ARTICLE
Auteur(s) : P
Sève1,2,3,4, J Mackey4, M
Sawyer4, T Lesimple5, C de
La Fouchardière6, C Broussolle1,2,3, C
Dumontet1,3, I Ray-Coquard1,6,3,7
1Université de Lyon, Lyon 69008, France
2Inserm U590, Lyon 69008, France
3Hospices civils de Lyon, Lyon 69003, France
4Department of oncology, university
of Alberta, Edmonton, T6G 1Z2 Alberta, Canada
5Department of oncology, centre Eugène-Marquis,
35042 Rennes, France
6Department of oncology, centre Léon-Bérard, 28,
rue Laennec, 69008 Lyon, France
7Inserm EA 4129 S-I-S, hospices civils de Lyon, 69676
Bron cedex Lyon, France
Article reçu le 2 Juin 2008, accepté le 9 Janvier 2009
Introduction
Cancers of unknown primary site (CUP) represent a group of
heterogeneous tumors that share the clinical characteristic of
being metastatic epithelial diseases with no identifiable origin at
the time of diagnosis; CUP comprise 2-5% of all new cancer cases
[1, 2]. In North America and Europe, organ specialists who referred
the patients to oncologists for definitive workup and management
usually made the diagnosis of CUP. Our impression was that the
patients reviewed had undergone an inappropriately high number of
diagnostic investigations, including endoscopies with potentially
harmful complications, high costs and adverse effects on quality of
life. This standard of practice is still being used although
several studies have demonstrated the relative futility of this
approach and, more importantly, its lack of impact on overall
survival and treatment [3, 4]. Although evidence-based medicine has
defined no standard for the systemic treatment of CUP that do not
belong to a specific anatomoclinical entity, chemotherapy is
generally considered for patients with a good performance status
(PS, 0 or 1) [5]. We have previously shown an age-related decline
in the percentage of adults with CUP who receive chemotherapy,
independently of other factors, suggesting that older CUP patients
might be under-treated [6].
The French national Clinical Practice Guidelines (CPGs) for CUP
management were published in 2002 and distributed to French
oncologists [7]. These guidelines, called Standards, Options,
Recommendations (SOR) are a nationwide project of the Fédération
nationale des centres de lutte contre le cancer (French Federation
of Comprehensive Cancer Centers, FNCLCC) designed to improve
patient care. We have previously shown that developing and
implementing regional CPGs could change medical practice for
localized breast and colon cancer [8]. More recently, we showed in
a controlled “before-after” study the existence of a positive
impact of CPGs and Regional cancer network organization on managed
care for breast and colon cancer patients [9]. Furthermore, we
demonstrated that conformity of the cancer network organization
with SOR guidelines is an important prognostic factor for optimal
clinical care, notably for patients with rare tumors [10].
To our knowledge, the conformity of CUP management with
published CPGs has never been investigated. We therefore conducted
a “before-after” study to evaluate the effectiveness of the SOR
guidelines. An external medical record audit was undertaken in two
countries to assess the impact of the SOR for CUP management on
diagnostic strategies. The experimental site was the Centre
Leon-Bérard, a comprehensive cancer center in Lyon, France. The
control site was the Cross Cancer institute, a comprehensive cancer
center in Edmonton, Canada, where no national or institutional CPGs
for CUP management were available at the time of data
collection.
The primary objective of this study was to describe the primary
diagnostic workup of CUP and assess the effects of introducing the
SOR guidelines. To do so, we determined the compliance of medical
practice with the SOR before and after SOR introduction at the
experimental site. Secondary objectives were to determine:
- – the rates of inappropriate investigations performed as
compared to the CPGs;
- – the usefulness of investigations not scheduled in the
CPGs;
- – whether investigations performed after CUP diagnosis
delayed treatment;
- – whether introducing CPGs changed clinical practice for
CUP treatment.
Patients, materials and methods
Clinical practice guidelines
The CPGs were a nationwide project of the FNCLCC. The first volume
of the SOR guidelines for the clinical management of cancer was
published in 1995. These SOR, and the following ones, were aimed at
board-certified oncologists only; no other dissemination of
information was planned. A summary of the CUP CPGs, intended
for patients, was prepared, based on the full text version
available on the Internet (http ://www.fnlcc.fr), and was published
in the French medical literature in August 2002 [7].
A translation of these CPGs was published in the
English-language medical literature in August 2003 [11]. The CPGs
concerned pathological examination, diagnostic strategies,
prognostic factors and CUP treatment. Diagnostic and treatment
strategies are summarized below.
Optimal workup for patients with CUP (outline from
the “Standards, Options et Recommandations pour la prise en
charge des patients atteints de carcinomes de site primitif
inconnu“ (SOR, 2002))
The diagnostic workup of CUP tumors as a function of their
histopathological and anatomic localization consists of three steps
[11]. The first step involves routine diagnostic workup including
pathological evaluation [11], clinical history and physical
examination, and a chest X-ray. Additional tests are not indicated
at this stage. The second step is a specific procedure based on
histopathology. In neuroendocrine and squamous cell carcinoma, no
specific workup is required to identify the primary site. In
adenocarcinoma and undifferentiated carcinoma, a mammography and an
ultrasound or pelvic CT scan are required for women and serum PSA,
αFP, βHCG assays are recommended for men. Step three consists in a
specific workup according to localization. For adenocarcinoma and
undifferentiated carcinoma, 11 different localizations have been
described (midline lymphadenopathies (LN); mediastinal LN; cervical
and/or supraclavicular LN; axillary LN; inguinal LN; liver
metastases; lung metastases; bone metastases; brain metastases;
single metastasis; pleural effusion; ascites), each with specific
clinical recommendations. For squamous cell carcinoma, four
localizations have been described (cervical and/or supraclavicular
LN; axillary LN; inguinal LN; bone metastases) with specific
clinical recommendations.
Treatment strategy for CUP patients who do not belong
to a favourable clinicopathologic entity
Favourable clinicopathologic entity includes men with poorly
differentiated carcinoma and “extragonadal syndrome” features,
women with axillary lymph-node metastases or peritoneal
adenocarcinomatosis, or patients with cervical lymph nodes,
squamous carcinoma and neuroendocrine carcinomas [11].
As most patients belong to this category, we restricted our
assessment of compliance with guidelines to this group. Several
systemic treatments can be envisaged (options): chemotherapy (level
of evidence B2) or symptomatic treatment only. If chemotherapy is
prescribed, a cisplatin-based combination of two drugs is
recommended (recommendations, expert agreement) for patients with a
good PS (0 or 1).
Data sources
Data on the primary workup of CUP patients were collected by one of
the investigators (PS) and analyzed by two of us (PS and IRC). PS
is a medical oncologist independent of practitioners caring for
patients in the different sites. It was impossible to perform the
analysis blinded to the year of treatment since data were obtained
directly from patients' records. To control the extracted data, the
same investigators, blinded to the previous decision, rechecked a
random 10% sample. Agreement with the original decision was 95%
(kappa test = 0.89).
Selection of patients' records
All newly diagnosed CUP patients registered at experimental and
control sites between January 2000 and December 2001 (“before”
implementation of the SOR) and between January 2003 and December
2004 (“after” implementation of the SOR) were eligible. Patients
meeting the following criteria were excluded from further analysis:
- – clinical diagnosis without pathologic evidence of
malignancy, excluding patients with cerebral metastases;
- – obvious primary tumor identified at time of initial
visit;
- – CUP of non-epithelial origin;
- – inappropriate data registration;
- – missing data.
At the experimental site, 90 and 128 patients were newly
referred during the “before” and “after” periods, respectively. At
the control site, 329 and 285 patients were newly referred during
the same periods. For this study, 50 patient records were necessary
to detect a statistically significant increase in compliance rates
[12]. Patients from the experimental center were randomly selected
by automatic extraction from the computerized database. Control
patients were recruited from the unknown primary clinic at the
Cross Cancer Institute. Because of high recruitment at the control
site, many CUP patients were potentially available; a computerized
procedure (random numbers) was used to select 50 among about 300
patients for each period. Randomly selected patients who were
deemed not eligible were ultimately excluded and replaced by
others. Fifty cases are considered to be sufficient to detect
differences in pre and post introduction of guidelines [12].
Measurements
Electronic and paper charts of selected patients were reviewed.
Research database entries included the following data: histology;
age; gender; PS; comorbidity evaluated using the
adult-comorbidity-evaluation-27 (ACE-27) index [6, 13, 14];
belonging to a specific entity; number of metastases; date of
diagnosis; characteristics of referring physician(s): age, gender,
specialty; site(s) of metastases; investigations performed after
pathologic diagnosis: tumor marker(s), radiologic examination(s),
endoscopy(ies), scintigraphy; utility of non-guideline
investigations, whether contributive or not; date of treatment.
The primary objective of the study was:
- – to describe the rate of overall diagnostic sequences
judged to conform to SOR. An overall diagnostic sequence included
decisions for each diagnostic step (step 1: routine diagnostic
workup; step 2: specific workup as a function of histopathology;
step 3: specific workup as a function of localization). Each
procedure was individually assessed for conformity with the SOR as
described above. Overall diagnostic evaluation was considered
compliant when all SOR recommended procedures were performed.
Univariate and multivariate analyses were performed to identify
patients' or physicians' characteristics possibly associated with
reduced compliance with SOR guidelines.
Secondary objectives were:
- – to determine the rates of inappropriate investigations
performed after pathologic diagnosis, according to the SOR in both
pre- and post-SOR publication groups;
- – to determine the usefulness of non-SOR recommended
investigations. To address this question, we scored non-SOR
procedures as to whether they provided diagnosis-relevant
information;
- – to investigate if investigations performed after the
diagnosis of CUP delayed treatment start. We determined whether the
presence of SOR-noncompliant investigations was associated with a
longer interval between CUP histological diagnosis (date of first
diagnostic biopsy) and first therapy;
- – to determine if introducing SOR changed clinical
practice for CUP treatment. We determined the proportion of
patients not belonging to a favorable clinicopathologic entity,
with a good PS, who received cisplatin-based combination
therapy.
Statistical analysis
Conformity was scored 1 when all the recommendations had been
followed (independently of the time of performing the tests), and 0
otherwise. Categorical data were analyzed using Pearson's
χ2 or Fisher's exact tests, as appropriate. Continuous
data were analyzed with Student's t-test. The statistical
significance level was set at P = 0.05 in a two-sided test.
Univariate analyses were performed to assess whether patients' or
physicians' characteristics were correlated with conformity or
non-conformity to the SOR. A logistic regression analysis
including parameters studied in the univariate analysis was
performed using SPSS Logistic software: a step-down regression
analysis of conformity rates was conducted using a P value inferior
to 0.05 for entry.
Patient consent
The prognosis of metastatic carcinomas of unknown primary site is
very poor, and even with modern taxane- and gemcitabine-based
chemotherapy regimens median survival is limited, in the range of
seven to ten months. It was therefore not possible to obtain
consent from patients surveyed in this study, since most of them
had died. This study had sufficient quality assurance and
scientific value to obtain approval from the local institutional
review board in the absence of specific consent from the patients.
Results
Patient characteristics
In the experimental group, 33 and 38 records selected within
2000-2001 and within 2003-2004, respectively, did not meet the
selection criteria and were ultimately replaced by randomly
selected, appropriate subjects. In the control group, 32 and 35
records selected within 2000-2001 and within 2003-2004,
respectively, were not assessable for overall diagnostic sequence.
At both sites, patient characteristics were similar for the 2
periods (table 1). In the experimental
group, an oncologist and a non-oncologist organized almost all
diagnostic workup procedures. The specific specialities of the
non-oncologist physicians were the following: 2 general
practitioners, 4 pneumonologists, 5 gastroenterologists, 5
rheumatologists, 4 internists, 14 general surgeons, 3 ENT surgeons,
between 2000 and 2001; and 2 general practitioners, 6
pneumonologists, 7 gastroenterologists, 5 rheumatologists, 3
internists, 8 general surgeons, 2 ENT surgeons, between 2003 and
2004.
Table 1 CUP patient characteristics.
|
Experimental group
|
Control group
|
|
2000-2001
|
2003-2004
|
P
|
2000-2001
|
2003-2004
|
P
|
|
Age
|
|
|
|
|
|
|
|
Median
|
64.5
|
62
|
0.54
|
63
|
64.5
|
0.62
|
|
Range
|
27-82
|
36-82
|
|
19-85
|
29-86
|
|
|
Sex
|
|
|
|
|
|
|
|
Women
|
23
|
23
|
1
|
24
|
31
|
0.16
|
|
Men
|
27
|
27
|
|
26
|
19
|
|
|
Overall comorbidities-score
|
|
|
|
|
|
|
|
ACE-27 grade 0-1
|
37
|
32
|
0.2
|
34
|
32
|
0.6
|
|
ACE-27 grade 2-3
|
13
|
18
|
8
|
16
|
18
|
7
|
|
Histology
|
|
|
|
|
|
|
|
Well differentiated adenocarcinoma
|
21
|
16
|
0.3
|
24
|
26
|
0.6
|
|
Poorly differentiated adenocarcinoma or carcinoma
|
10
|
17
|
|
14
|
12
|
9
|
|
Undifferentiated carcinoma
|
7
|
4
|
|
1
|
3
|
|
|
Squamous cell carcinoma
|
4
|
9
|
|
7
|
5
|
|
|
Othersa
|
8
|
4
|
|
4
|
4
|
|
|
PS
|
|
|
|
|
|
|
|
0-1
|
29
|
30
|
0.64
|
32
|
23
|
0.07
|
|
2-4
|
20
|
17
|
|
17
|
26
|
|
|
Unknown
|
1
|
3
|
|
1
|
1
|
|
|
Favorable clinicopathologic entities
|
|
|
|
|
|
|
|
Yes
|
9
|
9
|
1
|
9
|
5
|
0.2
|
|
No
|
41
|
41
|
|
41
|
45
|
5
|
|
Number of metastatic sites
|
|
|
|
|
|
|
|
1
|
27
|
25
|
0.6
|
23
|
29
|
0.2
|
|
2
|
14
|
18
|
9
|
19
|
15
|
3
|
|
>2
|
9
|
7
|
|
8
|
6
|
|
|
Diagnostic workup
|
|
|
|
|
|
|
|
Oncologist alone
|
13
|
17
|
0.3
|
3
|
1
|
0.3
|
|
Oncologist and non-oncologist
|
37
|
33
|
8
|
47
|
49
|
1
|
aOthers include: undifferentiated tumors, mixed
histological tumors and cerebral metastases without histological
proof of cancer.
Characteristics of oncologists
In the experimental group, 22 and 23 oncologists participated in
the diagnostic evaluation during the “before” and “after” periods,
respectively, versus 17 and 15 in the control group. Table 2 summarizes the characteristics of
physicians participating in the study. There was no difference
between the two periods in the control group, whereas in the
experimental group, a new medical oncologist managed more than half
of the patients (54%) in the second period.
Table 2 Characteristics of oncologists.
|
Experimental group
|
Control group
|
|
2000-2001
|
2003-2004
|
2000-2001
|
2003-2004
|
|
Number of oncologists involved
|
22
|
23
|
17
|
15
|
|
Age
|
|
|
|
|
|
Median
|
40
|
42
|
45
|
41
|
|
Sex
|
|
|
|
|
|
Women
|
8
|
8
|
7
|
5
|
|
Men
|
4
|
15
|
10
|
10
|
|
Specialty
|
|
|
|
|
|
Medical oncology
|
16
|
15
|
10
|
9
|
|
Radiation oncology
|
5
|
7
|
7
|
6
|
|
Surgery
|
1
|
1
|
0
|
0
|
Compliance with CPGs
Experimental group
We detected no difference in observed SOR compliance rates between
the 50 overall diagnosis sequences of the “before” (48%; 24 out of
50) and “after” (48%; 24 out of 50) periods (table 3). Compliances rate were similar whether
diagnosis was made by an oncologist alone or by both an oncologist
and a non-oncologist (data not shown). Observed SOR compliance
rates for each diagnosis step were also similar for the two
periods, except for step 3. For step 1, the main reasons for
non-compliance were inadequate pathologic examination (100%),
cytologic proof of cancer only (24 and 18%, in pre- and post-period
respectively), and/or lack of immunohistochemical investigations
(38 and 22%, in pre- and post-period respectively). For step 2, the
main reasons for non-compliance were the lack of αFP and βHCG
testing in men: 14 out of 18 (78%) within 2000-2001, and 10 out of
11 (91%) within 2003-2004.
Table 3 Compliance rates for medical decision as a
function of diagnosis step.
|
|
Experimental group
|
Control group
|
|
Number of decision assessed
|
Percentage of compliance with SORs (95% CI)
|
Percentage of compliance with SORs (95% CI)
|
|
2000-2001
|
2003-2004
|
2000-2001
|
2003-2004
|
P
|
2000-2001
|
2003-2004
|
P
|
|
Step 1: routine diagnostic workup and pathologic examination
|
50
|
50
|
30 (60%)
|
37 (74%)
|
0.14
|
27 (54%)
|
31 (62%)
|
0.42
|
|
Step 2: specific workup depending on histopathology
|
40 (39)
|
40 (42)
|
22 (55%)
|
29 (73.5%)
|
0.10
|
9 (23%)
|
20 (47%)
|
0.02
|
|
Step 3: specific workup as a function of localization
|
25 (31)
|
30 (39)
|
24 (96%)
|
23 (77%)
|
0.04
|
14 (45%)
|
21 (54%)
|
0.47
|
|
Overall diagnostic sequence
|
50
|
50
|
24 (48%)
|
24 (48%)
|
1
|
10 (20%)
|
15 (30%)
|
0.25
|
Control group
We detected no difference in observed SOR compliance rates between
the 50 overall diagnosis sequences within 2000-2001 (20%; 10 out of
50) and within 2003-2004 (30%; 15 out of 50). Observed SOR
compliance rates for each diagnosis step were also similar for the
two periods, except for step 2. For step 1, the main reasons for
non-compliance were: inadequate pathologic examination (100%),
cytologic proof of cancer only (26 and 34%, in pre- and post-period
respectively) and/or lack of immunohistochemical investigations (44
and 38%, in pre- and post-period respectively). For step 2, reasons
were the lack of αFP and βHCG testing in men (15 out of 30 [50%]
within 2000-2001, and 9 out of 22 [41%] within 2003-2004), and the
lack of pelvic examination (10 out of 30 [33%] within 2000-2001,
and 6 out of 22 [27%] within 2003-2004) and mammography (6 out of
30 [20%] within 2000-2001, 10 out of 22 [45%] within 2003-2004) in
women. We observed many heterogeneous reasons for non-compliance in
step 3 (data not shown).
Patient and physician characteristics and rate
of non-compliance with SOR guidelines
in experimental group
For patients in both pre and post periods, results from univariate
analyses showed that compliance with SOR guidelines was
significantly related to the following variables: patient age less
61 years (P = 0.01), PS < 2 vs. PS > 2 (P = 0.006),
non-adenocarcinoma vs. adenocarcinoma (P <0.001). Patient's
gender, ACE-27 index, the number of metastatic sites, belonging to
a favourable clinicopathologic entity, the specialities of the
non-oncologists physicians, the oncologists' gender, age and
category were not significant. Multivariate analysis of significant
factors found PS < 2 (P = 0.02) and non-adenocarcinoma histology
(P = 0.002) to be independently significant predictors of
compliance with SOR guidelines. PS < 2 yielded a hazard ratio of
3.12 (95% CI, 1.2-8.12) and non-adenocarcinoma histology yielded a
hazard ratio of 4.76 (95% CI, 1.8-12.65). Most adenocarcinoma
patients did not have adequate immunohistochemical analysis (51.4
vs. 22.6% for patients with non-adenocarcinoma histology; P =
0.003) or recommended germinal tumor markers assays (45.9 vs. 19%
for patients with non-adenocarcinoma histology; P = 0.01).
Inappropriate investigations after pathologic diagnosis
Experimental group
We detected no difference in rates of inappropriate investigations
for the 50 overall diagnostic sequences performed within 2000-2001
(74%; 37 out of 50) and within 2003-2004 (76%; 76 out of 50) (table 4). Rates of inappropriate
investigations were similar whether the diagnosis was made by an
oncologist alone or by both an oncologist and a non-oncologist.
Observed inappropriate investigation rates according to the SORs
for each diagnostic test were also similar for the two periods,
except for positron emission tomography (PET) which was more
frequent in the post-period.
Table 4 Inappropriate investigations according to the
SORs.
|
Experimental group
|
Control group
|
|
Percentage of inappropriate investigations according to the SORs
(95% CI)
|
Percentage of inappropriate investigations according to the SORs
(95% CI)
|
|
2000-2001
|
2003-2004
|
P
|
2000-2001
|
2003-2004
|
P
|
|
Tumor markers
|
22 (44%)
|
21 (42%)
|
0.84
|
22 (44%)
|
21 (42%)
|
0.84
|
|
Radiological investigations
|
24 (48%)
|
21 (42%)
|
0.42
|
17 (34%)
|
23 (46%)
|
0.22
|
|
Chest CT
|
13
|
13
|
NS
|
12
|
14
|
NS
|
|
Abdominal CT
|
8
|
10
|
NS
|
10
|
10
|
NS
|
|
Thyroid US
|
4
|
6
|
NS
|
1
|
4
|
NS
|
|
Others
|
12
|
14
|
NS
|
8
|
13
|
NS
|
|
Endoscopies
|
24 (48%)
|
20 (40%)
|
0.42
|
9 (18%)
|
17 (34%)
|
0.07
|
|
Gastroscopy
|
17
|
15
|
NS
|
4
|
13
|
0.02
|
|
Colonoscopy
|
16
|
9
|
NS
|
4
|
9
|
NS
|
|
Bronchoscopy
|
9
|
12
|
NS
|
2
|
5
|
NS
|
|
Others
|
5
|
2
|
NS
|
1
|
0
|
NS
|
|
Scintigraphy
|
9 (18%)
|
15 (30%)
|
0.16
|
9 (18%)
|
8 (16%)
|
0.79
|
|
Bone scintigraphy
|
6
|
6
|
NS
|
4
|
3
|
NS
|
|
PET
|
2
|
11
|
0.007
|
3
|
5
|
NS
|
|
Others
|
3
|
1
|
NS
|
2
|
2
|
NS
|
|
Overall inappropriate investigations
|
37 (74%)
|
38 (76%)
|
0.81
|
34 (68%)
|
39 (78%)
|
0.26
|
Control group
We detected no difference in inappropriate investigation rates for
the 50 overall diagnostic sequences performed within 2000-2001
(68%; 34 out of 50) and within 2003-2004 (78%; 39 out of 50) (table 4). Rates of inappropriate
investigations were similar whether the diagnosis was made by an
oncologist alone or by both an oncologist and a non-oncologist.
Observed inappropriate investigation rates according to the SORs
for each diagnostic test were also similar for the two periods,
except for endoscopies, which were more frequent in the
post-period. We observed a trend toward over-frequent prescription
of upper endoscopies. One medical oncologist ordered nine of the
thirteen upper endoscopies performed.
Usefulness of out-of-SOR guideline investigations
in the experimental group
By definition, tumor markers, endoscopies and radiologic tests were
unsuccessful in locating primary carcinomas. However, chest and
abdominal CT detected previously unrecognized metastases in
respectively 28 and 40% of the patients, and PET imaging in 28%.
Delayed start of treatment in the experimental
group
Investigations performed after the diagnosis of CUP did not delay
the start of treatment. The SOR-noncompliant investigations
performed in 64 patients were associated with a 36-day mean
interval between CUP histological diagnosis (date of first
diagnostic biopsy) and first therapy, which was not significantly
different from the 28-day mean interval observed in the 15 patients
with SOR-compliant investigations (P = 0.30).
Clinical practice for CUP treatment
Experimental group
Eighty-nine of the 100 patients of the experimental group received
at least one specific anticancer treatment (chemotherapy,
radiotherapy, or hormonal therapy). Among the 54 patients who did
not belong to a favorable clinicopathologic entity, with a PS of 0
or 1, 44 received chemotherapy (table
5). No difference in compliance with SOR guidelines for CUP
treatment was observed between periods: within 2000-2001, 6 of 22
patients (27%) with good PS received a cisplatin-based combination
therapy, whereas within 2003-2004, 9 of 24 patients (37.5%)
received the recommended treatment (P = 0.45). However, we detected
significant changes regarding chemotherapy regimens. Within
2000-2001, 11 patients with good PS received breast or colon cancer
regimens, one received gemcitabine-based chemotherapy and none
received taxanes, whereas in the post-period, 4 patients received
breast or colon cancer regimens (P = 0.02), 8 received
gemcitabine-based chemotherapy (P = 0.01) either as a single agent
(N = 2) or in combination with oxaliplatin (N = 2) or cisplatin (N
= 4), and 6 received taxane-based chemotherapy (P = 0.01).
Table 5 Chemotherapy for patients with good PS not
belonging to a favorable clinicopathologic entity.
|
Experimental group
|
Control group
|
|
2000-2001
|
2003-2004
|
2000-2001
|
2003-2004
|
|
Platinum-based regimens
|
7
|
11
|
2
|
0
|
|
Cisplatin/etoposide
|
5
|
1
|
1
|
|
|
Cisplatin/gemcitabine
|
1
|
4
|
0
|
|
|
Cisplatin/docetaxel
|
0
|
3
|
0
|
|
|
Cisplatin/5-fluorouracil (5-FU)
|
0
|
1
|
1
|
|
|
Cisplatin
|
0
|
0
|
0
|
|
|
Carboplatin
|
1
|
2
|
0
|
|
|
Taxane-based regimens
|
0
|
3
|
14
|
12
|
|
Carboplatin/paclitaxel
|
|
2
|
3
|
9
|
|
Carboplatin/paclitaxel/etoposide
|
|
0
|
11
|
3
|
|
Docetaxel
|
|
1
|
|
|
|
Paclitaxel
|
|
0
|
|
|
|
Gemcitabine-based regimens
|
0
|
4
|
0
|
0
|
|
Gemcitabine alone
|
|
2
|
|
|
|
Gemcitabine/oxaliplatin
|
|
2
|
|
|
|
Colon anticancer regimens
|
5
|
3
|
2
|
1
|
|
5-FU/leucovorin (LV)
|
1
|
1
|
1
|
0
|
|
5-FU/LV/irinotecan
|
0
|
0
|
1
|
1
|
|
5-FU/LV/oxaliplatin
|
3
|
2
|
0
|
0
|
|
Raltitrexed
|
1
|
0
|
0
|
0
|
|
Breast anticancer regimens
|
6
|
1
|
0
|
0
|
|
5-FU/epirubicin/cyclophosphamide
|
3
|
0
|
|
|
|
5-FU/doxorubicin/cyclophosphamide
|
2
|
1
|
|
|
|
Epirubicin/vinorelbine
|
1
|
|
|
|
|
Others
|
4
|
1
|
0
|
0
|
|
No chemotherapy
|
4
|
4
|
7
|
5
|
|
Total
|
26
|
28
|
25
|
18
|
Control group
Seventy-three of the 100 patients of the control group received at
least one specific anticancer treatment (chemotherapy,
radiotherapy, or hormonal therapy). Among the 43 patients who did
not belong to a favorable clinicopathologic entity, with a PS of 0
or 1.31 received chemotherapy (table 5).
We did not detect significant changes in regimens between the two
periods; most patients received taxane/carboplatin-based
combinations.
Discussion
Our study is the first to assess medical practices for CUP patients
in two different countries and their conformity with
“evidence-based medicine” promoted through CPGs. Since our main
purpose was to assess the impact of CPGs on CUP management, we did
not provide any statistical comparison between experimental and
control groups. However, we observed numerous differences between
the different countries, which could be related to their health
policy regulations, their physician training, etc. Overall, this
controlled “before-after” study of medical compliance did not allow
detecting significant practice changes after the publication of
CPGs in either the experimental or the control groups. Only 50% of
medical decisions for patient diagnostic workup were in accordance
with CPGs at the experimental site. Although several studies have
shown that morphology is the basis for CUP diagnosis and may be
usefully supplemented by immunohistochemistry, one third of our
patients had inadequate pathologic evaluation [15]. The main reason
for non-adherence to diagnostic recommendations for step 2 in the
experimental group was the lack of αFP and βHCG testing in men with
adenocarcinoma and poorly differentiated carcinoma. This
recommendation, which is of particular importance for young men
presenting with “midline syndrome”, could not be considered, in our
and other authors opinion, for all men whatever the clinical
presentation [5, 15]. Poor PS and adenocarcinoma histology were two
significant predictors of non-compliance to diagnostic
recommendations. Several studies have shown that PS and other
patient characteristics could influence medical decisions [16].
Most adenocarcinoma patients did not have adequate
immunohistochemical analysis and recommended tumor markers.
Approximately, 75% of our patients had inappropriate
investigations including tumor markers, radiological
investigations, endoscopic examinations and nuclear imaging. This
standard of practice is still being applied despite the results of
several studies that have demonstrated the relative futility of
extensive investigations, the high costs incurred and, more
importantly, the lack of impact on overall treatment and survival
[4, 15]. The SOR guidelines recommend systematic endoscopic
examination only for CUP patients with specific clinical
presentations: ENT endoscopy is recommended in patients with
isolated cervical node involvement, proctoscopy and/colposcopy in
patients with inguinal node involvement [7, 11]. Colonoscopy is
recommended in patients with isolated, resectable liver metastases.
In our opinion, colonoscopy should be performed in the presence of
isolated liver metastases in patients with PS < 2, since the
emergence of new chemotherapy protocols and targeted therapy has
improved metastatic colorectal cancer prognosis. Interestingly,
chest CT, abdominal CT and FDG-PET imaging led to the detection of
previously unrecognized metastases in respectively 28, 40 and 28%
of the patients. The SOR guidelines, which are mainly based on
Abbruzzese's work, do not recommend CT scan, except for patients
with specific clinical presentations such as midline syndrome [7,
11, 15]. However, several authors, based on their experience or old
studies, have recommended CT scan for detection of the primary,
determination of the extent of the metastatic disease and provision
of guidance in selecting the optimal biopsy site [17]. Studies of
PET in disseminated CUP patients have demonstrated that this new
imaging method leads to the detection of previously unrecognized
metastases in 37% of patients [18]. However, its benefit is
currently restricted to patients with a single metastatic site. In
our limited number of patients, investigations performed after CUP
diagnosis did not delay treatment.
The last point concerns clinical practice for CUP treatment. We
observed no difference in rates of compliance with SOR guidelines
for CUP treatment. Most of the patients who did not belong to a
favorable clinicopathologic entity with good PS received
chemotherapy, both in the experimental and the control groups. Most
patients in the control group received taxane/carboplatin regimens,
which are the standard of care for CUP patients in North America,
without significant changes between the two periods. Contrariwise,
we observed a significant change in the experimental group: most
patients treated in the pre-period received organ-specific regimens
while patients treated in the post-period generally received taxane
or gemcitabine-based regimens. This change could be the consequence
of the results published by the French Study Group on CUP and by
the Minnie Pearl Cancer Research Network in which such regimens
were used alone [19]. Another possible explanation is that
gemcitabine has shown significant activity in lung [20] and
pancreatic cancers [21], two frequent sources of CUP in necropsy
series [22]. Therefore, these data and some CUP clinical
presentation, suggesting lung or pancreatic origins, may have
prompted oncologists to favour gemcitabine-based regimens.
More recently, Varadhachary et al. showed that CUP patients with
a colon-cancer molecular profile, identified by
immmunohistochemestry or gene-expression studies, had better
responses to cancer-specific therapies with flurouracil,
oxaliplatin, irinotecan and targeted therapies with cetuximab and
bevacizumab, than they did to empiric therapy with taxane/platinum
regimens [23, 24]. These data should significantly change the
clinical practice for CUP treatment in a near feature.
Our study agrees with previous studies, which have shown that
simply distributing CPGs is insufficient to change physicians'
practice [25, 26]. These results are completely in contrast to our
previous similar research in breast and colon cancer [8, 9].
Successful introduction of CPGs leading to significant improvements
in clinical care depends on many factors including clinical
settings [27], methods of development, dissemination, and
implementation of the guidelines [28]. In our opinion, such a
difference is mainly due to the lack of implementation and regional
development of CPGs for the management of CUP patients. As has been
the case for other cancers, we think that CUP CPGs should be
included in a continuing medical education program (specific
meetings) using computer-based clinical decision support systems to
be able to change medical practice. Moreover, CPGs should be
disseminated and implemented with both oncologists and
organ-specialists who are often involved in the management of CUP
patients, as well as with pathologists. In the same way, a
centralized organization for CUP management could allow a better
conformity with SOR guidelines.
We have noted several limitations to our analysis. First,
although we performed a comprehensive evaluation of practice
patterns in two large institutions, it remains possible that other
institutions, both in France and in Canada, might have evolved
differently over time. Secondly, the retrospective design of this
study did produce a significant proportion of missing data.
Finally, we observed a staff turnover in both experimental and
control groups. Other important factors such as variations in
physician practice patterns and patient preferences, which may
constitute barriers to investigations on diagnosis and therapy,
have not been assessed.
Despite these limitations, our study led to several conclusions.
The most important one was that simply distributing CPGs did not
modify physicians' practice for CUP management. Even after CPG
publication, only 50% of medical decisions regarding diagnostic
workup were in accordance with the CPGs in the experimental group.
Approximately, 75% of our patients had inappropriately
investigations. We observed a significant change regarding
chemotherapy regimens in the experimental group, possibly due to
the publication of recent clinical research in CUP management. Our
results underline the necessity, if one wants to change medical
practice, to disseminate, implement and reinforce optimal
management recommendations to the attention of oncologists, as well
as organ specialists and pathologists, by means of continuing
medical education programs (specific meetings) using computer-based
clinical decision support systems.
Funding considerations and conflicts of interest: this work was
supported by a grant from the “hospices civils de Lyon”.
Acknowledgements
We thank Mary Burns and Estelle Papot for administrative
assistance, and Murielle Rabilloud for statistical analysis.
References
1 Pavlidis N, Briasoulis E, Hainsworth J,
Greco FA. Diagnostic and therapeutic management of cancer of
an unknown primary. Eur J Cancer 2003; 39: 1990-2005.
2 Seve P, Stankovic K, Charhon A,
Broussolle C. Carcinoma of unknown primary site. Rev Med
Interne 2006; 27: 532-45.
3 Farag SS, Green MD, Morstyn G,
Sheridan WP, Fox RM. Delay by internists in obtaining
diagnostic biopsies in patients with suspected cancer. Ann Intern
Med 1992; 116: 473-8.
4 Schapira DV, Jarrett AR. The need to consider
survival, outcome, and expense when evaluating and treating
patients with unknown primary carcinoma. Arch Intern Med 1995; 155:
2050-4.
5 National Comprehensive Cancer Network. NCCN practice
guidelines for occult primary tumors. Oncology 1998; 12: 226-309;
(Williston Park).
6 Seve P, Sawyer M, Hanson J, Broussolle C,
Dumontet C, Mackey JR. The influence of comorbidities,
age, and performance status on the prognosis and treatment of
patients with metastatic carcinomas of unknown primary site: a
population-based study. Cancer 2006; 106: 2058-66.
7 Bugat R, Bataillard A, Lesimple T,
Voigt JJ, Culine S, Lortholary A, et al.
Standards, Options and Recommendations for the management of
patient with carcinoma of unknown primary site. Bull Cancer 2002;
89: 869-75.
8 Ray-Coquard I, Philip T, Lehmann M,
Fervers B, Farsi F, Chauvin F. Impact of a clinical
guidelines program for breast and colon cancer in a French Cancer
Center. JAMA 1997; 278: 1591-5.
9 Ray-Coquard I, Philip T, de Laroche G,
Froger X, Suchaud JP, Voloch A, et al. A
controlled “before-after” study: impact of a clinical guidelines
programme and regional cancer network organization on medical
practice. Br J Cancer 2002; 86: 313-21.
10 Ray-Coquard I, Thiesse P, Ranchere-Vince D,
Chauvin F, Bobin JY, Sunyach MP, et al.
Conformity to clinical practice guidelines, multidisciplinary
management and outcome of treatment for soft tissue sarcomas. Ann
Oncol 2004; 15: 307-15.
11 Bugat R, Bataillard A, Lesimple T,
Voigt JJ, Culine S, Lortholary A, et al.
Summary of the Standards, Options and Recommendations for the
management of patients with carcinoma of unknown primary site
(2002). Br J Cancer 2003; 89 (Suppl 1): S59-S66.
12 Donabedian A. Criteria, norms and standards of quality:
what do they mean? Am J Public Health 1981; 71: 409-12.
13 Piccirillo JF, Tierney RM, Costas I,
Grove L, Spitznagel Jr. EL. Prognostic importance of
comorbidity in a hospital-based cancer registry. JAMA 2004; 291:
2441-7.
14 Seve P, Ray-Coquard I, Trillet-Lenoir V,
Sawyer M, Hanson J, Broussolle C, et al. Low
serum albumin levels and liver metastasis are powerful prognostic
markers for survival in patients with carcinomas of unknown primary
site. Cancer 2006; 107: 2698-705.
15 Abbruzzese JL, Abbruzzese MC, Hess KR,
Raber MN, Lenzi R, Frost P. Unknown primary
carcinoma: natural history and prognostic factors in 657
consecutive patients. J Clin Oncol 1994; 12: 1272-80.
16 Holzer S, Fremgen AM, Hundahl SA,
Dudeck J. Analysis of medical decision-making and the use of
standards of care in oncology. Proc AMIA Symp 2000: 364-8;
(volume?).
17 Pavlidis N, Fizazi K. Cancer of unknown primary
(CUP). Crit Rev Oncol Hematol 2005; 54: 243-50.
18 Seve P, Billotey C, Broussolle C,
Dumontet C, Mackey JR. The role of
2-deoxy-2-[F-18]fluoro-D-glucose positron emission tomography in
disseminated carcinoma of unknown primary site. Cancer 2006; 109:
292-9.
19 Greco FA, Culine S. Chemotherapy for patients with
metastatic carcinoma of unknown primary site. In: Fizazi K,
ed. Carcinoma of an Unknown Primary Site. New York, London: Taylor
and Francis, 2006: 93-103.
20 Johnson DH. Gemcitabine for the treatment of non-small
cell lung cancer. Oncology 2001; 15: 33-9; (Williston Park).
21 Crino L, Mosconi AM, Calandri C,
Corgna E, Porrozzi S, Chiara S, et al.
Gemcitabine in advanced pancreatic cancer: a phase II trial. Am J
Clin Oncol 2001; 24: 296-8.
22 Le Chevalier T, Cvitkovic E, Caille P,
Harvey J, Contesso G, Spielmann M, et al. A
clinical study of 302 consecutive autopsied patients. Arch Intern
Med 1988; 148: 2035-9.
23 Varadhachary GR, Raber MN, Matamoros A,
Abbruzzese JL. Carcinoma of unknown primary with a
colon-cancer profile-changing paradigm and emerging definitions.
Lancet Oncol 2008; 9: 596-9.
24 Varadhachary GR, Talantov D, Raber MN,
Meng C, Hess KR, Jatkoe T, et al. Molecular
profiling of carcinoma of unknown primary and correlation with
clinical evaluation. J Clin Oncol 2008; 26: 4442-8.
25 Davis DA, Thomson MA, Oxman AD,
Haynes RB. Evidence for the effectiveness of CME.
A review of 50 randomized controlled trials. JAMA 1992; 268:
1111-7.
26 Ferguson JH. The NIH consensus development program. Jt
Comm J Qual Improv 1995; 21: 332-6.
27 Grilli R, Lomas J. Evaluating the message: the
relationship between compliance rate and the subject of a practice
guideline. Med Care 1994; 32: 202-3.
28 Eddy DM. Guidelines for policy statements: the explicit
approach. JAMA 1990; 263: 2239-43.
|
Printable version
Version PDF