Key words
mammary carcinoma - mammography screening - ultrasound
Schlüsselwörter
Mammakarzinom - Mammografie-Screening - Ultraschall
Introduction
Germanyʼs comprehensive mammography screening, based on the “European guidelines for
quality assurance in
breast cancer screening and diagnosis”, [1] was agreed by the German
Parliament in 2002 and was introduced step-by-step, beginning in May 2005. The introduction
of
comprehensive screening was completed in 2008.
All women without clinical symptoms aged between 50 and 69 were invited to take part
in digital
mammographies two years apart, whereby a standard second examination was subsequently
carried out. The
complete participation of all eligible women was not achieved for various reasons.
A clinical
examination or additional mammography is not provided by the screening programme.
In the case of abnormalities in the mammography, patients are invited for additional
examinations in the
form of special imaging (enlarged images, spot images focussed on a densification
or additional
projections) or to an ultrasound scan. If clinical suspicion of a malignant tumour
arises from the
routine imaging as well as the additional examinations [2], the screening unit
will initiate a histological analysis through a minimally invasive punch biopsy or
vacuum-assisted
biopsy. Only in very rare cases, e.g. in the case of anatomical peculiarities (very
small breast,
calcifications close to the breast wall or in an axillary position), for which a minimally
invasive
biopsy is technically impossible, is an open biopsy required. The result of this histopathological
examination is checked for its correlation with the diagnosis given by the imaging
in a
multidisciplinary conference of X-ray-screening diagnosticians, pathologists, surgeons
and the partial
inclusion of the treating gynaecologists. In the case of unconformity between abnormalities
suspected to
be malignant in the imaging diagnostic procedures and benign histological findings
that cannot be
explained by the structure of these benign findings, a re-biopsy or early check will
take place [3].
Should an invasive carcinoma, a ductal carcinoma in situ (B5 lesion) or a B3 or B4
lesion (lesion with
unclear malignant potential or suspected presence of a carcinoma) appear histologically,
the patient
will be transferred to a certified breast centre for further treatment.
An assessment is therefore generally carried out as part of the mammography screening,
even when an
additional focus is detected by the imaging or has not yet been clarified histologically.
Recommendations for further treatment and diagnosis are given in the multidisciplinary
conference.
The results and recommendations are documented in certified software (‘Mammasoft’
or ‘MaSc’).
Structured quality assurance takes place in the postoperative conference regarding
serious deviations in
the size and type of tumour preoperatively and postoperatively, as well as through
the detection of
interval carcinomas. The most recently performed operation is also documented in conjunction
with the
initial recommendations, but the precise reasons for a deviation are not considered.
It is therefore unclear as to whether these differences occur as a result of the findings
of the
screening examinations and further assessment not being observable, or whether a change
in the surgical
methods resulted from additional preoperative examinations in the breast centres and
the new knowledge
regarding the extent or additional malignant findings obtained therefrom.
Whether postoperative or histological findings from imaging that lead to a surgical
extension were the
cause is also not considered.
The goal of the analysis was therefore to investigate the influence of further preoperative
examinations
and interventions in the breast centres indicated on the procedures suggested by the
screening units,
regardless of the screening results.
These additional examinations in the breast centres are not provided for in the federal
master agreement
for the training programme.
Material and Method
All 272 patients that came to our breast centre from screening units for further operative
clarification
between May 2007 and December 2009 were analysed retrospectively.
190 patients came from the primary cooperating screening unit. The patients analysed
also participated in
the screening programmes of three further screening units in the same federal state
(53, 21 and 6
patients from each respectively). In addition, each patientʼs data from two further
screening units was
analysed with respect to the initial recommendation of the multidisciplinary conference
and the ultimate
procedure. The data was taken from the screening unitsʼ written documentation in the
respective,
certified documentation software as well as the patient files and findings reports
in our breast
centre.
B categories and recommended procedure
The distribution of the B categories and the recommended type of operation for the
individual
patients was analysed and the type of operation planned was compared with the operation
that was
eventually performed as well as the histological results before and after the operation.
Additional examinations and their effects
The treatment recommendations as well as the number and type of further, non-operative
recommendations entered under the “Other” heading were also extracted from the “Mammasoft”
screening
software.
These further recommendations under the “Other” sub-category in the software are occasionally
documented in more detail in an additional free-text field.
Although the B3 lesions were not included in the actual assessment, they were included
in the
overview ([Table 1]) between the biopsy results and the final result in
order to illustrate the potential number of false positive B4 lesions. This case resulted
from the
uncertainty evident in the description, but which led to B4 classification.
Table 1 B classification biopsy cross-table * B-equivalent classification
postoperatively.
|
B classification
|
B-equivalent classification postoperatively
|
Total
|
|
Benign, but with uncertain biological potential
|
Non-invasive mammary carcinoma
|
Invasive mammary carcinoma
|
|
B3 – benign, but with uncertain biological potential
|
44
|
3
|
1
|
48
|
|
B4 – suspected malignancy
|
1
|
4
|
1
|
9
|
|
B5a – non-invasive mammary carcinoma
|
0
|
35
|
12
|
47
|
|
B5b – invasive mammary carcinoma
|
0
|
1
|
158
|
159
|
|
B5c – Disputably invasive mammary carcinoma
|
0
|
4
|
0
|
4
|
|
No punch biopsy
|
2
|
0
|
3
|
5
|
|
Total
|
47
|
47
|
178
|
272
|
The preoperative and postoperative tumour sizes were determined and, in addition,
the differences
between the tumour sizes determined initially and the final histological size were
calculated and
sorted into three groups of > + 10 mm, − 10 mm to + 10 mm and < − 10 mm.
The results of the lymph node staging were analysed.
In addition, the number of operations carried out per patient was determined.
Results
The median patient age was 63 (50–70) years old.
The results of the primary assessments of the patients detailed are presented in [Table 2].
Table 2 B classification from the screening unit.
|
B classification
|
n
|
%
|
|
B3
|
Benign, but with uncertain biological potential
|
48
|
17.6
|
|
B4
|
Suspected malignancy
|
9
|
3.3
|
|
B5a
|
Non-invasive mammary carcinoma
|
47
|
17.3
|
|
B5b
|
Invasive mammary carcinoma
|
159
|
58.5
|
|
B5c
|
Disputably invasive mammary carcinoma
|
4
|
1.5
|
|
No punch biopsy
|
5
|
1.8
|
|
Total
|
272
|
100.0
|
In total, 219 out of these 272 patients were classified as B4 or B5.
In 204 of these patients, a surgical procedure was directly recommended and further
diagnostic measures
were proposed for the other 15.
B categories and recommended procedure
A B3 lesion was diagnosed in 48 patients and an open biopsy was recommended. Four
(8.3 %) of these
patients then had a malignant tumour in the following histology, meaning that an additional
breast-conserving operation was ultimately necessary.
Of the five cases in which no punch biopsy could be carried out in the assessment
and an open biopsy
was then performed, three cases displayed an invasive carcinoma and so also required
a further
therapeutic operation. The findings for two of these patients were benign.
Lymph node metastasis was displayed in a total of 44 patients (pN1a: 35 = 12.9 %/pN2a:
6 = 2.2 %/pN3a: 3 = 1.1 %).
Through the primary and necessary re-operations as well as repeat operations due to
R1 resections or
two-stage axillary removals, a total of 306 operations were performed ([Table
3]).
Table 3 Number of surgical interventions per patient.
|
Number
|
n
|
%
|
|
1
|
240
|
88.2
|
|
2
|
30
|
11.0
|
|
3
|
2
|
0.7
|
|
Total
|
272
|
100.0
|
In as early as the screening stage, 28 (10.3 %) out of the 272 cases had multi-focal
abnormalities
and 8 (2.9 %) had bilateral abnormalities, whereby only the most biologically severe
side had an
influence on further evaluations for the latter group.
In eight cases (2.9 %), for which the screening units deemed both sides to be worth
clarifying,
bilateral clarification was finally achieved in nine cases (3.3 %), partly through
supplementary
examinations in the breast centre.
The further evaluations exclusively concerned the 219 B4 and B5 cases.
[Tables 4] and [5] show the relationship
between operations that were planned and operations that were performed. In 82.2 %
of cases
(n = 180), there was conformity between those planned and those performed, and 11.0 %
(n = 24)
displayed a deviation. Thereby, there was low total conformity with a Kappa of 0.294
[3]. The test as to whether Kappa = 0 (full conformity) was rejected with a
significance of 0.000.
Table 4 Cross-table of planned surgical screening * operation
performed.
|
Operation performed
|
Total
|
|
Breast-conserving surgery
|
Mastectomy
|
Secondary mastectomy
|
|
Planned operation
|
Preoperative MRI scan
|
„Other“ in Mammasoft
|
4
|
0
|
0
|
4
|
|
Preoperative 2nd focus punching
|
|
3
|
0
|
0
|
3
|
|
Open biopsy
|
|
8
|
0
|
0
|
8
|
|
Breast-conserving surgery
|
169
|
13
|
8
|
190
|
|
Mastectomy
|
3
|
11
|
0
|
14
|
|
Total
|
187
|
24
|
8
|
219
|
Table 5 Symmetrical measurements.
|
Value
|
Asymptotic standard errora
|
Approximate Tb
|
Approximate significance
|
|
a The null hypothesis was not accepted.
b When
assuming the null hypothesis, the asymptotic standard error is used.
|
|
Extent of conformity (Kappa)
|
0.294
|
0.080
|
6.274
|
0.000
|
|
Number of applicable cases
|
219
|
|
|
|
There was no assertion in 15 cases, as “other” had been entered into the Mammasoft
software and
therefore further clarification was recommended instead of surgical treatment.
Additional examinations and their effects
Documented under “Other”, the screening unit suggested an additional MRI on four occasions,
an
additional punch biopsy on three occasions and eight open biopsies. In total, an additional
three
ipsilateral and one contralateral ultrasounds, five ipsilateral mammographies as well
as seven
ipsilateral punch biopsies were performed on these patients.
Among the deviations due to additional examinations carried out with further suspect
findings
detected as a result, presented in [Table 6], was a change in size ([Table 7]) that made an expanded operation necessary, or also allowed for a
smaller operation, and serious histopathological results ([Table 8]).
Table 6 Cross-table between additional examinations and the comparison of the
treatment suggested in the screening unit with the treatment actually
provided.
|
Comparison of treatments
|
Total
|
|
Conformity
|
Deviation
|
No assertion, because “other”
|
|
Additional examination
|
Performed
|
60
|
16
|
12
|
88
|
|
Not performed
|
120
|
8
|
3
|
131
|
|
Total
|
180
|
24
|
15
|
219
|
Table 7 Difference in mm of the imaging diagnostics in the screening and
histopathological finding.
|
Difference in mm
|
Frequency
|
Percentage
|
|
≤ − 10
|
29
|
13.2
|
|
> − 10 < + 10
|
156
|
71.2
|
|
≥ + 10
|
31
|
14.2
|
|
None
|
3
|
1.4
|
|
Total
|
219
|
100.0
|
Table 8 Cross-table between histopathological diagnosis and the comparison of
the treatment suggested in the screening unit and treatment actually
given.
|
Pathology
|
Comparison of Treatments
|
Total
|
|
Conformity
|
Deviation
|
No assertion
|
|
Unambiguous histological assertion
|
156
|
22
|
10
|
188
|
|
“Consistent with”
|
14
|
1
|
4
|
19
|
|
Different histological differential diagnoses
|
10
|
|
|
|
|
Total
|
180
|
24
|
15
|
219
|
Additional examinations were conducted as a result of clinically occult findings that
were detected
while preparing for the operation or during marking or following the upgrade of the
lesion in the
histology and the therefore further indication for assessment in the case of lesions
that had
previously only been regarded as element to control.
The influence of additional examinations on patients for whom no clear treatment recommendation
was
initially given was considered in detail, showing that 76 additional examinations
carried out
resulted in a change of treatment in 16 cases, whereas there was a change of treatment
in only 8
cases out of 128 when no additional examinations were performed. In total, there was
treatment
conformity in 180 patients and treatment deviation in 24 patients. No assertion could
be made for 15
patients, as there was initially no treatment recommended and “other” was entered
([Table 6]).
For the ratio of conformity versus deviation regarding the initial treatment recommendation,
the
following pairs arise for the individual additional examinations carried out:
-
Ipsilateral ultrasound 9 : 4,
-
contralateral ultrasound 1 : 1,
-
ipsilateral mammography 3 : 1,
-
contralateral mammography 1 : 0,
-
ipsilateral punch biopsy 27 : 10,
-
contralateral punch biopsy 1 : 0,
-
MRI 35 : 11.
In addition, one patient in the group with conformity rejected the recommended MRI.
Multiple
additional examinations were conducted for some patients, which explains the total
of additional
examinations deviating from the number of patients.
There was a deviating postoperative result with regards to histology in 37 (16.9 %)
out of the 219
cases.
Here, it is to be noted that some of the descriptive histological assertions were
not formulated as
unambiguously as when they were finally documented under B classification. [Table 8] presents the comparison between the accuracy of the descriptive histology within
the groups and the conformity with the original treatment suggestion.
Discussion
Problem of B lesions
With this evaluation, the potential extent of further diagnostic measures before the
definitive
operative treatment outside the screening unit on the recommended treatment should
be
illuminated.
The reason why these further measures were not carried out before the patients went
to the breast
centres lies in the organisation of the mammography screening [4], [5]. An evaluation of how often additional preoperative diagnostics take
place following the completion of the assessment in the screening unit is not known
to us and we
therefore cannot make a comparison of this here.
In accordance with previous agreements, diagnostic investigation in the context of
a mammography
screening ends upon the confirmation of a B3–5 lesion and ideally the open biopsy
and/or definitive
treatment will have taken place in a certified breast centre in accordance with oncological
guidelines.
Depending on the existing histological result, the procedure and further consecutive
therapeutic
measures taken differ in the case of increasingly individualised treatment [6].
The proportion of malignant findings that make a further oncological operation necessary
then varies
in the case of open biopsies of B3 changes, for example, depending on the percentage
of ADH
(atypical ductal hyperplasia) or a potential histological result following the punch
biopsy that
corresponds to a B3 lesion, as there is only a quantitative difference between ADH
and a low grade
DCIS.
This is one reason for the deviation between the recommendation and the eventual treatment
provided
and the higher number of operations than number of patients.
In our group, this was the case in 8.3 % of our B3 cases, which corresponds to around
one third of
the number reported by the Weigel et al. working group [7].
Weigel et al. published a B3 lesion rate of 15.1 % of 37 178 patients across the German
mammography
screening and a minimally invasive clarification in 979 of these participants. The
presence of B4
lesions was proven in 0.4 % (n = 4) of cases, and B5 lesions were evinced in 36.3 %
(n = 355) of
cases. In the case of B3 lesions, for which the recommended surgical excision (91/109)
was
ultimately carried out, there were 25 malignant findings, which corresponded to a
PPV of 0.28 for a
malignant finding.
Deviations from the recommended procedure resulted from the patient rejecting the
operation or the
operation being delayed until the time of the evaluation [7].
12.5 % B3 lesions (n = 250) was reported by the Hannover/Schaumburg screening unit
for the years
2006–2009. Of these, an open biopsy was only carried out in 66.5 % (n = 183) of cases.
The authors
calculated a projected malignancy rate of 16 % for all B3 lesions, although a control
imaging
examination was only carried out in 33.5 % of cases [8].
This corresponds to roughly double the rates we calculated.
In the case studies we participated in, almost all B3 lesions led to an open biopsy,
also taking
possible overdiagnosis (quaternary prevention) into consideration. The prevalence
of malignant
tumours could not be presented as a percentage due to the different breast centres
taking part (one
screening unit cooperates with multiple breast centres) as well as partly due to us
not having data
available.
In comparison with the period from 1998 to 2000, a study from Nottingham showed that
there had been
an increase in B3 lesions from 3.1 to 4.5 % by 2007–2008, while at the same time the
positive
predictive value fell from 25 to 10 %. There were no significant changes of this kind
in the same
study for B4 lesions [9].
This evaluation is the view of one surgical facility, which had no influence on whether
all or a
particular representative percentage of patients with B3 lesions were admitted there
from the
screening units. It is also not known what percentage of patients with B3 lesions
were recommended
an open biopsy by their respective screening units in its evaluation. The B3 lesions
were therefore
addressed in detail.
Upgrade from punch biopsy to definitive histology
This investigation-related diagnostic uncertainty with regards to B3 lesions due to
the biopsies not
always being representative, also applies to a limited extent for the DCIS, and in
some cases leads
to an upgrade in its grading.
In the cases available to us, 6.8 % were upgraded to DCIS and 2.3 % to a carcinoma.
These results correspond to what is commonly experienced with limited diagnostic precision
from
different biopsy methods [10], [11], [12]. Kettritz et al., for example, reported an upgrade rate of 24 % from
ADH to DCIS and of 12 % from DCIS to carcinoma following a vacuum-assisted biopsy
[13]. However, the upgrade rate in the group of patients available to us
turned out to be much lower, and does not lend itself to a one-for-one comparison
due to the
incomplete data situation (see above). It is a point in support of a more precise
biopsy method.
Size deviations and multi-stage operations
There were more and more differences in the preoperative imaging diagnostics in clinical
practice
with regard to size, which can ultimately prevent a planned breast-conserving procedure
and force a
change in treatment, e.g. from a breast-conserving operation to a primary or secondary
mastectomy.
From a variety of comparative studies between mammography on its own and magnetic
resonance imaging
(MRI) of the breast with mammography, it is known that mammographyʼs ability to estimate
size alone
is limited, particularly with dense gland tissue [14], [15].
The studies of preoperative radiology also show that the extension is difficult to
estimate,
particularly when a lesion possesses an imaging correlate that is visible only to
a limited extent.
Here, the ability to assess the free edge is limited [16].
The deviation in size is also the main reason for multi-stage operations, as the follow-up
resections
that constitute part of the multi-stage operations can be explained by this. The further
part
resulted from the sentinel lymph node biopsies following an upgrade regarding the
grading or a
positive sentinel lymph node not recognised in the intra-operative quick section which
involved an
axilla dissection.
Supplementing the diagnosis through additional diagnostic investigation
What is decisive, however, is the assertion of to what extent additional diagnostic
measures lead to
further suspect lesions being discovered, and how far this additional diagnostics
leads to changed
treatment options. The examination presented here does not indicate this for surgery
([Table 3]). Changes in systematic treatment resulting therefrom were not
recorded. Finally, it must be verified in a long-term analysis whether more diagnostic
investigation
leads to life being extended. The unclear diagnostic possibilities in the imaging
examinations of
the mammary do not always lead to changes in the operation decision in the case of
additional
examinations and potentially further biopsies. Therefore, the subjective costs for
the patients
compared to the diagnostic gains in the decision for additional investigation steps
are to be
weighed up against each other.
In accordance with the S3 guidelines, ultrasound is the first-choice method in curative
situations in
the case of an unclear mammography finding. An ultrasound was recommended as a supplementary
method
from the outset for women with an ACR density index of between 3 and 4. When an ultrasound
scan and
an MRI were also added to the screening due to the presence of a dense parenchyma,
this resulted in
a higher false positive rate as well as a higher carcinoma detection rate [17] The comparative study of the American College of Radiology Imaging Network showed
that, for women aged forty years and older, almost 30 % more tumours were detected
through adding a
mammography ultrasound to the screening, but this also leads to false positive diagnoses
in almost
four times as many cases [18].
The QuaMaDi cohort (open, quality-assured breast diagnostics for all legally insured
women in
Schleswig-Holstein), who received an additional ultrasound in the case of grade 3
or 4 ACR gland
parenchyma thickness, showed that an additional 10 % of cases of breast cancer could
be identified
through a double reading of the mammographies or an additional ultrasound [19]. This emphasises the necessity for additional examinations for selected patients.
It is also worth discussing whether the potentially necessary additional examinations
should still
take place in a radiological facility (the screening unit) or be delegated to the
breast
centres.
Conclusion for Clinical Practice
Conclusion for Clinical Practice
A higher percentage of patients who received surgical treatment that deviated from
the original treatment
recommendation made by the screening units underwent additional examinations than
those who did not.
This does not mean that as many additional examinations should be carried out as possible,
but rather
that they should be in the case of ambiguous radiological findings. The surgeon ultimately
selected by
the patients must be aware of the limits of the screening programme and the therefore
sometimes
incomplete assessment of a patientʼs findings, and should thus carry out a preoperative
evaluation in
close cooperation with the clinical radiologists and order additional examinations
if necessary.
