Journal of Computer Assisted Tomography

27(1):48–55
© 2003 Lippincott Williams & Wilkins, Inc., Philadelphia

Accuracy in Differentiation of Mucinous and Nonmucinous

Rectal Carcinoma on MR Imaging

Myeong-Jin Kim, Jeong Seon Park, Sung Il Park, Nam Kyu Kim, Joo Hee Kim, Hee Jung Moon,
Young Nyun Park, and Won Ho Kim

Purpose: The purpose of this work was to determine the diagnostic accuracy of
MRI for the differentiation of mucinous and nonmucinous carcinomas in patients with
rectal carcinomas.
Method: Seventy-six patients with pathologically proven rectal adenocarcinomas
were evaluated with MRI. Both T2-weighted fast spin echo and T1-weighted spin echo
images were independently reviewed by two radiologists who were blinded for tumor
histology. Of the 76 cases, mucin pools were found in 28 (37%), and 16 had a mucin
pool occupying ⱖ50% of the tumor (mucinous carcinoma). The sensitivity, specificity,
accuracy, and positive and negative predictive values of distinguishing between mu-
cinous and nonmucinous carcinomas were calculated. The interobserver agreement
was estimated using ␬ statistics.
Results: The accuracies of the diagnosis of mucinous carcinoma were 97% for
Radiologist 1 and 96% for Radiologist 2. The sensitivities and specificities for diag-
nosing a mucinous carcinoma were 94 and 98% for Radiologist 1 and 100 and 95% for
Radiologist 2, respectively. Interobserver agreement was good (␬ ⳱ 0.61) for the
identification of a mucin pool and was excellent (␬ ⳱ 0.81) for the diagnosis of a
mucinous carcinoma.
Conclusion: Differentiation of mucinous and nonmucinous rectal carcinoma is
highly accurate on preoperative MRI.
Index Terms: Magnetic resonance imaging—Rectum—Neoplasms, rectal—
Adenocarcinoma, mucinous.

A mucinous carcinoma is a subtype of adenocarci-
noma containing abundant extracellular mucin within the
tumor (1–3). The amount of mucin required to warrant a
diagnosis of mucinous carcinoma, as set by the World
Health Organization, is at least ⱖ50% of the mucin pool
occupying the tumor mass (1). Previous reports have
shown that mucinous carcinomas of the rectum have a
higher rate of local recurrence and a less favorable prog-
From the Departments of Diagnostic Radiology (M.-J. Kim, J. S.
Park, S. I. Park, J. H. Kim, and H. J. Moon), Surgery (N. K. Kim), and
Pathology (Y. N. Park), Brain Korea 21 Project for Medical Science
(M.-J. Kim), and Division of Gastroenterology, Department of Internal
Medicine (W. H. Kim), Yonsei University College of Medicine, Seoul,
Republic of Korea. Address correspondence and reprint requests to Dr.
M.-J. Kim at Department of Diagnostic Radiology, Severance Hospital,
Yonsei University College of Medicine, Seodaemun-ku Shinchon-dong
134, Seoul, Republic of Korea, 120-752. E-mail: kimnex@yumc.yonsei.ac.kr
This work was supported by the Yonsei University Research Fund of
2002.
nosis than nonmucinous carcinomas (4–10) and that they
are also poor candidates for local excision even if con-
fined to the muscularis propria (11). Preoperative iden-
tification of the mucinous type, therefore, is useful for
treatment planning in cases of rectal carcinoma. How-
ever, it is sometimes difficult to locate the mucin or
estimate its content with a small amount of a biopsy
specimen (3,12). Therefore, it will be useful if preopera-
tive imaging can accurately predict the histologic type on
a preoperative imaging modality.
Previous reports have demonstrated that a mucinous
carcinoma may show a characteristic high signal inten-
sity on T2-weighted fast spin echo MR images and that
the high signal intensity corresponds to the tumor mucin
pool (12,13). However, the diagnostic accuracy of MRI
for distinguishing mucinous carcinomas from nonmuci-
nous carcinomas has not been assessed in a large series.
The purpose of this study was to determine the diagnos-
tic accuracy of MRI for the differentiation of mucinous
and nonmucinous rectal carcinomas on preoperative
MRI.

48
 MRI OF MUCINOUS/NONMUCINOUS RECTAL CARCINOMA 49

TABLE 1. Sensitivity, specificity, and accuracy of diagnosis of mucinous carcinoma

Diagnosis of mucinous carcinoma Identification of mucin pool
(n ⳱ 16) (n ⳱ 28)
Radiologist 1 Radiologist 2 Radiologist 1 Radiologist 2
True positive 15 16 19 28
True negative 59 57 45 39
False negative 1 0 9 0
False positive 1 3 3 9
Total 76 76 76 76
Accuracy 0.97 0.96 0.84 0.88
Sensitivity 0.94 1.00 0.68 1.00
Specificity 0.98 0.95 0.94 0.81
Positive predictive value 0.94 0.84 0.86 0.76
Negative predictive value 0.98 1.00 0.83 1.00

MATERIALS AND METHODS or suspected rectal carcinoma and subsequently under-

went surgery. There were 40 men and 36 women with
This study included 76 patients who were referred for ages ranging from 27 to 80 years (mean 59 years).
MR examination for the preoperative staging of known Two pathologists who had no knowledge of the MR

FIG. 1. A 57-year-old woman with a mucinous carcinoma initially diag-

nosed as nonmucinous adenocarcinoma. A: The axial fast spin echo T2-
weighted image depicts a large area of hyperintensity extending to the
perirectal space, suggestive of a mucin pool (arrow). Both radiologists in-
terpreted the tumor as a mucinous carcinoma. B: On the axial spin echo
T1-weighted image, the tumor shows signal intensity similar to the normal
rectum or muscles. C: Coronal fast spin echo T2-weighted images show the
tumor extending into the perirectal fat plane. D: Review of the pathologic
slide showed a large mucin pool volume (arrows); hence, the diagnosis was
changed to mucinous carcinoma. Hematoxylin/eosin; no magnification.

J Comput Assist Tomogr, Vol. 27, No. 1, 2003

50 M.-J. KIM ET AL.

findings jointly reviewed retrospectively the histopatho-
logic findings in all patients, with a special emphasis on
the mucin content of the mucinous carcinoma. Although
the original pathologic reports revealed mucin pools
within the tumors of 18 patients, a mucin pool was de-
tected in 10 additional cases upon reexamination. There-
fore, 28 patients (37%) were found to have mucin pools
within the tumor, and among these, 16 (21%) cases ful-
filled the criterion of mucinous rectal carcinoma, having
a content of ⱖ50% of extracellular mucin. Of these 16
patients, there were 5 women and 11 men between 27
and 79 years old, with a mean age 53 years.
MRI was performed with a 1.5 T system (Horizon; GE
Medical Systems, Milwaukee, WI, U.S.A.) in all pa-
tients, using a phased-array pelvic coil. The axial, sagit-
tal, and coronal T2-weighted (4,800–7,500/102 ms) im-
ages were obtained using a fast spin echo sequence with
the following parameters: a 20 to 24 × 20 to 24 cm field
of view, two signals averaged, a 512 × 256 matrix, an
echo train length of 16, a bandwidth of 31.3 kHz, and a
5 mm section thickness with a 1.5 mm gap. Axial T1-
weighted images were also obtained using a spin echo
sequence (400–500/8–10 ms, matrix size 256 × 192) with
the same field of view, scan thickness, and gap as used
for the T2-weighted axial images. Because the tumor was
better defined in the distended rectal lumen, tap water
was administered using a rectal tube, and the rectum was
filled until the patient indicated discomfort. The normal
volume used was between 200 and 500 mL. To minimize
peristalsis and to alleviate rectal spasm, 20 mg of sco-
polamine butylbromide (Buscopan; Boehringer Ingel-
heim, Germany) was injected intravenously.
Two radiologists with different experience (Radiolo-
gist 1, 8 years; Radiologist 2, 1 year) independently re-
viewed the films in a random order and without prior
knowledge of the histologic tumor type. Both radiolo-
gists regarded the areas of high signal intensity within
the tumor on T2-weighted images as being indicative of
a mucin pool (12,13). A high signal intensity was defined
as a signal intensity that was similar to or brighter than
the perirectal fat on the fast spin echo T2-weighted im-
ages. Each observer recorded whether the tumors con-
tained a mucin pool and determined the histologic type
of the tumor as “mucinous” or “nonmucinous” according
to the same criteria used for histologic diagnosis. Al-
though the assessments were made on the T2-weighted
image findings, the T1-weighted images were viewed
simultaneously.
Sensitivity, specificity, positive predictive value, and
negative predictive value for the diagnosis of mucinous
carcinoma and the identification of mucin pools in the
tumor were separately calculated for each radiologist. ␬
statistics were used to estimate the level of agreement
between the two radiologists in identification of a mucin
pool and diagnosing a mucinous carcinoma. ␬ values of
>0 were considered to indicate a positive correlation.
Values up to 0.4 were considered to indicate a marginal
correlation, values of 0.41–0.7 a good correlation, and
values of >0.7 an excellent correlation.
RESULTS
The accuracy, sensitivity, specificity, positive predic-
tive value, and negative predictive value for the identi-
fication of a mucin pool and the diagnosis of a mucinous
carcinoma are presented in Table 1. Interobserver agree-
ment was good (␬ ⳱ 0.61) for the detection of a mucin

FIG. 2. False-negative diagnosis of mucinous carcinoma by Radiologist 1 in a 67-year-old man. Axial (A) and sagittal (B) fast spin echo
T2-weighted images depict an area of hyperintense mucin pool (arrows). Both radiologists identified mucin pools in the tumor. Radiologist
2 estimated the volume of the mucin pool as >50% of the tumor, which agreed with the results of the pathologic examinations. Radiologist
1 judged that the mucin pool volume was insufficient for a diagnosis of mucinous carcinoma.

J Comput Assist Tomogr, Vol. 27, No. 1, 2003

 MRI OF MUCINOUS/NONMUCINOUS RECTAL CARCINOMA 51

pool and was excellent (␬ ⳱ 0.81) for the diagnosis of a
mucinous carcinoma (Fig. 1).
In the diagnosis of a mucinous carcinoma, false-
negative diagnosis was rendered in one case by Radiolo-
gist 1 and none by Radiologist 2. In the case that Radi-
ologist 1 rendered a false-negative diagnosis, the pres-
ence of the mucin pool was correctly identified;
however, the estimation of the mucin pool volume was
discordant with the pathologists’ determination (Fig. 2).
False-positive diagnoses of mucinous carcinoma were
rendered in three cases: one by Radiologist 1 and three
by Radiologist 2 (one case for both radiologists). Among
the three cases, mucin pools were detected in the patho-
logic specimens in two cases, but pathologists estimated
the mucin pool volume to be <50% of the tumor volume
(Fig. 3). Therefore, the estimation of the mucin pool
volume was discordant between the radiologists and pa-
thologists in these two cases. In the remaining one, sub-
mucosal inflammatory edema in the adjacent rectal wall,
not in the tumor, was regarded as a mucin pool by Ra-
diologist 2. Radiologist 1 correctly distinguished the tu-
mor and the peritumoral edema in this case (Fig. 4).
In regard to identifying the mucin pools in the tumors,
there were nine false-negative results for Radiologist 1
and none for Radiologist 2. In all nine cases, the propor-
tion of the mucin pools in the tumors was insufficient for
a diagnosis of mucinous carcinoma. Therefore, Radiolo-
gist 1 correctly judged all nine cases as nonmucinous
carcinomas, whereas Radiologist 2 falsely judged one of
them as mucinous carcinoma. A false-positive result for
the identification of mucin pool was made in 10 cases: 3
for Radiologist 1 and 9 for Radiologist 2 (2 cases for
both). In these cases, intratumoral congestion or abscess,
liquefactive necrosis, mural edema in the adjacent rectal
wall, and entrapped fluid between the tumor and the
adjacent rectal wall were the causes of the false-positive
identification of mucin pools (Fig. 5). However, the high
signal intensity portion that was falsely interpreted as a

FIG. 3. False-positive diagnosis of mucinous carcinoma in a 45-year-old man.

Axial (B), sagittal (B), and coronal (C) fast spin echo T2-weighted images depict an
area of hyperintense mucin pool (arrows) and a relatively hypointense nonmuci-
nous component (arrowheads). Both radiologists interpreted the tumor as mucin-
ous carcinoma. At the pathologic examination (D), foci of the mucin pools were
seen in the tumor (arrows), but pathologists judged that the mucin pool volume
was insufficient to define the tumor as mucinous. Hematoxylin–eosin stain; no
magnification.

J Comput Assist Tomogr, Vol. 27, No. 1, 2003

52 M.-J. KIM ET AL.

FIG. 4. A 70-year-old woman with a nonmucinous carcinoma. Sagittal (A) and coronal (B) fast spin echo T2-weighted images. Radiologist
2 considered the areas of the high signal intensity of submucosal edema in the adjacent rectal wall as a mucin pool (arrowheads) and
rendered a false-positive diagnosis of mucinous carcinoma. Radiologist 1 correctly identified the hypointense tumor narrowing a segment
of the rectum (arrows) in the absence of a hyperintense mucin pool. Hence, the tumor was correctly interpreted as a nonmucinous tumor
by Radiologist 1 because the extratumoral location of the high signal intensity was correctly identified.

mucin pool comprised only a small portion of the tumor
in most cases. Therefore, for Radiologist 1, a true-
negative diagnosis was made in determining the histo-
logic type of the tumor in all three cases. For Radiologist
2, a false-positive diagnosis of mucinous carcinoma was
made in one of them.
DISCUSSION
There has been some controversy regarding the prog-
nostic significance of mucinous carcinomas of colon and
rectum (14,15). The different results obtained may be
explained by inferences made from the site or based
upon the DNA microsatellite instability of the tumor
(16). Mucinous carcinomas of proximal colon are likely
to have high microsatellite instability and to be diploid
and therefore associated with a good prognosis (17,18).
However, mucinous carcinomas of the rectum are likely
to be DNA microsatellite stable and aneuploid. More-
over, they tend to show more extensive invasion at the
time of presentation and to have a poorer prognosis than
the nonmucinous type (19,20). The production of mucus
under pressure allows the cancers to separate tissue
planes in the bowel wall and thus to more frequently gain
access to the peritoneal cavity. In addition, the fluid pro-
duced by these tumors is taken up by lymphatics, which
helps propel the tumor into the regional lymph nodes
(21). Another implication of mucinous rectal carcinomas
is that they may present as a perirectal cystic mass be-
cause of their ability to easily penetrate the outside rectal
wall. Therefore, knowledge of the imaging features of
these tumors is important for radiologists and surgeons
during treatment planning.
The MR findings of colorectal mucinous carcinoma of
colorectum and other organs such as the liver, pancreas,
and prostate have been previously reported (22–25). Kim
et al. (12) studied 12 cases of colorectal carcinoma con-
taining various amounts of mucin pools. In that study, all
mucinous carcinomas, containing mucin pools of >50%
of the tumor by definition, showed hyperintense area on
T2-weighted MR images, which correlated well with the
area of the mucin pool within the tumor. Hussain et al.
(13) also studied 9 mucinous and 17 nonmucinous car-
cinomas with MRI. They found that the relative signal
intensity of the tumor was significantly higher in the
mucinous compared with the nonmucinous tumors. With
this finding, they suggested that MRI can help distin-
guish between mucinous and nonmucinous carcinomas.
However, this has not been demonstrated yet in a larger
study.
Our study shows that both experienced and less expe-
rienced radiologists accurately distinguish mucinous car-
cinoma from nonmucinous carcinoma with MRI. In-
terobserver agreement was also excellent in the differen-
tiation of mucinous and nonmucinous carcinoma. There
were three false-positive and one false-negative case for
the diagnosis of mucinous carcinomas. The causes of the
false diagnoses in three of the four cases were the dis-
cordance between the radiologists and the pathologists in
the visual estimates of the mucin pools in the tumors.
This was because the criterion of a mucinous carcinoma
is based on the visual estimation of the amount of the
extracellular mucin in the tumor. Therefore, careful in-
spection of the whole tumor volume may be required to
obtain correct diagnosis. In this regard, MRI may have
an advantage over a pathologic examination because the

J Comput Assist Tomogr, Vol. 27, No. 1, 2003

 MRI OF MUCINOUS/NONMUCINOUS RECTAL CARCINOMA 53

FIG. 5. A 59-year-old man with a nonmucinous carcinoma. A and B: Axial (A) and coronal (B) fast spin echo T2-weighted images. Both
radiologists considered the high signal intensity area near the outer section of the rectal carcinoma as a mucin pool (arrowheads). C: Axial
T1-weighted spin echo image depicted the area as a hypointensity (arrowhead). D: Upon pathologic examination, areas of intramural
abscess (thick arrow) and congestion in the perirectal vein (thin arrow) were noted. Hematoxylin–eosin; no magnification. Radiologist 1
correctly assessed the tumor as a nonmucinous carcinoma. However, Radiologist 2 interpreted the tumor as a mucinous carcinoma.

whole volume of the tumor can be inspected preopera-
tively or retrospectively at MRI. Thorough examination
of the entire volume of the tumor from the resected speci-
men may be difficult in clinical practice.
As compared with the accuracy in diagnosis of a mu-
cinous carcinoma, the accuracy in the identification of
the mucin pool was relatively lower. The interobserver
agreement for the identification of a mucin pool was also
relatively poorer than that for the histologic diagnosis.
For the radiologist with a high degree of experience, the
most common errors in the identification of the mucin
pools resulted from neglect of the small areas of high
signal intensity, whereas most of the errors by the radi-
ologist with less experience resulted from the false-
positive interpretation of areas of high signal intensities
mimicking mucin pools in nonmucinous carcinomas,
such as intratumoral congestion, abscess, necrosis, and
mural edema in the adjacent rectal wall or entrapped
fluid between the tumor and the adjacent rectal wall.
These areas were relatively small and scarce within the

J Comput Assist Tomogr, Vol. 27, No. 1, 2003

54 M.-J. KIM ET AL.
tumor or located mainly outside of the tumor. The radi-
ologist with more experience showed better ability to
differentiate those nonmucinous high signal intensities
from a true mucin pool. However, he showed a tendency
to ignore small mucin pools that were deemed to be
insignificant. The portion of the mucin pools that was
missed at MRI or the high signal intensity that was
falsely interpreted as a mucin pool usually comprised
only a small portion of the tumors. Therefore, false-
positive or -negative diagnosis of mucinous carcinoma
did not occur in the majority of cases for both radiolo-
gists. Additional use of gadolinium-enhanced sequences
may improve radiologists’ ability to differentiate be-
tween a mucin pool and edema, necrosis, or abscess be-
cause the tumor mucin pool may be enhanced, whereas
the others do not enhance (13). However, the results of
the current study showed that unenhanced images alone
showed high accuracy for the differentiation of mucinous
and nonmucinous tumors. In addition, gadolinium-
enhanced images have been reported not to be useful in
improving the staging accuracy of rectal carcinomas
(26).
One interesting finding was that the pathologic diag-
nosis could be altered at the retrospective examination.
As compared with the original pathologic reports, intra-
tumoral mucin pools were detected in 10 additional cases
and the diagnosis was changed to mucinous carcinoma in
5. The reason may be that the detection of the mucin
pools in the pathologic specimen was not a large concern
on routine pathologic examination. Previous reports also
noted that it might be difficult to diagnose mucinous
carcinomas based only on the preoperative biopsy speci-
mens (3,12). Our study additionally showed that a diag-
nosis of a mucinous carcinoma might be missed even on
resected surgical specimens. Therefore, radiologic detec-
tion of mucin pools in preoperative MRI may improve
the sensitivity of mucin pool detection in a pathologic
diagnosis.
In our series, mucin pools were frequently detected in
the portion of the tumor that extruded from the rectal
wall. In some cases where the tumor extended outside the
rectal wall, its extrarectal components were composed
largely of mucin pools (Fig. 1). This finding may reflect
previous reports suggesting that mucinous carcinoma has
a tendency to disseminate and infiltrate more rapidly
than nonmucinous carcinoma (7,21). It is believed that
this finding may be attributed to the greater ability of the
relatively softer mucin-containing part of tumor to infil-
trate the perirectal fat plane than its solid nonmucinous
component. Moreover, this perirectally extended mucin-
containing tumor would be difficult to observe during
surgery if it had not been noticed preoperatively. There-
fore, perirectal extension of the mucin pool may contrib-
ute to the relatively high rates of local failure and poorer
prognosis compared with nonmucinous tumors, as ob-
served in some previous surgical reports.
One of the limitations of this study was that MR ex-
aminations are not widely used in the preoperative evalu-
ation of colorectal carcinoma in current clinical practice.
J Comput Assist Tomogr, Vol. 27, No. 1, 2003
However, MR examinations are being used increasingly
for the preoperative evaluation of rectal carcinomas and
have been reported to be superior to CT at predicting
tumor infiltration (27–29).
In conclusion, our study shows that MRI is accurate in
preoperatively distinguishing between mucinous and
nonmucinous carcinoma by identification of the high sig-
nal intensity area of the mucin pools in the rectal carci-
nomas on T2-weighted images. Therefore, this ability of
MRI to histologically differentiate between rectal carci-
nomas should be incorporated into the preoperative MR
evaluation of rectal carcinomas.
REFERENCES
1. Hanski C. Is mucinous carcinoma of colorectum a distinct genetic
entity? Br J Cancer 1995;72:1350–6.
2. Connelly JH, Robey-Cafferty SS, Cleary KR. Mucinous carcino-
mas of the colon and rectum. An analysis of 62 stage B and C
lesions. Arch Pathol Lab Med 1991;115:1022–5.
3. Younes M, Katikaneni PR, Lechago J. The value of the preopera-
tive mucosal biopsy in the diagnosis of colorectal mucinous ad-
enocarcinoma. Cancer 1993;72:3588–92.
4. Pihl E, Nairn RC, Hughes ES, et al. Mucinous colorectal carci-
noma: immunopathology and prognosis. Pathology 1980;12:
439–47.
5. Okuno M, Ikehara T, Nagayama M, et al. Mucinous colorectal
carcinoma: clinical pathology and prognosis. Am Surg 1988;54:
681–5.
6. Sadahiro S, Ohmura T, Saito T, et al. An assessment of the mu-
cinous component in carcinoma of the colon and rectum. Cancer
1989;64:1113–6.
7. Yamamoto S, Mochizuki H, Hase K, et al. Assessment of clinico-
pathologic features of colorectal mucinous adenocarcinoma. Am J
Surg 1993;166:257–61.
8. Secco GB, Fardelli R, Campora E, et al. Primary mucinous adeno-
carcinomas and signet-ring cell carcinomas of colon and rectum.
Oncology 1994;51:30–4.
9. Consorti F, Lorenzotti A, Midiri G, et al. Prognostic significance of
mucinous carcinoma of colon and rectum: a prospective case-
control study. J Surg Oncol 2000;73:70–4.
10. Nozoe T, Anai H, Nasu S, et al. Clnicopathologic characteristics of
mucinous carcinoma of the colon and rectum. J Surg Oncol 2000;
75:103–7.
11. Masaki T, Sugiyama M, Atomi Y, et al. The indication of local
excision for T2 rectal carcinomas. Am J Surg 2001;181:133–7.
12. Kim MJ, Huh YM, Park YN, et al. Colorectal mucinous carci-
noma: findings on MRI. J Comput Assist Tomogr 1999;23:291–6.
13. Hussain SM, Outwater EK, Siegelman ES. Mucinous versus non-
mucinous rectal carcinomas: differentiation with MR imaging. Ra-
diology 1999;213:79–85.
14. Wu CS, Tung SY, Chen PC, et al. Clinicopathologic study of
colorectal mucinous carcinoma in Taiwan: a multivariate analysis.
J Gastroenterol Hepatol 1996;11:77–81.
15. Purdie CA, Piris J. Histopathologic grade, mucinous differentiation
and DNA ploidy in relation to prognosis in colorectal carcinoma.
Histopathology 2000;36:121–6.
16. Jass JR. Is mucinous differentiation of colorectal cancer a prog-
nostic feature? Histopathology 2000;37:474–5.
17. Jass JR, Do KA, Simms LA, et al. Morphology of sporadic colo-
rectal cancer with DNA replication errors. Gut 1998;42:673–9.
18. Halling KC, French AJ, McDonnell SK, et al. Microsatellite in-
stability and 8p allelic imbalance in stage B2 and C colorectal
cancers. JNCI 1999;91:1295–303.
19. Sasaki O, Atkin WS, Jass JR. Mucinous carcinoma of the rectum.
Histopathology 1987;11:259–72.
20. Symonds DA, Vickery AL. Mucinous carcinoma of the colon and
rectum. Cancer 1976;37:1891–900.
 MRI OF MUCINOUS/NONMUCINOUS RECTAL CARCINOMA
21. Sugarbaker PH. Mucinous colorectal carcinoma. J Surg Oncol
2001;77:282–3.
22. Outwater EK, Tomaszewski JE, Daly JM, et al. Hepatic colorectal
metastases: correlation of MR imaging and pathologic appearance.
Radiology 1991;180:327–32.
23. Outwater E, Schiebler ML, Tomaszewski JE, et al. Mucinous car-
cinomas involving the prostate: atypical findings at MR imaging. J
Magn Res Imag 1992;2:597–600.
24. Mergo PJ, Helmberger TK, Buetow PC, et al. Pancreatic neo-
plasms: MR imaging and pathologic correlation. Radiographics
1997;17:281–301.
25. Hayashi M, Matsui O, Ueda K, et al. Imaging findings of mucinous
type of cholangiocellular carcinoma. J Comput Assist Tomogr
1996;20:386–9.
55
26. Okizuka H, Sugimura K, Yoshizako T, et al. Rectal carcinoma:
prospective comparison of conventional and gadopentetate dime-
glumine enhanced fat-suppressed MR imaging. J Magn Res Imag
1996;6:465–71.
27. Kim NK, Kim MJ, Yun SH, et al. Comparative study of transrectal
ultrasonography, pelvic computerized tomography, and magnetic
resonance imaging in preoperative staging of rectal cancer. Dis
Colon Rectum 1999;42:770–5.
28. Wallengren NO, Holtas S, Andren-Sandberg A, et al. Rectal car-
cinoma: double-contrast MR imaging for preoperative staging. Ra-
diology 2000;215:108–14.
29. Beets-Tan RG, Beets GL, Borstlap AC, et al. Preoperative assess-
ment of local tumor extent in advanced rectal cancer: CT or high-
resolution MRI? Abdom Imag 2000;25:533–41.
J Comput Assist Tomogr, Vol. 27, No. 1, 2003