Efficacy and Safety of Imatinib Mesylate in Advanced Gastrointestinal Stromal Tumors

Abstract

Constitutive activation of KIT receptor tyrosine kinase is critical in the pathogenesis of gastrointestinal stromal tumors. Imatinib mesylate, a selective tyrosine kinase inhibitor, has been shown in preclinical models and preliminary clinical studies to have activity against such tumors.
We conducted an open-label, randomized, multicenter trial to evaluate the activity of imatinib in patients with advanced gastrointestinal stromal tumor. We assessed antitumor response and the safety and tolerability of the drug. Pharmacokinetics were assessed in a subgroup of patients.
A total of 147 patients were randomly assigned to receive 400 mg or 600 mg of imatinib daily. Overall, 79 patients (53.7 percent) had a partial response, 41 patients (27.9 percent) had stable disease, and for technical reasons, response could not be evaluated in 7 patients (4.8 percent). No patient had a complete response to the treatment. The median duration of response had not been reached after a median follow-up of 24 weeks after the onset of response. Early resistance to imatinib was noted in 20 patients (13.6 percent). Therapy was well tolerated, although mild-to-moderate edema, diarrhea, and fatigue were common. Gastrointestinal or intraabdominal hemorrhage occurred in approximately 5 percent of patients. There were no significant differences in toxic effects or response between the two doses. Imatinib was well absorbed, with pharmacokinetics similar to those reported in patients with chronic myeloid leukemia.
Imatinib induced a sustained objective response in more than half of patients with an advanced unresectable or metastatic gastrointestinal stromal tumor. Inhibition of the KIT signal-transduction pathway is a promising treatment for advanced gastrointestinal stromal tumors, which resist conventional chemotherapy.

Full text

472
·
N Engl J Med, Vol. 347, No. 7
·
August 15, 2002
·
www.nejm.org
The New England Journal of Medicine
EFFICACY AND SAFETY OF IMATINIB MESYLATE
IN ADVANCED GASTROINTESTINAL STROMAL TUMORS
G
EORGE
D. D
EMETRI
, M.D., M
ARGARET
VON
M
EHREN
, M.D., C
HARLES
D. B
LANKE
, M.D.,
A
NNICK
D. V
AN
DEN
A
BBEELE
, M.D., B
URTON
E
ISENBERG
, M.D., P
ETER
J. R
OBERTS
, M.D., M
ICHAEL
C. H
EINRICH
, M.D.,
D
AVID
A. T
UVESON
, M.D., P
H
.D., S
AMUEL
S
INGER
, M.D., M
ILOS
J
ANICEK
, M.D., P
H
.D., J
ONATHAN
A. F
LETCHER
, M.D.,
S
TUART
G. S
ILVERMAN
, M.D., S
ANDRA
L. S
ILBERMAN
, M.D., P
H
.D., R
ENAUD
C
APDEVILLE
, M.D., B
EATE
K
IESE
, M.S
C
.,
B
IN
P
ENG
, M.D., P
H
.D., S
ASA
D
IMITRIJEVIC
, P
H
.D., B
RIAN
J. D
RUKER
, M.D., C
HRISTOPHER
C
ORLESS
, M.D.,
C
HRISTOPHER
D.M. F
LETCHER
, M.D.,
AND
H
EIKKI
J
OENSUU
, M.D.
A
BSTRACT
Background
Constitutive activation of KIT receptor
tyrosine kinase is critical in the pathogenesis of gas-
trointestinal stromal tumors. Imatinib mesylate, a se-
lective tyrosine kinase inhibitor, has been shown in
preclinical models and preliminary clinical studies to
have activity against such tumors.
Methods
We conducted an open-label, randomized,
multicenter trial to evaluate the activity of imatinib in
patients with advanced gastrointestinal stromal tumor.
We assessed antitumor response and the safety and
tolerability of the drug. Pharmacokinetics were as-
sessed in a subgroup of patients.
Results
A total of 147 patients were randomly as-
signed to receive 400 mg or 600 mg of imatinib daily.
Overall, 79 patients (53.7 percent) had a partial re-
sponse, 41 patients (27.9 percent) had stable disease,
and for technical reasons, response could not be eval-
uated in 7 patients (4.8 percent). No patient had a com-
plete response to the treatment. The median duration
of response had not been reached after a median fol-
low-up of 24 weeks after the onset of response. Early
resistance to imatinib was noted in 20 patients (13.6
percent). Therapy was well tolerated, although mild-
to-moderate edema, diarrhea, and fatigue were com-
mon. Gastrointestinal or intraabdominal hemorrhage
occurred in approximately 5 percent of patients. There
were no significant differences in toxic effects or re-
sponse between the two doses. Imatinib was well ab-
sorbed, with pharmacokinetics similar to those report-
ed in patients with chronic myeloid leukemia.
Conclusions
Imatinib induced a sustained objec-
tive response in more than half of patients with an ad-
vanced unresectable or metastatic gastrointestinal
stromal tumor. Inhibition of the KIT signal-transduc-
tion pathway is a promising treatment for advanced
gastrointestinal stromal tumors, which resist conven-
tional chemotherapy. (N Engl J Med 2002;347:472-80.)
Copyright © 2002 Massachusetts Medical Society.
From the Dana–Farber Cancer Institute and Harvard Cancer Center,
Boston (G.D.D., A.D.V.A., D.A.T., S.S., M.J., J.A.F., S.G.S., C.D.M.F.); the
Fox Chase Cancer Center, Philadelphia (M.M., B.E.); the Oregon Health
and Science University and Portland Veterans Affairs Medical Center, Port-
land (C.D.B., M.C.H., B.J.D., C.C.); the University of Turku, Turku, Fin-
land (P.J.R.); Novartis Oncology, Basel, Switzerland (S.L.S., R.C., B.K., B.P.,
S.D.); and the University of Helsinki, Helsinki, Finland (H.J.). Address re-
print requests to Dr. Demetri at the Center for Sarcoma and Bone Oncol-
ogy, Dana–Farber Cancer Institute, SW 530, 44 Binney St., Boston, MA
02115, or at gdemetri@partners.org.
Drs. Demetri, von Mehren, Blanke, and Joensuu contributed equally to
the article.
ASTROINTESTINAL stromal tumors are
mesenchymal neoplasms that appear to be
related to the interstitial cells of Cajal of the
myenteric plexus, with which they share cer-
tain differentiation markers.
1,2
Gastrointestinal stromal
tumors express the cell-surface transmembrane recep-
tor KIT that has tyrosine kinase activity and is the pro-
tein product of the
KIT
proto-oncogene. There are
frequent gain-of-function mutations of
KIT
in gastro-
intestinal stromal tumors.
3,4
These mutations result in
the constitutive activation of KIT signaling, which
leads to uncontrolled cell proliferation and resistance
to apoptosis. It has recently been reported that KIT
activation occurs in all cases of gastrointestinal stromal
tumor, regardless of the mutational status of
KIT
.
4
Unresectable or metastatic gastrointestinal stromal
tumor is a fatal disease that resists conventional cyto-
toxic chemotherapy.
5,6
In a recently reported series, the
response rate to doxorubicin was less than 5 percent.
6
The effectiveness of radiation therapy for this disease
has not been proved.
5
The median duration of survival
for patients with a metastatic gastrointestinal stromal
tumor is approximately 20 months, and for patients
with local recurrence it is 9 to 12 months.
5
Imatinib mesylate (formerly STI571, now referred
to as Gleevec in the United States and Glivec in Europe
[Novartis]) is a selective inhibitor of certain protein
tyrosine kinases: the intracellular ABL kinase, the chi-
meric BCR-ABL fusion oncoprotein of chronic mye-
loid leukemia, the transmembrane receptor KIT, and
the platelet-derived growth factor receptors.
7-10
Ima-
tinib is highly active in patients with chronic myeloid
leukemia and other Philadelphia chromosome–pos-
G
Copyright © 2002 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org at HHS LIBRARIES CONSORTIUM on May 12, 2008 .

IMATINIB MESYLATE IN GASTROINTESTINAL STROMAL TUMORS
N Engl J Med, Vol. 347, No. 7
·
August 15, 2002
·
www.nejm.org
·
473
itive leukemias, in which it inhibits the dysregulated
kinase activity of the BCR-ABL fusion protein.
11,12
We
hypothesized that imatinib might also block the con-
stitutive activity of KIT receptor tyrosine kinase in the
cells of gastrointestinal stromal tumors. This hypoth-
esis was supported by experiments in human tumor-
cell lines that are dependent on the KIT pathway. Ex-
posure of these cells to imatinib blocked the kinase
activity of KIT, arrested proliferation, and caused ap-
optotic cell death.
9,10,13
Subsequently, a single patient
treated with imatinib for a chemotherapy-resistant
metastatic gastrointestinal stromal tumor had a rapid,
substantial, and durable response.
14
To build on these
results, we conducted a multicenter clinical trial to test
the efficacy and safety of imatinib in patients with
an unresectable or metastatic gastrointestinal stromal
tumor.
METHODS
Patients
Adults with a histologically confirmed, unresectable or metastatic
gastrointestinal stromal tumor that expressed CD117 (a marker of
KIT-receptor tyrosine kinase) were eligible for the study. Pathology
was reviewed centrally by a single pathologist. Criteria for inclusion
were at least one measurable tumor that had not previously been
treated with radiotherapy or embolization; adequate hepatic, renal,
and cardiac function; an adequate platelet count; and an Eastern
Cooperative Oncology Group (ECOG) performance status of 3 or
lower. Patients were allowed to have received any number of pre-
vious chemotherapeutic regimens (with the last administration of
chemotherapy at least four weeks before study entry) and to have
undergone radiotherapy, surgery, or both. The study was approved
by the institutional review board of each participating institution,
and written informed consent was obtained from all patients.
Study Design
We conducted a randomized, open-label, multicenter trial de-
signed to evaluate the activity of imatinib in inducing objective re-
sponses in gastrointestinal stromal tumors. Randomization was per-
formed centrally without stratification according to site or any other
factor. Blocking, with a block size of four, was used. Secondary ob-
jectives were the assessment of pharmacokinetics, safety, time to
treatment failure, and survival. Standard [
18
F]fluoro-2-deoxy-
D
-glu-
cose positron-emission tomographic (PET) scanning was performed
in 64 patients to complement standard computed tomographic
(CT) imaging and assess changes in the metabolic profiles of the
tumors. Histopathological and molecular changes during treatment
were evaluated in selected patients by means of serial biopsies of
the tumor.
Patients were randomly assigned to receive either 400 mg or
600 mg of imatinib orally, taken once daily with food, in the form
of 100-mg capsules. These doses were chosen on the basis of data
from patients with chronic myeloid leukemia
11
in order to achieve
target plasma concentrations that could be expected to inhibit KIT
activity.
9,13
Patients receiving 400 mg per day whose tumor pro-
gressed but who were otherwise in good clinical condition were el-
igible to increase the dose to 600 mg per day. Patients whose tumor
progressed despite treatment with 600 mg per day were withdrawn
from the study.
Patients had regular physical examinations and evaluations of per-
formance status, body weight, complete blood count, and serum
chemistry. The administration of each dose and any adverse events
were recorded in a diary for each patient.
The study was designed by the academic investigators in collab-
oration with Novartis. The academic investigators and their respec-
tive teams at the four centers collected and managed all of the data.
The data were then collected in a central data base and made fully
available to the principal investigators at each study site. The team
met regularly and had twice-monthly conference calls to discuss
study progress and results. The overall results were analyzed by the
principal investigators and employees of Novartis. This article was
written by Dr. Demetri with substantive collaboration from the
principal investigators and all other authors.
Efficacy and Safety Evaluation
The response of the tumor to imatinib was evaluated after one
month, three months, and six months, and every six months there-
after or whenever there was a medical need. Assessments were ac-
cording to the standard Southwest Oncology Group criteria and
were based solely on CT or magnetic resonance imaging (MRI).
15
Responses were classified as complete responses (disappearance of
all disease that could be measured and evaluated); partial responses
(»50 percent decrease in the sum of the products of the perpen-
dicular diameters of all measurable lesions, the absence of progres-
sion, and the absence of new lesions); stable disease (a response that
did not qualify as a complete response, a partial response, or disease
progression); or disease progression (»50 percent increase or an
increase of 10 cm
2
[whichever was smaller] in the sum of the prod-
ucts of the perpendicular diameters of all measurable lesions, wor-
sening of a lesion that could be evaluated, the reappearance of any
lesion or the presence of a new lesion, or failure of the patient to
return for evaluation because of disease progression). All responses
had to be confirmed by repeated imaging within 4 to 12 weeks.
Time to treatment failure was defined as the time from the first
dose of imatinib to the earliest occurrence of progression, death
from any cause, or withdrawal from the trial for any reason other
than that the condition no longer required therapy. Data for patients
who had not had disease progression or died or who were with-
drawn from the trial for any reason other than that their condition
no longer required therapy were censored at the time of the last
assessment of the tumor. Toxic effects were graded according to the
National Cancer Institute Common Toxicity Criteria.
16
Immunohistochemical analysis for the detection of CD117 was
performed with the use of polyclonal rabbit antiserum (A4502,
Dako) and routine methods for immunohistochemical analysis with-
out any antigen retrieval.
17
Biopsy specimens of the tumor were ob-
tained from selected consenting patients before and after treatment
for the histopathological assessment of treatment, mutational analy-
ses of
KIT,
and immunoblotting for detection of KIT phospho-
protein.
4
Pharmacokinetics
Plasma samples were collected from a subgroup of patients before
treatment and then 1, 2, 3, 8, 24, 48, and 72 hours after the admin-
istration of the drug. Sampling at the same intervals was repeated
after four weeks of treatment. The plasma imatinib concentration
was determined by liquid chromatography and mass spectrometry as
previously reported.
18
Statistical Analysis
The original sample size was based on a proof-of-concept ap-
proach, according to which we required at least 3 patients with a
response among 18 treated patients in each group in order to con-
tinue enrolling patients in the study. This rule resulted in a 94 per-
cent probability of rejection of the null hypothesis if either dose lev-
el had a true response rate of less than 5 percent and a probability
of rejection of the null hypothesis of less than 6 percent if either
dose level had a true response rate of 30 percent or greater. Because
of the promising results observed, the study was enlarged to allow
recruitment of up to 200 patients; 147 patients were recruited. With
Copyright © 2002 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org at HHS LIBRARIES CONSORTIUM on May 12, 2008 .

474
·
N Engl J Med, Vol. 347, No. 7
·
August 15, 2002
·
www.nejm.org
The New England Journal of Medicine
an intention-to-treat population of 147 patients, the 95 percent con-
fidence interval for response rate was no wider than ±8.4 percent.
Such a confidence interval was judged sufficient to allow a meaning-
ful comparison with historical data.
After the first 100 patients completed the six-month assessment,
an interim analysis was performed, and the evidence of efficacy and
safety was judged sufficient for submission to health authorities for
registration of the drug. This report provides updated results. All re-
ported P values are two-sided.
RESULTS
Patients
Between July 2000 and April 2001, 147 patients
were recruited at four study centers. Characteristics of
the patients are summarized in Table 1. The diagnosis
of CD117-positive gastrointestinal stromal tumor was
confirmed by central review in 135 of 137 cases (98.5
percent); 2 patients were judged to be ineligible be-
cause of the absence of CD117 expression in the prop-
er histopathological context,
19
and in 10 cases, patho-
logic material was unavailable for central review. The
analyses presented here include data from all 147 pa-
tients on an intention-to-treat basis.
Previous therapy included surgery in 144 patients
(98.0 percent), chemotherapy for metastatic or unre-
sectable disease in 75 patients (51.0 percent), and ra-
diotherapy in 22 patients (15.0 percent). Patients who
had previously undergone chemotherapy had received
between one and seven regimens (median, two). None
of them had exhibited an objective response to any
previous regimen. Patients generally had far-advanced,
bulky disease, and the mean total area of tumors was
173 cm
2
(range, 1 to 1130).
Pharmacokinetics
Imatinib was detectable in plasma soon after oral
administration of either a 400-mg dose or a 600-mg
dose, with a mean half-life in the circulation of approx-
imately 20 hours. The mean plasma concentration in-
creased with increases in the dose, with variability be-
tween patients similar to that described in patients
with chronic myeloid leukemia.
20
The mean (±SE)
area under the curve after four weeks of treatment was
61±25 µg-hr per milliliter for the 400-mg dose and
75±31 µg-hr per milliliter for the 600-mg dose.
Antitumor Response
With follow-up of more than 9 months for all pa-
tients (the median follow-up was 288 days as of Oc-
tober 15, 2001, the last date of data collection for
this report), 120 patients (81.6 percent) remained in
the study. Data on antitumor response are shown in
Table 2. No patient had a complete response. Over-
all, 53.7 percent of the patients had a partial response.
All these partial responses were confirmed by repeated
imaging at least 28 days later. The reduction in the
bulk of the tumor among patients who had a partial
response ranged from 50 percent to 96 percent. An
additional 27.9 percent of patients had stable disease,
and disease progression was noted in 13.6 percent of
patients between one and three months after study
entry. The median time to an objective response was
13 weeks. Responses have been durable for more than
46 weeks and the median duration of response has not
been reached as of this writing (median follow-up,
24 weeks after the onset of response). There were no
significant differences in the rate or duration of re-
sponse between the dose levels of imatinib mesylate
we tested.
The time to treatment failure and overall survival
are shown in Figure 1. Of nine patients who were as-
signed to receive the lower dose and who were later
given the higher dose because of disease progression,
one subsequently had a partial response, and two had
stable disease after the crossover to 600 mg per day.
Nine patients treated with 400 mg per day and five
*Eastern Cooperative Oncology Group (ECOG)
status is graded on a scale from 0 to 5, with 0 de-
noting fully active and 5 death. A score of 3 indicates
that the patient is capable of limited self-care but is
confined to a bed or chair more than 50 percent of his
or her waking hours.
†Some patients had tumors at more than one site;
patients with tumors in the liver had metastatic dis-
ease at initial presentation.
‡Some patients had recurrent disease at more than
one site.
T
ABLE
1.
C
HARACTERISTICS
OF
THE
P
ATIENTS
.
C
HARACTERISTIC
A
LL
P
ATIENTS
(N=147)
Age — yr
Median 54
Range 18–83
Sex — no. (%)
Male 83 (56.5)
Female 64 (43.5)
ECOG status — no. (%)*
Grade 0 62 (42.2)
Grade 1 57 (38.8)
Grade 2 27 (18.4)
Grade 3 1 (0.7)
Site of tumor at diagnosis — no. (%)†
Small intestine 72 (49.0)
Stomach 50 (34.0)
Liver 25 (17.0)
Peritoneum 19 (12.9)
Omentum 17 (11.6)
Site of tumor at recurrence — no. (%)
Any recurrent disease‡ 132 (89.8)
Liver 115 (78.2)
Peritoneum 56 (38.1)
Retroperitoneum 21 (14.3)
Previous treatment — no. (%)
Surgery 144 (98.0)
Chemotherapy 75 (51.0)
Radiotherapy 22 (15.0)
Copyright © 2002 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org at HHS LIBRARIES CONSORTIUM on May 12, 2008 .

IMATINIB MESYLATE IN GASTROINTESTINAL STROMAL TUMORS
N Engl J Med, Vol. 347, No. 7
·
August 15, 2002
·
www.nejm.org
·
475
patients treated with 600 mg per day died. Five pa-
tients in the 400-mg group and eight patients in the
600-mg group were withdrawn from the study. Dis-
ease progression during treatment occurred in 11 pa-
tients receiving 600 mg per day and 8 patients receiv-
ing 400 mg per day. The estimated one-year survival
rate for all patients was 88 percent. Median survival has
not been reached as of this writing.
Standard [
18
F]fluoro-2-deoxy-
D
-glucose PET
proved to be a sensitive, rapid, and reliable indicator of
response or resistance to imatinib. In all patients with
a response, the [
18
F]fluoro-2-deoxy-
D
-glucose up-
take in the tumor had decreased markedly from base
line as early as 24 hours after a single dose of imatinib.
Increases in tumor-related glycolytic activity, activity
at new sites, or both were seen in all patients with dis-
ease progression. PET results correlated with subse-
quent evidence of a response or progression on CT
or MRI. PET and CT scans from a representative
patient with a response are shown in Figure 2.
Performance status improved with imatinib treat-
ment. By month 4 of the study, with 144 patients still
receiving treatment, the number of patients with nor-
mal functional status (an ECOG performance status
*CI denotes confidence interval.
T
ABLE
2.
R
ESPONSES
TO IMATINIB IN PATIENTS WITH ADVANCED
GASTROINTESTINAL STROMAL TUMORS.*
BEST RESPONSE
400 mg
(N=73)
600 mg
(N=74)
EITHER DOSE
(N=147)
no. (% [95% CI])
Complete response 0 0 0
Partial response 36 (49.3 [37.4–61.3]) 43 (58.1 [46.1–69.5]) 79 (53.7 [45.3–62.0])
Stable disease 23 (31.5 [21.1–43.4]) 18 (24.3 [15.1–35.7]) 41 (27.9 [20.8–35.9])
Progressive disease 12 (16.4) 8 (10.8) 20 (13.6)
Could not be evaluated 2 (2.7) 5 (6.8) 7 (4.8)
Figure 1. Kaplan–Meier Estimates of Overall Survival and Time to Treatment Failure for All Patients.
Each arrowhead represents the point at which a patient’s data were censored.
0
1.0
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
686460565248444036322820 24161284
Weeks after the First Dose
Continued response to treatment
Survival
Probability
Copyright © 2002 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org at HHS LIBRARIES CONSORTIUM on May 12, 2008 .

476 · N Engl J Med, Vol. 347, No. 7 · August 15, 2002 · www.nejm.org
The New England Journal of Medicine
of 0) had increased to 64 percent from 42 percent
at study entry. Similarly, by month 4, the number of
patients with substantially impaired functional status
(a performance status of 2 to 3) had decreased to
5 percent from 19 percent at entry.
Safety
Treatment with imatinib was generally well tolerat-
ed, although virtually every patient had at least some
mild or moderate adverse events (grade 1 or 2) that
might have been related to therapy. The most com-
mon adverse events included edema (in 74.1 percent
of patients) that was most frequently periorbital, nau-
sea (in 52.4 percent), diarrhea (in 44.9 percent), my-
algia or musculoskeletal pain (in 39.5 percent), fatigue
(in 34.7 percent), dermatitis or rash (in 30.6 percent),
headache (in 25.9 percent), and abdominal pain (in
25.9 percent) (Table 3). Most of these adverse events
were mild or moderate. There was no hyperuricemia
or evidence of tumor lysis syndrome, even in patients
Figure 2. Sequential PET Scans Obtained in the Same Patient at Base Line (before Treatment, Panel A), 1 Month after Imatinib Treat-
ment Began (Panel B), and after 16 Months of Continuous Treatment (Panel C).
The images at each point include a two-dimensional PET scan of the body (top), an axial PET scan of a slice through the site of the
pelvic tumor (middle), and a correlating CT scan at the corresponding level. The standardized uptake values for the tumor at the
three time points were 4.5 (Panel A), 1.24 (Panel B), and 0.75 (Panel C). The uptake in the cardiac blood pool, the myocardium, the liver,
the bowel, the bilateral renal collecting system, and the bladder is within physiologic limits in this patient. Images were obtained
with the use of similar doses of [
18
F]fluoro-2-deoxy-D-glucose, acquisition times, and protocols at the three time points. The patient
also had similar blood glucose concentrations at each of these three time points.
A
BC
Copyright © 2002 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org at HHS LIBRARIES CONSORTIUM on May 12, 2008 .

IMATINIB MESYLATE IN GASTROINTESTINAL STROMAL TUMORS
N Engl J Med, Vol. 347, No. 7 · August 15, 2002 · www.nejm.org · 477
with very rapid decreases in tumor volume. Serious
adverse events (grade 3 or 4) occurred in 21.1 percent
of patients. The most serious adverse events were gas-
trointestinal or intraabdominal hemorrhages in pa-
tients with large, bulky tumors, which occurred in ap-
proximately 5 percent of patients.
Histopathological Changes
A subgroup of biopsies performed after treatment
showed reduced numbers of tumor cells and a hypo-
cellular myxohyaline stroma with small numbers of
scattered atypical nuclei and, frequently, prominent
stromal hemorrhage. Frank necrosis of the tumor was
rarely seen (Fig. 3). Other biopsies showed large num-
bers of residual CD117-positive tumor cells, even in
patients whose tumors showed a substantial reduction
in size on PET and CT scanning. These residual gas-
trointestinal stromal tumor cells often showed pyknot-
ic nuclei and reduced cytoplasmic volume, similar in
appearance to “crush artifact.”
DISCUSSION
There is compelling evidence from preclinical mod-
els that the constitutively activated KIT-receptor ty-
*Data are for the types of events that occurred in at least 5 percent of the patients in at least one of the two groups. GI denotes gastrointestinal.
TABLE 3. ADVERSE EFFECTS WITH A POSSIBLE OR SUSPECTED RELATION TO IMATINIB.*
ADVERSE EFFECT ANY GRADE GRADE 3 OR 4
400 mg
(
N=73)
600 mg
(N=74)
ALL PATIENTS
(N=147)
400 mg
(N=73)
600 mg
(N=74)
ALL PATIENTS
(N=147)
number (percent)
Any adverse effect with suspected relation
to study drug
71 (97.3) 73 (98.6) 144 (98.0) 15 (20.5) 16 (21.6) 31 (21.1)
Edema or fluid retention 52 (71.2) 57 (77.0) 109 (74.1) 1 (1.4) 1 (1.4) 2 (1.4)
Periorbital 33 (45.2) 37 (50.0) 70 (47.6) 0 0 0
Leg 19 (26.0) 11 (14.9) 30 (20.4) 0 0 0
Face 6 (8.2) 9 (12.2) 15 (10.2) 1 (1.4) 0 1 (0.7)
Other site 5 (6.8) 10 (13.5) 15 (10.2) 0 0 0
Eyelid 5 (6.8) 6 (8.1) 11 (7.5) 0 0 0
Nausea 37 (50.7) 40 (54.1) 77 (52.4) 1 (1.4) 1 (1.4) 2 (1.4)
Diarrhea 29 (39.7) 37 (50.0) 66 (44.9) 1 (1.4) 2 (2.7) 3 (2.0)
Myalgia or musculoskeletal pain 27 (37.0) 31 (41.9) 58 (39.5) 0 0 0
Fatigue 22 (30.1) 29 (39.2) 51 (34.7) 0 0 0
Dermatitis or rash 18 (24.7) 27 (36.5) 45 (30.6) 2 (2.7) 2 (2.7) 4 (2.7)
Headache 14 (19.2) 24 (32.4) 38 (25.9) 0 0 0
Abdominal pain 19 (26.0) 19 (25.7) 38 (25.9) 1 (1.4) 0 1 (0.7)
Flatulence 14 (19.2) 18 (24.3) 32 (21.8) 0 0 0
Vomiting 10 (13.7) 9 (12.2) 19 (12.9) 0 1 (1.4) 1 (0.7)
Hemorrhage 8 (11.0) 10 (13.5) 18 (12.2) 3 (4.1) 4 (5.4) 7 (4.8)
Tumor hemorrhage 1 (1.4) 3 (4.1) 4 (2.7) 1 (1.4) 3 (4.1) 4 (2.7)
Cerebral hemorrhage or subdural hematoma 0 0 0 0 0 0
Upper GI tract bleeding or perforation 3 (4.1) 2 (2.7) 5 (3.4) 3 (4.1) 1 (1.4) 4 (2.7)
Dyspepsia 7 (9.6) 9 (12.2) 16 (10.9) 0 0 0
Increased lacrimation 5 (6.8) 9 (12.2) 14 (9.5) 0 0 0
Anemia 4 (5.5) 9 (12.2) 13 (8.8) 1 (1.4) 2 (2.7) 3 (2.0)
Loose stools 5 (6.8) 7 (9.5) 12 (8.2) 0 0 0
Taste disturbance 2 (2.7) 10 (13.5) 12 (8.2) 0 0 0
Neutropenia 6 (8.2) 4 (5.4) 10 (6.8) 5 (6.8) 2 (2.7) 7 (4.8)
Abdominal distention 4 (5.5) 4 (5.4) 8 (5.4) 0 0 0
Abnormal liver-function results 4 (5.5) 4 (5.4) 8 (5.4) 2 (2.7) 2 (2.7) 4 (2.7)
Leukopenia 4 (5.5) 3 (4.1) 7 (4.8) 2 (2.7) 0 2 (1.4)
Arthralgia 1 (1.4) 5 (6.8) 6 (4.1) 0 0 0
Paresthesia 1 (1.4) 5 (6.8) 6 (4.1) 0 0 0
Esophageal reflux 1 (1.4) 5 (6.8) 6 (4.1) 0 0 0
Pruritus 2 (2.7) 4 (5.4) 6 (4.1) 0 0 0
Pain (in an extremity) 4 (5.5) 1 (1.4) 5 (3.4) 0 0 0
Blurred vision 4 (5.5) 1 (1.4) 5 (3.4) 0 0 0
Photosensitivity 0 4 (5.4) 4 (2.7) 0 0 0
Copyright © 2002 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org at HHS LIBRARIES CONSORTIUM on May 12, 2008 .

478 · N Engl J Med, Vol. 347, No. 7 · August 15, 2002 · www.nejm.org
The New England Journal of Medicine
rosine kinase stimulates the proliferation and enhanc-
es the survival of neoplastic gastrointestinal stromal
tumor cells. In both preclinical experiments
13
and a
previous case study,
14
inhibition by imatinib had con-
siderable antiproliferative and proapoptotic effects on
gastrointestinal stromal tumor cells. Our study dem-
onstrates, in a large series of patients with advanced
gastrointestinal stromal tumors, that imatinib is effec-
tive in most patients.
Advanced gastrointestinal stromal tumors are un-
responsive to conventional chemotherapy.
5,6
The high
rate of response to imatinib in these patients with
bulky disease who had no response to cytotoxic che-
motherapy is not only remarkable, but also supports
the hypothesis that dysregulated KIT kinase activity
is important in human gastrointestinal stromal tumors.
Responses, although partial, have lasted for many
months, as patients continue to receive daily treatment
in our ongoing clinical trial. Our results corroborate
the result obtained with imatinib in a single patient
with a gastrointestinal stromal tumor, who is still re-
ceiving therapy more than 22 months after its initi-
ation (unpublished data),
14
and the confirmed partial
responses in 19 patients in a phase 1 study of imatinib
in gastrointestinal stromal tumors.
21
Historical data
show a median survival of 19 months for all patients
with metastatic disease and 9 months for patients with
metastatic disease and local recurrence.
5
Despite the
extensive metastatic disease in the majority of our pa-
tients, 88 percent were alive one year after the initia-
tion of treatment with imatinib, with the median du-
ration of survival not yet reached.
This phase 2 trial was not adequately powered to
distinguish between the efficacy of the 400-mg and
600-mg doses. Although there was no statistically sig-
nificant difference between the dose levels, three of the
nine patients who received the higher dose after ev-
idence of disease progression was uncovered had a
sustained partial response or stable disease after the
crossover. The optimally effective dose of imatinib in
Figure 3. Biopsy Specimens of Metastatic Malignant Gastrointestinal Stromal Tumors before Treatment and after Four
Weeks of Daily Treatment with Imatinib.
At base line, the CD117 immunostaining shows a dot-like pattern (top image, Panel A). Only rare CD117-positive cells
are noted on immunostaining in the specimen obtained after treatment (top image, Panel B). Scattered atypical nuclei
and inflammatory cells present in a myxohyaline stroma are visible on the specimen that was obtained after treatment
and stained with hematoxylin and eosin (bottom image, Panel B). (The bottom image in Panel A shows hematoxylin and
eosin staining of the tumor before treatment.)
A
Before Treatment
B
After Treatment
Copyright © 2002 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org at HHS LIBRARIES CONSORTIUM on May 12, 2008 .

IMATINIB MESYLATE IN GASTROINTESTINAL STROMAL TUMORS
N Engl J Med, Vol. 347, No. 7 · August 15, 2002 · www.nejm.org · 479
patients with a gastrointestinal stromal tumor is the
subject of large, appropriately powered, randomized
studies that are now under way.
The high bioavailability of orally administered ima-
tinib in our study was generally similar to that report-
ed in patients with chronic myeloid leukemia.
11
Ther-
apeutic plasma concentrations were attained despite
the presence of altered gastrointestinal anatomy from
previous, and often extensive, resections.
Overall, imatinib was well tolerated, with adverse
effects similar to those reported in a large population
of patients with chronic myeloid leukemia.
11,12
My-
elotoxicity was less frequent in patients with gastro-
intestinal stromal tumors, suggesting that the myelo-
suppression associated with imatinib in hematologic
cancers may be related to the pathophysiology of the
leukemic bone marrow. An important finding was se-
rious gastrointestinal and tumor hemorrhage in about
5 percent of our patients. These hemorrhages could
be related to the underlying disease, but they were
probably related to tumor degeneration induced by
imatinib.
The close relation between clinical outcome and the
findings on [
18
F]fluoro-2-deoxy-D-glucose PET scan-
ning indicates that such scanning is a useful comple-
ment to standard anatomical imaging with CT or MRI
for monitoring the therapeutic effect of imatinib in pa-
tients with gastrointestinal stromal tumors. The mo-
lecular mechanisms responsible for the rapid decreases
in glycolytic activity associated with imatinib treatment
remain unknown, particularly given that some biop-
sies demonstrated the continued presence of substan-
tial numbers of viable CD117-bearing cells.
Although our results indicate that imatinib is effec-
tive for many patients with advanced gastrointestinal
stromal tumors, resistance of tumors to single-agent
therapy is common. In 5 percent of our patients, the
tumor exhibited primary resistance to imatinib with-
in the first two months. In other patients with disease
progression, resistance became evident only after sev-
eral months of treatment. Nonetheless, patients with
an objective response and the majority of patients with
stable disease had durable evidence of a treatment ben-
efit lasting more than six months. These findings con-
trast somewhat with the experience in patients with
advanced chronic myeloid leukemia. In patients treat-
ed for blast crisis, most cases of secondary resistance
appeared within four months after the initial re-
sponse.
12
Resistance in patients with chronic myeloid
leukemia is caused by more than one molecular mech-
anism, including amplification of the gene encoding
the aberrant kinase and mutation of the drug-binding
site in the kinase domain.
22,23
Molecular mechanisms
responsible for resistance in patients with gastrointes-
tinal stromal tumors may be quite different. Constitu-
tive activation of KIT receptor tyrosine kinase in
gastrointestinal stromal tumors can be caused by mu-
tations in any of several exons, and in a subgroup of
patients there is no detectable KIT mutation.
4
Even
at the most common site of mutations (exon 11), a
wide variety of in-frame deletions and substitutions
has been reported.
24
Careful study of molecular mech-
anisms will be needed in order to develop rational
strategies for preventing or overcoming the emergence
of resistance to imatinib in patients with gastrointes-
tinal stromal tumors.
Supported in part by grants from Novartis Oncology, grants from the
Katz Foundation, the Rubenstein Foundation, and the Quick Family (to
Dr. Demetri), and grants from the Veterans Affairs Merit Review Program
and Northwest Health Foundation (to Dr. Heinrich).
Drs. Demetri, von Mehren, Blanke, Van den Abbeele, Eisenberg, Hein-
rich, J. Fletcher, Druker, Corless, C. Fletcher, and Joensuu have consulted
for or received research grants from Novartis. Drs. Silberman, Capdeville,
Kiese, Peng, and Dimitrijevic are employees of and hold equity in Novartis.
We are indebted to the following persons for their contributions to
the study: Suzanne George, M.D., Jeffrey Morgan, M.D., David P.
Ryan, M.D., Priscilla Merriam, Amy Potter, M. Travis Quigley, R.N.,
B.S.N., Margaret Buonanno, Adriana Torre, Ellen Bosnak, Tricia
Spangler, C.N.M.T., Richard J. Tetrault, C.N.M.T., Ramsey D.
Badawi, Ph.D., Joan Canniff, R.N., Jean-Pierre Cliche, M.D., Da-
vid A. Israel, M.D., Jeanne Griffin, R.N., David Harmon, M.D.,
Athena Moutsiolis, Judith Manola, M.Sc., Dana–Farber Cancer In-
stitute and Harvard Cancer Center, Boston; Monica Davey, R.N.,
B.S.N., Barton N. Milestone, M.D., Rosaleen Parsons, M.D., Fox
Chase Cancer Center, Philadelphia; Diana J. Griffith, Andrea Ha-
ley, Laura McGreevey, Lea Herndon Smith, C.C.R.P., Cecily L. Wait,
Lora Wilson, R.N., B.S.N., Oregon Health and Science University,
Portland; and Inkeri Elomaa, M.D., Carl Blomqvist, M.D., Pekka
Virkkunen, M.D., University of Helsinki, Helsinki, Finland.
REFERENCES
1. Miettinen M, Sarlomo-Rikala M, Lasota J. Gastrointestinal stromal tu-
mors: recent advances in understanding of their biology. Hum Pathol
1999;30:1213-20.
2. Rubin BP, Fletcher JA, Fletcher CDM. Molecular insights into the his-
togenesis and pathogenesis of gastrointestinal stromal tumors. Int J Surg
Pathol 2000;8:5-10.
3. Lux M, Rubin BP, Biase TL, et al. KIT extracellular and kinase domain
mutations in gastrointestinal stromal tumors. Am J Pathol 2000;156:791-5.
4. Rubin BP, Singer S, Tsao C, et al. KIT activation is a ubiquitous feature
of gastrointestinal stromal tumors. Cancer Res 2001;61:8118-21.
5. DeMatteo RP, Lewis JJ, Leung D, Mudan SS, Woodruff JM, Brennan
MF. Two hundred gastrointestinal stromal tumors: recurrence patterns and
prognostic factors for survival. Ann Surg 2000;231:51-8.
6. Goss GA, Merriam P, Manola J, Singer S, Fletcher CD, Demetri GD.
Clinical and pathological characteristics of gastrointestinal stromal tumors
(GIST). Prog Proc Am Soc Clin Oncol 2000;19:599a. abstract.
7. Druker BJ, Tamura S, Buchdunger E, et al. Effects of a selective inhib-
itor of the Abl tyrosine kinase on the growth of Bcr-Abl positive cells. Nat
Med 1996;2:561-6.
8. Buchdunger E, Cioffi CL, Law N, et al. Abl protein-tyrosine kinase in-
hibitor STI571 inhibits in vitro signal transduction mediated by c-kit and
platelet-derived growth factor receptors. J Pharmacol Exp Ther 2000;295:
139-45.
9. Heinrich MC, Griffith DJ, Druker BJ, Wait CL, Oh KA, Zigler AJ. In-
hibition of c-kit receptor tyrosine kinase activity by STI 571, a selective ty-
rosine kinase inhibitor. Blood 2000;96:925-32.
10. Wang WL, Healy ME, Sattler M, et al. Growth inhibition and modu-
lation of kinase pathways of small cell lung cancer cell lines by the novel
tyrosine kinase inhibitor STI 571. Oncogene 2000;19:3521-8.
11. Druker BJ, Talpaz M, Resta DJ, et al. Efficacy and safety of a specific
inhibitor of the BCR-ABL tyrosine kinase in chronic myeloid leukemia.
N Engl J Med 2001;344:1031-7.
Copyright © 2002 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org at HHS LIBRARIES CONSORTIUM on May 12, 2008 .

480 · N Engl J Med, Vol. 347, No. 7 · August 15, 2002 · www.nejm.org
The New England Journal of Medicine
12. Druker BJ, Sawyers CL, Kantarjian H, et al. Activity of a specific in-
hibitor of the BCR-ABL tyrosine kinase in the blast crisis of chronic mye-
loid leukemia and acute lymphoblastic leukemia with the Philadelphia
chromosome. N Engl J Med 2001;344:1038-42. [Erratum, N Engl J Med
2001;345:232.]
13. Tuveson DA, Willis NA, Jacks T, et al. STI571 inactivation of the gas-
trointestinal stromal tumor c-KIT oncoprotein: biological and clinical im-
plications. Oncogene 2001;20:5054-8.
14. Joensuu H, Roberts PJ, Sarlomo-Rikala M, et al. Effect of the tyrosine
kinase inhibitor STI571 in a patient with a metastatic gastrointestinal stro-
mal tumor. N Engl J Med 2001;344:1052-6.
15. Green S, Weiss GR. Southwest Oncology Group standard response
criteria, endpoint definitions and toxicity criteria. Invest New Drugs 1992;
10:239-53.
16. Cancer Therapy Evaluation Program. Common toxicity criteria man-
ual: common toxicity criteria, version 2.0. Bethesda, Md.: National Cancer
Institute, June 1999. (Also available at http://ctep.cancer.gov/forms/
CTCManual_v4_10-4-99.pdf.)
17. Hornick JL, Fletcher CDM. Immunohistochemical staining for KIT
(CD117) in soft tissue sarcomas is very limited in distribution. Am J Clin
Pathol 2002;117:188-93.
18. Bakhtiar R, Lohne J, Ramos L, Khemani L, Hayes M, Tse F. High-
throughput quantification of the anti-leukemia drug STI571 (Gleevec) and
its main metabolite (CGP 74588) in human plasma using liquid chroma-
tography-tandem mass spectrometry. J Chromatogr B Biomed Sci Appl
2002;768:325-40.
19. Fletcher CD, Berman JJ, Corless C, et al. Diagnosis of gastrointestinal
stromal tumors: a consensus approach. Hum Pathol 2002;33:459-65.
20. Peng B, Hayes M, Racine-Poon A, et al. Clinical investigation of the
pharmacokinetic/pharmacodynamic relationship for Glivec (STI571): a
novel inhibitor of signal transduction. Prog Proc Am Soc Clin Oncol 2001;
20:71a. abstract.
21. van Oosterom AT, Judson I, Verweij J, et al. Safety and efficacy of ima-
tinib (STI571) in metastatic gastrointestinal tumors: a phase I study. Lan-
cet 2001;358:1421-3.
22. le Coutre P, Tassi E, Varella-Garcia M, et al. Induction of resistance to
Abelson inhibitor STI571 in human leukemic cells through gene amplifi-
cation. Blood 2000;95:1758-66.
23. Gorre ME, Mohammed M, Ellwood K, et al. Clinical resistance to
STI-571 cancer therapy caused by BCR-ABL gene mutation or amplifica-
tion. Science 2001;293:876-80.
24. Miettinen M, Furlong M, Sarlomo-Rikala M, Burke A, Sobin LH, La-
sota J. Gastrointestinal stromal tumors, intramural leiomyomas, and leio-
myosarcomas in the rectum and anus: a clinicopathologic, immunohisto-
chemical, and molecular genetic study of 144 cases. Am J Surg Pathol
2001;25:1121-33.
Copyright © 2002 Massachusetts Medical Society.
FULL TEXT OF ALL JOURNAL ARTICLES ON THE WORLD WIDE WEB
Access to the complete text of the Journal on the Internet is free to all subscribers. To use
this Web site, subscribers should go to the Journal’s home page (http://www.nejm.org)
and register by entering their names and subscriber numbers as they appear on their
mailing labels. After this one-time registration, subscribers can use their passwords to log
on for electronic access to the entire Journal from any computer that is connected to the
Internet. Features include a library of all issues since January 1993 and abstracts since
January 1975, a full-text search capacity, and a personal archive for saving articles and
search results of interest. All articles can be printed in a format that is virtually identical
to that of the typeset pages. Beginning six months after publication the full text of all
original articles and special articles is available free to nonsubscribers who have completed
a brief registration.
Copyright © 2002 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org at HHS LIBRARIES CONSORTIUM on May 12, 2008 .