Piyaporn Limanond, MD
Steven S. Raman, MD
Charles Lassman, MD
James Sayre, MD
R. Mark Ghobrial, MD, PhD
Ronald W. Busuttil, MD,
PhD
Sammy Saab, MD
David S. K. Lu, MD
Index terms:
Liver, CT, 761.12111, 761.12115
Liver, fatty, 761.50
Liver, transplantation
Published online
10.1148/radiol.2301021176
Radiology 2004; 230:276–280
Abbreviations:
BMI
body mass index
LAI
liver attenuation index
ROI
region of interest
1
From the Departments of Radiology
(P.L., S.S.R., J.S., D.S.K.L.), Pathology
(C.L.), Surgery (R.M.G., R.W.B.), and
Digestive Diseases (S.S.), David Geffen
School of Medicine at UCLA, 10833 Le
Conte Ave, Los Angeles, CA 90095-
1721. Received September 13, 2002;
revision requested November 22; final
revision received April 30, 2003; ac-
cepted May 20. Address correspon-
dence to S.S.R. (e-mail: SRaman
@mednet.ucla.edu).
Author contributions:
Guarantors of integrity of entire study,
S.S.R., D.S.K.L.; study concepts and
design, S.S.R., D.S.K.L.; literature re-
search, P.L., S.S.R.; clinical studies,
R.M.G., R.W.B., S.S., C.L., P.L.; data
acquisition, P.L.; data analysis/in-
terpretation, P.L., D.S.K.L.; statistical
analysis, J.S.; manuscript preparation
and definition of intellectual content,
P.L., S.S.R.; manuscript editing and
revision/review, P.L., S.S.R., D.S.K.L.;
manuscript final version approval,
D.S.K.L.
?
RSNA, 2004
Macrovesicular Hepatic
Steatosis in Living Related
Liver Donors: Correlation
between CT and Histologic
Findings
1
PURPOSE: To assess degree of macrovesicular steatosis with unenhanced com-
puted tomography (CT) and correlate it with histologic findings in potential donors
for living related liver transplantation.
MATERIALS AND METHODS: Forty-two candidates underwent unenhanced CT
within 4 weeks of core liver biopsy. An experienced liver pathologist, blinded to both
CT and surgical findings, retrospectively reviewed biopsy specimens and deter-
mined degree of macrovesicular steatosis. A radiologist blinded to histologic grad-
ing calculated mean hepatic attenuation in each donor liver by averaging 25
region-of-interest (ROI) measurements on five sections (five ROIs per section). Mean
splenic attenuation was calculated with three separate ROI measurements. Liver
attenuation index (LAI) was derived and defined as the difference between mean
hepatic and mean splenic attenuation. Body mass index (BMI) was determined for
each patient. Linear regression analysis was used to correlate degree of macrove-
sicular steatosis with both LAI and BMI.
RESULTS: LAI correctly predicted degree of macrovesicular steatosis in 38 (90%) of
42 cases. In four of four livers, LAI below 10 HU correlated with greater than 30%
macrovesicular steatosis (unacceptable for liver transplantation). In nine of 11 livers,
LAI was between 10 and 5 HU and correctly predicted 6%–30% steatosis (relative
contraindication). In two of 11 cases, LAI overestimated degree of hepatic steatosis.
LAI above 5 HU correctly predicted 0%–5% steatosis in 25 of 27 livers. In two of 27
cases, parenchymal hemosiderin deposition led to an increase in LAI into the normal
range, despite mild histologically confirmed steatosis. Degree of histologic macrove-
sicular steatosis correlated well with LAI (r
0.92) and marginally with BMI (r

0.45). Of 27 potential donors with normal livers at CT and acceptable LAI levels, four
(15%) were deemed poor donor candidates because core biopsy revealed subtle
hepatic necrosis and nonspecific hepatitis.
CONCLUSION: Although unenhanced CT quantifies the degree of macrovesicular
steatosis relatively well, it may preclude a liver biopsy only in a small percentage of
potential donors with low LAI (unacceptable degree of steatosis). Core liver biopsy is still
necessary in the majority of donors with normal LAI to identify those with both fatty liver
and coexistent hemosiderin deposition or radiologically occult diffuse liver diseases.
?
RSNA, 2004
Interest in living related liver transplantation has substantially increased in the United
States in recent years, in part because of the shortage of cadaveric liver donors. However,
careful donor selection must be performed to minimize risks of complications to both
donor and recipient (1–3). Preoperative evaluation of hepatic steatosis, especially the
macrovesicular subtype, is critical for donor selection (4–6). Severe macrovesicular steato-
sis (60%) in the donor liver has been associated with a greater than 60% risk of primary
nonfunction after transplantation. Moderate degrees of macrovesicular steatosis (30%–
276
R
a
d
i
o
l
o
g
y

60%) in donor livers may also result in
decreased hepatocyte regeneration and
higher rates of graft dysfunction, non-
function, and ischemic injury (7–9). The
clinical importance of microvesicular ste-
atosis is still controversial; however, most
authors believe that it has less clinical
importance than does the macrovesicular
subtype (5,6,10). At our institution, do-
nors with greater than 30% histologically
determined macrovesicular steatosis are
generally rejected because of these risks.
Macrovesicular hepatic steatosis, the
most common cause of low-attenuating
liver at computed tomography (CT), is
linked to a variety of causes that include
obesity; hyperlipidemia; and alcohol, ste-
roid, and chemotherapeutic drug use. Al-
though preoperative hepatic core biopsy is
currently the standard method for accurate
quantification and characterization of ma-
crovesicular steatosis, it is invasive, is asso-
ciated with risk, and contributes to overall
cost and morbidity (11–15). Various unen-
hanced CT techniques have been proposed
to help noninvasively determine the de-
gree of hepatic macrovesicular steatosis
(16–18). However, because of ethical con-
siderations in obtaining specimens from a
disease-free population, no study of a cor-
relation between findings at unenhanced
CT and the histologic degree of hepatic
steatosis has been performed in normal
subjects. Body mass index (BMI) has also
been used to predict degree of macrove-
sicular steatosis although results vary (19).
The purpose of our study was to assess de-
gree of macrovesicular steatosis with unen-
hanced CT and correlate it with histologic
findings in potential donors for living re-
lated liver transplantation.
MATERIALS AND METHODS
Patients
This study was approved by our insti-
tutional review board, which did not re-
quire informed consent. Between July
1993 and April 2002, 113 potential do-
nors were evaluated for living related
liver transplantation at our institution.
Of these candidates, 50 potential donors
underwent both unenhanced CT and
core liver biopsy with a 16-gauge needle.
Because of the potential for rapid change
in the degree of hepatic steatosis, we in-
cluded only those donors who under-
went both CT and liver biopsy within 4
weeks of each other. Of the 50 potential
donors, 42 satisfied our criteria and eight
were excluded because the time between
CT and liver biopsy exceeded the maxi-
mum 4-week interval. In eligible candi-
dates, the mean age was 37.4 years, with
a range of 21–57 years. These included 29
men (mean age, 36.7 years; SD, 9.66;
range, 21–57 years) and 13 women (mean
age, 39 years; SD, 7.39; range, 29–49
years). There were no statistically signifi-
cant differences between both sex distri-
butions (P
.45).
CT Imaging
In 15 donors, CT was performed with a
single–detector row helical scanner
(CT/i; GE Medical Systems, Milwaukee,
Wis); in 27 donors, CT was performed
with a multi–detector row helical scan-
ner (LightSpeed Qx/i; GE Medical Sys-
tems). Contiguous transverse images
were acquired through the liver with
10-mm collimation during a single
breath hold without intravenous con-
trast agent administration prior to donor
hepatic CT angiography. The unen-
hanced CT images were retrospectively
reviewed with a picture archiving and
communication system workstation
(Radstation; UCLA) by a radiologist (P.L.,
S.S.R., D.S.K.L.) blinded to pathologic
and surgical findings. For each case, the
hepatic attenuation was measured by
means of a random selection of 25 circu-
lar regions of interest (ROIs) on both
lobes on five transverse sections at differ-
ent hepatic levels (five ROIs per section).
For each ROI, we selected the largest pos-
sible ROI and avoided areas of visible he-
patic vascular and biliary structures to
represent liver parenchymal attenuation.
Our ROIs ranged from 200 to 400 mm
2
.
The ROI values were averaged as a mean
hepatic attenuation. To provide an inter-
nal control, the mean splenic attenua-
tion was also calculated by averaging
three random ROI values of splenic atten-
uation measurement on three transverse
sections at different splenic levels (one
ROI per section). The largest possible ROI
(size range, 200–400 mm
2
) was also se-
lected to represent splenic parenchymal
attenuation. The liver attenuation index
(LAI), derived from the difference be-
tween mean hepatic attenuation and
mean splenic attenuation, was used as a
parameter for prediction of the degree of
macrovesicular steatosis.
BMI Data
We assessed the utility of BMI, defined
as donor weight (in kilograms) divided
by height (in square meters), in the pre-
diction of the degree of histologically de-
termined steatosis. The BMI was available
in 37 (88%) of 42 potential donors.
Histologic Analysis
All the histologic specimens were ob-
tained by an experienced hepatologist
during the living related liver transplan-
tation preoperative evaluation. The spec-
imens were obtained from a single biopsy
of the right lobe by using a large 16-gauge
biopsy needle (Klatskin; Becton Dickin-
son, Franklin Lakes, NJ), which provided
the adequate hepatic specimens (at least
25 different portal triads for each speci-
men). An experienced liver pathologist
(C.L.), blinded to radiologic and surgical
findings, retrospectively reviewed the
histologic findings of selected donor liv-
ers. The slides for histologic review were
prepared with methods that included he-
matoxylin-eosin staining for routine his-
tologic analysis and quantification of fat
content, reticulin staining for identifica-
Figure 1. Graph demonstrates high degree of correlation between
LAI on y axis as a function of percentage of histologically confirmed
macrovesicular steatosis on x axis across a range of values.
Volume 230


Number 1 Pretransplantation CT for Steatosis in Living Related Liver Donors


277
R
a
d
i
o
l
o
g
y