Jpn J Clin Oncol 2008;38(10)689 – 694
doi:10.1093/jjco/hyn086
Impact of Baseline Sum of Longest Diameter in Target Lesions by
RECIST on Survival of Patients with Metastatic Colorectal Cancer
Nozomu Machida1, Takayuki Yoshino1, Narikazu Boku1, Shuichi Hironaka1, Yusuke Onozawa1, Akira Fukutomi1,
Kentaro Yamazaki1, Hirofumi Yasui1, Keisei Taku1 and Masahiro Asaka2
1
Division of Gastrointestinal Oncology, Shizuoka Cancer Center, Sunto-gun, Shizuoka and 2Department of
Gastroenterology, Hokkaido University Graduate School of Medicine, Sapporo, Japan
Received April 11, 2008; accepted July 27, 2008
Objective: In patients with metastatic colorectal cancer (mCRC), several prognostic factors
such as performance status (PS), number of metastatic sites, carcinoembryonic antigen
(CEA), alkaline phosphatase (ALP) and lactate dehydrogenase (LDH) have been reported.
The objective of this study was to clarify the prognostic impact of Baseline Sum of Longest
Diameter (BSLD) of target lesions by Response Evaluation Criteria in Solid Tumor (RECIST)
in patients with mCRC.
Methods: The subjects of this study were consecutive 103 patients with mCRC who had
been received the first line systemic chemotherapy between September 2002 and March
2005.
Results: The chemotherapy regimens included leucovorin-modulated 5-fluorouracil (5-FU)
(n ¼ 27) and 5-FU plus irinotecan (n ¼ 76). The median overall survival time was 547 days.
The median BSLD was 14.3 cm (range, 1.1 – 54.7). In univariate analysis, identified prognostic variables on survival were PS (0, 1 versus 2), number of metastatic sites (1 versus .1),
peritoneal dissemination (þ versus 2), pleural effusion (PE) and/or ascites, white blood cell
( versus ,10 000/mm 3), ALP ( versus ,300 IU), LDH ( versus ,300 IU), CEA
( versus ,5 ng/ml), chemotherapy regimen, presence of liver metastasis, and BSLD. In
multivariate analysis with covariates of the above significant factors, BSLD ( versus
,10 cm) (HR 0.431, 95% CI 0.237 – 0.785, P ¼ 0.0059), PS (HR 0.248, 95% CI
0.107 – 0.577, P ¼ 0.0012), PE and/or ascites (HR 0.402, 95% CI 0.228 – 0.708, P ¼ 0.0016)
were independent prognostic factors.
Conclusion: BSLD of target lesions by RECIST representing tumor volume might be an
independent prognostic factor of patients with mCRC after systemic chemotherapy.
Key words: RECIST – colorectal cancer – prognostic factor
INTRODUCTION
Colorectal cancer (CRC) is one of the most common causes
of cancer death in Western countries and Japan (1 – 3). More
than 50% of the patients have unresectable locally advanced
or metastatic disease. Chemotherapy of CRC has improved
substantially over the last 10 years. Combination chemotherapy with modulated 5-fluorouracil (5-FU) plus irinotecan
and oxaliplatin along the treatment course of the first and the
second line can achieve response rates over 50% and median
For reprints and all correspondence: Narikazu Boku, Division of
Gastrointestinal Oncology, Shizuoka Cancer Center, 1007 Shimonagakubo,
Nagaizumi-cho, Sunto-gun, Shizuoka, Japan. E-mail: [email protected]
overall survival times reach in excess of 20 months (4 – 11).
Moreover, bevacizumab has demonstrated an additional
survival advantage (11,12).
Several studies have shown that performance status (PS)
(14 – 20), lactate dehydrogenase (LDH), white blood cell
(WBC) (20), carcinoembryonic antigen (CEA) (24) and
alkaline phosphatase (ALP) are prognostic factors after
chemotherapy (13 – 15,21,22). Recently, Köhne et al. (15)
reported prognostic factors on the multivariate analysis,
using databases of 3825 metastatic CRC (mCRC) patients
enrolled in prospective trials. Four independent baseline
prognostic parameters, PS, WBC, ALP and number of metastatic sites were identified. Combination of these factors
# The Author (2008). Published by Oxford University Press. All rights reserved.
690
Survival impact of baseline sum of longest diameter
could categorize patients with mCRC into three groups
according to survival time: high risk, 6.1 – 6.4 months; intermediate risk, 10.7 – 10.9 months and low risk, 14.7 – 15.0
months (15). However, the subject had been treated with
only 5-FU with leucovorin (LV) because other active agents,
such as irinotecan, oxaliplatin and molecular targeting
agents, had not been available at that time. Hurwitz et al.
(23) reported that this model may be applicable to
irinotecan-based regimen.
However, the above independent variables never reflect
tumor volume (TV), although patients with large TV tended
to have short survival times. RECIST (response evaluation
criteria in solid tumor) was established for the purpose to
uniform reporting of outcomes of clinical trials (25). Overall
response rate has usually been adopted for primary endpoint
of a phase II study, in which documentation of the baseline
sum of longest diameter (BSLD) of target lesions is mandatory. BSLD can be considered to represent TV. The aims of
this analysis are to clarify the prognostic significance of
baseline sum longest diameter of target lesions by RECIST
for mCRC patients after chemotherapy.
PATIENTS AND METHODS
SUBJECTS
A total of 125 patients with mCRC treated with
LV-modulated 5-FU or irinotecan in combination with 5-FU
as the first-line treatment initiated from September 2002 to
March 2005 at our institution were included. The selection
criteria for the inclusion on this retrospective analysis were
(i) histologically or cytologically proven adenocarcinoma of
the colon or rectum, (ii) unresectable or recurrent cases after
previous surgery, (iii) having at least one unidimensionally
measurable lesion (.10 by 5 mm slice CT), (iv) age of
20, (v) Eastern Cooperative Oncology Group PS (ECOG
PS) of 2, (vi) WBC of 3000 cells/ml, (vii) hemoglobin
of 8 g/dl, (viii) platelet of 100 000/ml, (ix) bilirubin of
1.1 mg/dl, (x) serum aspartate aminotransferase (AST) and
alanine aminotransferase (ALT) of 100 U/l; (xi) serum
creatinine level of ,1.1 mg/dl (for men) or 0.8 mg/dl (for
women), (xii) available survival data and (xiii) no prior chemotherapy except prior adjuvant chemotherapy after resection of the primary tumor completed more than 6 months
before recurrence. Patients with symptomatic central nervous
system metastasis and any active extracolonic cancer were
excluded. Patients who were treated with oxaliplatincontaining regimen in the first-line setting were not included
because oxaliplatin had not been approved for mCRC in
Japan until April 2005.
intravenous (IV) infusion with 5-FU of 600 mg/m2 as an IV
bolus 1 h after the l-LV had begun, given once every week
for 4 weeks, and repeated every 6 weeks, (ii) tefafure-uracil
(UFT)/LV: UFT of 300 mg/m2 and LV of 75 mg/body were
orally administered daily for 4 weeks followed by 1-week
rest and repeated every 5 weeks, (iii) bolus 5-FU plus
irinotecan, fluorouracil and levofolinate (IFL): irinotecan
of 100 mg/m2 as a 90-min IV infusion, l-LV of 10 mg/m2 as
an IV bolus, and then 5-FU of 500 mg/m 2 as an IV bolus
infusion, given once every week for 4 weeks, and repeated
every 6 weeks, (iv) infusional 5-FU plus irinotecan FOLic
acid, Fluorouracil and IRInotecan (FOLFIRI): l-LV of
200 mg/m2 as a 2-h IV infusion, irinotecan of 180 mg/m2
as a 90-min IV infusion followed by IV bolus 5-FU of
400 mg/m2, and a 46-h infusion of 5-FU of 2400 mg/m2
given every other week. All patients routinely received 3 mg
of granisetron plus 8 mg of dexamethasone in a 30-min IV
infusion before administration of irinotecan. Treatment continued until disease progression, unacceptable toxicity, or
patient refusal. Whenever severe adverse events occurred,
the doses were adjusted to appropriate levels of tolerance
based on the physician’s assessment. Any Grade 3 or 4
adverse event resulted in 20% dose reduction of
bolus 5-FU for subsequent cycles for RPMI and UFT/LV,
20% dose reduction of irinotecan and bolus 5-FU for IFL,
20% dose reduction of irinotecan and infusional 5-FU and
50% dose reduction of bolus 5-FU for FOLFIRI. Persistent
Grade 2 or worse adverse events delayed therapy until
recovery. The use of colony-stimulating factors was allowed
if medically justified. Intensive treatment with loperamide, if
needed, was used for diarrhea. Other supportive treatments
were given if required.
EVALUATION
OF
PATIENTS
We retrospectively reviewed medical records of patients
including characteristics [age, sex, ECOG PS, inoperative/
recurrence, primary tumor location, history of primary resection, the existence of peritoneal dissemination, the presence
of pleural effusion (PE) and/or ascites, number of metastatic
sites, the existence of liver metastasis, metastatic site, chemotherapy regimen, WBC, ALP, LDH and serum CEA]. All
target lesions (10 mm or larger in longest diameter, up to a
maximum of five lesions per organ and 10 lesions in total)
and non-target lesions at baseline were identified and
measured by CT scans using 5 mm slice within 28 days
before the initial treatment of chemotherapy. The overall
survival was calculated from the date of starting initial
chemotherapy to death or last contact.
STATISTICAL METHODS
TREATMENT
The treatment schedules consisted of the following six types
of regimens: (i) 5-FUþl-LV [Roswell Park Memorial
Institute (RPMI) regimen]: l-LV of 500 mg/m2 in a 2-h
The Kaplan – Meier method was used to evaluate median
overall survival. The univariate and multivariate analyses of
prognostic factors using a Cox proportional hazard model
were carried out with categorized variables to calculate risks
Jpn J Clin Oncol 2008;38(10)
and their 95% confidence interval (95% CI). The factors
with substantial impacts (P , 0.10) in the univariate analysis
were introduced to a Cox proportional hazard model in the
multivariate analysis with backward selection. The following
18 categories were examined: age; gender, ECOG PS; 0 or 1
versus 2, recurrence or unresectable, primary tumor
resection; 2 versus þ, tumor location; rectum versus colon,
peritoneal dissemination; þ versus 2, PE and/or ascites; þ
versus 2, number of metastatic sites; 1 versus .1, treatment
regimen: LV-modulated 5-FU regimen versus irinotecan
containing regimen, WBC; 10 000 versus ,10 000/mm3,
ALP; 300 versus ,300 IU, LDH; 300 versus ,300 IU,
CEA; 5 versus ,5 ng/ml, liver metastasis; þversus 2,
BSLD (cm), BSLD; 5 versus ,5 cm, BSLD; 10 versus
,10 cm, BSLD; median versus ,median. The calculations were performed using the SAS Version 8.02 (SAS
Inc, USA).
691
Table 1. Patient characteristics
Number of patients
Number included
%
103
Age (years)
Range
29–80
Median
62
62
55
53.4
,62
48
46.6
Male
63
61.2
Female
40
38.8
0
68
66
1
28
27.2
2
7
6.8
Sex
ECOG PS
Disease status
RESULTS
PATIENT BACKGROUND
The subjects selected for the retrospective study were 103
patients. Twenty-two patients (18%) were excluded according to the selection criteria. Patients were excluded by the
following reason: four for histologically proven neuroendocrine tumors, seven for unavailable survival data, three for
multiple cancers, six for no measurable lesion, one for no
CT scan before chemotherapy and one for oxaliplatin
regimen.
Patient characteristics are given in Table 1. Their median
age was 62 years. Sixty-six percent of the patients had an
ECOG PS of 0 at base line. Seventy-six percent of the patients
had at least two organs involved, liver being the most
common site of metastasis. Forty-seven percent of the patients
received IFL regimen, and 27% did FOLFIRI regimen.
UNIVARIATE ANALYSIS
Median overall survival was 547 days, with a median
follow-up time among survivors of 694 days (Fig. 1). BSLD
ranged from 1.1 to 54.7 cm and the median is 14.3 cm.
Variables significantly associated with overall survival in the
univariate analysis were ECOG PS, number of metastatic
sites, peritoneal dissemination, PE and/or ascites, treatment
regimen, WBC, LDH, liver metastasis, BSLD (5, 10 and
14.3 cm) (Table 2). The BSLD (10 cm) showed a hazard
ratio (HR), the most significant variable compared with
BSLD (5 cm, 14.3 cm) in the univariate. Figure 2 shows
the overall survival curves divided by BSLD (10 cm), which
were clearly separated.
MULTIVARIATE ANALYSIS
The results of the multivariate analysis are given in Table 3.
Variables significantly associated with overall survival in the
Inoperable
81
78.6
Recurrence
22
21.4
Colon
66
64.1
Rectum
37
35.9
þ
43
41.7
2
60
58.3
þ
20
19.4
2
83
80.6
Primary tumor location
Primary lesion
Peritoneal dissemination
Pleural effusion (PE) and/or ascites
þ
31
30.1
2
72
69.9
25
24.3
78
75.7
Yes
72
69.9
No
31
30.1
Number of metastatic sites
1
2
Liver metastasis
BSLD (cm)
Range
1.1–54.7
Median
14.3
,5
18
17.5
5
85
82.5
,10
40
38.8
10
63
61.2
,14.3
51
49.5
14.3
52
50.5
16
15.5
Treatment regimen
5-FUþLV
Continued
692
Survival impact of baseline sum of longest diameter
Table 1. Continued
Table 2. Factors associated with overall survival (OS) in the univariate
analysis
Number of patients
%
UFTþLV
11
10.7
IFL
48
46.6
FOLFIRI
28
27.2
P value
,10 000
94
91.3
10 000
9
8.7
,300
43
41.7
62
0.4087
300
60
58.3
1.237
0.747– 2.049
MST
(days)
544
1.000
563
PS
2
,0.0001
0, 1
Alcaline phosphatase (IU/l)
95% CI
Age (years)
,62
While blood cells count (/ml)
HR
Recurrence
5.155
2.288– 11.628
1.000
0.4424
Inoperable
0.800
145
553
0.442– 1.412
1.000
569
543
Tumor location
LDH (IU/l)
,300
55
53.4
Rectum
300
48
46.6
Colon
,5
15
14.6
5
87
85.4
0.2528
0.730
0.426– 1.251
1.000
565
543
Primary resection
CEA (ng/ml)
2
0.3439
þ
1.280
0.763– 2.146
1.000
548
547
Peritoneal dissemination
Köhne’s index
Low
24
23.3
þ
Intermediate
28
27.2
2
High
51
49.5
2.217
1.217– 4.032
1.000
356
553
Pleural effusion (PE) and/or ascites
þ
ECOG PS, Eastern Cooperative Oncology Group performance status;
BSLD, Baseline Sum of Longest Diameter; 5-FU, 5-fluorouracil;
LV, leucovorin; LDH, lactate dehydrogenase; CEA, carcinoembryonic antigen.
0.0093
,0.0001
2
3.472
2.049– 5.917
1.000
264
565
Number of metastatic sites
2
0.0227
1
2.288
1.122– 4.651
1.000
495
596
Liver metastasis
þ
0.0182
2
2.020
1.127– 3.610
1.000
512
586
Treatment regimen
5-FU/LV
0.061
5-FU plus irinotecam
1.689
0.976– 2.922
1.000
474
553
White blood cells count (/ml)
10 000
0.0214
,10 000
2.427
0.194– 0.877
1.000
351
553
Alcaline phosphatase (IU/l)
Figure 1. Overall survival of all 103 patients.
300
0.0643
,300
multivariate analysis were PS, PE and/or ascites and BSLD
(10 cm) (Table 3). PE and/or ascites could become the most
significant variable in the present study.
1.658
0.970– 2.833
1.000
512
586
LDH (IU/l)
300
0.0139
,300
1.919
1.142– 3.226
1.000
399
596
CEA (ng/ml)
DISCUSSION
5
0.0554
,5
Although we have experienced that the patients with huge
TV would have poor prognosis and have used RECIST in
many clinical trials, it is not clear whether BSLD of target
lesion could be one of prognostic factors on overall survival
2.457
0.979– 6.173
1.000
521
960
BSLD (cm)
Continuous
0.0004
1.003
1.001– 1.005
Continued
Jpn J Clin Oncol 2008;38(10)
Table 2. Continued
P value
5
0.0132
,5
10
95% CI
MST
(days)
2.519
1.212–5.236
1.000
0.0005
,10
14.3 (median)
HR
2.770
1.562–4.902
1.000
0.0161
,14.3 (median)
1.898
512
730
373
708
1.126–3.195
1.000
386
596
HR, hazard ratio; 95% CI, 95% confidential interval; MST, median overall
survival time.
Figure 2. Survival of patients. BSLD ,10 cm and BSLD 10 cm.
693
In this study, PE and/or ascites was the most significant
variable and was not in the previous reports. Although
patients having PE and/or ascites are usually excluded from
clinical trials, they were not excluded from the selection criteria in the present study, because they are routinely treated
in clinical practice setting. The reason that PE and/or ascites
could be the most significant variable might be derived from
the difference of the patient population. And this study
suggests that it may be reasonable that usual clinical trials
exclude patients with PE and/or ascites.
The univariate and subsequent multivariate analyses
demonstrated that ECOG PS, PE and/or ascites and BSLD
(10 cm) before the initial chemotherapy could be the independent prognostic factors on overall survival. The ECOG
PS had previously been reported as one of the prognostic
factors in mCRC patients in several studies (14 – 20). The
results of the present study were compatible with those of
the past studies.
It is suggested that the BSLD (10 cm) might be an independent prognostic factor.
As our findings were based on a retrospective analysis in
the single institution, validation of the usefulness of BSLD
of target lesions is necessary and ongoing, using the other
prospective cohorts.
In conclusion, the retrospective study might suggest that
the BSLD of target lesions could become an independent
prognostic factor for mCRC.
Conflict of interest statement
None declared.
Table 3. Factors associated with overall survival (OS) in the multivariate
analysis
Variables
P value
HR
HR 95% CI
ECOG PS
0.0012
0.248
0.107–0.577
PE and/or ascites
0.0016
0.402
0.228–0.708
BSLD (10 cm)
0.0059
0.431
0.237–0.785
Analysis including individual parameters of Köhne index.
in patients with unresectable or mCRC. We retrospectively
investigated the correlation of overall survival with clinical
characteristics of patients who underwent the front-line chemotherapy at our institution. In the present study, patients
treated with oxaliplatin-containing regimen in the first-line
setting were excluded because the application of Köhne’s
index to oxaliplatin-based regimen was not validated and
oxaliplatin had not been approved for mCRC in Japan until
April 2005. On the other hand, irinotecan-containing
regimen was included because the Köhne’s index can be
applicable to irinotecan-based regimen and to 5-FU
monotherapy (23).
References
1. Jemal A, Siegel R, Ward E, Murray T, Xu J, Thun MJ, et al. Cancer
statistics, 2007. CA Cancer J Clin 2007;57:43– 66.
2. August DA, Sugarbaker PH, Ottow RT, et al. Hepatic resection of
colorectal metastases: influence of clinical factors and adjuvant
intraperitoneal 5-fluorouracil via Tenckhoff catheter on survival. Ann
Surg 1985;201:210– 8.
3. Yoshimi I, Sobue T. Current status and trends in cancer mortality in
Japan. Gan To Kagaku Ryoho 2004;31:832– 9.
4. de Gramont A, Figer A, Seymour M, et al. Leucovorin and fluorouracil
with or without oxaliplatin as first-line treatment in advanced colorectal
cancer. J Clin Oncol 2000;18:2938 –47.
5. Giacchetti S, Perpoint B, Zidani R, et al. Phase III multicenter
randomized trial of oxaliplatin added to chronomodulated fluorouracil
leucovorin as first-line treatment of metastatic colorectal cancer. J Clin
Oncol 2000;18:136–47.
6. Grothey A, Deschler B, Kroening H, et al. Phase III study of bolus
5-fluorouracil (5-FU)/folinic acid (FA) (Mayo) vs weekly high-dose
24 h 5-FU infusion/FAþoxaliplatin (OXA) (FUFOX) in advanced
colorectal cancer (ACRC). Proc Am Soc Clin Oncol 2002;21:129a.
7. Tournigand C, Andre T, Achille E, et al. FOLFIRI Followed by
FOLFOX6 or the reverse sequence in advanced colorectal cancer: a
randomized GERCOR Study. J Clin Oncol 2004;22:229– 37.
8. Goldberg RM, Sargent DJ, Morton RF, et al. A randomized controlled
trial of Fluorouracil plus leucovorin, irinotecan, and oxaliplatin
combinations in patients with previously untreated metastatic colorectal
cancer. J Clin Oncol 2004;22:23– 30.
694
Survival impact of baseline sum of longest diameter
9. Grothey A, Sargent D, Goldberg RM, Schmoll HJ. Survival of patients
with advanced colorectal cancer improves with the availability of
fluorouracil-leucovorin, irinotecan, and oxaliplatin in the course
of treatment. J Clin Oncol 2004;22:1209– 14.
10. Köhne CH, Van Cutsem E, Wils JA, et al. Phase III study of weekly
high-dose infusional fluorouracil plus folinic acid with or without
irinotecan in patients with metastatic colorectal cancer: European
Organisation for Research and Treatment of Cancer Gastrointestinal
Group Study 40986. J Clin Oncol 2005;23:4856–65.
11. Hurwitz H, Fehrenbacher L, Novotny W, et al. Bevacizumab plus
irinotecan, fluorouracil, and leucovorin for metastatic colorectal cancer.
N Engl J Med 2004;350:2335–42.
12. Hurwitz Herbert I, Fehrenbacher L, Hainsworth JD, et al. Bevacizumab
in combination with fluorouracil and leucovorin: an active regimen
for first-line metastatic colorectal cancer. J Clin Oncol 2005;23:
3502–8.
13. Köhne CH, Cunningham D, Di CF, et al. Clinical determinants of
survival in patients with 5-fluorouracil-based treatment for metastatic
colorectal cancer: results of a multivariate analysis of 3825 patients.
Ann Oncol 2002;13:308–17.
14. Cunningham D, Pyrhonen S, James RD, et al. Randomised trial of
irinotecan plus supportive care versus supportive care alone after
fluorouracil failure for patients with metastatic colorectal cancer. Lancet
1998;352:1413– 8.
15. Knight R, Miller L, Pirotta N, et al. First-line irinotecan (C),
fluorouracil (F), leucovorin (L) especially improves survival (OS) in
metastatic colorectal cancer (MCRC) patients (PT) with favorable
prognostic indicators. Proc Am Soc Clin Oncol 2000;19:255a.
(Abstr 991).
16. Assersohn L, Norman A, Cunningham D, et al. Influence of metastatic
site as an additional predictor for response and outcome in advanced
colorectal carcinoma. Br J Cancer 1999;79:1800– 5.
17. Kouri M, Pyrhonen S, Kuusela P, et al. Elevated CA19-9 as the most
significant prognostic factor in advanced colorectal carcinoma. J Surg
Oncol 1992;49:78– 85.
18. Steinberg J, Erlichman C, Gadalla T, et al. Prognostic factors in patients
with metastatic colorectal cancer receiving 5-fluorouracil and folinic
acid. Eur J Cancer 1992;28A:1817–20.
19. Steinberg SM, Barkin JS, Kaplan RS, Stablein DM, et al. Prognostic
indicators of colon tumors. The Gastrointestinal Tumor Study Group
experience. Cancer 1986;57:1866 –70.
20. Kemeny N, Braun DW, Jr, et al. Prognostic factors in advanced
colorectal carcinoma. Importance of lactic dehydrogenase level,
performance status, and white blood cell count. Am J Med 1983;
74:786– 94.
21. Mitry E, Douillard J-Y, et al. Predictive factors of survival in patients
with advanced colorectal cancer: an individual data analysis of 602
patients included in irinotecan phase III trials. Ann Oncol 2004;15:
1013– 7.
22. Chang AE, Steinberg SM, Culnane M, White DE, et al. Determinants
of survival in patients with unresectable colorectal liver metastases.
J Surg Oncol 1989;40:245– 51.
23. Hurwitz Zurlo A, et al. Analysis of outcomes of patients with metastatic
colorectal cancer (mCRC) treated with IFL with or without
bevacizumab (BV) in a phase III clinical trial based on baseline risk.
Proc Am Soc Clin Oncol 2006; (Abstr 249).
24. Webb A, Scott-Mackie P, Cunningham D, et al. The prognostic value of
CEA, beta HCG, AFP, CA125, CA19-9 and C-erb B-2, beta HCG
immunohistochemistry in advanced colorectal cancer. Ann Oncol
1995;6:581 –7.
25. Therasse P, Arbuck SG, Eisenhauer EA, Wanders J, et al. New
guidelines to evaluate the response to treatment in solid tumors.
European Organization for Research and Treatment of Cancer, National
Cancer Institute of the United States, National Cancer Institute of
Canada. J Natl Cancer Inst. 2000;92:205–16.