Cystic Changes in Intraabdominal Extrahepatic Metastases ... · Cystic Changes in Intraabdominal...
Transcript of Cystic Changes in Intraabdominal Extrahepatic Metastases ... · Cystic Changes in Intraabdominal...
Korean J Radiol 5(3), September 2004 157
Cystic Changes in IntraabdominalExtrahepatic Metastases fromGastrointestinal Stromal Tumors Treatedwith Imatinib
Objective: This study was undertaken for the purpose of describing the CTfeatures of intra-abdominal extra-hepatic metastases from gastrointestinal stro-mal tumors in patients who were treated with imatinib.
Materials and Methods: Eleven patients with intra-abdominal extra-hepaticmetastases from gastrointestinal stromal tumors, who were treated with imatinibbetween May 2001 and December 2003, were included in this study. The clinicalfindings and CT scans were retrospectively reviewed. The metastatic lesionswere assessed according to the location, size (greatest diameter), attenuation,and the enhancing pattern before and after imatinib treatment.
Results: Prior to the treatment, the sizes and attenuation values of themetastatic lesions ranged from 5 to 20 cm and from 63 to 131 H, respectively.The metastatic lesions showed a heterogeneous enhancement pattern on thecontrast-enhanced CT scans. After the treatment, the metastatic lesions becamesmaller in all 11 patients, and the corresponding attenuation value ranged from15 to 51 H. The metastatic lesions became homogeneous and cystic in appear-ance on the follow-up CT scans, mimicking ascites.
Conclusion: Intra-abdominal extra-hepatic metastases of patients with gas-trointestinal stromal tumors treated with imatinib may appear as well-circum-scribed cystic lesions on contrast-enhanced CT. These metastases are likely tobecome smaller and resemble ascites, but may persist indefinitely on the follow-up CT.
astrointestinal stromal tumors, formerly classified as leiomyomas orleiomyosarcomas, constitute the most common form of mesenchymaltumor in the gastrointestinal tract, with the stomach being the most
common site of origin (1). The diagnosis of these tumors became feasible through theapplication of CD117 immunohistochemistry, which allows these neoplasms to bedistinguished from leiomyomas or leiomyosarcomas (2). Moreover, the recentlydeveloped KIT-tyrosine kinase inhibitor (STI 571, imatinib [Gleevec], Novartis,Basel, Switzerland) has dramatically improved the treatment of gastrointestinalstromal tumors (3).
The radiologic findings of gastrointestinal stromal tumors were recently described inthe radiologic literature (1, 4 8), and are ostensibly similar to those of the previouslydescribed leiomyomas or leiomyosarcomas (9). The characteristic features of thisdisease are its recurrence at the primary site and the presence of metastases, primarilyin the liver and the peritoneum (4). Recently, it was reported that the hepaticmetastases in patients treated with imatinib resembled cystic lesions (10 12). In linewith these observations, our findings indicate that when gastrointestinal stromal
Hyo-Cheol Kim, MD1
Jeong Min Lee, MD1
Seung Hong Choi, MD1
Heon Han, MD2
Sam Soo Kim, MD2
Sang Hyun Lee, MD3
Joon Koo Han, MD1
Byung Ihn Choi, MD1
Index terms:Gastrointestinal stromal tumorNeoplasms, metastasesAbdomen, CT
Korean J Radiol 2004;5:157-163Received October 19, 2003; accepted after revision February 9, 2004.
1Department of Radiology, Seoul NationalUniversity College of Medicine, Instituteof Radiation Medicine, SNUMRC, andClinical Research Institute, SeoulNational University Hospital; 2Departmentof Radiology, Kangwon NationalUniversity College of Medicine; 3RadiationMedicine Branch, National Cancer Center
This study was supported in part by the2003 BK21 Project for Medicine,Dentistry and Pharmacy.
Address reprint requests to:Jeong Min Lee, MD, Department ofRadiology, Seoul National UniversityHospital, 28 Yongon-dong, Chongno-gu,Seoul, 110-744, Korea.Tel. (822) 760-2584Fax. (822) 743-6385e-mail: [email protected]
G
Kim et al.
158 Korean J Radiol 5(3), September 2004
Tabl
e 1.
Sum
mar
y of
the
11 P
atie
nts
with
Intr
aabd
omin
al E
xtra
hepa
tic M
etas
tasi
s fr
om G
astr
oint
estin
al S
trom
al T
umor
Tre
ated
with
Imat
inib
Patie
ntAg
e/Si
te o
fSi
te o
fC
linic
alIn
itial F
indi
ngs
of M
etas
tasi
s1s
t F/U
Last
F/U
und
er Im
atin
ib th
erap
yC
omm
ent
Num
ber
Sex
Prim
ary
Met
asta
ticSy
mpt
omM
etas
tatic
In
terv
alH
ouns
field
Atte
nuat
ion
Size
Inte
rval
Hou
nsfie
ldAt
tenu
atio
nSi
ze
Inte
rval
Hou
nsfie
ldAt
tenu
atio
n
Tum
orLe
sion
sLe
sion
betw
een
Mea
sure
men
tsPa
ttern
afte
rbe
twee
nM
easu
rem
ents
Patte
rnaf
ter
betw
een
Mea
sure
men
tsPa
ttern
Size
(cm
)O
pera
tion
and
Imat
inib
Initia
lIm
atin
ibIn
itial
Det
ectio
n of
Ther
apy
Det
ectio
nTh
erap
yD
etec
tion
Met
asta
sis
(cm
)an
d 1s
t F/U
(cm
)an
d La
st F
/U
(mon
ths)
(wee
ks)
(mon
ths)
0159
/MSt
omac
hLo
cal
Abdo
min
al15
0773
Het
ero
0308
36H
omo
< 1
22H
omo
recu
rpa
in
0246
/MSt
omac
hPe
riton
eal
Rou
tine
0512
72H
eter
o03
0825
Hom
o<
110
Hom
oAg
grav
ated
seed
ing,
F/U
11 m
onth
s
liver
late
r
0365
/MSt
omac
hPe
riton
eal
Abdo
min
al20
0988
Het
ero
1504
39H
eter
oF/
U lo
ss
seed
ing,
pain
liver
0426
/MSt
omac
hLo
cal
Rou
tine
0925
106
Het
ero
0310
25H
omo
205
18H
omo
recu
rF/
U
0568
/MSt
omac
hPe
riton
eal
Rou
tine
1512
81H
eter
o08
0828
Hom
o5
1320
Hom
o
seed
ing
F/U
0642
/MR
ectu
mPe
riton
eal
Rou
tine
0538
67H
eter
o03
0839
Hom
o1
1017
Hom
o
seed
ing,
F/U
liver
0752
/MM
esen
tery
Perit
onea
lR
outin
e07
2813
1H
eter
o05
0451
Het
ero
1.5
0822
Hom
o
seed
ing
F/U
0858
/MM
esen
tery
Perit
onea
lAb
dom
inal
1015
79H
eter
o04
1033
Het
ero
404
21H
omo
seed
ing
pain
0941
/MSm
all
Perit
onea
l R
outin
e 13
3574
Het
ero
1004
15H
omo
306
15H
omo
bow
else
edin
gF/
U
1058
/MSm
all
Perit
onea
lAb
dom
inal
0518
63H
eter
o03
0844
Hom
o<
114
Hom
o
bow
else
edin
gpa
in
1134
/MSm
all
Perit
onea
lR
outin
e10
1682
Het
ero
0508
35H
omo
311
28H
omo
Aggr
avat
ed
bow
else
edin
g,F/
U2
mon
ths
liver
late
r
tumors are treated with imatinib, the intra-abdominalextra-hepatic metastases come to have a homogeneous oralmost homogeneous low density, simulating cystic massesor ascites on CT.
MATERIALS AND METHODS
From May 2001 to December 2003, 11 patients withintra-abdominal extra-hepatic metastases from gastroin-testinal stromal tumor were treated with imatinib. Theirmean age was 50 years, ranging from 34 to 68 years, andall were men. The mean time between the initial diagnosisof the gastrointestinal stromal tumor and the diagnosis ofmetastases was 19 months (range, 7 38 months). Theprimary sites were the stomach (n=5), the small bowel(n=3), the rectum (n=1), and the mesentery (n=2). In allpatients, the primary tumors were removed by curativesurgery. The presence of metastatic lesions was confirmedby biopsy (n=5) or radiologic studies and clinical follow-up(n=6). The criterion for the diagnosis of metastasis was thepresence of new lesions detected in the follow-up CT scan,which were not detected in the initial CT scan and which
changed in size on the follow-up CT scan after imatinibtreatment. In all patients, CD117 expression in the primarygastrointestinal stromal tumors (n=6) or metastatic lesions(n=5) was proved by immunohistochemical studies. Allpatients were treated with the oral administration of 400mg of imatinib daily. The clinical symptoms, duration ofimatinib therapy, and CT scan reports were reviewed. Theinstitutional review board at our hospital did not requireapproval or informed patient consent for the review of themedical records and images.
All patients underwent CT scans prior to the administra-tion of imatinib and 4 10 weeks after the initiation of theimatinib therapy. Ten patients underwent two or morefollow-up CT scans 4 22 months after the initiation of thetreatment with imatinib. The CT scan data were availableon a picture archiving and communications system (PACS;Marotech, Seoul, Korea) for all patients. The CT scanswere performed using a Somatom Plus-4 (Siemens MedicalSystems, Erlangen, Germany), a HiSpeed Advantagescanner (General Electric Medical Systems, Milwaukee,WI), or a MX8000 four-detector row CT scanner (PhilipsMedical Systems, Cleveland, OH). Each patient received
Imatinib Causes Cystic Changes in Intraabdominal Extrahepatic Metastases from GI Tract Stromal Tumors
Korean J Radiol 5(3), September 2004 159
A B
Fig. 1. A 34-year-old man with peritoneal seeding and livermetastases after resection of gastrointestinal stromal tumor of theduodenum.A. CT scan before treatment shows multiple heterogeneousmasses (82 H) (M) compressing inferior vena cava (arrowhead)and superior mesenteric vein (arrow). Metastatic lesion in liver(curved arrow).B. CT scan obtained after 8 weeks of treatment with imatinibshows peritoneal and hepatic metastases that have decreased insize and density (35 H). Note the decompressed inferior vena cavaand superior mesenteric vein and shrunk metastatic lesions in theliver.C. CT scan obtained 2 months after stopping imatinib therapyshows increased size and density of peritoneal and hepaticmetastases. After resumption of imatinib therapy, the metastaticlesions showed cystic changes again (now shown).
C
120 mL of a nonionic contrast material (Iopromide,Ultravist 370; Schering Korea, Seoul, Korea) through an18-gauge angiographic catheter inserted into a forearmvein. The contrast material was injected at a rate of 2.5mL/sec using an automatic injector. In the case of thesingle-detector scanner, a helical CT scan was performedwith the following parameters: 5 7 mm collimation, 1:1table pitch, and 5 7 mm reconstruction intervals. In thecase of the MX8000 scanner, the parameters were 2.5 mmdetector collimation, 20 mm/sec table speed, 5 mm slicethickness, and a 5 mm reconstruction interval. The delaybetween the contrast material administration and scanningwas 55 70 seconds.
Two radiologists reviewed all of the CT scans retrospec-tively, and the final interpretations were reached byconsensus. All images were reviewed on a 2,000 2,000PACS monitor. The presence of the metastatic lesion andits size before and after the imatinib treatment werecompared. The metastatic lesions were assessed accordingto their location, size (greatest diameter), attenuation, and
enhancing pattern. If multiple metastatic lesions weredetected, the largest lesion was recorded. For the objectiveanalysis, the CT attenuation value was measured in acircular region of interest with a diameter of 10 mm. TheCT attenuation value was measured three times by a singleradiologist and the mean value was recorded. In the case ofa heterogeneous mass, the CT attenuation value wasmeasured in the solid portion of the tumor. The CT attenu-ation value before and after the imatinib treatment wascompared using the paired t-test. Statistical analyses wereperformed using a computer software package (SPSS,version 10.0; SPSS, Chicago, Ill). A p value of less than .05was considered to indicate a statistically significant differ-ence.
RESULTS
The clinical and radiologic findings are summarized inTable 1. One patient (patient 3) was lost to follow-up after1 month of imatinib therapy. Two patients (patients 2 and
Kim et al.
160 Korean J Radiol 5(3), September 2004
A B
Fig. 2. A 68-year-old man with peritoneal seeding after resection ofgastrointestinal stromal tumor of the stomach.A. CT scan before treatment shows 15 13 cm heterogeneousmetastatic lesion (81 H) (G) in left subphrenic space. Note smallmetastatic nodule (arrow) in right subphrenic space and ascites (15H).B. CT scan obtained after 8 weeks of treatment with imatinibshows metastatic lesion (G) that has decreased in size to 8 6 cmand is cyst-like in appearance lesion (28 H) around spleen (S).Note the disappearance of the ascites.C. CT scan obtained after 13 months of treatment with imatinibshows 5 3.5 cm cystic lesion (20 H).
C
11) stopped imatinib therapy 10 and 11 months, respec-tively, after the initiation of the treatment. The remaining8 patients were under imatinib therapy for periods rangingfrom 4 to 22 months at the time this article was written.
On the contrast-enhanced CT, the metastatic lesionswere detected in the peritoneal cavity (n=9), and at thesurgical bed of the primary site (n=2). In four patients,metastasis was also detected in the liver. Prior to thetreatment, the mean size of the metastatic lesions was 10.4
4.9, ranging from 5 to 20 cm, and they showed a hetero-geneous enhancement pattern on the contrast-enhancedCT scans.
After the treatment, the mean size of the metastaticlesions was 5.8 3.6, ranging from 3 to 15 cm, on the firstfollow-up CT scan, showing a reduction in size for all 11patients. On the first follow-up CT scan, the attenuation ofthe metastatic lesions was homogeneous in eight patients(Figs. 1 and 2), and heterogeneous in three patients (Fig.3). In cases of peritoneal seeding, the metastatic lesionsdeveloped a cystic appearance, mimicking ascites (Figs. 2and 3). In reviewing the original CT reports, it was foundthat the cystic change of the tumor was described as ascitesor fluid collection in three patients.
Prior to the treatment, the mean attenuation value of themetastatic lesions was 83 20 H, ranging from 63 to 131H. On the first follow-up CT scan, the mean attenuationvalue was 34 13 H, ranging from 15 to 51 H. This differ-ence in the mean CT attenuation value was statisticallysignificant (p < 0.01).
On the subsequent CT scans, the metastatic lesionsbecame smaller, homogeneous and cystic during imatinibtherapy. However, they did not disappear completely andwere always detected throughout the study in all patients.
In two patients who showed a heterogeneous enhancementpattern on the first follow-up CT scan, the metastaticlesions became homogeneous on the second follow-up CTscan obtained 3 and 4 months, respectively, after the initia-tion of treatment.
Two patients (patients 2 and 11) stopped imatinibtherapy 10 and 11 months, respectively, after the initiationof the treatment. The disease was found to haveprogressed in these 2 patients 11 and 2 months, respec-tively, after the termination of the treatment, with themetastatic lesions increasing in size and attenuation, andshowing a heterogeneous enhancement pattern on the CTscans (Fig. 1C). These two patients resumed imatinibtherapy, and their metastatic lesions subsequently becamesmaller and homogeneous on the follow-up CT scans.
DISCUSSION
Conventional chemotherapeutic agents are rarelyeffective against gastrointestinal stromal tumors. The newchemotherapeutic agent, imatinib, has been applied andthe results are extremely encouraging. The rationalebehind imatinib treatment for gastrointestinal stromaltumors lies in the fact that the KIT (encodes the humanhomolog of the proto-oncogene c-kit) gene mutation hasbeen detected frequently in gastrointestinal stromaltumors. This mutation induces the constitutive activationof the tyrosine kinase receptor, causing the proliferation oftumor cells (2). Imatinib is highly effective in bringingabout a reduction in KIT tyrosine kinase activity.
Gastrointestinal stromal tumors frequently spread to theliver and the peritoneum (4). On the CT scan of the portalvenous phase, the metastases within the liver are usually
Imatinib Causes Cystic Changes in Intraabdominal Extrahepatic Metastases from GI Tract Stromal Tumors
Korean J Radiol 5(3), September 2004 161
Fig. 3. A 52-year-old man with peritoneal seeding after resection of gastrointestinal stromal tumor of the mesentery.A. CT scan before treatment shows multiple peritoneal implants (arrows) in both paracolic gutters.B. CT scan obtained after 4 weeks of treatment with imatinib shows that the metastatic lesions in the right paracolic gutter have somesolid components, while that in the left paracolic gutter resembles ascites.
A B
heterogeneous and peripherally enhanced, similar toprimary tumors (4). The low attenuation in the center ofthese metastatic lesions often indicates the presence ofnecrosis in the center of the solid mass. The peripheralenhanced portion represents viable solid tumor. Peritonealmetastasis shows a CT appearance similar to that ofmetastasis in the liver.
In the peritoneum, metastatic lesions treated withimatinib may appear as ascites or fluid collection. Inreviewing the original CT reports, we found that the cysticchange of the tumor was described as ascites or fluidcollection in three patients. Although long-term follow-upis needed, metastatic lesions in the peritoneum graduallydecrease in size, although they may persist for months oryears, which is not the case for ascites. The density of themetastases decreased to 15 51 H on the first CT scan afterthe treatment and then to 15 28 H on the follow-up CTscan, which is close to that of ascites. Metastases can bedistinguished from ascites by reviewing the change in theattenuation value and the previous CT scan. Ideally, thescans should be interpreted by a radiologist who is familiarwith scans of peritoneal metastases from gastrointestinalstromal tumors following imatinib treatment, in order toavoid the underestimation of the extent of the tumors.
The mechanism that induces the cystic change afterimatinib treatment is not clear. In several reported cases,histological examination of the tumors treated withimatinib showed areas with extensive necrosis, hyalinizedareas with sparse, scattered tumor cells containing small,condensed nuclei and areas with viable tumor cells (1012).
The optimal duration of the treatment is not yet known(13). It is not clear whether viable tumor cells withmalignant potential persist within the cystic lesions and,consequently, the continuous maintenance of imatinibtreatment is required. In this study, two patients whosemetastatic lesions became small and cystic after imatinibtherapy, experienced aggravation of metastasis aftertermination of the imatinib treatment.
Traditionally, the response to cancer treatment in solidtumors is evaluated by subsequent clinical or radiologicalassessments, and is defined as a significant decrease in themeasurable tumor dimensions. A reduction in the viabletumor cell fraction, however, does not always result in avolume reduction, since tumor tissue can be replaced bynecrotic or fibrotic tissue, and morphological images areoften unable to differentiate between these different tissuetypes. In recent years, metabolic imaging with positronemission tomography (PET) has become increasinglyimportant in cancer management. Although the perfor-mances of PET and CT are comparable in terms of the
process of staging before the initiation of imatinib therapy,PET can evaluate the tumor response as early as 1 weekafter the start of treatment, preceding the CT response byseveral weeks (14). Treatment-induced changes resulting intumor cell death or growth arrest should therefore result ina subsequent reduction in glucose uptake, making thistechnique a sensitive and early marker for responseevaluation.
There are several limitations to this study. First, this wasa retrospective review of cases collected over a number ofyears for which CT scans were performed irregularly,depending on the condition of the patients. Second,unenhanced images were not obtained in all patients, andit is unclear whether or not any subtle enhancementchanges are present in the metastatic lesions. Third, we didnot provide any pathologic correlation in any of the cases.Pathologic correlation with the radiologic findings formetastatic lesions is helpful for clinicians as well as forradiologists.
In conclusion, after treatment with imatinib, responsiveintra-abdominal extra-hepatic metastases of gastrointesti-nal stromal tumors appear as well-defined cystic lesions oncontrast-enhanced CT. These metastases become smallerand resemble ascites, but may be detected for a long timeon the follow-up CT scans.
References1. Levy AD, Remotti HE, Thompson WM, Sobin LH, Miettinen M.
Gastrointestinal stromal tumors: radiologic features withpathologic correlation. RadioGraphics 2003;23:283-304
2. Hirota S, Isozaki K, Moriyama Y, et al. Gain-of-functionmutations of c-kit in human gastrointestinal stromal tumors.Science 1998;279:577-580
3. Demetri GD, von Mehren M, Blanke CD, et al. Efficacy andsafety of imatinib mesylate in advanced gastrointestinal stromaltumors. N Engl J Med 2002;347:472-480
4. Burkill GJ, Badran M, Al-Muderis O, et al. Malignant gastroin-testinal stromal tumor: distribution, imaging features, andpattern of metastatic spread. Radiology 2003;226:527-532
5. Levy AD, Remotti HE, Thompson WM, Sobin LH, Miettinen M.Anorectal gastrointestinal stromal tumors: CT and MR imagingfeatures with clinical and pathologic correlation. AJR Am JRoentgenol 2003;180:1607-1612
6. Nishida T, Kumano S, Sugiura T, et al. Multidetector CT ofhigh-risk patients with occult gastrointestinal stromal tumors.AJR Am J Roentgenol 2003;180:185-189
7. Kim H-C, Lee JM, Kim SH, et al. Primary gastrointestinalstromal tumors in the omentum and mesentery: CT findings andpathologic correlations. AJR Am J Roentgenol 2004;182:1463-1467
8. Kim H-C, Lee JM, Son KR, et al. Gastrointestinal stromaltumors of the duodenum: CT and barium study findings. AJRAm J Roentgenol 2004;183:415-419
9. Megibow AJ, Balthazar EJ, Hulnick DH, Naidich DP, BosniakMA. CT evaluation of gastrointestinal leiomyomas andleiomyosarcomas. AJR Am J Roentgenol 1985;144:727-731
Kim et al.
162 Korean J Radiol 5(3), September 2004
10. Joensuu H, Roberts PJ, Sarlomo-Rikala M, et al. Effect of thetyrosine kinase inhibitor STI571 in a patient with a metastaticgastrointestinal stromal tumor. N Engl J Med 2001;344:1052-1056
11. Chen MY, Bechtold RE, Savage PD. Cystic changes in hepaticmetastases from gastrointestinal stromal tumors (GISTs) treatedwith Gleevec (imatinib mesylate). AJR Am J Roentgenol2002;179:1059-1062
12. Bechtold RE, Chen MY, Stanton CA, Savage PD, Levine EA.Cystic changes in hepatic and peritoneal metastases from
gastrointestinal stromal tumors treated with Gleevec. AbdomImaging 2003;28:808-814
13. Bumming P, Andersson J, Meis-Kindblom JM, et al.Neoadjuvant, adjuvant and palliative treatment of gastrointesti-nal stromal tumours (GIST) with imatinib: a centre-based studyof 17 patients. Br J Cancer 2003;89:460-464
14. Gayed I, Vu T, Iyer R, et al. The role of 18F-FDG PET in stagingand early prediction of response to therapy of recurrentgastrointestinal stromal tumors. J Nucl Med 2004;45:17-21
Imatinib Causes Cystic Changes in Intraabdominal Extrahepatic Metastases from GI Tract Stromal Tumors
Korean J Radiol 5(3), September 2004 163