| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Report |
1 Department of Medical Imaging, University Health Network and Mount Sinai
Hospitals, 200 Elizabeth St., Toronto, Ontario M4G 2C4, Canada.
2 Department of Medical Imaging, 3-964, Princess Margaret Hospital, University
Health Network, 610 University Ave., Toronto, Ontario M5G 2M9, Canada.
3 Department of Internal Medicine, University Health Network and Mount Sinai
Hospitals, Toronto, Ontario M4G 2C4, Canada.
4 Department of Laboratory Medicine and Pathobiology, University Health Network
and Mount Sinai Hospitals, Toronto, Ontario M4G 2C4, Canada.
Received August 26, 2002;
accepted after revision October 28, 2002.
Address correspondence to K. Khalili.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CONCLUSION. Hepatic subcapsular steatosis appeared as subcapsular nodules and often rindlike areas of low attenuation in seven (18%) of 39 patients who received intraperitoneal insulin with their peritoneal dialysate. Cessation of intraperitoneal insulin therapy led to reversal of the steatosis in three patients.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Disadvantages to intraperitoneal insulin delivery include an increased rate of peritonitis and a worsening serum cholesterol profile [1]. In addition, this route exposes the subcapsular hepatocytes to a higher concentration of insulin than the remainder of the liver. Insulin blocks the usual oxidation of free fatty acids in the hepatocytes, leading to preferential esterification into triglycerides, which then accumulate in the cell [2]. The result is a unique pattern of fatty infiltration in a subcapsular location known as hepatic subcapsular steatosis. This process was first described by Wanless et al. [2] in 1989 in a pathohistologic series of 11 postmortem livers (Figs. 1 and 2).
|
|
With the improved resolution of cross-sectional imaging modalities, hepatic subcapsular steatosis has recently become recognized in the imaging literature [3, 4]. To our knowledge, only one systematic study using sonography has appeared in the literature [3], and no reports have been published on the features and prevalence of hepatic subcapsular steatosis using CT or MR imaging. The purpose of this study was to investigate the appearance and prevalence of hepatic subcapsular steatosis using CT.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Thirty-nine (60%) of these patients received intraperitoneal insulin with their peritoneal dialysate and constituted the study group (age range, 36–80 years; mean age, 57 years). Twenty-six patients (40%) who did not receive intraperitoneal insulin were used as control subjects (age range, 30–85 years; mean age, 54 years). A total of 82 CT scans (mean, 2.1 scans per patient) in the study group and 39 scans (mean, 1.6 scan per patient) in the control group were available. All scans were performed with a helical or multidetector CT scanner (HiSpeed or LightSpeed, General Electric Medical Systems, Milwaukee, WI). The collimation varied between 5 and 10 mm with a 50% overlap.
In the study group, 55 CT scans (67%) were enhanced and 25 (31%) were unenhanced; in the control group, 21 CT scans (54%) were enhanced and 11 (28%) were unenhanced. In addition, one patient (two CT scans, 2%) in the study group and three patients (seven CT scans, 18%) in the control group underwent both contrast-enhanced and unenhanced CT. All patients were randomized and their CT scans independently reviewed by two dedicated abdominal imagers. The reviewers were unaware of whether the patients received intraperitoneal insulin at the time of their CT studies. Thirty-seven CT scans (46%) in the study group and seven (18%) in the control group were on hard copy, whereas 44 (54%) in the study group and 32 (82%) in the control group were on soft copy. Hard-copy images had standard soft-tissue and liver window settings; and both were examined for the purpose of this study. Studies available on PACS (picture archiving and communication system) were individually adjusted by the reviewers to obtain sufficiently narrow windows for increased sensitivity. The Fisher's exact test was used for statistical analysis.
The criteria used for diagnosis of hepatic subcapsular steatosis were based on the location and morphology of hepatic subcapsular steatosis reported in the literature [2, 3]. The diagnosis of hepatic subcapsular steatosis was made by the presence of low-attenuation lesions in subcapsular locations. To avoid overdiagnosis of hepatic subcapsular steatosis, we excluded patients with a single lesion, including those with an area of low attenuation adjacent to the falciform ligament (a typical location of focal fat). Care was also taken to exclude pseudolesions caused by high-attenuation ribs producing a low-attenuation artifact (beam-hardening artifact) on the adjacent liver tissue or perfusion defect caused by rib compression. For the patients whose images were positive for hepatic subcapsular steatosis, a search of all available clinical data was performed to exclude underlying liver disease. In each patient with hepatic subcapsular steatosis, the pattern, number, and maximal two-dimensional measurements of the lesions were recorded. We also performed a search of the available cross-sectional imaging records of these patients. Seventeen sonograms in five patients were available and were reviewed by consensus for the presence of hepatic subcapsular steatosis manifest by subcapsular, echogenic nodules or rinds.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Hepatic subcapsular steatosis was manifest by two patterns. All patients had discrete, nodular, subcapsular low-attenuation lesions (Figs. 3A, 3B). There was a mean of 3.9 discrete lesions (range, 3–6 discrete lesions) per patient, with a mean subcapsular depth of 2.0 cm (range, 0.4–7.3 cm) and a mean length of 2.2 cm (range, 0.7–13.3 cm). In three of seven patients, a relatively thin, diffuse, subcapsular rind of low attenuation was also noted (Figs. 4A, 4B).
|
|
|
|
In five patients with hepatic subcapsular steatosis, previous sonograms of the abdomen were available for retrospective analysis; four of the five patients had findings typical of hepatic subcapsular steatosis. In one of these patients, the diagnosis of hepatic subcapsular steatosis had been made prospectively with biopsy proof (Fig. 2). In this patient, continuous ambulatory peritoneal dialysis was discontinued after the initial investigations, and multiple sonographic follow-up studies showed decreasing size and eventual resolution of most lesions over 27 months. Two patients with hepatic subcapsular steatosis had multiple CT scans. In both of these patients, continuous ambulatory peritoneal dialysis had been discontinued after the first scan because of peritoneal complications. The findings decreased in severity in the follow-up studies obtained 4 months in one patient and 28 months in the other patient after the positive CT findings (Figs. 5A, 5B).
|
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
To avoid overdiagnosis of hepatic subcapsular steatosis and thus improve the specificity, we excluded patients who had a single hypoattenuating subcapsular lesion. Use of this criterion would prevent attributing the common areas of subcapsular fat deposition or perfusion abnormality to intraperitoneal insulin. These areas include segments III and IVB adjacent to the falciform ligament, where there is often a mixing of portal venous and systemic venous blood supplied by the epigastric–paraumbilical venous system, or anterior to the portal vein in segment IV because of the potential supply of the parabiliary venous plexus [5]. Although many patterns of fatty deposition have been previously described in the imaging literature [6], both the multinodular and rindlike areas of subcapsular fatty deposition seem too specific within the setting of intraperitoneal insulin delivery. We also took care to exclude pseudolesions of the liver that may be caused either by beam-hardening artifacts or perfusion defects due to rib compression [5].
To our knowledge, the natural history of hepatic subcapsular steatosis has not been previously described. In our subset of two patients with multiple follow-up CT scans and one patient with follow-up sonograms, the findings of hepatic subcapsular steatosis decreased in severity with the discontinuation of intraperitoneal insulin and continuous ambulatory peritoneal dialysis. This decrease suggests that the changes are reversible once the subcapsular hepatocytes are no longer exposed to high insulin concentrations.
Five patients in our study whose images were positive for hepatic subcapsular steatosis had previous sonograms available for review, and four showed evidence of hepatic subcapsular steatosis. The significant difference in acoustic impedance of fatty infiltrated liver tissue makes sonography relatively sensitive for detection of hepatic steatosis [7, 8]. However, CT may not be as sensitive, especially in the setting of contrast enhancement [9]. This difference may partly explain both the low prevalence of hepatic subcapsular steatosis in our study and its lack of recognition in the CT literature. We diagnosed hepatic subcapsular steatosis in seven (18%) of the 39 patients in our study group, whereas both the histopathologic and sonographic series reported a much higher prevalence involving 10 (91%) of 11 patients [2] and seven (88%) of eight patients [3]. A large proportion of our CT examinations were contrast-enhanced (54% in the study group), potentially decreasing the conspicuity of the fatty areas.
The small size of the lesions may be another factor causing decreased detection of hepatic subcapsular steatosis on CT. In the histopathologic series of Wanless et al. [2], five of 10 patients had only a mild degree of hepatic subcapsular steatosis that was not visible on gross examination and that was as thin as 0.05 mm (four cells thick) [2]. We would not expect to be able to detect such subtle changes on cross-sectional imaging. However, the series by Wanless et al. was a retrospective study in which, in most patients, the entire liver specimen was not available for microscopic examination. It is therefore possible that regions of thicker hepatic subcapsular steatosis, which would be visible on cross-sectional imaging, were not examined.
We were unable to definitively prove the presence of fat in most patients with hepatic subcapsular steatosis because MR imaging of the liver was not performed, and only one biopsy specimen was obtained. We diagnosed fatty infiltration because of the typical findings, their similarity to the described pathologic and sonographic appearances [2, 3], and the presence of the typical findings on unenhanced CT (one patient) and sonography (four patients). Further evidence was provided by the fact that we found hepatic subcapsular steatosis in only the study group. MR imaging should be an excellent noninvasive method for proving subcapsular fatty infiltration [4]. However, most patients treated with continuous ambulatory peritoneal dialysis are imaged with CT or sonography for complications related to peritoneal dialysis. Furthermore, the typical CT finding of hepatic subcapsular steatosis and its lack of clinical significance decrease the indication for MR imaging. Therefore, the role of MR imaging may be as a confirmatory test in patients with atypical findings.
Apart from hepatic subcapsular steatosis, local effects of high concentrations of insulin within the liver have been documented elsewhere. The parabiliary venous system, partly draining the pancreatic head, often joins the portal circulation before entering the liver. When such communication does not occur, or when an aberrant pancreatoduodenal vein is present, the result is exposure of the dorsal portion of segment IV to insulin-rich blood from the pancreas. This result is believed to be the cause of the typical region of fatty infiltration seen in segment IV of the liver anterior to the portal vein [5, 10]. Also, in a well-illustrated case report, a focus of fatty infiltration was shown around a liver metastasis from an insulin-producing primary pancreatic neoplasm [4]. More recently, nonalcoholic steatohepatitis and its relationship to insulin has become the subject of much interest. Nonalcoholic steatohepatitis is typically seen in patients with type 2 diabetes with insulin resistance who have high serum insulin concentrations. The discovery of hepatic subcapsular steatosis was one of the earliest clues to the pathogenesis of nonalcoholic steatohepatitis [2, 11, 12].
In summary, the CT appearance of hepatic subcapsular steatosis is similar to that described in the pathology and sonography literature and is easily identifiable. The prevalence of hepatic subcapsular steatosis as detected on CT is lower than that described by other studies. Recognition of hepatic subcapsular steatosis in the correct clinical setting will prevent misinterpretation and further unnecessary investigations.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  D. T. Boll and E. M. Merkle Diffuse Liver Disease: Strategies for Hepatic CT and MR Imaging RadioGraphics, October 1, 2009; 29(6): 1591 - 1614. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Stuart, T. C. Booth, C. J. C. Cash, A. Hameeduddin, J. A. Goode, C. Harvey, and A. Malhotra Complications of Continuous Ambulatory Peritoneal Dialysis1 RadioGraphics, March 1, 2009; 29(2): 441 - 460. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Demir, D. Torun, N. Tokmak, U. Ozkan, and L. Oguzkurt Resolution of hepatic subcapsular steatosis after discontinuation of CAPD Nephrol. Dial. Transplant., April 1, 2007; 22(4): 1247 - 1249. [Full Text] [PDF]
|
|
|
|
|
|
O. W. Hamer, D. A. Aguirre, G. Casola, J. E. Lavine, M. Woenckhaus, and C. B. Sirlin Fatty Liver: Imaging Patterns and Pitfalls RadioGraphics, November 1, 2006; 26(6): 1637 - 1653. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
