically important finding on biopsy. Equally disturbing was the correlation between fibrosis on sonography and the presence of histologic inflammation. Fibrosis is usually a permanent change, whereas inflammation implies an active Figure 2, Ultrasonogram: Mild fatty liver, some loss of portal venule walls, but preservation of diaphragm. Pathologic findings: Moderate inflammation, mild fibrosis, no steatosis., Ultrasonogram: Moderate fatty liver, loss of portal venule walls, tightly packed echoes, and some loss of diaphragm. Pathologic findings: Mild inflammation, mild fibrosis, mild steatosis.

4 324 CHRONIC HEPTITIS C VIRUS INFECTION J Ultrasound Med 17: , 1998 Figure 3, Ultrasonogram: Mild fibrotic liver, coarsened texture with areas of pinhead echoes. Pathologic findings: Moderate inflammation, mild fibrosis, mild steatosis., Ultrasonogram: Severe fibrosis (immediately post biopsy), nodular pattern. Pathologic findings (two readers): mild to severe inflammation, no fibrosis, no steatosis. but potentially reversible process that may resolve without leaving any residual structural abnormality. Two limitations of this study are that it was a retrospective study, and prospective liver evaluation during scanning would likely allow for more accurate parenchymal analysis. Over the 4 year period of data accumulation, equipment and transducers improved, with superior imaging of the liver being possible in the more recent cases. Nevertheless, this study demonstrates that, at the present time, gray scale ultrasonography cannot be used as a method of grading or staging chronic HCV infection, and, by inference, other inflammatory liver diseases with the same histologic abnormalities. ppropriate management of HCV infection requires accurate grading and staging; chronic inflammation with only minimal fibrosis and small amounts of hepatic iron favors a good response to therapy with alpha-interferon 2. iopsy is equally important in the care of patients with other inflammatory liver diseases, many of which cannot be diagnosed accurately without examination of a liver biopsy specimen. correlation was found between elevated serum LT levels and the presence of inflammation on histologic examination. This finding is consistent with the results of other studies, although it is considered to be unreliable by hepatologists, as active disease can be found in patients with low or normal LT levels. It is widely recognized that some variation will occur in interpretation of liver biopsy findings from observer to observer. This was certainly demonstrated in the 12 biopsy specimens that were evaluated by the two pathologists in our study (Table 2). sampling error of up to 10% is accepted for liver biopsies in some diseases. 4 This fact and the findings in our study do raise some concern regarding use of pathologic findings as the gold standard. Since a large portion of our study population had end-stage renal disease, we examined this group as a subset to determine if they had any unique characteristics. Previous studies have shown that HCV infection is not made worse by hemodialyisis or renal transplantation and associated immunosuppression. 10 Findings in the subset of patients with end-stage renal disease evaluated in our study were not different from those in the rest of the study population. Slightly more fat was detected by ultrasonography and biopsy in these patients, a finding that might be related to chronic immunosuppressive therapy with prednisone. Table 1: Comparison of Sonographic Pattern with Histologic Pattern Histologic Pattern (n) Ultrasound (mild moderate severe) Pattern (n) Normal Steatosis Inflammation Fibrosis Normal (14) (5 0 0) (8 2 2) (4 3 0) Fatty (6) (2 2 0) (1 4 0) (3 1 0) Fibrofatty (16) (6 2 2) (7 6 1) (8 3 2) Fibrotic (28) (10 3 0) (9 10 6) (7 7 7)

5 J Ultrasound Med 17: , 1998 KUTCHER ET L 325 Table 2: Interobserver Variation of Histologic Interpretation in 12 Patients fter Double Reading ICC* Path (I) = 0.07 ICC Path (F) = 0.21 ICC Path (S) = 0.78 P < *Intraclass correlation coefficient. I, Inflammation; F, fibrosis; S, fat. In conclusion, although microscopic features of liver disease in patients with chronic HCV infection were detected by ultrasonography, no consistent correlation between sonographic and histologic findings was found, suggesting that ultrasonographic imaging is unreliable for grading and staging the degree of liver damage in chronic HCV infection. REFERENCES Very poor Not correlative Very high 3. Lindnor KD, ru C, Jorgensen R, et al: The role of ultrasonography and automatic needle biopsy in outpatient percutaneous liver biopsy. Hepatology 23:1079, Joseph E: Diffuse liver disease. Clin Diagn Ultrasound 29:1, Taylor KJW, Riely C, Hammers L, et al: Quantitative US attenuation in normal liver and in patients with diffuse liver disease: Importance of fat. Radiology 160:65, Kester NL, LaFortune M, ubin, et al: Focal sparing of liver parenchyma. J Ultrasound Med 15:89, Needleman L, Kurtz, Rifkin MD, et al: Sonography of diffuse benign liver disease. JR 146:1011, Taylor KJW, Gorelick FS, Rosenfield T, et al: Ultrasonography of alcoholic liver disease with histological correlation. Radiology 141:157, Zwiebel WJ: Sonographic diagnosis of diffuse liver disease. Semin Ultrasound CT MRI 16:8, Rosen HR, Friedman LS, Martin P: Hepatitis C and the renal transplant patient. Semin Dialysis 9:39, Desmet VJ, Gerber M, Hoofnagle JH, et al: Classification of chronic hepatitis: Diagnosis, grading and staging. Hepatology 19:1513, McQuillan GM, lter MJ, Everhart JE: Viral hepatitis. In Everhart JE (Ed): Digestive Diseases in the United States: Epidemiology and Impact. US Department of Health and Human Services, Public Health Service, National Institutes of Health, National Institute of Diabetes, and Digestive and Kidney Diseases. Washington, DC, US Government Printing Office, 1994; NIH publication no p National Institutes of Health Consensus Development Conference Statement: Management of hepatitis C. US Department of Health and Human Services, Public Health Service, National Institutes of Health, National Institute of Diabetes, and Digestive and Kidney Diseases. Washington, DC, US Government Printing Office, 1997 (in press)