- Email: [email protected]
Diffusion-weighted MRI of the transplanted liver
Clinical Radiology 66 (2011) 820e825
Contents lists available at ScienceDirect
Clinical Radiology journal homepage: www.elsevierhealth.com/journals/crad
Original Paper
Diffusion-weighted MRI of the transplanted liver K. Sandrasegaran a, *, R. Ramaswamy a, S. Ghosh b, B. Tahir a, F.M. Akisik a, R. Saxena c, P. Kwo d a
Department of Radiology, Indiana University School of Medicine, Indianapolis, IN, USA Department of Mathematics, Indiana University, IN, USA c Department of Pathology, Indiana University School of Medicine, Indianapolis, IN, USA d Department of Medicine, Indiana University School of Medicine, Indianapolis, IN, USA b
article in formation Article history: Received 16 June 2010 Received in revised form 2 December 2010 Accepted 14 December 2010
AIM: To assess the value of diffusion-weighted imaging (DWI) in evaluating parenchymal disorders following orthotopic liver transplantation (OLT). MATERIALS AND METHODS: This institutional review board-approved, retrospective study measured the hepatic apparent diffusion coefficients (ADC) in patients following OLT. Those with vascular complications or within 3 months of OLT were excluded. A single-shot echoplanar sequence with b values of 50, 400 (or 500), and 800 s/mm2 was performed. Liver biopsy specimens [performed with a median of 17 days after magnetic resonance imaging (MRI)] were recorded for the presence and severity of parenchymal disorders, such as acute cellular rejection, and recurrence of fibrosis in all patients, and the recurrence of viral hepatitis in patients with hepatitis C. ADC values were measured blinded to histology in 41 patients (33 males) who had 56 MRI scans. RESULTS: There was a significant difference in ADC values associated with a histological abnormality seen on core biopsy [n ¼ 43, mean (SD) ADC of 0.91 (0.15)103 mm2/s] and those associated with no histological abnormality [n ¼ 13, mean (SD) ADC of 1.11 (0.17)103 mm2/s; (p ¼ 0.003)]. ADC values did not predict any of the individual parenchymal disorders on logistic regression analysis. When the ADC value was <0.99103 mm2/s, there was a sensitivity and specificity of 85% and 72%, respectively, in predicting a parenchymal disorder (area under ROC curve ¼ 0.84; 95% CI 0.72 to 0.92). CONCLUSION: ADC measurements may help in deciding which patients require core liver biopsy after OLT. However, ADC values are not likely to be reliable in differentiating between the various parenchymal disorders. Ó 2011 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.
Introduction Orthotopic liver transplantation (OLT) is a well-accepted technique for the treatment of end-stage chronic liver disease and some cases of acute liver failure. Current 1 year survival rates are 85e90%, and 10-year survival rates are * Guarantor and correspondent: K. Sandrasegaran, 550 N University Blvd, UH 0279, Indianapolis, IN 46202, USA. Tel.: þ1 317 2741837; fax: þ1 317 2741848. E-mail address: [email protected] (K. Sandrasegaran).
50e65%.1,2 Imaging studies have been used to assess vascular and biliary complications of OLT.3e5 Another important group of post-transplant disorders may be termed “hepatic parenchymal disorders”. This group includes acute cellular rejection, chronic ductopenic rejection, recurrent viral hepatitis, and fibrosis. Conventional imaging studies are not considered to be reliable for diagnosing these disorders, and core liver biopsies are used for diagnosis. However, liver biopsy has a small (1e5%) but definite rate of complications, and is subject to sampling variability related to the heterogeneity of the distribution of liver fibrosis.6 Patients with viral hepatitis
0009-9260/$ e see front matter Ó 2011 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.crad.2010.12.020
K. Sandrasegaran et al. / Clinical Radiology 66 (2011) 820e825
often undergo repeated liver biopsies as the correlation between liver biochemical tests and histological findings may be unreliable,7 and because serial liver biopsies may be necessary to reliably differentiate acute rejection from recurrent viral hepatitis.8e10 Therefore, non-invasive methods for diagnosing such parenchymal disorders would be useful. The present study was undertaken to assess the value of diffusion-weighted imaging (DWI) in evaluating parenchymal disorders following OLT.
Materials and methods Patients For this retrospective Health Insurance Portability and Accountability Act (HIPAA)-compliant study, the magnetic resonance imaging (MRI) database was searched for DWI examinations performed between October 2006 and March 2009. Institutional review board permission was obtained for retrospective assessment of imaging and clinical data with a waiver of informed consent. Ninety-four patients were found who had at least one DWI study after undergoing OLT. Patients were excluded from analysis if there was clinical or radiological evidence of vascular or biliary abnormalities (n ¼ 7), if core liver biopsy was not performed within 3 months of the MRI examination (n ¼ 42), or if the DWI contained artefacts that precluded apparent diffusion coefficients (ADC) measurements (n ¼ 4). The remaining 41 patients, 33 males and eight females, were included in the study. The median age was 54.5 years (range 33e77 years). Fifty-six DWI studies were analysed in the study patients. Each imaging study was correlated with a separate liver biopsy. Twenty-seven patients had hepatitis C-related cirrhosis, prior to OLT. Thirty-one patients had hepatocellular cancer (HCC) in the explanted liver, and most MRI studies (n ¼ 40) were performed as a screening test to exclude the recurrence of HCC in the transplanted liver. No HCC were found on the MRI studies evaluated. The indications of the remaining MRI studies were determination of vascular patency (n ¼ 3) or assessment of recurrent cirrhosis (n ¼ 13). As per exclusion criteria, there was no evidence of main or branch portal vein thrombosis on the MRI studies.
MRI examination MRI examinations were performed using a 1.5 T 18channel MRI machine (Magnetom Avanto, Siemens, Erlangen, Germany) with high-performance gradients (maximum gradient 45 mT/m, maximum slew rate of 200 T/m/s). Transverse echoplanar imaging was performed using six-element phased-array coil. Frequency-selective fat saturation was used to reduce chemical shift artefacts. The parameters used were repetition time (TR)/echo time (TE) 1400e1500 ms/70 ms; field of view 36e40 cm; matrix 192115; number of excitations 2; section width/gap, 6 mm/ 7.8 mm; receiver bandwidth 1736 Hz/pixel. Parallel imaging was performed using generalized autocalibrating partially parallel acquisitions (GRAPPA) with an acceleration factor (r) of 2 as well as 75% partial phase Fourier and 80% phase
821
resolution to reduce the TE and minimize artefacts. Respiratory-triggering (navigator correction) was undertaken and the typical acquisition time was 4e6 min. B values of 50, 400 (or 500), and 800 s/mm2 were used. A single ADC was calculated mono-exponentially using the three b values images, on a voxel-by-voxel basis, by the scanner software (Syngo vB13A or vB15A, Siemens Medical Solutions). A fellowship-trained abdominal radiologist with 12-years experience of reading MRI (K.S.), blinded to histology, placed a region-of-interest (ROI) in each lobe of the liver. The ROI were between 2 to 5 cm2 in size. The central vascular region, potential masses, and peripheral 2 cm of the liver were avoided. All patients had routine MRI including postgadolinium sequences. These were used for clarification of liver outline and detection of focal liver lesions.
Histopathological analysis Core liver biopsies were performed at the request of hepatologists, when there was clinical or biochemical suspicion of a parenchymal liver disorder. The core liver biopsies were graded for fat infiltration (none, present, i.e., >10%), acute cellular rejection (none, mild, and severe), recurrence of viral hepatitis (none, mild, and severe), chronic ductopenic rejection (none, mild, and severe), and fibrosis, using established criteria11,12 by a specialist hepatic histopathologist. Fibrosis was scored using the METAVIR system,13 which classifies fibrosis into five stages: F0, no fibrosis; F1, portal fibrosis without septa; F2, few septa; F3, bridging fibrosis without cirrhosis; F4, cirrhosis with architectural distortion. The equivalent METAVIR stage of fibrosis was recorded in patients in whom the cause for chronic liver disease was not hepatitis C.
Statistical analysis The agreement between the left and right lobe ADC values was analysed using intraclass correlation coefficients. Logistic regression analysis was performed to evaluate the predictive association of ADC values on the histological features, such as fibrosis and presence of viral hepatitis. Within the framework of the logistic regression method, the likelihood ratio test was used to assess the significance of predictors. Odds ratios and 95% confidence intervals were determined. P < 0.05 was considered to indicate a significant difference. Receiver operating characteristic (ROC) curves were used to evaluate the sensitivity and specify of ADC measurements for the presence and severity of parenchymal disorders. Statistical analysis was performed using SAS v9.2 software (SAS, Cary, NC, USA) and MedCalc v 10.3 (MedCalc, Mariakerke, Belgium).
Results DWI and histopathology examinations Of the 41 patients, two separate DWI studies were performed in five patients and three studies in another five
822
K. Sandrasegaran et al. / Clinical Radiology 66 (2011) 820e825
patients, resulting in a total of 56 DWI studies analysed. The median time interval between the DWI study and biopsy was 17 days (range 0e2.8 months). Forty-two of the MRI studies were performed within 1 month of liver biopsy, and 14 studies were performed on the same day as biopsy. In 27 cases, biopsy was performed before MRI, and in 29 cases biopsy was performed after MRI. Each DWI study was associated with a separate histological examination. The presence of acute rejection, chronic rejection, recurrent viral hepatitis, fibrosis, and steatosis were seen, respectively, in 11, 10, 26, 26, and seven liver biopsy samples, with some samples showing more than one parenchymal disorder. Forty-three DWI studies were associated with histological examinations showing one or more abnormalities, and 13 studies were associated with negative histological findings. The median time interval between OLT and the DWI study was 12.3 months (range 3.2e195 months).
ADC values The intraclass correlation of the ADC values of the right and left lobes was 0.51, corresponding to a moderate correlation. Fig 1 is a box-and-whisker plot of the ADC values according to parenchymal disorders. There was a significantly lower ADC associated with at least one positive histological finding [n ¼ 43, mean (SD) ADC ¼ 0.91 (0.15)103 mm2/s] compared to those associated a negative histology [n ¼ 13, mean (SD) ADC ¼ 1.11 (0.17)103 mm2/s, p ¼ 0.003]. The results of logistic regression analysis are given in Table 1. This table gives the significance values of each of the histological parameters on logistic regression analysis. In addition, the odds ratios and their 95% confidence intervals are presented. The predictive effects of ADC on the histological parameters are given. None of the individual histological parameters could be independently predicted by ADC. This is evident from the p-values of more than 0.05
Figure 1 Box-and-whisker plot of ADC values. The boxes represent the interquartile range. Whiskers represent range of values. Horizontal line within box is the median apparent diffusion coefficient (ADC) value. The small square represents outliers. y-axis gives ADC values in 103 mm2/s. None, no parenchymal disorder; AR, acute rejection; CR, chronic rejection; Viral, recurrent viral hepatitis; Fib, recurrence of fibrosis; Fat, fat infiltration.
Table 1 Results of logistic regression analysis. Histological parameter
p-Value
Odds ratio (95% CI)
Viral hepatitis Fat Fibrosis Acute rejection Chronic rejection Inflammationa Diseaseb
0.28 0.08 0.11 0.64 0.88 0.06 0.003
0.17 0.01 0.06 0.40 1.34 0.03 0.001
(0.01e4.14) (0.001e1.63) (0.002e1.82) (0.01e18.71) (0.03e68.85) (0.001e1.14) (0.001e0.029)
a Inflammation was a composite parameter made of cases that either had recurrence of viral hepatitis or acute rejection. Those without either condition were taken as not showing inflammation. b Disease is a composite parameter that was positive if any of the parameters was present and negative if there was no evidence of any of the histological parameters.
and the fact that the 95% confidence interval of the odds ratio includes 1.0. ADC values were able to predict the presence or absence of a parenchymal disorder. ADC values showed a trend towards significance in differentiating the presence of inflammation, i.e., acute rejection or recurrent hepatitis (n ¼ 34), from the absence of inflammation [n ¼ 22; mean (SD) ADC 0.92 (0.15) 103 mm2/s versus 1.02 (0.20)103 mm2/s, p ¼ 0.06; Fig 2]. The ADC values associated with more severe viral inflammation (n ¼ 6) were not significantly different with those associated with mild viral hepatitis [n ¼ 20; mean (SD) ADC of 0.88 (0.20)103 mm2/s versus 0.95 (0.15)103 mm2/s, p ¼ 0.62].
ROC curve analysis The ROC curve plotting the sensitivity and specificity of ADC values in determining the presence of a parenchymal disorder is shown in Fig 3. The area-under-curve was 0.84 (95% CI 0.72 to 0.92). ADC values less than 0.99103 mm2/s indicated the presence of a parenchymal disorder with a sensitivity of 85% (95% CI of 71% to 95%) and specificity of 72% (95% CI of 56% to 85%). Twenty-seven patients in the cohort had hepatitis C and underwent 41 biopsies. This subgroup comprised 24 male and three female patients with a mean age of 54.1 years (range 34e77 years). In patients with hepatitis C, reinfection of the allograft is almost inevitable,14 but the severity and time of onset vary. Antiviral therapy is usually started in patients with severe recurrent hepatitis (necro-inflammatory score of 3 or 4) or with recurrent fibrosis (stage 2 and above).15,16 The ROC curve plotting the sensitivity and specificity of ADC values in determining the subgroup of patients with hepatitis C who would benefit from antiviral therapy is shown in Fig 4. The area-under-curve was 0.68 (95% CI 0.51 to 0.83). An ADC value of less than 0.98103 mm2/s was 78% sensitive and 58% specific in determining this subgroup of patients who should be started on antiviral therapy.
Discussion Post-transplant hepatic parenchymal disorders often present with non-specific symptoms and variable elevation
K. Sandrasegaran et al. / Clinical Radiology 66 (2011) 820e825
823
Figure 3 ROC curve plotting sensitivity (y-axis) and 1-specificity (x-axis) of ADC values in determining the presence or absence of a parenchymal disorder. The area-under-curve (Az) was 0.84 (95% CI 0.72 to 0.92). The diagonal line represents Az of 0.50.
Figure 2 A 67-year-old man with hepatitis C and hepatocellular cancer (HCC) who underwent orthotopic liver transplantation (OLT). Axial diffusion-weighted images (b ¼ 500 s/mm2) performed (a) 6 months and (b) 15 months later are shown. The MRI studies, performed as routine screening for recurrent HCC, did not show morphological abnormalities. The signal intensities of the liver in the two studies appear similar. However, the left lobe ADC values at 6 and 15 month studies were 1.30 and 0.97103 mm2/s, respectively. Liver biopsies showed no recurrent viral hepatitis at 7 months, and presence of moderately severe viral hepatitis at 15 months, after OLT. No other parenchymal disorders were seen on the biopsies. In this patient, there was a reduction of ADC coinciding with recurrent viral hepatitis. We did not have sufficient patients with multiple MRI studies to determine the value of serial ADC measurements.
in liver function tests. Diagnosis usually requires a core liver biopsy, sometimes performed serially. Acute rejection is usually seen in the first month following OLT. Rejection occurring after 4 weeks is seen in 15e20% of patients and is less responsive to immunosuppressive therapy, partly because of delay in diagnosis.17 Thus, reliable, non-invasive techniques for diagnosing parenchymal disorders are useful. The authors are not aware of a prior publication assessing the usefulness of DWI in determining parenchymal disorders after OLT. In general, a significantly lower ADC was found in patients with at least one parenchymal disorder following OLT
(p ¼ 0.003). It may be possible to explain some of these findings based on the pathophysiology of these conditions. For instance, acute rejection is characterized by a mixed, predominantly mononuclear, infiltrate containing lymphocytes, neutrophils, and eosinophils in the portal tracts, with endothelitis or perivenulitis.18 It would be expected that the cellular infiltrate would restrict diffusion by increasing the intracellular volume per unit tissue volume. Lobular inflammation, hepatocytes ballooning, and apoptotic cytoplasmic bodies are seen in recurrent viral hepatitis,19 and would also be expected to reduce diffusion. However, severe acute rejection or hepatitis may result in centrilobular necrosis,
Figure 4 ROC curve plotting sensitivity (y-axis) and 1-specificity (x-axis) of ADC values in determining the subgroup of patients with hepatitis C who would benefit from antiviral therapy (see text). The area-under-curve (Az) was 0.68 (95% CI 0.51 to 0.83). The diagonal line represents Az of 0.50.
824
K. Sandrasegaran et al. / Clinical Radiology 66 (2011) 820e825
which may potentially increase ADC as a result of cell membrane lysis. These opposing effects may be a reason why there was no clear difference in ADC between mild and severe acute rejection. Nevertheless, there may be a role for serial ADC values in monitoring the progression of acute rejection or recurrent viral hepatitis. In the present cohort, the number of patients with more than two DWI tests was small and the value of serial DWI measurements could not be reliably assessed. Chronic rejection occurs in 5e10% of patients undergoing OLT. Histologically, chronic rejection is characterized by atrophy and loss of interlobular and septal bile ducts, foam cell arteriopathy, perivenular or bridging fibrosis, and a relative absence of inflammatory infiltrates.12 It is possible that fibrosis in this condition would lead to a reduction in ADC. Recurrence of fibrosis and cirrhosis occurs in 15e30% of patients 5 years after OLT, regardless of the original disease causing liver failure.20,21 Deposition of extracellular collagen is likely to reduce diffusion. Several studies have confirmed a reduction in ADC in cirrhotic, non-transplanted, livers.22e24 In macrovesicular hepatic steatosis, there is intracytoplasmic accumulation of fat-containing vacuoles, leading to ballooning of hepatocytes.25 As a result, ADC values may reduce in the presence of substantial steatosis. Given that most of the parenchymal disorders result in a reduction in diffusion, it is not surprising that no differences were found in the ADC values for these entities. A finding of this study was the assessment of patients with hepatitis C. Antiviral therapy is usually started when there is with severe recurrent hepatitis or with substantial fibrosis.15,16 ADC values were reasonably sensitive (78%) in determining the subgroup of patients who should be started on antiviral therapy. However, the specificity is low (58%), and ADC values cannot be used on the only criterion for determining the need for antiviral therapy. There were some limitations of the present study. The study was retrospective and hence the time interval between MRI and histological confirmation of liver fibrosis varied with a median interval of 17 days and a maximum interval of 2.8 months. Most biopsies (n ¼ 42) were performed within 1 month of the MRI study. Despite this relatively small time interval, it is possible that medical therapy for hepatitis or rejection may have led to differences in the severity of disease between the time of the MRI and that of the biopsy. Liver biopsy was used as the reference standard. Biopsy results are prone to interobserver variability in interpretation and may yield erroneous results if the parenchymal disease has a heterogeneous distribution. The technique of DWI is not standardized, and the present findings may not necessarily be comparable with studies performed with other DWI techniques, e.g. breath-hold. In conclusion, ADC values may be used to distinguish between those who have parenchymal disorders after OLT and those who do not. They may help to decide which patients with hepatitis C should be started on antiviral therapy after OLT. In this respect, DWI may be a useful screening tool in determining patients who require liver biopsy after OLT. In general, most parenchymal disorders resulted in the lowering of the ADC values. It was not
possible to differentiate between these disorders using ADC values, and DWI is unlikely to replace core biopsy for the assessment of parenchymal disorders following OLT.
Supplementary data Supplementary data associated with this article can be found in online version at doi:10.1016/j.crad.2010.12.020.
References 1. Wiesner RH, Menon KV. Late hepatic allograft dysfunction. Liver Transpl 2001;7:S60e73. 2. Charlton M, Ruppert K, Belle SH, et al. Long-term results and modeling to predict outcomes in recipients with HCV infection: results of the NIDDK liver transplantation database. Liver Transpl 2004;10:1120e30. 3. Caiado AH, Blasbalg R, Marcelino AS, et al. Complications of liver transplantation: multimodality imaging approach. RadioGraphics 2007;27:1401e17. 4. Chong WK. Ultrasound evaluation of liver transplants. Abdom Imaging 2004;29:180e8. 5. Girometti R, Cereser L, Como G, et al. Biliary complications after orthotopic liver transplantation: MRCP findings. Abdom Imaging 2008;33:542e54. 6. Afdhal NH, Nunes D. Evaluation of liver fibrosis: a concise review. Am J Gastroenterol 2004;99:1160e74. 7. Wiesner RH, Rakela J, Ishitani MB, et al. Recent advances in liver transplantation. Mayo Clin Proc 2003;78:197e210. 8. Regev A, Molina E, Moura R, et al. Reliability of histopathologic assessment for the differentiation of recurrent hepatitis C from acute rejection after liver transplantation. Liver Transpl 2004;10:1233e9. 9. Petrovic LM, Villamil FG, Vierling JM, et al. Comparison of histopathology in acute allograft rejection and recurrent hepatitis C infection after liver transplantation. Liver Transpl Surg 1997;3:398e406. 10. Younossi ZM, Boparai N, Gramlich T, et al. Agreement in pathologic interpretation of liver biopsy specimens in posttransplant hepatitis C infection. Arch Pathol Lab Med 1999;123:143e5. 11. Demetris AJ, Batts KP, Dhillon AP, et al. Banff schema for grading liver allograft rejection: an international consensus document. Hepatology 1997;25:658e63. 12. Demetris A, Adams D, Bellamy C, et al. Update of the International Banff Schema for Liver Allograft Rejection: working recommendations for the histopathologic staging and reporting of chronic rejection. An international panel. Hepatology 2000;31:792e9. 13. The French METAVIR Cooperative Study Group. Intraobserver and interobserver variations in liver biopsy interpretation in patients with chronic hepatitis C. Hepatology 1994;20:15e20. 14. Wright TL, Donegan E, Hsu HH, et al. Recurrent and acquired hepatitis C viral infection in liver transplant recipients. Gastroenterology 1992;103:317e22. 15. Terrault NA. Hepatitis C therapy before and after liver transplantation. Liver Transpl 2008;14(Suppl. 2):S58e66. 16. Wiesner RH, Sorrell M, Villamil F. Report of the first International Liver Transplantation Society expert panel consensus conference on liver transplantation and hepatitis C. Liver Transpl 2003;9:S1e9. 17. Anand AC, Hubscher SG, Gunson BK, et al. Timing, significance, and prognosis of late acute liver allograft rejection. Transplantation 1995;60:1098e103. 18. Snover DC, Sibley RK, Freese DK, et al. Orthotopic liver transplantation: a pathological study of 63 serial liver biopsies from 17 patients with special reference to the diagnostic features and natural history of rejection. Hepatology 1984;4:1212e22. 19. Ziarkiewicz-Wroblewska B, Gornicka B, Oldakowska-Jedynak U, et al. Morphologic features of hepatitis C recurrence in patients after orthotopic liver transplantationdpreliminary analysis of our case observations. Transplant Proc 2006;38:226e30. 20. Davern TJ, Lake JR. Recurrent disease after liver transplantation. Semin Gastrointest Dis 1998;9:86e109.
K. Sandrasegaran et al. / Clinical Radiology 66 (2011) 820e825 21. Feray C, Caccamo L, Alexander GJ, et al. European collaborative study on factors influencing outcome after liver transplantation for hepatitis C. European Concerted Action on Viral Hepatitis (EUROHEP) Group. Gastroenterology 1999;117:619e25. 22. Lewin M, Poujol-Robert A, Boelle PY, et al. Diffusion-weighted magnetic resonance imaging for the assessment of fibrosis in chronic hepatitis C. Hepatology 2007;46:658e65.
825
23. Taouli B, Tolia AJ, Losada M, et al. Diffusion-weighted MRI for quantification of liver fibrosis: preliminary experience. AJR Am J Roentgenol 2007;189:799e806. 24. Sandrasegaran K, Akisik FM, Lin C, et al. Value of diffusion-weighted MRI for assessing liver fibrosis and cirrhosis. AJR Am J Roentgenol 2009;193:1556e60. 25. Lall CG, Aisen AM, Bansal N, et al. Nonalcoholic fatty liver disease. AJR Am J Roentgenol 2008;190:993e1002.
Contents lists available at ScienceDirect
Clinical Radiology journal homepage: www.elsevierhealth.com/journals/crad
Original Paper
Diffusion-weighted MRI of the transplanted liver K. Sandrasegaran a, *, R. Ramaswamy a, S. Ghosh b, B. Tahir a, F.M. Akisik a, R. Saxena c, P. Kwo d a
Department of Radiology, Indiana University School of Medicine, Indianapolis, IN, USA Department of Mathematics, Indiana University, IN, USA c Department of Pathology, Indiana University School of Medicine, Indianapolis, IN, USA d Department of Medicine, Indiana University School of Medicine, Indianapolis, IN, USA b
article in formation Article history: Received 16 June 2010 Received in revised form 2 December 2010 Accepted 14 December 2010
AIM: To assess the value of diffusion-weighted imaging (DWI) in evaluating parenchymal disorders following orthotopic liver transplantation (OLT). MATERIALS AND METHODS: This institutional review board-approved, retrospective study measured the hepatic apparent diffusion coefficients (ADC) in patients following OLT. Those with vascular complications or within 3 months of OLT were excluded. A single-shot echoplanar sequence with b values of 50, 400 (or 500), and 800 s/mm2 was performed. Liver biopsy specimens [performed with a median of 17 days after magnetic resonance imaging (MRI)] were recorded for the presence and severity of parenchymal disorders, such as acute cellular rejection, and recurrence of fibrosis in all patients, and the recurrence of viral hepatitis in patients with hepatitis C. ADC values were measured blinded to histology in 41 patients (33 males) who had 56 MRI scans. RESULTS: There was a significant difference in ADC values associated with a histological abnormality seen on core biopsy [n ¼ 43, mean (SD) ADC of 0.91 (0.15)103 mm2/s] and those associated with no histological abnormality [n ¼ 13, mean (SD) ADC of 1.11 (0.17)103 mm2/s; (p ¼ 0.003)]. ADC values did not predict any of the individual parenchymal disorders on logistic regression analysis. When the ADC value was <0.99103 mm2/s, there was a sensitivity and specificity of 85% and 72%, respectively, in predicting a parenchymal disorder (area under ROC curve ¼ 0.84; 95% CI 0.72 to 0.92). CONCLUSION: ADC measurements may help in deciding which patients require core liver biopsy after OLT. However, ADC values are not likely to be reliable in differentiating between the various parenchymal disorders. Ó 2011 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.
Introduction Orthotopic liver transplantation (OLT) is a well-accepted technique for the treatment of end-stage chronic liver disease and some cases of acute liver failure. Current 1 year survival rates are 85e90%, and 10-year survival rates are * Guarantor and correspondent: K. Sandrasegaran, 550 N University Blvd, UH 0279, Indianapolis, IN 46202, USA. Tel.: þ1 317 2741837; fax: þ1 317 2741848. E-mail address: [email protected] (K. Sandrasegaran).
50e65%.1,2 Imaging studies have been used to assess vascular and biliary complications of OLT.3e5 Another important group of post-transplant disorders may be termed “hepatic parenchymal disorders”. This group includes acute cellular rejection, chronic ductopenic rejection, recurrent viral hepatitis, and fibrosis. Conventional imaging studies are not considered to be reliable for diagnosing these disorders, and core liver biopsies are used for diagnosis. However, liver biopsy has a small (1e5%) but definite rate of complications, and is subject to sampling variability related to the heterogeneity of the distribution of liver fibrosis.6 Patients with viral hepatitis
0009-9260/$ e see front matter Ó 2011 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.crad.2010.12.020
K. Sandrasegaran et al. / Clinical Radiology 66 (2011) 820e825
often undergo repeated liver biopsies as the correlation between liver biochemical tests and histological findings may be unreliable,7 and because serial liver biopsies may be necessary to reliably differentiate acute rejection from recurrent viral hepatitis.8e10 Therefore, non-invasive methods for diagnosing such parenchymal disorders would be useful. The present study was undertaken to assess the value of diffusion-weighted imaging (DWI) in evaluating parenchymal disorders following OLT.
Materials and methods Patients For this retrospective Health Insurance Portability and Accountability Act (HIPAA)-compliant study, the magnetic resonance imaging (MRI) database was searched for DWI examinations performed between October 2006 and March 2009. Institutional review board permission was obtained for retrospective assessment of imaging and clinical data with a waiver of informed consent. Ninety-four patients were found who had at least one DWI study after undergoing OLT. Patients were excluded from analysis if there was clinical or radiological evidence of vascular or biliary abnormalities (n ¼ 7), if core liver biopsy was not performed within 3 months of the MRI examination (n ¼ 42), or if the DWI contained artefacts that precluded apparent diffusion coefficients (ADC) measurements (n ¼ 4). The remaining 41 patients, 33 males and eight females, were included in the study. The median age was 54.5 years (range 33e77 years). Fifty-six DWI studies were analysed in the study patients. Each imaging study was correlated with a separate liver biopsy. Twenty-seven patients had hepatitis C-related cirrhosis, prior to OLT. Thirty-one patients had hepatocellular cancer (HCC) in the explanted liver, and most MRI studies (n ¼ 40) were performed as a screening test to exclude the recurrence of HCC in the transplanted liver. No HCC were found on the MRI studies evaluated. The indications of the remaining MRI studies were determination of vascular patency (n ¼ 3) or assessment of recurrent cirrhosis (n ¼ 13). As per exclusion criteria, there was no evidence of main or branch portal vein thrombosis on the MRI studies.
MRI examination MRI examinations were performed using a 1.5 T 18channel MRI machine (Magnetom Avanto, Siemens, Erlangen, Germany) with high-performance gradients (maximum gradient 45 mT/m, maximum slew rate of 200 T/m/s). Transverse echoplanar imaging was performed using six-element phased-array coil. Frequency-selective fat saturation was used to reduce chemical shift artefacts. The parameters used were repetition time (TR)/echo time (TE) 1400e1500 ms/70 ms; field of view 36e40 cm; matrix 192115; number of excitations 2; section width/gap, 6 mm/ 7.8 mm; receiver bandwidth 1736 Hz/pixel. Parallel imaging was performed using generalized autocalibrating partially parallel acquisitions (GRAPPA) with an acceleration factor (r) of 2 as well as 75% partial phase Fourier and 80% phase
821
resolution to reduce the TE and minimize artefacts. Respiratory-triggering (navigator correction) was undertaken and the typical acquisition time was 4e6 min. B values of 50, 400 (or 500), and 800 s/mm2 were used. A single ADC was calculated mono-exponentially using the three b values images, on a voxel-by-voxel basis, by the scanner software (Syngo vB13A or vB15A, Siemens Medical Solutions). A fellowship-trained abdominal radiologist with 12-years experience of reading MRI (K.S.), blinded to histology, placed a region-of-interest (ROI) in each lobe of the liver. The ROI were between 2 to 5 cm2 in size. The central vascular region, potential masses, and peripheral 2 cm of the liver were avoided. All patients had routine MRI including postgadolinium sequences. These were used for clarification of liver outline and detection of focal liver lesions.
Histopathological analysis Core liver biopsies were performed at the request of hepatologists, when there was clinical or biochemical suspicion of a parenchymal liver disorder. The core liver biopsies were graded for fat infiltration (none, present, i.e., >10%), acute cellular rejection (none, mild, and severe), recurrence of viral hepatitis (none, mild, and severe), chronic ductopenic rejection (none, mild, and severe), and fibrosis, using established criteria11,12 by a specialist hepatic histopathologist. Fibrosis was scored using the METAVIR system,13 which classifies fibrosis into five stages: F0, no fibrosis; F1, portal fibrosis without septa; F2, few septa; F3, bridging fibrosis without cirrhosis; F4, cirrhosis with architectural distortion. The equivalent METAVIR stage of fibrosis was recorded in patients in whom the cause for chronic liver disease was not hepatitis C.
Statistical analysis The agreement between the left and right lobe ADC values was analysed using intraclass correlation coefficients. Logistic regression analysis was performed to evaluate the predictive association of ADC values on the histological features, such as fibrosis and presence of viral hepatitis. Within the framework of the logistic regression method, the likelihood ratio test was used to assess the significance of predictors. Odds ratios and 95% confidence intervals were determined. P < 0.05 was considered to indicate a significant difference. Receiver operating characteristic (ROC) curves were used to evaluate the sensitivity and specify of ADC measurements for the presence and severity of parenchymal disorders. Statistical analysis was performed using SAS v9.2 software (SAS, Cary, NC, USA) and MedCalc v 10.3 (MedCalc, Mariakerke, Belgium).
Results DWI and histopathology examinations Of the 41 patients, two separate DWI studies were performed in five patients and three studies in another five
822
K. Sandrasegaran et al. / Clinical Radiology 66 (2011) 820e825
patients, resulting in a total of 56 DWI studies analysed. The median time interval between the DWI study and biopsy was 17 days (range 0e2.8 months). Forty-two of the MRI studies were performed within 1 month of liver biopsy, and 14 studies were performed on the same day as biopsy. In 27 cases, biopsy was performed before MRI, and in 29 cases biopsy was performed after MRI. Each DWI study was associated with a separate histological examination. The presence of acute rejection, chronic rejection, recurrent viral hepatitis, fibrosis, and steatosis were seen, respectively, in 11, 10, 26, 26, and seven liver biopsy samples, with some samples showing more than one parenchymal disorder. Forty-three DWI studies were associated with histological examinations showing one or more abnormalities, and 13 studies were associated with negative histological findings. The median time interval between OLT and the DWI study was 12.3 months (range 3.2e195 months).
ADC values The intraclass correlation of the ADC values of the right and left lobes was 0.51, corresponding to a moderate correlation. Fig 1 is a box-and-whisker plot of the ADC values according to parenchymal disorders. There was a significantly lower ADC associated with at least one positive histological finding [n ¼ 43, mean (SD) ADC ¼ 0.91 (0.15)103 mm2/s] compared to those associated a negative histology [n ¼ 13, mean (SD) ADC ¼ 1.11 (0.17)103 mm2/s, p ¼ 0.003]. The results of logistic regression analysis are given in Table 1. This table gives the significance values of each of the histological parameters on logistic regression analysis. In addition, the odds ratios and their 95% confidence intervals are presented. The predictive effects of ADC on the histological parameters are given. None of the individual histological parameters could be independently predicted by ADC. This is evident from the p-values of more than 0.05
Figure 1 Box-and-whisker plot of ADC values. The boxes represent the interquartile range. Whiskers represent range of values. Horizontal line within box is the median apparent diffusion coefficient (ADC) value. The small square represents outliers. y-axis gives ADC values in 103 mm2/s. None, no parenchymal disorder; AR, acute rejection; CR, chronic rejection; Viral, recurrent viral hepatitis; Fib, recurrence of fibrosis; Fat, fat infiltration.
Table 1 Results of logistic regression analysis. Histological parameter
p-Value
Odds ratio (95% CI)
Viral hepatitis Fat Fibrosis Acute rejection Chronic rejection Inflammationa Diseaseb
0.28 0.08 0.11 0.64 0.88 0.06 0.003
0.17 0.01 0.06 0.40 1.34 0.03 0.001
(0.01e4.14) (0.001e1.63) (0.002e1.82) (0.01e18.71) (0.03e68.85) (0.001e1.14) (0.001e0.029)
a Inflammation was a composite parameter made of cases that either had recurrence of viral hepatitis or acute rejection. Those without either condition were taken as not showing inflammation. b Disease is a composite parameter that was positive if any of the parameters was present and negative if there was no evidence of any of the histological parameters.
and the fact that the 95% confidence interval of the odds ratio includes 1.0. ADC values were able to predict the presence or absence of a parenchymal disorder. ADC values showed a trend towards significance in differentiating the presence of inflammation, i.e., acute rejection or recurrent hepatitis (n ¼ 34), from the absence of inflammation [n ¼ 22; mean (SD) ADC 0.92 (0.15) 103 mm2/s versus 1.02 (0.20)103 mm2/s, p ¼ 0.06; Fig 2]. The ADC values associated with more severe viral inflammation (n ¼ 6) were not significantly different with those associated with mild viral hepatitis [n ¼ 20; mean (SD) ADC of 0.88 (0.20)103 mm2/s versus 0.95 (0.15)103 mm2/s, p ¼ 0.62].
ROC curve analysis The ROC curve plotting the sensitivity and specificity of ADC values in determining the presence of a parenchymal disorder is shown in Fig 3. The area-under-curve was 0.84 (95% CI 0.72 to 0.92). ADC values less than 0.99103 mm2/s indicated the presence of a parenchymal disorder with a sensitivity of 85% (95% CI of 71% to 95%) and specificity of 72% (95% CI of 56% to 85%). Twenty-seven patients in the cohort had hepatitis C and underwent 41 biopsies. This subgroup comprised 24 male and three female patients with a mean age of 54.1 years (range 34e77 years). In patients with hepatitis C, reinfection of the allograft is almost inevitable,14 but the severity and time of onset vary. Antiviral therapy is usually started in patients with severe recurrent hepatitis (necro-inflammatory score of 3 or 4) or with recurrent fibrosis (stage 2 and above).15,16 The ROC curve plotting the sensitivity and specificity of ADC values in determining the subgroup of patients with hepatitis C who would benefit from antiviral therapy is shown in Fig 4. The area-under-curve was 0.68 (95% CI 0.51 to 0.83). An ADC value of less than 0.98103 mm2/s was 78% sensitive and 58% specific in determining this subgroup of patients who should be started on antiviral therapy.
Discussion Post-transplant hepatic parenchymal disorders often present with non-specific symptoms and variable elevation
K. Sandrasegaran et al. / Clinical Radiology 66 (2011) 820e825
823
Figure 3 ROC curve plotting sensitivity (y-axis) and 1-specificity (x-axis) of ADC values in determining the presence or absence of a parenchymal disorder. The area-under-curve (Az) was 0.84 (95% CI 0.72 to 0.92). The diagonal line represents Az of 0.50.
Figure 2 A 67-year-old man with hepatitis C and hepatocellular cancer (HCC) who underwent orthotopic liver transplantation (OLT). Axial diffusion-weighted images (b ¼ 500 s/mm2) performed (a) 6 months and (b) 15 months later are shown. The MRI studies, performed as routine screening for recurrent HCC, did not show morphological abnormalities. The signal intensities of the liver in the two studies appear similar. However, the left lobe ADC values at 6 and 15 month studies were 1.30 and 0.97103 mm2/s, respectively. Liver biopsies showed no recurrent viral hepatitis at 7 months, and presence of moderately severe viral hepatitis at 15 months, after OLT. No other parenchymal disorders were seen on the biopsies. In this patient, there was a reduction of ADC coinciding with recurrent viral hepatitis. We did not have sufficient patients with multiple MRI studies to determine the value of serial ADC measurements.
in liver function tests. Diagnosis usually requires a core liver biopsy, sometimes performed serially. Acute rejection is usually seen in the first month following OLT. Rejection occurring after 4 weeks is seen in 15e20% of patients and is less responsive to immunosuppressive therapy, partly because of delay in diagnosis.17 Thus, reliable, non-invasive techniques for diagnosing parenchymal disorders are useful. The authors are not aware of a prior publication assessing the usefulness of DWI in determining parenchymal disorders after OLT. In general, a significantly lower ADC was found in patients with at least one parenchymal disorder following OLT
(p ¼ 0.003). It may be possible to explain some of these findings based on the pathophysiology of these conditions. For instance, acute rejection is characterized by a mixed, predominantly mononuclear, infiltrate containing lymphocytes, neutrophils, and eosinophils in the portal tracts, with endothelitis or perivenulitis.18 It would be expected that the cellular infiltrate would restrict diffusion by increasing the intracellular volume per unit tissue volume. Lobular inflammation, hepatocytes ballooning, and apoptotic cytoplasmic bodies are seen in recurrent viral hepatitis,19 and would also be expected to reduce diffusion. However, severe acute rejection or hepatitis may result in centrilobular necrosis,
Figure 4 ROC curve plotting sensitivity (y-axis) and 1-specificity (x-axis) of ADC values in determining the subgroup of patients with hepatitis C who would benefit from antiviral therapy (see text). The area-under-curve (Az) was 0.68 (95% CI 0.51 to 0.83). The diagonal line represents Az of 0.50.
824
K. Sandrasegaran et al. / Clinical Radiology 66 (2011) 820e825
which may potentially increase ADC as a result of cell membrane lysis. These opposing effects may be a reason why there was no clear difference in ADC between mild and severe acute rejection. Nevertheless, there may be a role for serial ADC values in monitoring the progression of acute rejection or recurrent viral hepatitis. In the present cohort, the number of patients with more than two DWI tests was small and the value of serial DWI measurements could not be reliably assessed. Chronic rejection occurs in 5e10% of patients undergoing OLT. Histologically, chronic rejection is characterized by atrophy and loss of interlobular and septal bile ducts, foam cell arteriopathy, perivenular or bridging fibrosis, and a relative absence of inflammatory infiltrates.12 It is possible that fibrosis in this condition would lead to a reduction in ADC. Recurrence of fibrosis and cirrhosis occurs in 15e30% of patients 5 years after OLT, regardless of the original disease causing liver failure.20,21 Deposition of extracellular collagen is likely to reduce diffusion. Several studies have confirmed a reduction in ADC in cirrhotic, non-transplanted, livers.22e24 In macrovesicular hepatic steatosis, there is intracytoplasmic accumulation of fat-containing vacuoles, leading to ballooning of hepatocytes.25 As a result, ADC values may reduce in the presence of substantial steatosis. Given that most of the parenchymal disorders result in a reduction in diffusion, it is not surprising that no differences were found in the ADC values for these entities. A finding of this study was the assessment of patients with hepatitis C. Antiviral therapy is usually started when there is with severe recurrent hepatitis or with substantial fibrosis.15,16 ADC values were reasonably sensitive (78%) in determining the subgroup of patients who should be started on antiviral therapy. However, the specificity is low (58%), and ADC values cannot be used on the only criterion for determining the need for antiviral therapy. There were some limitations of the present study. The study was retrospective and hence the time interval between MRI and histological confirmation of liver fibrosis varied with a median interval of 17 days and a maximum interval of 2.8 months. Most biopsies (n ¼ 42) were performed within 1 month of the MRI study. Despite this relatively small time interval, it is possible that medical therapy for hepatitis or rejection may have led to differences in the severity of disease between the time of the MRI and that of the biopsy. Liver biopsy was used as the reference standard. Biopsy results are prone to interobserver variability in interpretation and may yield erroneous results if the parenchymal disease has a heterogeneous distribution. The technique of DWI is not standardized, and the present findings may not necessarily be comparable with studies performed with other DWI techniques, e.g. breath-hold. In conclusion, ADC values may be used to distinguish between those who have parenchymal disorders after OLT and those who do not. They may help to decide which patients with hepatitis C should be started on antiviral therapy after OLT. In this respect, DWI may be a useful screening tool in determining patients who require liver biopsy after OLT. In general, most parenchymal disorders resulted in the lowering of the ADC values. It was not
possible to differentiate between these disorders using ADC values, and DWI is unlikely to replace core biopsy for the assessment of parenchymal disorders following OLT.
Supplementary data Supplementary data associated with this article can be found in online version at doi:10.1016/j.crad.2010.12.020.
References 1. Wiesner RH, Menon KV. Late hepatic allograft dysfunction. Liver Transpl 2001;7:S60e73. 2. Charlton M, Ruppert K, Belle SH, et al. Long-term results and modeling to predict outcomes in recipients with HCV infection: results of the NIDDK liver transplantation database. Liver Transpl 2004;10:1120e30. 3. Caiado AH, Blasbalg R, Marcelino AS, et al. Complications of liver transplantation: multimodality imaging approach. RadioGraphics 2007;27:1401e17. 4. Chong WK. Ultrasound evaluation of liver transplants. Abdom Imaging 2004;29:180e8. 5. Girometti R, Cereser L, Como G, et al. Biliary complications after orthotopic liver transplantation: MRCP findings. Abdom Imaging 2008;33:542e54. 6. Afdhal NH, Nunes D. Evaluation of liver fibrosis: a concise review. Am J Gastroenterol 2004;99:1160e74. 7. Wiesner RH, Rakela J, Ishitani MB, et al. Recent advances in liver transplantation. Mayo Clin Proc 2003;78:197e210. 8. Regev A, Molina E, Moura R, et al. Reliability of histopathologic assessment for the differentiation of recurrent hepatitis C from acute rejection after liver transplantation. Liver Transpl 2004;10:1233e9. 9. Petrovic LM, Villamil FG, Vierling JM, et al. Comparison of histopathology in acute allograft rejection and recurrent hepatitis C infection after liver transplantation. Liver Transpl Surg 1997;3:398e406. 10. Younossi ZM, Boparai N, Gramlich T, et al. Agreement in pathologic interpretation of liver biopsy specimens in posttransplant hepatitis C infection. Arch Pathol Lab Med 1999;123:143e5. 11. Demetris AJ, Batts KP, Dhillon AP, et al. Banff schema for grading liver allograft rejection: an international consensus document. Hepatology 1997;25:658e63. 12. Demetris A, Adams D, Bellamy C, et al. Update of the International Banff Schema for Liver Allograft Rejection: working recommendations for the histopathologic staging and reporting of chronic rejection. An international panel. Hepatology 2000;31:792e9. 13. The French METAVIR Cooperative Study Group. Intraobserver and interobserver variations in liver biopsy interpretation in patients with chronic hepatitis C. Hepatology 1994;20:15e20. 14. Wright TL, Donegan E, Hsu HH, et al. Recurrent and acquired hepatitis C viral infection in liver transplant recipients. Gastroenterology 1992;103:317e22. 15. Terrault NA. Hepatitis C therapy before and after liver transplantation. Liver Transpl 2008;14(Suppl. 2):S58e66. 16. Wiesner RH, Sorrell M, Villamil F. Report of the first International Liver Transplantation Society expert panel consensus conference on liver transplantation and hepatitis C. Liver Transpl 2003;9:S1e9. 17. Anand AC, Hubscher SG, Gunson BK, et al. Timing, significance, and prognosis of late acute liver allograft rejection. Transplantation 1995;60:1098e103. 18. Snover DC, Sibley RK, Freese DK, et al. Orthotopic liver transplantation: a pathological study of 63 serial liver biopsies from 17 patients with special reference to the diagnostic features and natural history of rejection. Hepatology 1984;4:1212e22. 19. Ziarkiewicz-Wroblewska B, Gornicka B, Oldakowska-Jedynak U, et al. Morphologic features of hepatitis C recurrence in patients after orthotopic liver transplantationdpreliminary analysis of our case observations. Transplant Proc 2006;38:226e30. 20. Davern TJ, Lake JR. Recurrent disease after liver transplantation. Semin Gastrointest Dis 1998;9:86e109.
K. Sandrasegaran et al. / Clinical Radiology 66 (2011) 820e825 21. Feray C, Caccamo L, Alexander GJ, et al. European collaborative study on factors influencing outcome after liver transplantation for hepatitis C. European Concerted Action on Viral Hepatitis (EUROHEP) Group. Gastroenterology 1999;117:619e25. 22. Lewin M, Poujol-Robert A, Boelle PY, et al. Diffusion-weighted magnetic resonance imaging for the assessment of fibrosis in chronic hepatitis C. Hepatology 2007;46:658e65.
825
23. Taouli B, Tolia AJ, Losada M, et al. Diffusion-weighted MRI for quantification of liver fibrosis: preliminary experience. AJR Am J Roentgenol 2007;189:799e806. 24. Sandrasegaran K, Akisik FM, Lin C, et al. Value of diffusion-weighted MRI for assessing liver fibrosis and cirrhosis. AJR Am J Roentgenol 2009;193:1556e60. 25. Lall CG, Aisen AM, Bansal N, et al. Nonalcoholic fatty liver disease. AJR Am J Roentgenol 2008;190:993e1002.