close

Вход

Забыли?

вход по аккаунту

?

580

код для вставкиСкачать
Int. J. Cancer: 66,191-196 (1996)
0 1996 Wiley-Liss, Inc.
Publication of the International Union Against Cancer
Publication de I Union lnternationaleContre le Cancer
CORRELATION BETWEEN KARYOTYPIC PATTERN AND
CLINICOPATHOLOGIC FEATURES IN 125 BREAST CANCER CASES
Nikos PANDIS',S,R,
Ingrid IDVALL~,
Georgia BARDI',', Yuesheng JIN', Ludmila GORUNOVA',
Fredrik MERTENS',
HHkan oLSSON3, Christian INGVAR4,Konstantine BEROUKAS~,
Felix MITELMAN' and Sverre HEIM1x7
Departments of 'Clinical Genetics, 2Pathology, 30nco10gy and 4Surgeiy, University Hospital, Lund, Sweden;
Department of SGenetics,Saint Savas Hospital and hOncology-RadiotherapyCenter, Institution of Social Insurance,
Athens, Greece; and 'Department of Genetics, The Norwegian Radium Hospital and Institutefor Cancer Research, Oslo, Norway.
A correlation analysis was performed on I25 cytogenetically
characterized breast cancer cases to assess the relationship
between the tumor karyotype and clinicopathologic features.
The carcinomas of young women had a higher modal chromosome number than those of older women. The number of
chromosomal aberrations and modal chromosome number
were also found to correlate with the histologic type, grade and
mitotic activity of the tumor. Whereas all lobular carcinomas
were karyotypicallynormal or near-diploid, more than 3 aberrations and sometimes near-triploid or near-tetraploid karyotypes were common findings in ductal carcinomas, especially in
grade-Ill tumors and in tumors showing high mitotic activity in
vivo. Karyotypes with cytogenetically unrelated clones and
unbalanced structural chromosomal rearrangementswere more
frequent in infiltratingthan in in situ carcinomas but, at least as
far as the second of these 2 characteristics is concerned,
especially in infiltrating carcinomas that also had an in situ
component. The presence of cytogenetic polyclonality correlated with tumor grade. Although recurrent chromosomeaberrations were significantly more common in ductal than in
lobular carcinomas, none of these breast cancer-associated
anomalies seemed to be specific for any particular clinicopathologic parameter. The associations between modal chromosome
number and mitotic activity and betweencytogenetic polyclonality and tumor grade were found to be statistically significant in
multivariate models. No correlation was seen between the
karyotypic findings and tumor size or the presence of axillarylymph-node metastases.
o 1996 Wiley-Liss,Inc.
Although the histopathologic picture remains and is likely to
remain the main criterion for the diagnosis and classification of
breast cancer, recent approaches in tumor biology and genetics
have offered a host of new parameters providing insight into
the tumorigenic process and the intrinsic aggressiveness of
individual neoplasms. Particularly important are those reflecting the extent and type of genomic alterations acquired by the
tumor cells, i.e., the changes that according to the somatic
mutation theory are the effective causes in carcinogenesis.
These aberrations can be assessed at 3 main levels of resolution: as changes in total DNA content, using flow-cytometric
techniques, as chromosomal abnormalities, using cytogenetic
techniques, and at the gene or primary-DNA-structure level,
using molecular genetic techniques. Each methodology has its
own advantages and disadvantages. Particular to the cytogenetic approach is its cytological nature; since individual cells
are studied, karyotypic variation among them can be detected,
enabling assessment of intratumor heterogeneity. Chromosome banding analysis is, furthermore, a screening method
which, unlike techniques dependent on molecular-level probing, is not dependent on a priori knowledge as to what might
constitute the relevant genomic area(s) to examine. Finally,
unlike what is the case with flow-cytometric investigations,
balanced and near-balanced genomic alterations are equally
visible in the karyotype as those leading to substantial gains or
losses of genetic material.
Large series of breast carcinomas have been studied using
molecular genetic or flow-cytometric techniques and clinicopathologic tumor features have been correlated with DNA
ploidy (eg., Cornelisse and Tanke, 1991; Gnant et al., 1993)
and specific oncogene or tumor suppressor gene aberration
patterns ( e g , Bikche and Lidereau, 1995). In contrast, only
some 400 breast carcinomas with abnormal karyotypes have
been described (Mitelman, 1994), and fewer still have been
reported in studies attempting to correlate cytogenetic and
more traditional disease parameters (Hainsworth et al., 1992;
Zafrani et al., 1992). We present a correlation analysis between
karyotypic and clinicopathologic features based on 125 breastcancer cases. This is the most extensive series reported so far,
both with regard to the number of tumors included and to the
karyotypic and histopathologic parameters compared.
MATERIAL AND METHODS
Tumors
Samples from 125 breast carcinomas were processed for
histopathologic and cytogenetic investigation. They were all
but one from primary tumors and unselected in the sense that
they were consecutive in our laboratory. The tumors were
histopathologically classified in accordance with WHO recommendations (Scarff and Torloni, 1968) without prior knowledge of their karyotypic characteristics. There were 11 ductal
carcinoma in situ (CIS) and 114 infiltrating carcinomas. The
latter included 88 ductal carcinomas not otherwise specified
(NOS), 15 lobular carcinomas, 7 mucinous carcinomas, 2
tubular carcinomas, 1 adenoid cystic carcinoma, and 1 papillary carcinoma. The clinical and pathologic data are summarized in Table I.
Cytogenetic analysis
The tumor cells were short-term cultured and analyzed
cytogenetically after G-banding of the chromosomes. The
detailed karyotypes have been reported elsewhere (Pandis et
al., 1992, 1993a,b, 1995u,b).
Correlation analysis
To enable correlation analyses between the cytogenetic and
clinicopathologic parameters, the series was broken down first
according to the cytogenetic features of the cases and then
according to their clinicopathologic characteristics. The former
included modal chromosome number, the number and type of
chromosomal aberrations, and the overall clonal composition
of the tumor. The latter included patient age, lymph node
metastasis status, and the tumor's size, histopathologic type
and differentiation, mitotic activity and infiltration pattern.
Modal chromosome number. Tumors were classified as nearhaploid, near-diploid, near-triploid or near-tetraploid. Since
only one tumor had a near-haploid karyotype (case 42), this
case and group had to be excluded from the correlation
analysis.
T o whom correspondence and reprint request should be addressed,
at: Department of Clinical Genetics, University Hospital, S-221 85
Lund, Sweden. Fax: (+46) 4613-1061.
Received: October 18,1995 and in revised form December 14,1995.
PANDIS ETAL.
192
TABLE I - CLINICAL AND PATHOLOGIC FINDINGS
Case
number'
1
2
3
4
1
2
3
5
6
7
8
9
10
11
12
13
14
16
17
19
20
1
2
3
4
5
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
Age
(years)
71
62
69
70
81
48
58
72
48
73
43
65
78
62
58
72
62
46
62
48
62
45
54
42
51
62
94
66
44
62
56
78
75
76
55
65
31
63
53
45
61
62
72
39
87
37
49
90
40
57
56
81
63
61
42
77
41
63
45
61
62
63
55
53
Size
(mm)*
20
17
25
28
10
40
40
9
12
20
150
18
40
12
15
20
30
12
25
15
30
35
26
13
16
20
30
4
120
30
20
15
40
14
13
14
50
20
18
18
20
20
15
10
30
16
19
30
35
10
32
30
13
35
50
40
12
15
20
5
25
20
19
20
50
15
69
58
41
38
35
41
33
11
25
18
14
65
35
11
10
50
Type and
grade3
M, I
D, 111
D, I11
D, 111
D. I
D; 111
D, 111
D, 111
D, I1
D, I1
DCIS
D, I1
D, I1
M, I
D, I1
D, 111
D, I1
D, 111
D, 111
D, I1
D. 111
D: 111 comedo
D. 111 comedo
D; I1
M, I
L, 1
M, I1
DCIS
DCIS comedo
D, I1
D, 111
M, I
DCIS comedo
L, I1
D, I1
D, I1
D, I11
D, I1
D, I1
AC, I
D, 111
DCIS comedo
D, I1
DCIS comedo
D, 11
D, 111
D, I1
M, 1
D, I11 comedo
L, I
D, I1
D, I1
D, I1
T, I
D, 111
D, I1
D, I
L, 1
D, I
D, 111 comedo
DCIS
L, 1
DCIS
D, I1
D, III
L, I1
DCIS comedo
D, I11 comedo
L, I
D, 111 comedo
D, I1
DCIS comedo
T. I
017
0120
6/13
0111
015
6113
0111
10116
015
217
0110
018
4/12
01 10
n.d.
12115
Oil2
0110
018
01 10
014
17/17
n.d.
1
5
3
1
1
3
2
1
1
1
1
1
1
1
1
1
1
2
6
2
8
4
5
017
2
0124
16/16
015
017
013
016
5
3
0110
2110
n.d.
n.d.
018
013
219
3110
114
01 12
10/10
012
213
0116
n.d.
216
018
n.d.
1
multifocal
5
1
2
3
4
2
3
1
2
1
2
3
1
2
2
3
2
1
017
01 17
013
4
n.d.
1
017
519
2110
017
015
0130
0113
2
1
4
4
1
1
1
2
n.d.
1
018
multifocal
1
4
3
1
1
015
015
016
3/12
0113
2
013
3
518
0118
10121
115
0112
018
8
rnultifocal
multifocal
1
7
3
3
1
(continued)
CYTOGENETIC-CLINICOPATHOLOGIC CORRELATIONS IN BREAST CANCER
193
TABLE I - CLINICAL AND PATHOLOGIC FINDINGS (CONTINUED)
Case
number’
Age
(years)
Size
Type and
grade’
Lymph node
metastases4
48
49
50
51
52
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
1
2
3
4
5
6
7
8
9
10
61
49
66
75
80
78
76
84
51
35
79
61
49
82
65
46
46
43
46
42
74
68
52
38
67
90
68
49
84
52
73
63
62
47
38
46
46
76
87
60
50
33
62
75
86
80
53
49
54
55
55
55
15
11
10
38
28
50
23
30
60
60
20
22
20
42
18
25
35
15
20
15
13
21
18
70
16
30
15
25
25
5
35
22
23
21
30
25
25
21
25
15
12
60
38
130
20
70
25
19
35
10
100
20
D. I11 comedo
D; I1
P, I
D, I11 comedo
D, I11
D, I11 comedo
D, I11
M, I1
D. I11 comedo
D: I11 comedo
D. I1
016
0110
0110
3/17
1/15
n.d.
119
n.d.
117
618
n.d.
0110
0111
115
0111
215
013
0121
016
016
0
118
5/15
12115
618
n.d.
013
41 19
n.d.
018
016
017
115
014
0115
1/17
118
017
n.d.
0112
216
n.d.
119
313
0115
616
2116
018
8113
n.d.
313
10/10
11
12
13
14
15
16
17
1
1
2
2
~
L,’I
L. I1
D, I1
D, I11 comedo
D. I11
L,’I
D. I1
L,’I
D, I1
D. I11 comedo
Di I
D. I1
D; I11 comedo
D, I11
D, I11 comedo
D, I
L. I
D, I
D, I1
L, 1
D, I1
D, I1
D, 1
D, I11
D. I1
D; I1
D. I11 comedo
D; I11 comedo
D, I1
D, I11 comedo
L, 1
L, I1
D, I11 comedo
D, I11 comedo
D. I11 comedo
D; I11 comedo
D, I11
D, 111
cIss
Mitoses
HPF6
Remarks
6
2
1
5
4
8
5
1
4
6
2
1
1
3
1
2
3
4
5
1
2
1
2
6
1
2
bilateral
multifocal
multifocal
multifocal
5
-
3
3
1
1
2
2
1
1
1
1
3
2
1
3
5
2
3
3
1
7
4
2
2
3
3
multifocal
multifocal
multifocal
bilateral7
bilateral7
‘Case numbers correspond to the original numbers used in the articles where the karyotypes were
published; see reference list. The first 4 cases (numbers 1 4 ) were described by Pandis et a/. (1992),
the next 20 (numbers 1-20) by Pandis et a/. (1993a), the next 5 (numbers 1-5) by Pandis et al.
(19936 ;the next 79 (numbers 1-79) are the karyotypically abnormal cases described by Pandis et nl.
199% , the next 17 (numbers 1-17) the karyotypically normal cases described by Pandis et al.
199% , and the last 4 tumors (numbers 1 and 2) correspond to bilateral breast-cancer cases
described by Pandis et a/. (1995b).-2Largest diameter.-3WH0 classification; A, adenoid cystic; D,
ductal; DCIS, ductal carcinoma in situ; L, lobular; M, mucinous; P, papillary; T, t~bular.-~n.d.,not
determined.-5Carcinoma in situ within the surgical sample.-6HPF = 12.5 x 40 high-power field
magnifi~ation.-~One
tumor analyzed from each breast.
i I
Number of aberrations. Tumors were classified according to 2
partly separate schemes. In one, they were grouped as having
no aberrations (ie., a normal karyotype), 1 aberration, 2 to 3
aberrations, or more than 3 aberrations. In the second, they
were dichotomized into those having 3 or fewer than 3
aberrations, and those having more than 3 aberrations.
Type of aberrations. Cytogenetically abnormal tumors were
classified according to whether they had numerical changes
only or at least one structural chromosomal rearrangement, to
whether their structural abnormalities were balanced or unbalanced, and to whether or not certain abnormalities recurrently
associated with breast cancer (Pandis et al., 1995a), ie.,
der(l;l6), i(lq), del(lq), del(3p), del(6q), +7, +18, and +20,
were present.
Clonal composition. Tumors were classified according to
whether they contained cytogenetically unrelated abnormal
clones (multiclonal tumors) or only one or cytogenetically
related clones (monoclonal tumors).
194
PANDIS ETAL.
Age. Cases were classified according to whether the patient
was older than 40 years, or 40 years or younger.
Metastasis status. Two groups were recognized, patients with
and patients without axillary lymph node metastases.
Tumor size. Tumors were grouped as small (less than 10 mm
in diameter), medium sized (10 to 20 mm) or large (more than
20 mm in diameter).
Histologic type. The numbers were sufficient for statistical
comparisons of 2 histologic groups of infiltrating carcinomas,
namely, lobular carcinomas (n = 15) and ductal carcinomas
(n = 99).
DifSerentiation. Depending on whether the neoplastic breast
tissue was highly, moderately or poorly differentiated, tumors
were classified as belonging to grades I, I1 or 111 (Scarff and
Torloni, 1968). Ductal grade I11 carcinomas were further
sub-divided into comedo and non-comedo carcinomas (Line11
et al., 1980).
Mitotic activity. The number of mitoses per high-power field
(HPF; 12.5 X 40) in vivo was evaluated by counting the mitoses
in each of 10 HPF and registering the number found in the
most mitosis-rich field as representative of the tissue. This
enabled classification of the tumors into 3 groups: those with
none or only one (0-1) mitosis, those with few (2-3) mitoses,
and those with many mitoses (more than 3) per HPF.
Infiltration pattern. Two main groups were recognized: infiltrating and in situ carcinomas. The former were further
sub-divided depending on whether or not CIS lesions were also
present within the carcinoma or in its vicinity.
Spearman correlation coefficients were used for painvise
comparisons between the karyotypic and clinicopathologic
features. Multivariate models were analyzed using linear and
logistic regression;p values of less than 0.05 were accepted as
statistically significant.
RESULTS
The distribution of the clinicopathologic parameters in the
various cytogenetic sub-groups is given in Table I1 and the
statistically significant values reached in the various correlation analyses in Table 111.
The modal chromosome number was significantly correlated
with the tumors’ grade and mitotic activity and with patient
age. The comparison with histologic type gave a borderline
value ( p = 0.05). Near-triploid karyotypes were found only in
TABLE I1 - KARYOTYPIC-CLINICOPATHOLOGIC COMPARISONS
Number of chromosome
aberrations
Modal number1
Ag: 40
< 40
Size
< 10
10-20
> 20
Grade
I
I1
111
Type
Ductal
Lobular
Mitoses
0-1 / HPF
2-31 HPF
> 3IHPF
Growth
DCIS
Infiltration
Both
Lymph node metastases
Piesent
Absent
Type of chromosome
aberrationsZ
Clonality’
n-dip
n-trip
n-tetr
0
1
2-3
>3
num
struct
bal
unbal
rec
non-rec
mono
poly
95
6
12
5
6
0
13
8
40
16
10
3
21
14
12
2
59
31
20
5
39
26
38
20
33
13
41
24
30
9
10
45
46
0
8
9
0
1
5
1
4
16
6
29
21
1
7
5
2
15
18
2
5
7
7
46
37
2
15
8
5
31
29
5
27
26
4
24
18
6
31
28
3
20
16
26
36
31
2
1
12
0
3
2
5
7
9
15
24
11
3
2
7
5
3
6
4
20
28
32
6
9
4
14
19
28
13
I8
24
10
16
12
16
23
16
7
11
20
18
3
31
1
7
2
63
1
28
9
39
8
31
4
8
4
2
34
32
24
15
7
3
19
25
21
22
15
21
20
21
5
29
23
13
13
1 1 0
13
80
1
41
6
19
7
4
61
34
3
55
26
8
38
16
10
55
20
1
38
21
7
16
22
35
20
29
14
26
22
36
12
19
67
1
15
5
5
0
0
8
18
9
1
30
1
0
46
38
17
3
3
11
0
5
1
7
11
3
30
15
11
4
6
3
8
15
12
8
93
41
2
15
7
1
5
4
0
21
11
6
50
21
1
12
6
4
31
15
35
52
7
9
1
3
10
9
16
32
5
6
13
17
5
4
29
51
13
50
42
0
5
5
3
13
13
‘n-dip, near-diploid; n-trip, near-triploid; n-tetr, near-tetraploid.-2num, numerical only; struct, structural; bal, balanced; unbal,
unbalanced; rec, recurrent; non-rec, non-rec~rrent.-~mono,cytogenetically monoclonal; poly, cytogenetically polyclonal.
TABLE 111- SPEARMAN CORRELATION COEFFICIENTS’
P
Grade-Modal
number
0.2428 0.009 Type-Modal
number
rs
P
0.1842 0.050 Type-Recurrent
aberrations
r,
P
r3
P
0.2379 0.011 Growth-Nu0.3266 0.001
merical aberrations
0.3283 0.000 Growth-Balanced 0.2543 0.016
aberrations
0.1 920 0.032 Growth-Clonality 0.2018 0.040
Grade-Number 0.1594 0.090 Type-Number of 0.2061 0.037 Mitoses-Modal
of aberrationsZ
aberrations2
number
Grade-Number 0.2080 0.026 Type-Number of 0.2379 0.011 Mitoses-Number
of aberrations3
aberrations3
of aberrations*
0.1861 0.038
Grade-Clonality 0.2204 0.034 Type-Balanced 0.2331 0.037 Mitoses-Balanced 0.2907 0.005 Age-Modal
aberrations
aberrations
number
‘See “Material and Methods” and “Results” for further information about the correlation analyses.-2Four groups of tumors
recognized: 0,1,2-3, and > 3 aberrati~ns.-~Two
groups of tumors recognized: I 3 and > 3 aberrations.
CYTOGENETIC-CLINICOPATHOLOGIC CORRELATIONS IN BREAST CANCER
ductal carcinomas and more often in grade I11 carcinomas and
in tumors with high mitotic activity than in those that were
grade I or grade I1 and had low mitotic activity. All lobular
carcinomas were diploid or near-diploid. The tumors of
women I40 years of age generally had higher modal chromosome numbers than those in older women; whereas 45% of
women in the former group had near-triploid carcinomas, the
corresponding number in the latter group was only 10%.
The number of chromosome aberrations was significantly
correlated with histologic tumor type and mitotic activity.
Grade likewise showed a significant correlation with chromosome number when tumors were grouped as having > 3 and
< 3 changes, whereas only a tendency (p = 0.09) was evident
with the more detailed sub-division (see above). Whereas
lobular carcinomas and ductal carcinomas of grades I and I1
typically had simple aberrations or normal karyotypes, poorly
differentiated ductal carcinomas (grade 111) often had more
than 3 chromosomal abnormalities. Tumors with many (> 3)
aberrations had more mitoses per HPF than did the karyotypically more simple tumors.
Infiltrating carcinomas which also had a CIS component
carried structural chromosomal rearrangements significantly
more often than did infiltrating carcinomas with no detectable
in situ lesions.
The presence of balanced vs. unbalanced structural chromosomal rearrangements correlated in a statistically significant
manner with the mitotic activity, histologic type and infiltration
pattern of the tumors. Unbalanced structural changes were
more often present in ductal than in lobular carcinomas, in
carcinomas with high mitotic activity than in those with few
mitoses, and in infiltrating than in in situ carcinomas.
Cytogenetically unrelated clones were more often found in
infiltrating than in in situ carcinomas. This type of polyclonality
was more common in grade I1 and grade I11 tumors than in
carcinomas of grade I.
The 8 recurrent chromosome anomalies assessed were
significantly more frequently detected in ductal than in lobular
carcinomas, although no particular aberration was found to be
restricted to any of the clinicopathologic sub-groups. No
statistically significant correlations were observed between the
karyotype and the size of the tumor or the presence or absence
of lymph node metastases.
Finally, in the multivariate models comparing modal chromosome number with tumor grade, type and mitotic activity, only
the correlation between chromosome number and mitotic
activity turned out to be statistically significant ( p = 0.039). In
the similar comparison of tumor size, type and grade with
cytogenetic polyclonality, the latter was shown to correlate
significantly only with grade (alsop = 0.039).
DISCUSSION
The cytogenetic results obtained when analyzing any given
tumor may not necessarily reflect all important alterations that
in principle should be detectable at the chromosomal level; the
choices made with regard to how samples are cultured, when
and how they are harvested, and how the analysis is performed
may significantly affect the final outcome. When cytogeneticpathologic correlations are the issue, furthermore, additional
uncertainties arise from the fact that it is never possible to
determine both the karyotype and the phenotype of a tumor on
exactly the same piece of tissue. Neighboring samples of tumor
material may differ cytogenetically (Teixeira et al., 1995), and
the same goes for their histologic characteristics. Some breast
carcinomas, for example, may consist of 80% ductal NOS, 10%
mucinous, and 10% lobular areas (Sakamoto, 1987), and it is
therefore obvious that how they are classified depends to some
extent on stochastic sampling factors. Other potential biases
may be related to the sometimes practical rather than strictly
195
logical classification criteria used, eg., tumors with 2 to 3
chromosome aberrations constituted one sub-set and those
with more than 3 another, or to the fact that some sub-groups
were too small to compare. In spite of all these difficulties,
statistically significant associations could nevertheless be found
for most of the karyotypic-clinicopathologicrelationships tested.
The correlation between patient age and modal chromosome number in our series may well have a parallel in, and
indeed be causally contributing to, the worse prognosis for
breast cancer in young compared with older women. Both Host
and Lund (1986) and Adami et al. (1986) showed, in large
series of Scandinavian breast cancer patients, that the youngest women ( < 3 0 and <34 years of age, respectively) had
shorter relative survival than did middle-aged women.
Of all the clinicopathologic parameters examined in this
study, the histologic tumor type came across as the one
showing the best overall correlation with karyotypic features.
Statistical significance was reached in the comparisons with
modal chromosome number, the number of chromosome
aberrations detected, the presence of balanced vs. unbalanced
chromosomal changes, and the presence of recurrent vs.
non-recurrent anomalies. All lobular carcinomas were karyotypically normal or had 1 to 3, often balanced, aberrations
yielding near-diploid karyotypes. The ductal carcinomas often
had more than 3 aberrations, the karyotypes were typically
unbalanced, and the chromosome number could be neartriploid to near-tetraploid. Although ours appears to be the
first statistical analysis along these lines, a perusal of cytogenetically characterized breast cancer cases (Mitelman, 1994)
strengthens the impression that lobular carcinomas usually
have no or only few chromosomal changes resulting in neardiploid karyotypes. This observation fits well with the fact that
lobular carcinomas typically are fairly homogeneous, smallcell, low-grade tumors that, in spite of their often considerable
size, are not as biologically aggressive as their ductal counterparts (Silverstein et al., 1994).
The finding of statistically significant relationships between
tumor grade and mitotic activity and the number of chromosome aberrations as well as modal chromosome number was
not unexpected, since grading is intimately dependent on the
tumor cells’ nuclear characteristics (Sakamoto, 1987). Similar
relationships have been found in ovarian, colorectal, and
pancreatic carcinomas (Heim and Mitelman, 1995). These
cytogenetic conclusions tally with the results of DNA flow
cytometric studies, which show a strong correlation between
total DNA content and tumor grade (Cornelisse and Tanke,
1991; Gnant et al., 1993). The fact that the karyotypically more
complex tumors had many mitoses in vivo is in agreement with
the observation that breast cancers analyzed in direct preparations generally show more extensive changes than those
analyzed after short-term in vitro culturing; evidently, to be
successful, the former procedure requires that the tumors have
high mitotic activity (Pandis et al., 1995~).
The tumor infiltration pattern correlated with karyotypic
characteristics, but in a manner that was not entirely expected.
The fact that infiltrating tumors more often carried unbalanced structural chromosomal rearrangements than did in situ
carcinomas was not surprising, since loss of heterozygosity
studies have shown that allelic imbalances parallel tumor
aggressiveness (Sato et al., 1991). Presumably, the imbalances
are pathogenetically important, e.g., through loss of tumor
suppressor genes. When the group of infiltrating carcinomas
was further divided into tumors showing or not showing a CIS
component, however, it was in the former sub-set that these
structural anomalies were most common. It is not clear how
this finding fits into a scenario in which ever-more complex
genetic changes are the main factor causing breast carcinomas
to follow a linear, unidirectional path from in situ to infiltrating
lesions. Perhaps selection biases in the material explain the
196
PANDISETAL.
apparent discrepancy, or perhaps some CIS develop from,
not into, infiltrating carcinomas, as our cytogenetic data on
bilateral breast cancers appear to indicate (Pandis et a[.,
19956).
The presence of cytogenetically unrelated clones was shown
to correlate positively both with the ability to infiltrate and
with tumor grade; the latter correlation was statistically significant even in the multivariate analysis. The relationship, whether
causal or coincidental, between tumor clonality and infiltrative
behavior is far from simple, as is evidenced by the fact that
benign mammary neoplasms also harbor cytogenetically unrelated clones (Dietrich et al., 1995). Polyclonal neoplasms may
be predisposed to transform malignantly, although we have no
facts suggesting how such a propensity might be acquired.
Whatever mechanisms are involved, it appears to us that
recruitment hypotheses, stipulating that the parenchymatous
cells of carcinomas produce not only collagenase and other
proteases facilitating invasion but also mutagens that transform cells of the neighboring epithelium, fail to account for the
detection of multiclonality in cases of unquestionably benign
breast proliferation. The biological importance of cytogenetic
polyclonality in breast tumors thus remains unclear.
ACKNOWLEDGEMENTS
This work was supported by grants from the Swedish and
Norwegian Cancer Societies and the Medical Faculty of Lund
University. L.G. is on leave from the Institute of Cytology of
the Russian Academy of Sciences, St. Petersburg, Russia.
REFERENCES
ADAMI,H.O., MALKER,
B., MEIRIK,O., PERSSON,
I., BERGKVIST,
L. and netic multiclonality and karyotypic-pathologic correlations. Genes
STONE,G., Age as a prognostic factor in breast cancer. New Engl. J. Chromosom. Cancer, 6,51-57 (1993a).
Med., 315,559-563 (1986).
L., PETERSON,C., BARDI,G., IDVALL,
PANDIS,N., JIN,Y., GORUNOVA,
B., INGVAR,
C., MANDAHL,
N., MITELMAN,
F.and HEIM,
R., Genetic alterations in breast cancer. I., JOHANSSON,
BIECHE,I. and LIDEREAU,
S., Chromosome analysis of 97 primary breast carcinomas: identificaGenes Chromosom. Cancer, 14,227-251 (1995).
12,
CORNELISSE,
C,J, and TAN=, H.J,, Flow cytometry, In; M. Bid& (ed,), tion Of eight karyotypic sub-groups. Genes Chromosorn.
Comprehensive cytopathology, pp. 984-1010, W.B. Saunders, Philadel- 173-185
PANDIS,N., JIN,Y., LIMON,J., BARDI,G., IDVALL,I., MANDAHL,
N.,
phia (1991).
MITELMAN,
and
HEIM,, S., Interstitial deletion of the short arm of
DIETRICH,
c., pANDIS,
N,, T
~M.R,, B~ ~G., ~G~~~~~
~
~~A -M,
~ ~ chromosome
,
~ F.3 as
~a primary
chromosome abnormality in carcinomas of
ANDERSEN,
J. and HEIM, s,, Chromosome abnormalities in'
69151-155 (1993b).
hyperproliferative disorders of epithelial and stromal breast tissue. Int. the breast. Genes Chromosom.
PANDIS,N., TEIXEIRA,
M.R., GERDES,A.-G., LIMON,J., BARDI,G.,
J. Cancer, 60,49-53 (1995).
F. and
J.A., IDVALL>
I., MANDAHL,
N.,
G.H., REINER,A., SCHEMPER,
M., REYN- *DERSEN,
GNANT,M.F.X., BILIJHAM,
S., Chromosome abnormalities in bilateral breast carcinomas. CytogeDERS, M., SCHUTTE,
B,, VAN ASCHE,c.,STEGER,G, and JAKESZ,
R:,
heuploidy fraction but not DNA index is important for the prognosis netic
Of the
Origin Of
primary tumors'
of patients with stage-I and -11 breast cancer-10-year results. Ann. Cancer, 76,250-258 (1995b).
SAKAMOTO,
G., Infiltrating carcinoma: major histological types, In:
Oncol., 4,643-650 (1993).
D.L. Page and T.J. Anderson (eds.), Diagnostic histopathology of the
HAINSWORTH,
P.J., RAPHAEL,
K.L., STILLWELL, R.G., BENNETT,
R.C.
and GARSON,O.M., Rearrangements of chromosome l p in breast breast, pp, 193-235, Churchill Livingstone, Edinburgh (1987).
cancer correlates with poor prognostic features. Brit. J. Cancer, 66, SATO,T., AKIYAMA, F., SAKAMOTO, G., G S U M L F. and NAKAMUW
Y., Accumulation of genetic alterations and progression of primary
131-135 (1992).
HEIM,S. and MITELMAN,
F., Cancer cytogenetics (2nd ed.), Wiley-Liss, breast cancer. Cancer Res., 51,5794-5799 (1991).
SCARFF,R.W. and TORLONI,
H., Histological typing of breast turnours,
New York (1995).
World Health Organization, Geneva (1968).
H ~ s TH.
, and LUND,E., Age as a prognostic factor in breast cancer.
SILVERSTEIN,
M.J., LEWINSKY,
B.S., WAISMAN,
J.R., GIERSON,
E.D.,
Cancer, 57,2217-2221 (1986).
W.J., SENOFSKY,G.M. and CAMAGAMI,
P., Infiltrating
COLBURN,
F., LJUNGBERG,
0. and ANDERSSON,I., Breast carcinoma.Acta
LINELL,
lobular carcinoma: is it different from infiltrating duct carcinoma?
pathol. microbiol. Scand., sect. A, Suppl. 272 (1980).
Cancer, 73,1673-1677 (1994).
MITELMAN,
F., Catalog of chromosome aberrations in cancer (5th ed.), TEIXEIRA,
,
BARDI,G., ~ D E R S EJ.A.,
N , MITELMAN,
M.R., P ~ D I S N.,
Wiley-Liss, New York (1994).
F. and HEIM,S., Cytogenetic analysis of clonal heterogeneity in breast
PANDIS,N., HEIM, S., BARDI,G., IDVALL,I., MANDAHL,
N. and cancer: karyotypic comparisons of multiple intra- and extratumorous
MITELMAN,
F., Whole-arm t(1;16) and i(1q) as sole anomalies identify samples from three patients. Int. J. Cancer, 63,63-68 (1995).
gain of l q as a primary chromosomal abnormality in breast cancer. ZAFRANI,
V. and DUTRILLAUX,
B., GERBAULT-SEUREAU,
M., MOSSERI,
Genes Chromosom. Cancer, 5,235-238 (1992).
B., Cytogenetic study of breast cancer: clinicopathologic significance of
PANDIS, N., HEIM, S., BARDI,G., IDVALL,I., MANDAHL,N. and homogeneously staining regions in 84 patients. Hum. Pathol., 23,
MITELMAN,
F., Chromosome analysis of 20 breast carcinomas: cytoge- 542-547 (1992).
Документ
Категория
Без категории
Просмотров
2
Размер файла
613 Кб
Теги
580
1/--страниц
Пожаловаться на содержимое документа