| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Radiology, University of Pittsburgh Medical Center (Presbyterian
Campus), Pittsburgh, PA.
2 Present address: Instituto di Radiologia, Azienda Ospedaliero-Universitaria
"Santa Maria della Misericordia" di Udine, Via Colugna 50, 33100
Udine (UD), Italy.
3 Department of Surgery, Erasmus University Medical Center Rotterdam, Rotterdam,
The Netherlands.
4 Department of Pathology, University of Pittsburgh Medical Center, Pittsburgh,
PA.
Received November 13, 2007;
accepted after revision May 9, 2008.
A. Furlan and D. J. van der Windt contributed equally to this study.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Radiology and pathology records from January 1999 to March 2007 were reviewed to identify 24 patients (22 women and two men; mean age, 40 years) with a pathology-proven diagnosis of hepatic adenoma (mean size ± SD, 7.2 ± 3.7 cm) who underwent helical contrast-enhanced CT (n = 23) and/or gadolinium-enhanced MRI (n = 8). The control group was composed of 24 patients of similar age and sex (21 women and three men; mean age, 43 years) with hepatic hemangioma who underwent CT or MR evaluation during the same time period. Two radiologists independently interpreted the imaging studies to determine the number of lesions and whether steatosis was present. The difference in prevalence of steatosis between the adenoma group versus the control group and the difference between patients with a single hepatic adenoma versus those with multiple hepatic adenomas were assessed (chi-square test).
RESULTS. Hepatic steatosis was present in 14 of 24 patients (58%) with hepatic adenoma versus seven of 24 patients (29%) with hemangioma (p = 0.042). Steatosis was more common in patients with multiple hepatic adenomas (9/11, 82%) than in those with a single hepatic adenoma (5/13, 38%) (p = 0.047).
CONCLUSION. Hepatic adenomas occur more frequently and more often are multiple in patients with hepatic steatosis.
Keywords: CT • hepatic adenomas • hepatic adenomatosis • liver adenomas • liver adenomatosis • liver steatosis • MRI
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In recent studies, investigators have described cases of multiple adenomas in patients with liver steatosis [8–11]. The rapidly rising prevalence of nonalcoholic fatty liver disease in Western societies [12], often detected at CT and MRI of the liver, as well as its potential role for the development of cirrhosis and even hepatic malignancy has been described [13]. However, to our knowledge, less is known about the relationship between intrahepatocyte fat deposition and the development of benign liver tumors. The possible existence of this association may have consequences for the management of patients with multiple adenomas because segmental or lobar resection of a fatty liver can be technically difficult and can lead to increased postoperative morbidity, including perihepatic abscesses, infectious complications, and hepatic insufficiency [14, 15].
The purpose of our study was to investigate the association between liver steatosis and hepatic adenomas. To provide statistical support for this hypothesis, the prevalence of steatosis in patients with hepatic adenoma was compared with the prevalence of steatosis in patients with hepatic hemangioma, the most common benign liver tumor [16], in a case-control setting.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Study Group
A certified research coordinator searched the medical, radiology, and
pathology databases for information obtained between January 1999 and March
2007 using the terms "hepatic adenoma," "liver
adenoma," "hepatic adenomatosis," and "liver
adenomatosis." The search yielded the records for 39 patients with a
proven diagnosis of hepatic adenoma. Fifteen patients were excluded because of
lack of multiphasic CT or MR scans available for review; these images were
considered essential for optimal lesion detection and identification of liver
steatosis. Therefore, our study group included 24 patients (22 women, two men;
age range, 23–57 years; mean age, 40 years); the mean size (± SD)
of the largest hepatic adenoma was 7.2 ± 3.7 cm. All had a
histologically confirmed diagnosis of hepatic adenoma. The histologic specimen
for the diagnosis was obtained at core needle biopsy in five patients (21%),
liver resection in 16 (67%), and liver explantation in three (13%).
Twelve patients (50%) underwent CT or MRI performed specifically to evaluate a known or suspected liver mass, whereas 12 patients (50%) had hepatic adenoma discovered on CT or MRI performed to evaluate abdominal pain or other indications. Nine patients (38%) presented with symptoms that could be related to hepatic adenoma: five with pain or discomfort of the upper right abdominal quadrant and four with signs and symptoms of acute or recent hemorrhage. In three patients (13%), hepatic adenoma was detected on imaging performed as routine follow-up for GSD type I. No patients had clin ical, radiologic, or histologic evidence of viral hepatitis or cirrhosis, and no patient was imaged for elevated liver function test results or for a clinical suspicion of fatty infiltration of the liver. The clinical management of patients included observation in five (21%) after core needle biopsy results had excluded malignancy. Three patients (13%) underwent orthotopic liver transplantation, two because of adenomatosis and one because of end-stage alcoholic liver disease, and hepatic adenoma was revealed in the liver explant.
Control Group
The CT and MRI reports in the same time period (i.e., from January 1999 to
March 2007) were reviewed by a certified research coordinator to identify
patients with a diagnosis of hepatic hemangioma on multiphasic CT or MRI
series. Patients with hepatic hemangioma were chosen as control subjects
because this tumor is considered the most common hepatic primary lesion,
reported in up to 20% of the general adult population
[17], and the diagnostic
criteria are well established both at CT and MRI
[18,
19]. In addition, there is, to
our knowledge, no known relation between the prevalence of hemangioma and
steatosis.
Imaging criteria to document hemangioma included marked hyperintensity on T2-weighted MR images, hypointensity on T1-weighted MR images, and nodular peripheral and centripetal contrast material enhancement on dynamic CT and MR images obtained after IV contrast administration [18, 19]. In our institution, for detection and characterization of both hepatic adenoma and hemangioma, a multiphasic CT or MR examination is indicated. For each case, the closest matching control subject for age at diagnosis, sex, and imaging technique was selected (Table 1). Among the 24 patients in the control group, 21 (88%) were women and three (12%) were men (p = 1.00, compared with the study group), with the age at diagnosis ranging from 26 to 54 years (mean age, 43 years; p = 0.227).
|
CT and MRI
For 23 study patients (96%) and 22 control subjects (92%), abdominal CT
images were available for review (p = 1.00), whereas eight patients
(33%) in the study group and two patients (8%) in the control group had MRI
series avail able (p = 0.033). Among these, both CT and MRI were
available for review in seven patients (29%) in the study group and none (0%)
in the control group. In those cases, data obtained from the CT images were
included in the analyses to achieve maximum comparability with the control
group. Comparison of CT and MR images showed concordant results for steatosis
and number of lesions in seven of the seven patients (data not shown).
CT series consisted of biphasic abdominal heli cal CT, which included contrast-enhanced im aging through the liver with both hepatic arter ial and portal venous phases with delays of 25–35 seconds and 60–70 seconds, respectively, after initiation of IV injection of contrast material. All CT scans were obtained on MDCT scanners. All patients received nonionic IV contrast material (ioversol [Optiray 350, Mallinckrodt Imaging]) that was administered at 3 or 4 mL/s at a volume of 125–150 mL. In addition, for 16 patients and 15 control subjects, unenhanced scanning of the liver before contrast material administration was performed.
MRI was performed on a 1.5-T system (Signa Excite HD, GE Healthcare) using a variety of software upgrades that continuously evolved during the study period. Standard liver imaging sequences included T1-weighted in-phase and opposed-phase gradient-echo and T2-weighted fast spin-echo sequences. T1-weighted imaging was repeated in all patients after contrast material administration during the hepatic arterial (delay = 20–25 seconds), portal venous (delay = 60–70 seconds), and delayed (delay = 3–5 minutes) phases. All patients received a gadolinium chelate at a dose of 0.1 mmol/kg of body weight followed by a 20-mL saline flush.
Qualitative Analysis
Images were reviewed on a PACS workstation (Stentor, Philips Healthcare)
independently by two abdominal radiologists with 5 and 4 years of experience,
respectively, and discordance was resolved by consensus. Because we were not
attempting to determine the accuracy of CT and MRI for the diagnosis of focal
hepatic lesions, the readers had knowledge of the diagnosis of hepatic adenoma
or hemangioma but were blinded to all clinical information and to the purpose
of the study. In each patient, the CT and MR images obtained before and after
contrast injection were reviewed to assess whether steatosis was present and,
if so, the distribution of steatosis in the surrounding liver parenchyma, the
total number of lesions, and the maximum diameter of the index lesion. In
addition, every lesion with imaging features different from the characteristic
findings of hepatic adenoma or hemangioma was recorded.
For the assessment of liver steatosis on CT images, each reader obtained attenuation measurements in Hounsfield units (HU) from the average of two regions of interest (ROIs) drawn in the most representative fatty infiltrated area of the liver and one ROI in the spleen. Steatosis was considered present when the difference in attenuation between the liver and spleen (measured as liver – spleen) was less than –10 HU on unenhanced images or less than –30 HU on portal venous phase images [20, 21].
On MRI, the presence of steatosis was qualitatively evaluated; the readers assessed for substantial signal intensity loss on the T1-weighted opposed-phase gradient-echo images in comparison with the in-phase images [22]. The distribution of fat in the liver was described using the following terms: diffuse, lobar, or focal or multifocal [23].
Additional Data Collection
The clinical reports of each patient were reviewed by three investigators
for the patient's weight and height to calculate body mass index (BMI) and
history of OC use, steroid intake, diabetes, and GSD.
Statistical Analysis
Continuous variables were described as a mean ± SD or as a median
and range and categoric variables, as a number (percentage). The chi-square
test was performed to compare the prevalence of steatosis in the patients
versus the control subjects. Subsequently, the chi-square test was used to
compare the prevalence of steatosis in the patients with a single adenoma
versus those with multiple adenomas in a subgroup analysis. No attempt was
made to test the interobserver differences or accuracy because, in most cases,
no absolute standard was available to establish the exact number, size, and
nature of the hepatic lesions. A p value of less than 0.05 was
indicative of a statistically significant difference. Data analyses were
performed with commercially avail able software (SPSS 2004 for Microsoft
Windows, version 13.0, SPSS).
|
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
|
|
|
Thirteen patients (54%) had a single adenoma and 11 (46%) had multiple adenomas (range, 1–20 lesions; median, 4 lesions) identified at imaging. The maximum diameter of the index lesion ranged from 2 to 17 cm with a median of 7.5 cm. In one patient, one additional lesion with imaging characteristics of focal nodular hyperplasia (FNH), which has a known association with hepatic adenoma [24], was identified. In two other patients with hepatic adenoma, a small hemangioma (< 2 cm) was also observed in nonsteatotic liver parenchyma.
Control Group
Among the 24 patients in the control group, liver steatosis was identified
in seven subjects (29%) (Table
2). The diagnosis was achieved by means of unenhanced CT images in
six subjects and MR in-phase and opposed-phase gradient-echo images in one
subject.
Fifteen control subjects (62%) had a single hemangioma and nine (38%) had multiple hemangiomas (range, 1–5 lesions; median, 1 lesion) identified at imaging. The maximum diameter of the index lesion ranged from 1 to 24 cm with a median of 3 cm. No lesions with imaging findings different from hemangioma could be identified in the studies available.
Study and Control Groups
The prevalence of liver steatosis was significantly higher in the patients
with hepatic adenoma (58%) than in the control subjects (29%) (p =
0.042) (Table 2).
Information about the use of OCs was available in 17 patients and 11 control subjects. A positive history for OC use was present in 14 (82%) of the 17 hepatic adenoma patients (median time of use, 15 years) and in four (36%) of the 11 control subjects (median time of use, 15 years) (p = 0.02). The BMI was calculated in 16 patients with adenoma and in 10 patients with hemangioma (mean ± SD, 30.1 ± 7.24 vs 28.1 ± 5.28, respectively; p = 0.456). Three study patients (13%) and no control subjects were affected by GSD type I (p = 0.023). Eight patients (33%) with hepatic adenoma and two control subjects (8%) with hemangioma had diabetes (p = 0.033).
Study Group: Single Versus Multiple Adenomas
Liver steatosis was present in five patients with a single hepatic adenoma
(38%) and in nine (82%) with multiple hepatic adenomas (Figs.
1A,
1B and
2A,
2B,
2C,
2D). This difference was
statistically significant (p = 0.047)
(Table 3).
|
Nine (82%) of the 11 patients with a single adenoma and five (83%) of six patients with multiple adenomas and available clinical information had a positive history of OC use (p = 1.00). BMI was calculated in seven patients with a single adenoma and in nine patients with multiple adenomas (mean ± SD, 32.5 ± 7.3 vs 28.3 ± 7.0, respectively; p = 0.265). All three patients in our population with a diagnosis of GSD type I had multiple adenomas at imaging. Among the eight patients with a history of diabetes, five had a single adenoma and three had multiple adenomas (p = 0.679) (Table 3).
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Long-term exposure to exogenous estrogens by means of OCs is recognized as the main risk factor for the development of the classic single hepatic adenoma, or "pill adenoma," as first described by Baum et al. [25] and supported by others [4, 26]. However, the conditions that predispose patients to develop multiple adenomas are not fully understood to our knowledge. In 1985, liver adenomatosis was described as a separate entity that is characterized by the presence of multiple adenomas (> 10) and lack of correlation with steroid medication intake or GSD that affects both men and women [27]. However, recent studies have not been able to confirm this classification and have reported neither histologic nor radiologic differences among the adenomas arising as a single lesion, those arising as multiple lesions (2–10 lesions), and liver adenomatosis (> 10 lesions) [11, 28]. Therefore, in our study we considered every patient to have multiple adenomas if imaging showed multiple lesions with identical characteristics in a patient with at least one pathologically confirmed adenoma. We did not consider "adenomatosis" as a distinct entity.
Recently, several groups of investigators reported or suggested an increased incidence of multiple adenomas in the setting of liver steatosis [8–11]. Hepatic steatosis is due to fat accumulation within the hepatocytes mainly resulting from alcoholic liver disease and nonalcoholic fatty liver disease. The incidence of nonalcoholic fatty liver disease is rapidly increasing in the Western countries along with obesity [12, 29]. Although histologic analysis is considered the reference standard, CT and MRI are well-established noninvasive tests that accurately detect steatosis [20–22].
We hypothesize two possible mechanisms to explain the significant association between multiple hepatic adenomas and liver steatosis: The first hypothesis is that an increase in the intracellular content of lipid may lead to a hyperplastic reaction of the hepatocytes and finally to the formation of hepatic adenoma. Oxidative and inflammatory effects as well as alteration in hepatocyte regeneration have been described to occur in fatty liver cells [30] and could possibly be involved in a hyperplastic reaction. The second hypothesis is that the fatty tissue may account for a continuous local estrogen generation, similar to what has been shown for peripheral fatty tissues [31] through an increase in aromatase [32]. If this is indeed the case, one could speculate that a continuous local stimulus may alter the hepatocyte growth rates more substantially than systemically elevated levels in OC users, resulting in the development of multiple tumors.
Other factors have been postulated in the development of multiple hepatic adenomas. Portal vein abnormalities (i.e., portal vein agenesis or portal hepatic shunts) have been reported to facilitate the development of benign lesions, especially hepatic adenoma [29] and FNH [33], through a focal disturbance of the hepatic blood supply [34]. Even hepatocellular carcinoma in a child has been reported as possibly related to hepatic vascular abnormalities [35]. Although in our series hepatic adenoma was present in association with FNH and hemangioma (one and two cases, respectively), no patient had any identifiable vascular abnormality.
Investigators have reported that, in patients with diabetes, the accumulation of glycogen within the cells, related to the defect in hepatic catabolism, can play a role in the hyperplastic reaction leading to the genesis of hepatic adenomas [5]. The obesity epidemic has paralleled the rapid increase in the prevalence of type 2 diabetes. In our study, there was a significant increased incidence of diabetes in the adenoma group versus the control group; however, no difference could be found between patients with a single adenoma and those with multiple adenomas. Of interest, one man in the patient group had a history of diabetes mellitus as a unique potential explanation for the presence of adenoma and a negative history of steroid intake or GSD.
There are limitations of our study that need to be acknowledged. Because of its retrospective nature, complete data on all variables could not be retrieved. The lack of height and weight data limited the possibility of correlating BMI to the presence of steatosis in patients and control subjects. This correlation has been described as rather strong [36]; however, our data did not suggest a higher BMI in study patients versus control subjects or in patients with multiple versus single adenomas. With respect to the history of OC use, the unavailability of complete data prevented us from performing a multivariate analysis adjusting for OC use. Investigators have reported that steroid hormones can be an independent risk factor in the development of liver steatosis [37], and OC use should, therefore, be treated as a confounder in the relation between steatosis and hepatic adenoma. However, the difference in availability of these data (i.e., data available for 71% of patients vs 46% of control subjects) could bias a multivariate analysis because it is likely that physicians asked more specifically about OC use in patients with adenoma than in patients with hemangioma. Nevertheless, the observation that multiple hepatic adenomas occur more frequently in fatty livers is evident. The fact that this possibly includes a proportion of patients in whom steatosis is an intermediate in the pathway between OC and hepatic adenoma is perhaps not of direct cli nical relevance.
Other limitations of our study design could include underrepresentation of small hepatic adenomas in our series because those tumors are less likely to be biopsied or resected. State-of-the-art imaging, especially dynamic MR series, can be used to accurately diagnose hepatic adenoma and distinguish this tumor from FNH or malignancies on the basis of the described criteria in many cases [38]. However, histology results remain the reference standard and therefore only patients with a histologic diagnosis were included in this analysis. Another concern that may arise is whether hemangiomas in steatotic livers could be underrepresented because fatty infiltration may complicate the accurate detection of lesions. By including only patients with dynamic imaging series, on which hypervascular lesions are clearly visible, we tried to avoid the disproportional exclusion of control subjects with liver steatosis.
The treatment for hepatic adenoma is not uniform. Some clinicians prefer a conservative approach; however, because of the risk of hemorrhage and malignant degeneration, surgery is often advocated [39, 40]. The co-prevalence of multiple adenomas and steatosis may complicate the decision making for surgery versus observation in those patients because tumors are spread through the liver and steatosis increases the risk of surgical morbidity. Current MRI techniques can be of great value in providing the information needed to balance the risks versus the benefits of surgical intervention. In addition, the use of standardized MRI protocols for patients with benign liver lesions could resolve many of the limitations of our study in future prospectively conducted research. These studies are needed to confirm the association between hepatic adenoma and liver steatosis as suggested by our data.
In conclusion, the results of our study show the association of liver steatosis and hepatic adenoma in a case-control setting, which is a suitable study design for uncommon conditions such as hepatic adenoma. Multiple adenomas were especially correlated with steatosis. In view of the increasing incidence of hepatic steatosis and obesity, this group of hepatic adenoma patients might be of growing importance and needs special attention.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
