C ANCER TREATMENT REVIEWS 2000; 26: 91–102

doi: 10.1053/ctr v. 1999.0151, available online at http://www.idealibr ar y.com on

CONTROVERSY

Tumour marker measurements in the diagnosis and

monitoring of breast cancer
K. L. Cheung, C. R. L. Graves and J. F. R. Robertson

Professorial Unit of Surgery, City Hospital, Nottingham, UK

Elevation of established blood tumour markers correlates with the stage of breast cancer.The major role of current blood
markers is therefore in the diagnosis and monitoring of metastatic disease. A combination of markers is better than a single
marker with the most widely adopted combination being CEA and one MUC1 mucin, commonly detected as either CA15.3
or CA27.29.Tumour marker measurement is now used as a complementary test in the diagnosis of symptomatic metastases.
In the monitoring of therapeutic response to both endocrine and cytotoxic therapies in advanced disease, biochemical
assessment using blood markers not only correlates with conventional UICC criteria but has a lot of advantages which make
it a potentially superior way of assessment. In this regard, CA15.3, CEA and ESR are the best validated combination.
Studies are ongoing to evaluate the use of sequential blood tumour marker measurements in the follow-up of patients
after treatment for their primary breast cancer, in terms of both early detection and early therapeutic intervention. Further
randomized studies are also required to ascertain that marker-directed therapy is superior to the current practice
for metastatic disease. In line with clinical studies, intensive laboratory work is being carried out to optimize the use of
blood markers in advanced disease as well as to exploit their use in screening and diagnosis of early primary breast cancer.
© 2000 Harcourt Publishers Ltd

Keywords: Blood tumour markers; measurements; assays; breast cancer; diagnosis, monitoring; therapeutic response;
biochemical assessment.

discussion to blood tumour markers, covers the

INTRODUCTION basic principles in tumour marker measurements,
their established roles in current clinical practice
Blood tumour markers in breast cancer have been and their potential roles in future development in
known for decades. In contrast to markers in the pri- the diagnosis and monitoring of breast cancer.
mary tumour tissue, blood tumour markers reflect a
dynamic situation and their measurements can be
repeated as required. The use of blood tumour PRINCIPLES OF BLOOD TUMOUR MARKER
markers is most established in the diagnosis and MEASUREMENTS
monitoring of symptomatic metastatic disease.
Their role in early diagnosis and treatment of breast A large number of blood tumour markers have been
cancer remains to be elucidated. Active laboratory proposed for breast cancer, including mucins, onco-
research is also ongoing to refine the assays of blood foetal proteins [e.g., carcinoembryonic antigen
tumour markers. The present review, confining the (CEA)], oncoproteins (e.g. c-erbB2, c-myc and p53),
cytokeratins [e.g., tissue polypeptide antigen (TPA)
and tissue polypeptide specific antigen (TPS) which
Address correspondence to: Professor JFR Robertson,
Professorial Unit of Surgery, City Hospital, Hucknall Road,
refers to one specific epitope on the TPA molecule
Nottingham NG5 1PB, UK. Tel: +44 (0) 115 9627951; Fax: +44 (0) and erythrocyte sedimentation rate (ESR). Of
115 8402618; E-mail:[email protected] these CEA and MUC1 mucin are most widely used

0305-7372/00/020091 + 12 $35.00/0 © 2000 HARCOURT PUBLISHERS LTD

92 K.L. CHEUNG ET AL.

in clinical practice. At present their use in breast can- MUC1 mucin has a high molecular weight of
cer is confined to metastatic disease, both in its diag- 250–1000 kD with a protein core and carbohydrate
nosis and in the monitoring of response to systemic side chains. The extracellular domain of MUC1 con-
therapy. sists of a heavily O-linked glycosylated peptide core
made up in the main of a variable number of multi-
ple repeats of a 20-amino acid sequence referred to as
Carcinoembryonic antigen the variable number tandem repeat (VNTR) domain
with each VNTR having five potential O-linkage
Carcinoembryonic antigen has been the marker most sites. Glycosylation is under enzymatic control and
investigated. It is a glycoprotein normally found in occurs post-translationally and sequentially, leading
embryonic endodermal epithelium. In the mid 1960s, potentially to the production of four sugar core types
Gold and Freeman isolated CEA from extracts of capable of elongation into polysaccharide chains.
malignant tissue (1, 2). It can be measured by a num- Alterations to these enzymes and to enzymes
ber of commercially available assays of which the involved in post-translational modification, which
automated chemiluminescence system (ACS: 180®, can be associated with disease, will obviously affect
Chiron Diagnostics, UK) is one example. This is a the glycosylation state of MUC1. In general, the poly-
two-site sandwich immunoassay using two antibod- saccharide side chains of tumour associated MUC1
ies: a purified polyclonal rabbit anti-CEA antibody are shorter than those on the normally expressed
labelled with acridinium ester in the lite reagent, and molecule. Both aberrant and up-regulated expression
in the solid phase, a monoclonal mouse anti-CEA of MUC1 are features of malignancy. It is apparent,
antibody covalently coupled to paramagnetic parti- therefore, that MUC1 is highly heterogeneous owing
cles. Increased serum CEA levels have been detected to its polymorphic nature (the VNTRs) and the high
in patients with primary colorectal cancer and in degree of variation in glycosylation. A large number
patients with other malignancies including gastroin- and diversity of distinct monoclonal antibodies have
testinal, breast, lung, ovarian, prostatic, liver and therefore been raised against these different epitopes.
pancreatic cancers (1–5). It might also be raised in It is this finding which has been the basis for the
individuals with inflammatory as well as various development of numerous immunoassay kits such as
other conditions (e.g., diverticulitis, gastritis, gastric cancer antigen 15.3 (CA15.3), mucin-like carcinoma
ulcer, bronchitis, cholangitis, liver abscess and alco- associated antigen (MCA), CA549, CA27.29, breast
holic cirrhosis), especially in the elderly and in cancer mucin (BCM), EMCA, M26 and M29.
smokers (5,6). MUC1 mucin as detected by CA15.3 sandwich
Serum CEA is elevated in 30–50% of patients with capture assay using specific monoclonal antibodies
symptomatic metastatic breast cancer (7–9). A num- 115D8 (raised against human milk fat globule mem-
ber of studies have reported a positive correlation branes) and DF3 (raised against a membrane
between changes in serum CEA and therapeutic enriched fraction of metastatic human breast carci-
response in patients with metastatic breast cancer noma) was the first mucin marker in which sequen-
(10–13). tial changes were reported to correlate with
therapeutic response (16–19). CA15.3 is most widely
used and is regarded as the “gold” standard against
MUC1 mucin which other assays are compared. Serum CA15.3 is
elevated in 54–80% of patients with metastatic breast
Polymorphic epithelial mucin (PEM) is one of a num- cancer (16,19–23). It may also be increased in other
ber of descriptive terms for the large glycoprotein benign (e.g., chronic hepatitis, liver cirrhosis, tuber-
encoded by the MUC1 gene. The glycoprotein MUC1 culosis, sarcoidosis and systemic lupus erythemato-
mucin is expressed by most “wet” epithelia e.g., sus) and malignant (e.g., lung, ovarian, endometrial,
bladder, breast, stomach, pancreas, ovary and respi- gastrointestinal and bladder carcinomas) conditions.
ratory tract. In normal breast tissue, MUC1 is These should normally be considered and excluded
expressed on the apical surface of epithelial cells in as causes of elevated CA15.3 by a detailed history
the ducts and acini from where the molecule is shed and examination as well as investigations.
via milk fat globules and in soluble form into the Owing to the presence of numerous epitopes
milk. In case of tumours, cell polarization is lost and against which a variety of monoclonal antibodies can
this altered cell surface expression, coupled with the be raised in the detection of circulating MUC1
disruption of the normal tissue architecture caused mucin, use of different monoclonal antibodies can in
by the growing tumour, allows MUC1 mucin to be a minority of individual patients produce different
shed into the circulation where it can be measured by results in the monitoring of breast cancer. As men-
means of immunoassays (14, 15). tioned the CA15.3 test has been most widely used
TUMOUR MARKERS IN BREAST C ANCER 93

until recently when the CA27.29 monoclonal anti- no value in employing a combination of MUC1
body has been utilized in a competitive radioim- mucin markers (28–30). CA15.3 has a higher sensitiv-
munoassay to detect MUC1 mucin (24). Different ity than CEA but with a similar specificity. However,
limits in the normal range have been reported when combining different markers has been shown to be
using different assays and the results were found to better than any single marker in the diagnosis and
be not interchangeable (23). Variations have also monitoring of metastatic breast cancer. The most
been encountered between manual and automated widely adopted recommendation has been the com-
methods. It is therefore necessary to emphasize the bination of CEA and one MUC1 mucin, most com-
use of the same MUC1 mucin assay for interpreting monly detected as CA15.3 (31, 32). The sensitivity of
sequential results. If for some reason it is necessary to tumour marker measurements for detecting metasta-
change test methods during patient follow-up, a new tic disease can be increased to over 80% when both
baseline has to be established (25). CA15.3 and CEA are used (33).
When these two markers are used together with
ESR, the sensitivity can reach beyond 90% (33–37). It
Tumour marker spike has been known that ESR, a test frequently used in
clinical medicine, tends to increase in patients with
Another phenomenon worth noting during the mon- cancer, particularly as the disease progresses.
itoring process is the occurrence of “tumour marker Elevation of ESR has been reported in patients with
spike” (26, 27). It is a transient increase from baseline metastatic breast cancer (9, 38–40). Our group has
with a subsequent return to or below baseline levels devised a biochemical index score using CA15.3, CEA
in up to 30% of patients who show a response to ther- and ESR in the monitoring of therapy in advanced
apy. The peaks usually occur within 30 days of com- breast cancer (34, 38). One criticism raised against the
mencing a new therapy although spikes may last as use of ESR is that it is not tumour specific. In fact all
long as 90 days. It is therefore important to take note patients have already had the diagnosis of breast can-
of this phenomenon rather than interpreting it as dis- cer and the presence of metastatic disease has been
ease progression in the early course of therapy. diagnosed also by methods other than blood tumour
Spikes can be distinguished from a true progressive markers (i.e., clinically and/or imaging). Concurrent
rise in a tumour marker either by waiting to repeat malignancies, particularly with both showing dissem-
marker measurements until 2–3 months after starting inated disease, are uncommon. The chance therefore
treatment or alternatively by measuring marker lev- of having a second malignancy (even asymptomatic)
els after both 1 and 2 months of treatment (25). will have been greatly reduced by clinical assessment
not to mention radiological investigations. Hence the
use of ESR in monitoring patients with advanced
Sensitivity and specificity breast cancer is specific for the reasons noted above.
The sensitivity of the blood markers can be
Breast cancer is a heterogeneous disease. From early increased by using a combination of markers. The
studies it was clear that no single blood marker specificity of tumour marker measurement has been
would suffice for monitoring the therapeutic improved by using cut-off based criteria (41). A num-
response for advanced breast cancer. Different ber of cut-off values have been used in the various
MUC1 mucin assays appear equivalent and there is reported studies (Table 1) (20, 22, 34, 42–45). A higher

TABLE 1 Cut-off values in blood tumour marker measurements

CEA (ng/ml) References

3 Safi 1991 (43)
5 Colomer 1989 (20)
6 Robertson 1991 (34), Jäger 1995 (45)
7 Al-Jarallah 1993 (44)
10 Cantwell 1982 (42), Molina 1995 (22)

CA15.3 (U/ml) References

25 Safi 1991 (43)
33 Robertson 1991 (34)
40 Colomer 1989 (20), Al-Jarallah 1993 (44), Jäger 1995 (45)
60 Molina 1995 (22)
94 K.L. CHEUNG ET AL.

cut-off value is chosen since it is unusual to see ele- advanced disease. Elevation of markers can be seen
vation of the markers in patients with in benign con- more in locally advanced primary disease though
ditions. Molina et al. achieved a specificity of 99% not as many as in patients with metastatic disease.
when the cut-off values were taken as 10 ng/ml for Blood tumour markers were reported to be raised in
CEA and 60 U/ml for CA15.3 (22). They also incorpo- 33% and 68.2% of patients with locally advanced pri-
rated the criterion of having a serial rise of >15% in mary and metastatic breast cancers respectively in a
relation to the previous value to attain such results. large series which included patients with benign dis-
Comparable results have also been obtained by our ease as well as normal controls. Statistical analyses
group using the mean +2 SD of the normal controls to showed that blood tumour markers were not useful
calculate the cut-off value for each marker. in diagnosing primary cancers (stage I/II or stage III)
To qualify a change, an increase or decrease more (9). Moreover, their role in diagnosing locally
than the inter-assay coefficient of variation (i.e., advanced primary disease is minimal since the diag-
>10%) was required (34, 37, 38). We found compara- nosis can easily be made clinically but their measure-
ble results when using either 10% or 20% as a change ment after primary therapy may be clinically useful.
although various percentages were employed in Blood tumour marker measurements have minimal
other studies (Table 2) (22, 30, 33, 38, 46). role in diagnosing locoregional recurrence after pri-
mary treatment for early breast cancer. Nevertheless,
the measurement of CA15.3 at the time of local recur-
BLOOD TUMOUR MARKERS IN CURRENT rence has been found to be useful. Elevation of
CA15.3 at that time indicated either the presence of
CLINICAL PRACTICE radiologically visible metastases or predicted the
development of symptomatic metastases sooner than
Diagnosis in patients who had local recurrence and a normal
CA15.3 measurement (51).
Most studies consistently showed that elevation of In the diagnosis of metastatic breast cancer,
blood tumour markers correlated with the stage of CA15.3 assay has been shown to be superior with
breast cancer. In the majority of studies no significant CEA being the next most clinically useful marker
increase in marker values was found in patients with (47). High sensitivity up to 87% with high specificity
primary breast cancer compared to controls reaching 96% have been reported when using
(9,19,43,47). However some groups have claimed that CA15.3 alone (52–54). Various reports of sensitivity
MUC1 mucin is significantly elevated in primary of common markers are summarised in Table 3 (22,
breast cancer compared to normal controls (48,49). 33–35, 43, 44, 52, 55). As mentioned a sensitivity of
Even if CA15.3 is statistically raised in patients with >90% could be reached when a panel of three mark-
primary breast cancer it raises the question as to the ers i.e., CA15.3, CEA and ESR are used (34–37). Blood
value of a preoperative CA15.3 for the individual tumour markers are specific “alarm” indicators of
patient. CA15.3 and CEA were found to be increased metastatic disease and preoperative elevation might
respectively in only 31% and 26% of patients with pri- indicate under-staging (41). O’Hanlon et al. found
mary breast cancer when lower than usual cut-off val-
ues were used (25 U/ml and 3 ng/ml respectively) TABLE 3 Sensitivities of blood tumour marker measurements in
(42). However the increased sensitivity is at the cost metastatic breast cancer
of lower specificity. With the low sensitivity of the
current marker assays, blood tumour markers have Marker(s) Sensitivity References
yet been proven to be useful in screening and CEA alone 50% Safi 1991 (43)
diagnosis of early breast cancer (9, 19, 41, 43, 47, 50). 46% Molina 1995 (22)
The major role of current blood tumour markers 53% Jäger 1995 (45)
in the diagnosis of breast cancer is therefore in CA15.3 alone 73% Safi 1991 (43)
81% O’Brien 1992 (55)
54% Molina 1995 (22)
TABLE 2 Percentages of change in blood tumour marker 56% Jäger 1995 (45)
concentrations (See text) 87% Coveney 1995 (52)
CEA or CA15.3 80–90% Al-Jarallah 1993 (44)
% change References
64% Molina 1995 (22)
10 Williams 1990 (38) 81% Jäger 1995 (45)
10 or 20 Robertson 1999 (33) 94% Coveney 1995 (52)
15 Molina 1995 (22) CEA or CA15.3 or ESR 92% Robertson 1991 (34)
20 Iwase 1994 (46) 100% Dixon 1991 (35)
25 Deprés-Brummer 1995 (30) 96% Robertson 1999 (33)
TUMOUR MARKERS IN BREAST C ANCER 95

that CA15.3 correlated with tumour burden. In a have been shown to be the best combination vali-
prospective study of the role of preoperative meas- dated retrospectively and prospectively. A biochemi-
urement of CA15.3 in 500 patients with breast cancer, cal index score using these three markers was
patients with a value >40 U/ml had an 83% chance retrospectively derived, prospectively validated, both
of having at least stage III disease. The authors there- in a single unit and in a multicentre study (33, 34,
fore recommended a routine use of tumour markers 36–38, 60). Namely, any change in marker while the
in the preoperative assessment of patients with a patient is on therapy is related to the pre-treatment
presumptive diagnosis of primary breast cancer (53). value. A cut-off for each marker of the mean +2 SD of
A normal tumour marker level generally suggests the normal controls is calculated. Patients who never
the absence of metastatic disease. This was substanti- show an elevation of the marker above this level are
ated by a study of over 150 patients revealing that a regarded as biochemically unassessable for that par-
normal CA15.3 value was strongly predictive of a ticular marker. Patients with an initial pre-treatment
negative bone scan and vice versa (56). In the recently value below the cut-off which subsequently rises
published British Association of Surgical Oncology above the cut-off or patients with an initial value
(BASO) Guidelines for the Management of Metastatic above the cut-off which subsequently increases above
Bone Disease in Breast Cancer in the United the inter-assay coefficient of variation (i.e., >10%) are
Kingdom, tumour marker (CA15.3 and CEA) meas- regarded as showing an increasing marker level
urement is recommended as one of the diagnostic (scored +2), indicative of biochemical progression.
tests in addition to the time-honoured algorithm Patients who start with an initially high value which
using plain radiographs, serum calcium and skeletal falls to below the cut-off or patients with an initial
scintigraphy for a woman with clinical suspicion of value above the cut-off which subsequently decreases
bony metastases (57). by more than the inter-assay coefficient of variation
At present blood tumour marker elevation is not for that marker (i.e., >10%) are regarded as showing a
used as the sole means of diagnosis of metastatic decreasing marker level (scored –2), indicating a bio-
breast cancer. Nevertheless, as will be outlined chemical response; ESR falling by >10% is scored –1.
below, there are currently studies ongoing looking to Patients with levels which start and remain above the
identify patterns of marker changes and marker ele- cut-off but which move by less than the inter-assay
vation which would allow early therapeutic inter- coefficient of variation (i.e., ±10%) are regarded as
vention by clinicians for occult metastatic disease. If being biochemically stable and scored +1. These
these studies are successful it could be envisaged in changes and the scoring attached to them are sum-
future that tumour marker measurement might then marised in Table 4. Scores for individual markers are
be accepted as the method for monitoring patients in then added together to produce an overall biochemi-
the follow-up period after primary treatment and cal index score. Total scores >0 are considered as bio-
that elevated blood markers might be an indication chemical progression while an index score of ≤0 is
to change systemic therapy. considered a biochemical response.
The role of blood tumour marker measurements
has been shown to be useful in the monitoring of
Monitoring response to endocrine therapy and cytotoxic therapy
(including standard regimen using cyclophos-
The use of blood tumour markers in the monitoring phamide, methotrexate and 5-fluorouracil, as well as
of therapeutic response in patients with metastatic mitozantrone, anthracycline and even docetaxel con-
breast cancer is well established. Various markers taining regimens) (33, 34, 36–38, 60, 63).
have been employed starting from more than a Traditionally the response to systemic therapy is
decade ago, including CEA, C-reactive protein, fer- monitored using criteria laid down by the
ritin, orosomucoid, CA15.3, HMFG1 and HMFG2 International Union Against Cancer (UICC) (64).
(monoclonal antibodies raised to human milk fat Biochemical assessment using blood markers has
globule membrane fractions), NCRC-11, BCA225,
MCA, TPS etc (9, 18, 19, 30, 33–37, 43, 44, 46, 52,
58–61). CA15.3 was found to be superior to other TABLE 4 Scores for changes in blood tumour marker
concentrations
markers (18, 19, 43, 52) although there has been some
recent work suggesting that TPS may be of use (59). Upper limit Normal Decrease Stable Increase
However others have suggested that TPS does not of normal limits (>10%) (±10%) (>10%)
add to current markers (62).
CEA 6 ng/ml 0 –2 +1 +2
As mentioned previously, a combination is better
CA15.3 33 U/ml 0 –2 +1 +2
than a single marker when blood tumour marker
ESR 20 mm/hr 0 –1 +1 +2
measurements are considered. CA15.3, CEA and ESR
96 K.L. CHEUNG ET AL.

been shown to correlate with conventional UICC cri- marker measurements at 3 and 6 months of therapy.
teria with a lot of advantages which make it a supe- Furthermore, biochemical assessment may result in
rior way of monitoring patients with metastatic breast at least 50% cost-savings when compared with con-
cancer. The UICC criteria assess structural changes as ventional assessment by clinical/radiological criteria
visualised on imaging while tumour markers assess which often require expensive imaging techniques
the dynamics of the tumour which could be detected such as CT or MRI scans (68).
earlier. The imaging tests according to UICC criteria
reflect structural change – a late event because it
depends on at least one metastasis reaching a signifi- BLOOD TUMOUR MARKERS IN FUTURE
cant size. Changes in blood tumour markers appear CLINICAL PRACTICE
to reflect the total tumour burden which may be
measurable from the summation of numerous sub-
clinical metastases. Blood tumour marker measure- Diagnosis
ments are more objective and are reproducible while
interpretation of imaging can be subjective. Bio- Despite recent advances in breast cancer care (e.g.,
chemical response or progression occurs in line with mammographic screening, breast conservation and
and often pre-dates UICC assessed response or pro- adjuvant therapy), the majority of patients will still
gression (41). A lead time of 1–10 months ahead of develop metastases from which they will ultimately
UICC assessed response or progression has been die. Early detection of metastases by intensive imag-
reported when assessment was made using blood ing tests has been reported to be of no benefit over
markers (22, 35, 46, 52, 63). This allows more effective routine follow-up in terms of patients survival
palliation which is especially important in patients (69–71). However, tumour markers are known to be
with advanced breast cancer. Early discontinuation of elevated in some patients even before radiological
ineffective therapy, early change of treatment for non- evidence of metastases. Some evidence can be cited
responders and further continuation of effective treat- in support of the view that earlier detection using
ment are made feasible. Using a combination of blood tumour markers may be clinically worthwhile.
CA15.3, CEA and ESR, practically 100% of patients It is now established that adjuvant therapy prolongs
are biochemically assessable (34, 39, 42). In a study survival compared to waiting and using the same
conducted by the European Group for Serum Tumour therapy when symptomatic metastases are diag-
Markers in Breast Cancer, 83 patients with metastatic nosed (72). The concept could be extended to treat-
breast cancer assessable for CA15.3 and CEA (with 67 ment of patients post-surgery who appear disease-
patients assessable for ESR as well) were recruited free clinically but in whom there is other evidence of
and prospectively evaluated in 11 centres from six metastases (i.e., elevation of blood tumour markers).
European countries. Among the 67 patients who had This could apply irrespective or not of whether
all three markers assessed in the form of the biochem- patients received adjuvant therapy.
ical index score, 84% of patients had elevation of one Blood tumour markers have been recognized to be
or more of these three markers while during therapy potentially more useful than clinical and/or radiolog-
the number rose to 96%. The other 4% remained in ical assessment since 1980s (42). Blood tumour marker
remission throughout the study and all markers measurements can identify metastatic relapse in
remained below the cut-off levels (33). 50–80% of patients during follow-up after treatment
Biochemical assessment provides the only vali- of primary breast cancer with a lead time of 2–12
dated method of assessing the response to systemic months (22, 35, 41, 55, 73). A large series was reported
therapy for disease unassessable by UICC criteria by Molina et al. in Spain (22). During the follow-up of
e.g., irradiated lesions, hilar enlargement, pleural 1023 patients after treatment of primary breast cancer,
effusion, ascites, bone marrow infiltration, sclerotic 246 of them developed metastases. Elevation of blood
bone metastases, lytic bone metastases (in patients tumour markers were seen prior to the relapse in 46%
receiving bisphosphonates) which account for and 54% with a lead time of 4.9 months and 4.2
10–40% of all patients (60, 65–67). In the mentioned months respectively for CEA and CA15.3 with a speci-
BASO Guidelines for the Management of Metastatic ficity of 99% for both markers. Higher levels of mark-
Bone Disease in Breast Cancer, blood tumour marker ers, higher sensitivity and longer lead time were
(CA15.3, CEA and ESR) measurement is also recom- found in patients with oestrogen receptor or proges-
mended a valuable tool in monitoring therapy (57). terone receptor positive primary tumours. To improve
In addition to subjective assessment by symptoms, the accuracy of detection, diagnosis should not be
radiological assessment (plain radiographs with based on a single value of elevation. Either using
or without skeletal scintigraphy or MRI scan), serial measurements to confirm a persistent elevation
patients should have sequential blood tumour or watching out for the trend rather than discrete
TUMOUR MARKERS IN BREAST C ANCER 97

values has been recommended (21,47). Most studies The American Society of Clinical Oncology chose
employed CA15.3 with or without CEA while recently a cautious policy to recommend that present data
other markers such as serum c-erbB2 protein were also regarding both CEA and CA15.3 were insufficient to
found to be useful e.g., better sensitivity was achieved recommend their routine use in the diagnosis of
when using c-erbB2 rather than CA15.3 in patients recurrent breast cancer during follow-up (66). As
with over-expression of c-erbB2 in the primary tumour mentioned, at the time of occult metastasis which is
tissue (74). When compared with other diagnostic only detected by blood marker measurements, the
methods, the tumour volume is about 2–8 times tumour volume is much less than when the patient
smaller if detected only by blood markers (75). Two develops symptoms. The prognosis of patients at the
studies published at around the same time in Spain time of symptomatic metastasis is poor with a
and Germany have found a high sensitivity (64–81%) median survival of two years. It would seem logical
and specificity (99%) when using a combination of to postulate a potential beneficial effect when treat-
CA15.3 and CEA in the diagnosis of symptomatic ment is started earlier, which is possible with blood
metastases in over 9000 patients with primary breast tumour markers. In fact, results of two small pilot
cancer (46, 73). studies supported this hypothesis. In a preliminary
The use of blood markers for measuring tumour randomized study in Germany, Jäger et al., showed
burden and monitoring therapy is supported by that treatment of relapse based only on increased
studies in advanced disease as mentioned early in CEA and CA15.3 reduced the risk of developing
this review. The various advantages of objectivity, metastasis from 88% to 39% at 12 months. A longer
reproducibility, cost-savings etc seen in the manage- disease-free survival was also suggested (76).
ment of patients with metastatic disease could be Nicolini et al., from Italy, in a retrospective, non-ran-
extended to the diagnosis of recurrence in the follow- domized study of 384 patients attending a postoper-
up of primary breast cancer patients. Being a simple ative follow-up clinic after treatment of their primary
and an efficient test, blood tumour marker measure- breast cancer, evaluated the role of early therapeutic
ments could even be carried out at the primary intervention based on blood tumour markers (CEA,
healthcare level which will further enhance the cost- CA15.3 and TPA) (77). Among all patients who
effectiveness of the follow-up programme. relapsed, 28 patients were treated “early” based on
Despite such encouraging results in various stud- rising tumour markers while 22 patients were treated
ies the medical community is still unconvinced that after radiological diagnosis of metastases as in the
regular blood tumour marker measurements will in usual practice. The authors found a lengthening of
fact provide significant lead time and therapeutic the time from marker increase to the appearance of
advantage in the diagnosis of metastases. They there- clinical or radiological signs of metastasis (13.5 ver-
fore continue to recommend clinical follow-up with sus 3.4 months) as well as improved overall survival
imaging investigations when patients develop suspi- (42.9% versus 22.7% at 72 months) when patients
cious symptoms despite the fact that the imaging were treated “earlier”.
tests according to the UICC criteria reflect structural As mentioned, studies are currently ongoing to
change (a late event) while changes in blood tumour identify individual patterns of changes in markers
markers appear to reflect the dynamics of change in including the lead time in the diagnosis of recurrent
tumour burden. Our group is at present co-ordinat- breast cancer. This should not prevent pilot studies of
ing a prospective, international, multicentre study of the potential value of early therapeutic intervention
early detection of recurrent disease after primary based on rising blood tumour markers. Our group
surgery for operable breast cancer – current follow- has therefore started a prospective, multicentre, ran-
up protocols versus the use of sequential blood domized, pilot study of early intervention based on
tumour marker measurements. All patients are fol- tumour markers in the follow-up of patients with
lowed up 3-monthly for two visits after surgery, primary breast cancer. This study is seen as a pilot for
thereafter 6-monthly up to 5 years and annually a future large prospective randomized study which
afterwards. Diagnostic investigations for possible would compare standard practice with early treat-
recurrence are carried out as directed by symptoms ment intervention based on blood tumour marker
and clinical findings. Blood samples are collected for measurements.
the measurements of CA15.3 and CEA at each visit.
The purpose of the study is to establish the presence
of a significant lead time to symptomatic metastasis Monitoring
in a well-designed research context. Hopefully the
future follow-up protocols for primary breast cancer Although the usefulness of blood tumour markers
patients will incorporate regular blood tumour has been well established in the diagnosis and mon-
marker measurements. itoring of advanced breast cancer, blood marker
98 K.L. CHEUNG ET AL.

measurements are at present only complementary to

UICC assessment in most centres. Change of ther- Current markers
apy is still usually based on UICC evidence of dis-
ease progression despite the fact that biochemical Although CA15.3 has been the ‘gold’ standard for
progression often occurs ahead of clinical and/or blood tumour marker measurements, studies are
radiological progression. Therapy as directed by currently ongoing to assess whether different com-
tumour markers rather than by UICC criteria has mercial assays e.g., for MUC1 mucin have different
definite potential in improving clinical outcome, sensitivities. Recent studies measuring CA27.29 tend
quality of life and health economics. Significant cost- to suggest comparable results with current CA15.3
savings are achievable when expensive treatment assays (23). Other assays may have higher sensitivity
regimens are directed by tumour markers allowing in locally advanced primary disease as well as in
earlier discontinuation of an ineffective treatment. A multiple metastases when compared with single
saving of £1600/cycle is possible in the case of using metastasis (78).
docetaxel as a single-agent regimen (63). In addition Cytokeratins, including TPA and TPS, have been
to saving costs on drug therapy, patients themselves reported to be useful in monitoring therapeutic
also benefit from earlier discontinuation of treat- response in patients with metastatic breast cancer
ment as directed by tumour markers since they (79–83). In a multicentre study of 129 patients with
would not have to suffer prolongation of side-effects advanced breast cancer, TPS appeared to indicate
from an ineffective treatment. On the other hand, for outcome better than CA15.3 or CEA (59). Further
patients whose tumour markers have not decreased prospective studies of sequential measurements are
to below the cut-off values after completion of a necessary to validate the role of TPA and/or TPS in
course of cytotoxic therapy, continuation of further the diagnosis and monitoring of breast cancer.
treatment is made feasible using marker-directed Part of the extracellular domain of the c-erbB2
approach. Expensive therapy could therefore be bet- receptor has been detected in the circulating blood of
ter targeted. All these mean that patients are receiv- some breast cancer patients. The value of serum c-
ing the best choice of therapy as well as having the erbB2 measurements in terms of diagnosis, progno-
best palliation. sis, early detection of recurrent disease, selection for
Our group has previously demonstrated better endocrine therapy or chemotherapy, and monitoring
disease stabilization and survival as well as possibly therapy has recently been reviewed (84, 85).
improved quality of life in two small randomized Elevation of serum c-erbB2 was detected in 80% of
studies where marker-directed chemotherapy was patients with tumours which over-expressed c-erbB2
evaluated (35, 37). Further randomized studies incor- compared to 3.3% of those which did not over-
porating large patient number should be carried out express the oncoprotein. Correlation between c-erbB2
to ascertain that marker-directed therapy is superior expression in tissue and low oestrogen receptor sta-
to the current practice in patients receiving endocrine tus also seemed to be found when serum c-erbB2 was
or cytotoxic therapy for advanced breast cancer. elevated. It would appear that measurement of the c-
These studies should proceed along with early inter- erbB2 extracellular domain in the serum might be
vention studies based on tumour markers in the fol- useful in monitoring for tumour recurrence and in
low-up of primary breast cancer when the diagnosis predicting resistance to endocrine therapy but it has
of advanced disease is first made with blood marker yet been shown to be promising in predicting
measurements. response to chemotherapy. With the introduction of
trastuzumab (Herceptin®), the humanized anti-c-
erbB2 monoclonal antibody, as a therapeutic option
for treatment of c-erbB2 positive advanced breast
REFINEMENT OF BLOOD TUMOUR cancer, clinicians will have to decide how to monitor
MARKER MEASUREMENTS its therapy e.g., in combination with chemotherapy
such as taxanes. The measurement of c-erbB2 in the
While the usefulness of blood tumour markers is serum appeared to be useful in the monitoring of
well established in advanced breast cancer, active patients receiving fractionated paclitaxel chemother-
research, both clinical and laboratory, is ongoing to apy, particularly in those with over-expression of c-
refine the measurements of existing markers, to erbB2 in the tumour (86). The use of different markers
explore newer markers and to develop better marker may be required to tailor different therapies in the
assays, aiming to optimize their use in advanced dis- monitoring of breast cancer patients. Further studies
ease as well as to exploit their use in screening and to validate the role of serum c-erbB2 in the monitor-
diagnosis of early primary breast cancer. ing of Herceptin®-containing therapy are necessary.
TUMOUR MARKERS IN BREAST C ANCER 99

CEA reflect the efficacy of anti-cancer therapy on

New markers tumour mass. In these ways new markers have a
complementary rather than an exclusive role in the
New markers which are being investigated include diagnosis and monitoring of breast cancer.
various markers of angiogenesis, bone metabolism,
and growth factors. A lot of studies have been carried
out on angiogenesis which has been believed to per- CONCLUSION
mit metastases. Correlation of angiogenesis with
prognosis has been found (87, 88) but it has yet to Blood tumour marker measurements in breast cancer
be reliably proven to add significant information are most established in advanced disease. Their roles
to established prognostic factors (89). However in current clinical practice include the diagnosis of
there was evidence suggesting that apoptosis and symptomatic metastases and the monitoring of
angiogenesis might be valuable as markers for response to treatment in patients with metastatic dis-
response in patients receiving systemic therapy for ease. Their potential roles in future clinical practice
breast cancer (90). are in the early detection and therapeutic interven-
Traditional markers of bone metabolism include tion for occult metastases in the follow-up of patients
serum alkaline phosphatase, serum and urinary cal- with primary breast cancer, as well as in the institu-
cium, urinary hydroxproline etc. Markers of collagen tion of marker-directed systemic therapy for
synthesis have been evaluated as bone markers for metastatic breast cancer. Although the most widely
metastatic bone disease due to breast cancer. The adopted markers are the combination use of CEA
propeptide of Type III procollagen (PIIINP) was and a MUC1 mucin (often detected as CA15.3), other
reported to be a promising marker reflecting treat- markers (e.g., TPS, c-erbB2) and assays (e.g.,
ment response in 1980s (91). Related new novel CA27.29) of potential value are being evaluated to
markers were further investigated. The most abun- optimise their use in advanced disease. Furthermore,
dant protein in bone is type I collagen. During its for- research studies are ongoing to exploit the use of
mation two extension peptides from the procollagen blood tumour markers in the diagnosis and screening
molecule, carboxy- and aminoterminal propeptides of early primary breast cancer.
(PICP and PINP) are released into the circulation and
they are markers of bone formation. Type I collagen
carboxyterminal telopeptide (ICTP) is formed during REFERENCES
bone collagen breakdown and is again liberated into
the circulation. Its level in the serum reflects bone 1. Statland BE, Winkel P. Neoplasia. In: Kaplan LA, Pesc AJ, eds.
Clinical chemistry: theory, analysis, and correlation: 2nd edi-
resorption and/or bone turnover. The same group in
tion. St Louis: CV Mosby, 1989; 734–5.
Finland has found a significant correlation of ICTP 2. Sikorska H, Shuster J, Gold P. Clinical applications of carci-
levels with therapeutic response in 36 patients with noembryonic antigen. Cancer Detect Prevent 1988; 12: 321–355.
bone metastases due to breast cancer (92). Almost 3. Lahousen M, Stettner H, Pickel H, Urdl W, Purstner P. The
identical findings to support the distinct role of ICTP predictive value of a combination of tumour markers in mon-
itoring patients with ovarian cancer. Cancer 1987; 60:
were reported by other centres (93–96). Although
2228–2232.
ICTP level seemed to correlate well with response to 4. Go VLW, Zamcheck N. The role of tumor markers in the man-
systemic therapy, it was criticized to be not very use- agement of colorectal cancer. Cancer 1982; 50: 2618–2623.
ful in the diagnosis of bone metastases. Despite hav- 5. McNeely MDD. Gastrointestinal function and digestive dis-
ing a high specificity over 90% and being the best ease. In: Kaplan LA, Pesc AJ, eds. Clinical chemistry: theory,
analysis, and correlation: 2nd edition. St Louis: CV Mosby,
bone metabolism marker evaluated (93, 94), it has a
1989; 411–412.
relatively low sensitivity (e.g., 56% in Tahtela’s series 6. Fletcher RH. Carcinoembryonic antigen. Ann Int Med 1986;
(94). Serum concentrations of ICTP and some other 104: 66–73.
markers have also been recently shown to correlate 7. Mughal AW, Hortogagyi GN, Fritsche HA, Buzdar AU, Yap
with the disease extent in the bone as well as the clin- HY, Blumenschein GR. Serial plasma carcinoembryonic anti-
gen measurements during treatment of metastatic breast can-
ical symptom (i.e., bone pain) (97). Further studies to
cer. JAMA 1983; 249: 1881–1886.
evaluate the cost-effectiveness of measuring these 8. Williams MR, Turkes A, Pearson D, Twining P, Griffiths K,
markers and to explore newer markers of bone Blamey RW. The use of serum carcinoembryonic antigen to
metabolism are required. In the present era when the assess therapeutic response in locally advanced and metasta-
use of bisphosphonates has been popularised in the tic breast cancer: a prospective study with external review. Eur
J Surg Oncol 1988; 14: 417–422.
management of bone metastases for breast cancer,
9. Robertson JFR, Pearson D, Price MR, Selby C, Pearson J,
markers of bone metabolism might provide a meas- Blamey RW, Howell A. Prospective assessment of the role of
urement of the effect of sclerosis on the bone while five tumour markers in breast cancer. Cancer Immunol
conventional blood markers such as CA15.3 and Immunother 1991; 33: 403–410.
100 K.L. CHEUNG ET AL.

10. Lokich JJ, Zamcheck N, Lowenstein M. Sequential carcinoem- 29. Garcia MB, Blankenstein MA, van der Wall E, Nortier JWR,
bryonic antigen levels in the therapy of metastatic breast can- Schornagel JH, Thijssen JHH. Comparison of breast cancer
cer. Ann Int Med 1978; 89: 902–906. mucin (BCM) and CA15.3 in human breast cancer. Breast
11. Haagensen DE, Kister SJ, Vandevoorde JP, Gates JB, Smart EK, Cancer Res Treat 1990; 17: 69–76.
Hansen HJ, Wells SA. Evaluation of carcinoembryonic antigen 30. Deprés-Brummer P, Itzhaki M, Bakker PJM, Hoek FJ, Veenhof
as a plasma monitor for human breast carcinoma. Cancer 1978; KHN, de Wit R. The usefulness of CA15.3, mucin-like carci-
42: 1512–1519. noma-associated antigen (MCA) and carcinoembryonic antigen
12. Doyle PJ, Nicholson RI, Groome GV, Blamey RW. in determining the clinical course in patients with metastatic
Carcinoembryonic antigen (CEA): its role as a tumour marker breast cancer. J Cancer Res Clin Oncol 1995; 121: 419–422.
in breast cancer. Clin Oncol 1981; 7: 53–58. 31. Price MR. High molecular weight epithelial mucins as markers
13. Lee YN. Serial tests of carcinoembryonic antigen in patients in breast cancer. Eur J Cancer Clin Oncol 1988; 24: 1799–1804.
with breast cancer. Am J Clin Oncol 1983; 6: 287–293. 32. Robertson JFR. The role of serum tumour markers for moni-
14. Koldovsky U, Wargalla U, Hilkens J, Taylor-Papadimitrou J, toring therapy in metastatic breast cancer. Ligand Assay 1996;
Schnuerch HG, Hilgers J. Reactions of monoclonal antibodies 1: 257–262.
against membranes of human milk fat globules with uterine 33. Robertson JFR, Jaeger W, Syzmendera JJ, Selby C, Coleman R,
and cervical epithelia during proliferative and secretory Howell A, Winstanley J, Jonssen PE, Bombardieri E, Sainsbury
stages. Prot Biol Fluids 1984; 32: 239–242. JRC, Gronberg H, Kumpulainen E, Blamey RW. The objective
15. Hayes D, Sekine H, Ohao T, Abe M, Keefe K, Kufe DW. Use of measurement of remission and progression in metastatic
murine monoclonal antibody for detection of circulating breast cancer by use of serum tumour markers. Eur J Cancer
plasma DF3 antigen levels in breast cancer patients. J Clin 1999; 35: 47–53.
Invest 1985; 75: 1671–1678. 34. Robertson JFR, Pearson D, Price MR, Selby C, Blamey RW,
16. Hayes DF, Zurawski VR, Kufe DW. Comparison of circulating Howell A. Objective measurement of therapeutic response in
CA15.3 and carcinoembryonic antigen levels in patients with breast cancer using tumour markers. Br J Cancer 1991; 64:
breast cancer. J Clin Oncol 1986; 10: 1542–1550. 757–763.
17. Pons-Anicet DMF, Krebs BP, Namer M. Value of CA15.3 in the 35. Dixon AR. Tumour markers – a logical approach to the guid-
follow-up of breast cancer patients. Br J Surg 1987; 55: 567–569. ance of therapy in advanced breast cancer? Doctor of
18. Todini C, Hayes DF, Gelman R, Henderson IC, Kufe DW. Medicine Thesis, University of Nottingham 1991.
Comparison of CA15.3 and carcinoembryonic antigen in mon- 36. Dixon AR, Jönrup I, Jackson L, Chan SY, Badley RA, Blamey
itoring of the clinical course of patients with metastatic breast RW. Serological monitoring of advanced breast cancer treated
cancer. Cancer Res 1988; 48: 4107–4112. by systemic cytotoxic using a combination of ESR, CEA, and
19. Robertson JFR, Pearson D, Price MR, Selby C, Badley RA, CA15.3: fact or fiction? Disease Markers 1991; 9: 167–174.
Pearson J, Blamey RW, Howell A. Assessment of four mono- 37. Dixon AR, Jackson L, Chan SY, Badley RA, Blamey RW.
clonal antibodies as serum markers in breast cancer. Eur J Continuous chemotherapy in responsive metastatic breast
Cancer 1990; 26: 1127–1132. cancer: a role for tumour markers? Br J Cancer 1993; 68:
20. Colomer R, Ruibal A, Salvador L. Circulating tumour marker 181–185.
levels in advanced breast carcinoma correlate with the extent 38. Williams MR, Turkes A, Pearson D, Griffiths K, Blamey RW.
of metastatic disease. Cancer 1989; 64: 1674–1681. An objective biochemical assessment of therapeutic response
21. Colomer R, Ruibal A, Genolla J, Rubio D, Del Campo JM, Bodi in metastatic breast cancer: a study with external review of
R, Salvador L. Circulating CA15.3 levels in the post surgical clinical data. Br J Cancer 1990; 61: 126–132.
follow-up of breast cancer patients and in non-malignant dis- 39. Coombes RC, Powles TJ, Gazet JC, Ford HT, McKinna A,
eases. Breast Cancer Res Treat 1989; 13: 123–133. Abbott Z, Gehrke CW, Keyser JW, Mitchell PEG, Patel S,
22. Molina R, Zanon G, Filella X, Moreno F, Jo J, Daniels M, Latre Stimson WH, Worwood M, Jones M, Neville AM. Screening
ML, Gimenez N, Pahisa J, Velasco M, Ballesta AM. Use of for metastases in breast cancer: an assessment of biochemical
serial carcinoembryonic antigen and CA15.3 assays in detect- and physical methods. Cancer 1981; 48: 310–315.
ing relapses in breast cancer patients. Breast Cancer Res Treat 40. Lamoreux G, Mandeville R, Poisson R, Legault-Poisson S,
1995; 36: 41–48. Jolicoeur R. Biologic markers in breast cancer: a multipara-
23. Van Dalen A. New markers for breast carcinoma-associated metric study. 1. Increased serum protein levels. Cancer 1982;
antigen in comparison with CA15.3. Anticancer Res 1996; 16: 49: 502–512.
2339–2343. 41. Gion M. Serum markers in breast cancer management. The
24. Reddish MA, Helbrecht N, Almeida AF, Madiyalakan R, Suresh Breast 1992; 1: 173–178.
MR, Longenecker BM. Epitope mapping of Mab B27.29 within 42. Cantwell B, Fennelly JJ, Jones M. Evaluation of follow-up
the protein core of the malignant breast carcinoma-associated methods to detect relapse mastectomy in breast cancer
mucin antigen, MUC1. J Tumor Marker Oncol 1992; 7: 19–27. patients. Ir J Med Sc 1982; 151: 1–5.
25. Graves R, Hilgers J, Fritsche H, Hayes D, Robertson JFR. 43. Safi F, Kohler I, Röttinger E, Beger HG. The value of the
MUC1 mucin assays for monitoring therapy in metastatic tumour marker CA15.3 in diagnosing and monitoring breast
breast cancer. The Breast 1998; 7: 181–186. cancer. Cancer 1991; 68: 574–582.
26. Hayes DF, Kiang DT, Korzun A, Tondini C, Wood W, Kufe D. 44. Al-Jarallah MA, Behbehani AE, El-Nass SA, Temim L,
CA15.3 and CEA spikes during chemotherapy for metastatic Ebraheem AK, Ali MAA, Szymendera JJ. Serum CA-15.3 and
breast cancer. Proc Am Soc Clin Oncol 1988; 7: 38a. CEA patters in postsurgical follow-up, and in monitoring clin-
27. Yasasever V, Camlica H, Karaloglu D, Dalay N. Utility of ical course of metastatic cancer in patients with breast carci-
CA15.3 and CEA in monitoring breast cancer patients with noma. Eur J Surg Oncol 1993; 19: 74–79.
bone metastases: special emphasis on “spiking” phenonomen. 45. Jäger W, Krämer S, Palapelas V, Norbert L. Breast cancer and
Clin Biochem 1997; 30: 53–56. clinical utility of CA15.3 and CEA. Scand J Clin Lab Invest 1995;
28. Martoni A, Zamagni C, Bellanova B, Zanichelli L, Vecchi F, 55: 87–92.
Cacciari N, Strocchi E, Pannuti F. CEA, MCA, CA15.3 and 46. Iwase H, Kobayashi S, Itoh Y, Fukuoka H, Kuzushima T, Iwata
CA549 and their combinations in expressing and monitoring H, Yamashita T, Naitoh A, Itoh K, Masaoka A. Evaluation of
metastatic breast cancer: a prospective comparative study. Eur serum tumor markers in patients with advanced or recurrent
J Cancer 1995; 31A: 1615–1621. breast cancer. Breast Cancer Res Treat 1994; 33: 83–88.
TUMOUR MARKERS IN BREAST C ANCER 101

47. Kleist SV, Bombardieri E, Buraggi G, Gion M, Hertel A, Hör G, 66. American Society of Clinical Oncology. Clinical practice
Noujaim a, Schwartz M, Senekowitsch R, Wittekind C. guidelines for the use of tumour markers in breast and col-
Immunodiagnosis of tumours. Eur J Cancer 1993; 29A: orectal cancer. J Clin Oncol 1996; 14: 2843–2877.
1622–1630. 67. Whitlock JPL, Evans AJ, Jackson L, Chan SY, Robertson JFR.
48. Gion M, Mione R, Nascimben O, Valsecchi M, Gatti C, Leon A, Imaging of metastatic breast cancer: distribution and radio-
Bruscagnin G. The tumour associated antigen CA15.3 in pri- logical assessment of involved sites at presentation [In
mary breast cancer. Evaluation of 667 cases. Br J Cancer 1991; preparation].
63: 809–813. 68. Robertson JFR, Whynes DK, Dixon A, Blamey RW. Potential
49. O’Hanlon DM, Kerin MJ, Kent PJ, Skehill R, Maher D, Grimes for cost economics in guiding therapy in patients with
H, Given HF. A prospective evaluation of CA15.3 in stage I metastatic breast cancer. Br J Cancer 1995; 72: 174–177.
carcinoma of the breast. J Am Coll Surg 1995; 180: 210–212. 69. Rosselli Del Turco M, Palli D, Cariddi A, Ciatto S, Pacini P,
50. Robertson JFR. Blood tumour markers in breast cancer. Distante V. Intensive diagnostic follow-up after treatment of
Tumour Marker Update 1998; 10: 31–37. primary breast cancer: a randomized trial. JAMA 1994; 271:
51. Geraghty JG, Coveney EC, Sherry F, O’Higgins NJ, Duffy MJ. 1593–1597.
CA15.3 in patients with locoregional and metastatic breast 70. The GIVIO Investigators. Impact of follow-up testing on sur-
carcinoma. Cancer 1992; 70: 2831–2834. vival and health-related quality of life in breast cancer
52. Coveney EC, Geraghty JG, Sherry F, McDermott EW, Fennelly patients: a multicenter randomized controlled trial. JAMA
JJ, O’Higgins NJ, Duffy MJ. The clinical value of CEA and 1994; 271: 1587–1592.
CA15.3 in breast cancer management. Int J Biol Markers 1995; 71. Palli D, Russo A, Saieva C, Ciatto S, Rosselli Del Turco M,
10: 35–41. Distante V, Pacini P. Intensive vs clinical follow-up after treat-
53. O’Hanlon DM, Kerin MJ, Kent P, Maher D, Grimes H, Given ment of primary breast cancer: 10-year update of a random-
HF. An evaluation of preoperative CA15.3 measurement in ized trial. JAMA 1999; 281: 1586.
primary breast carcinoma. Br J Cancer 1995; 71: 1288–1291. 72. Early Breast Cancer Trialists’ Collaborative Group. Systemic
54. Tomlinson IPM, Whyman A, Barrett JA, Kremer JK. Tumour treatment of early breast cancer by hormonal, cytotoxic and
marker CA15-3: possible uses in the routine management of immune therapy. Lancet 1992; 1: 1–15, 71–85.
breast cancer. Eur J Cancer 1995; 31A: 899–902. 73. Molina R, Zanón G, Filella X, Moreno F, Jo J, Daniels M, Latre
55. O’Brien DP, Skehill R, Horgan PG, Grimes H, Gough DB, Given ML, Giménez N, Pahisa J, Velasco M, Ballesta AM. Use of
HF. CA15.3: a reliable indicator of metastatic bone disease in serial carcinoembryonic antigen and CA15.3 assays in detect-
breast cancer patients. Ann R Coll Surg Engl 1992; 74: 9–12. ing relapses in breast cancer patients. Breast Cancer Res Treat
56. Buffaz PD, Gauchez AS, Caravel JP, Vuillez JP, Cura C, 1995; 36: 41–48.
Agnius-Delord C, Fagret D. Can tumour marker assays be a 74. Molina R, Jo J, Zanon G, Filella X, Farrus B, Muñoz M, Latre
guide in the prescription of bone scan for breast and lung can- ML, Pahisa J, Velasco M, Fernandez P, Estapé J, Ballesta AM.
cers? Eur J Nu Med 1999; 26: 8–11. Utility of c-erbB2 in tissue and in serum in the early diagnosis
57. Breast Specialty Group of the British Association of Surgical of recurrence in breast cancer patients: comparison with carci-
Oncology. The guidelines for the management of metastatic noembryonic antigen and CA15.3. Br J Cancer 1996; 74:
bone disease in breast cancer in the United Kingdom. Eur J 1126–1131.
Surg Oncol 1999; 25: 3–23. 75. Molina R, Gion M. Use of blood tumour markers in the detec-
58. Krieger G, Kehl A, Wander HE, Salo AM, Rauschecker HF, tion of recurrent breast cancer. The Breast 1998; 7: 187–109.
Nagel GA. Clinical significance of circulating immune com- 76. Jäger W, Merkle E, Lang N. Increasing serum tumour markers
plexes (CIC) in patients with metastatic breast cancer. Int J as decision criteria for hormone-therapy of metastatic breast
Cancer 1983; 31: 207–211. cancer. Tumor Biol 1994; 12: 60–66.
59. Van Dalen A, Heering KJ, Barak V, Peretz T, Cremaschi A, 77. Nicolini A, Anselmi L, Michelassi C, Carpi A. Prolonged sur-
Geroni P, Gion M, Saracchini S, Molina R, Namer M, Stieber P, vival by “early” salvage treatment of breast cancer patients: a
Sturgeon C, Leonard RCF, Einarsson R. Treatment response in retrospective 6-year study. Br J Cancer 1997; 76: 1106–1111.
metastatic breast cancer: a multicentre study comparing UICC 78. Murray A, Willsher P, Price MR, Dixon AR, Robertson JFR.
criteria and tumour marker changes. The Breast 1996; 5: 82–88. Evaluation of the IMMULITE BR-MA and CEA assays and
60. Robertson JFR. Prospective confirmation of a biochemical comparison with immunoradiometric assays for CA15.3 and
index for measuring therapeutic efficacy in metastatic breast CEA in breast cancer. Anticancer Res 1997; 17: 1945–1949.
cancer in a multicentre study. The Breast 1996; 5: 372–373. 79. Nicolini A, Colombini C, Luciani L, Carpi A, Giuliani L.
61. Wojtacki J, Dziewulska-Bokiniec A, Rolka-Stempniewicz G. Evaluation of serum CA15.3 determination with CEA and
Usefulness of CA15–3 antigen determination for evaluation of TPA in the postoperative follow-up of breast cancer patients.
response to second-line chemotherapy in patients with breast Br J Cancer 1991; 64: 154–158.
cancer: preliminary study. Polskie Archiwum Medycyny 80. Sonoo H, Kuerbayashi J. Serum tumour marker kinetics and
Wewnetrznej 1997; 97: 343–350. the clinical course of patients with advanced breast cancer.
62. Willsher PC, Beaver J, Blamey RW, Robertson JFR. Serum tis- Surg Today 1996; 26: 250–257.
sue polypeptide specific antigen (TPS) in breast cancer 81. Van Dalen A. TPS in breast cancer – a comparative study with
patients: comparison with CA15.3 and CEA. Anticancer Res carcinoembryonic antigen and CA15.3. Tumour Biol 1992; 13:
1995; 15: 1609–1611. 10–17.
63. Cheung KL, Sadozye AH, Chan SY, Blamey RW, Evans AJ, 82. Einarsson R. TPS – a cytokeratin marker for therapy control in
Robertson JFR. The role of blood tumour markers in directing breast cancer. Scan J Clin Lab Invest 1995; 55: 113–115.
cytotoxic therapy in advanced breast cancer [In preparation]. 83. Van Dalen A. Is TPS able to assist in decision making in breast
64. Hayward JL, Carbone PP, Heuson JC, Kumaoka S, Segaloff A, cancer? In: Tumour associated antigens, oncogenes, receptors,
Rubens RD. Assessment of response to therapy in advanced cytokines in tumour diagnosis and therapy at the beginning
breast cancer. Cancer 1977; 39: 1289–1293. of the nineties. 6th Symposium on Tumour Markers,
65. Brown AL, Middleton G, MacVicar AD, Husband JES. T1 Hamburg, 1991.
weighted magnetic resonance imaging in breast cancer verte- 84. Harris L, Lüftner D, Jager W, Robertson JFR. C-erbB2 in
bral metastases: changes on treatment and correlation with serum of patients with breast cancer. Int J Biol Markers 1999;
response to therapy. Clin Radiol 1998; 53: 493–501. 14: 8–15.
102 K.L. CHEUNG ET AL.

85. Tagliabue E, Ménard S, Robertson JFR, Harris L. C-erbB2 monitoring of treatment response in bone metastases from
expression in primary breast cancer. Int J Biol Markers 1999; 14: breast carcinoma. Br J Cancer 1996; 73: 1074–1079.
16–26. 93. Plebani M, Bernardi D, Zaninotto M, De Paoli M, Secchiero S,
86. Lüftner D, Schinabel S, Possinger K. C-erbB2 in serum of Sciacovelli L. New and traditional serum markers of bone
patients receiving fractionated paclitaxel chemotherapy. Int J metabolism in the detection of skeletal metastases. Clin
Biol Markers 1999; 14: 55–59. Biochem 1996; 29: 67–72.
87. Weidner N, Semple JP, Welch WR, Folkman J. Tumour angio- 94. Tahtela R, Tholix E. Serum concentrations of type I collagen
genesis and metastases – correlation in invasive breast carci- carboxyterminal telopeptide (ICTP) and type I procollagen
noma. N Engl J Med 1991; 324: 1–8. carboxy- and aminoterminal propeptides (PICP, PINP) as
88. Weidner N, Folkman J, Pozza F, Bevilacqua P, Allred EN, markers of metastatic bone disease in breast cancer. Anticancer
Moore DH, Meli S, Gasparini G. Tumour angiogenesis: a new Res 1996; 16: 2289–2293.
significant and independent prognostic indicator in early- 95. Yamamoto I, Miura H, Kigami Y, Yamamura Y, Ohta T, Yuu I,
stage breast carcinoma. J Natl Cancer Inst 1992; 84: 1875–1887. Yamamoto Y, Morita R. Detection of bone metastases by serial
89. Magennis DP. Angiogenesis: a new prognostic marker for measurement of C-terminal telopeptide of type I collagen in
breast cancer. Br J Biomed Sc 1998; 55: 214–220. patients with malignancy. Kakuigaku 1996; 33: 423–429.
90. Wu J. Apoptosis and angiogenesis: two promising tumor 96. Shimozuma K, Sonoo H, Fukunaga M, Ichihara K, Aoyama T,
markers in breast cancer. Anticancer Res 1996; 16: 2233–2239. Tanaka K. Biochemical markers of bone turnover in breast
91. Blomqvist C, Elomaa I, Virkkunen P, Porkka L, Karonen SL, cancer patients with bone metastases: a preliminary report.
Risteli, Risteli J. The response evaluation of bone metastases in Jpn J Clin Oncol 1999; 29: 16–22.
mammary carcinoma. The value of radiology, scintigraphy 97. Berruti A, Dogliotti L, Gorzegno G, Torta M, Tampellini M,
and biochemical markers of bone metabolism. Cancer 1987; 60: Tucci M, Cerutti S, Frezet MM, Stivanello M, Sacchetto G,
2907–2912. Angeli A. Differential patterns of bone turnover in relation to
92. Blomqvist C, Risteli L, Risteli J, Virkkunen P, Sarna S, Elomaa bone pain and disease extent in bone in cancer patients with
I. Markers of type I collagen degradation and synthesis in the skeletal metastases. Clin Chem 1999; 45: 1240–1247.