Tumor Markers in Cancer Care: Overview
Tumor Markers in Cancer Care: Overview
3. Biochemical indicators of the presence of a tumor.[7] Immunological detection usually relies on monoclonal
However, in common clinical practice, the term usually antibodies that specifically bind to epitopes on tumor
refers to a molecule that can be detected in plasma or markers and are in turn tagged for identification with
other body fluids. dyes in immunohistochemistry (IHC), radioactive tags in
radio-immuno assay (RIA), or enzymes in enzyme-linked
Ideal tumor marker immunosorbent assay (ELISA).[2,5,6,10,11] Alternatively, in a
suspension, flow cytometry can analyze the presence and
Only a few tumor markers have stood the test of time and percentage of antibody-tagged cells.[9-11] These methods are
entered in the diagnostic or management algorithms for highly sensitive and can detect quantities in the nanogram
clinicians.[1,2,5]
The three most important characteristics of an ideal tumor Table 2: Molecular basis of tumor markers[8]
marker are (a) it should be highly specific to a given Levels of classification Examples
tumor type, (b) it should provide a lead-time over clinical DNA
diagnosis and (c) it should be highly sensitive to avoid Epigenetic Promoter Hyper-methylation, e.g.,
false positive results. Additionally, the levels of the marker GSP1, DAP in lung cancer; p15, p16 in
should correlate reliably with the tumor burden, accurately liver cancer
Endogenous Mutations, e.g., NADH dehydrogenase 4
reflecting any tumor progression or regression, along with Mitochondrial genetic (ND4) in urine in bladder cancer
a short half-life allowing frequent serial measurements. Oncogene Mutation, e.g., K-ras in pancreatic
The test used for detection should be cheap for screening cancer; micro-satellite alterations in
application at mass level and should be of such nature as to head and neck cancers
be acceptable to the target population [Table 1]. In reality Exogenous viral EBV in NPC, Burkitt’s lymphoma; HPV in
cervical cancer
an ideal tumor marker does not exist.
RNA
Cell based Tissue-specific markers, e.g., PSA mRNA
Molecular basis of tumor markers endogenous in prostate cancer, cytokeratin 20 mRNA
in breast cancer
Cell free Circulating mRNA, e.g., Tyrosinase
Genetic alteration in a tumor cell affects directly or
mRNA in melanoma
indirectly the gene expression pattern of the tumor cell Exogenous viral Viral RNA, e.g., EBV-coded RNA in NPC
or the surrounding tissue.[7] These genetic alterations can Translational protein
be reflected at various levels [Table 2], from viral genomic Native protein PSA in prostate cancer, CEA in colonic
incorporation to genetic defects, forming the molecular (Conventional cancer
basis of tumor markers.[8] markers)
Glycan Aberrant glycosylation, e.g.,
monosialytactec AFP in HCC
Methods of detection
AFP: Alfa fetoprotein; CEA: Carcinoembryonic antigen; EBV: Epstein-Barr virus;
HCC: Hepatocellular carcinoma; HPV: Human papilloma virus; mRNA: Messenger
The methods of detection can be classified into 6 major RNA; NPC: Nasopharyngeal carcinoma; PSA: Prostate-specific antigen
groups [Table 3].[2,5,6,9,10] The most common method in use
today is serological enzyme assays. Table 3: Methods of detection of tumor
markers[2,5,6,9,10]
Serology Enzyme assays
Table 1: Characteristics of an ideal tumor
Immunological Immuno histo chemistry
marker[5]
Radio immuno assay
Characteristics Remarks
Enzyme-linked immuno sorbent assay
Highly specific Detectable only in one tumor type
Flow cytometry
Highly sensitive Non-detectable in physiological or
Cytogenetic analysis Fluorescent in-situ hybridization
benign disease states
Spectral karyotyping
Long lead-time Sufficient time for alteration of
natural course of disease Comparative genomic hybridization
Levels correlate with tumor Prognostic and predictive utility of Genetic analysis Sequencing (automated)
burden the tumor marker Reverse transcription
Short half-life Frequent serial monitoring of the Gel electrophoresis
marker levels after 5-6 half lives DNA micro-array analysis
Simple and cheap test Applicability as screening test Proteomics Surface-enhanced laser desorption/
Easily obtainable specimens Acceptability by target population Ionization
to picogram range (1026 to 1029 g). Of these, the most Table 4: Classification scheme for tumor
commonly used technique today is IHC. Uses of IHC markers[2]
in oncology include categorization of undifferentiated Category Subcategory Examples
malignant tumors, categorization of leukemias and Oncofetal antigens AFP, CEA
lymphomas, determination of site of origin of metastatic Hormones Catecholamines,
tumors and detection of molecules of prognostic or Calcitonin, b-hCG
therapeutic significance (e.g., Estrogen/progesterone Glycoproteins CA 125, CA 15-3, CA
receptors (ER/PR) in breast cancer). [2,5- 7,10] 19-9, CA 72-4, PSA
Metabolites VMA, HIAA
Tumor-associated Viral antigens Polyoma, SV 40
Classification and uses antigens
MHC-related H-2 k antigen
Tumor markers can be detected either in tissue (tissue antigens
Enzymes PAP, NSE, PLAP
tumor markers; for example, in solid tumors, lymph nodes,
Oncogene products c-myc, c-erbB2
bone marrow or circulating tumor cells in the blood) or in
Cytogenetic Philadelphia
body fluids like ascitic or pleural fluid or serum (serological products chromosome
tumor markers).[9,10] Tissue tumor markers are of prime Tumor-associated Proteins Immunoglobulins,
importance to a diagnostic pathologist, while the serological markers b-2M
tumor markers are more often used by a clinician and will Enzymes Lactate
be discussed in more detail in this review. dehydrogenase,
alkaline phosphatase,
pteridines, pterines
Many classification schemes exist based on differences in Acute-phase C-reactive protein,
origin, structure, biological function or their relationship proteins ferritin
to the event in tumor growth or formation [Table 4].[1,11] Inflammatory ESR, viscosity
makers
Clinical application of tumor markers can be broadly Ultrastructural Intermediate Desmin, vimentin
components filament
classified into 4 groups: Screening and early detection, components
diagnostic confirmation, prognosis and prediction AFP: Alfa fetoprotein; CEA: Carcinoembryonic antigen; ESR: Erythrocyte
of therapeutic response and monitoring disease and sedimentation rate; HIAA: Hydroxy indole acetic acid; NSE: Neuron-specific enolase;
recurrence.[5,7,9,10] Some of the recommended uses of PAP: Prostatic acid phosphatase; PLAP: Placental alkaline phosphatase; PSA:
Prostate-specific antigen; SV: Simian virus; VMA: Vanillmandelic acid
tumor markers in routine clinical practice are summarized
in Table 5. Canadian Task Force,[29] American Association for Clinical
Chemistry,[1] etc., but the only tumor marker that finds a place
Although extremely appealing, the concept of screening a in any screening algorithm is prostate-specific antigen (PSA).
large apparently healthy population for occult tumors using
a tumor marker and thereby enabling early therapeutic Attempts to improve the sensitivity and/ or specificity
intervention is not currently possible because no test yet of tumor markers have led to combination of tumor
devised is 100 percent specific and many tumor markers markers with other procedures (e.g., combination of
may be elevated in benign conditions [Table 6].[2,7,10,11] Carbohydrate antigen (CA) 125 with ultrasonography for
early detection of ovarian malignancy) or to refining the
Importance and use of a particular tumor marker may evaluation criteria for tumor markers (e.g., PSA density
change depending upon the clinical scenario, ranging or PSA velocity or age-specific PSA cut off ranges for
from initial presentation to differential diagnosis to early detection of prostate cancer). However, these have
recurrence. [2] Thus, for example, while tissue tumor markers either not stood the rigorous evaluation of randomized
like Cytokeratin, smooth muscle antigen (SMA), etc., may trials or have still not received widespread approval of
be extremely useful in categorization of malignancy, they professional clinical organizations.[1,12-14]
are of no use in prognosis or monitoring; on the other
hand, markers like Ki-67 (proliferation index) may help in In addition to variable sensitivity and specificity of tumor
prognostication or choice of therapy but have no role in markers, the prevalence of a particular malignancy may be a
the diagnostic arena. major determinant in the application of a tumor marker as a
screening tool. A stark example of this effect is seen in China,
A large body of literatures exists on guidelines suggested for where serum alfa fetoprotein (AFP) has been successfully
clinical application in various malignancies by professional used as a screening tool for primary hepatoma in endemic
bodies like American Society of Clinical Oncology,[12,13] regions in contrast to its failure in rest of the world.[15]
clinical or radiological evidence of disease is apparent Tumor marker kinetics should always be factored before
(“biochemical recurrence”). repeating the tests. Too frequent estimation of the tumor
marker may misrepresent the course of the disease due to
Determination of risk usually involves genetic probes distribution and elimination kinetics. As a general guideline,
or tools to evaluate any specific genetic abnormality or the time interval between serial determinations should
mutation noted to confer an increased risk of a particular be 3 months; but in case of an abnormal value, a repeat
malignancy. Examples of such abnormalities would include estimate can be ordered within 2 to 4 weeks irrespective
carriers of Philadelphia chromosome for hematological of the initial reading. The success of surgical removal of
malignancies; and BRCA 1 or 2 genes, which confer a a tumor as determined by tumor marker concentrations is
higher risk of breast or ovarian malignancies.[22] ideally ascertained after a period not less than 5-6 half-lives,
to allow tumor marker levels to make a plateau or fall to
Recommendations for ordering tumor marker normal. This period may be even longer in case of treatment
with chemotherapy or radiotherapy, wherein the therapeutic
tests
effects themselves are manifested after a lag period.[20,21]
It is imperative to remember that though an aggressive
Patient characteristics affect the tumor marker values to
investigative approach may be warranted on the basis of raised
a significant degree. The anticipated fall in levels may not
tumor marker values, treatment cannot be initiated without
be evident in situations where the metabolism or excretion
undisputable documentation (often histological) of the
of the tumor marker is altered, like in patients with renal
disease.[23] There may be instances such as initiation of therapy
or liver disease, depending on whether the tumor marker
on the basis of b-hCG levels in cases of choriocarcinoma,
is removed through glomerular filtration or metabolized
but these are few, exceptional and clearly defined.
by the liver. For example, serum CEA is often elevated
A single value or test is unreliable in itself.[6] It is noteworthy in patients with liver diseases because the metabolism of
that in most situations, elevations of markers in nonmalignant CEA by the diseased liver is subnormal. False tumor marker
elevation is also known to occur in other confounding
diseases are often transient, whereas elevations associated
situations like smoking, ethanol consumption, COPD,
with cancer either remain constant or continuously rise.
etc., especially if there has been a recent change in habits.[5]
Ordering serial testing can help detect falsely elevated
levels due to transient elevation. Establishing a diagnosis
Usually multiple tumor markers are associated with
on the basis of tumor markers (especially a single result) individual malignancy [Table 7]; vice versa, individual
is fraught with associated pitfalls because of the problem tumor markers may be associated with various malignancies
of nonspecificity.[24-31] [Table 8].[28] Thus the use of multiple markers based on
the combination pattern for the selected malignancy will
Knowledge of the assay method is important in improve sensitivity and specificity of the detection.[32,33]
interpretation of either an abnormal value or a serial change However, tumor markers that run parallel to each other,
in tumor marker values.[20,29,30] Various methods of detection when correlated with tumor behavior, should not be
have their own specific cut off values and sensitivities.[20,21] selected for this purpose.[32-34]
Thus, for any set of serial values to be meaningful, they
have to come from the same assay methods and preferably Nonspecific tumor markers are a good choice for
from the same laboratory. In certain situations of so-called monitoring disease activity. Although nonspecific tumor
biochemical recurrences, it is always useful to go back to markers, by definition, have a poor sensitivity, nevertheless
the laboratory to confirm this before beginning (at times) their concentrations are sensitive to any alterations in tumor
a frustrating search for the elusive recurrence. volume. They are usually inexpensive and simple to measure.
For example, lipid-associated sialic acid P (LASA-P) can be
It is imperative to be certain that the marker in question quantified with a simple, rapid and inexpensive calorimetric
was, in fact, elevated before relying on it for monitoring procedure, and its serum concentration is closely parallel
disease activity, the reason being that none of the tumor to the serum concentrations of many tumor markers of
markers are 100% sensitive (may not be elevated in some higher specificity in various malignancies.[34]
cases).[2,5,7,10] In tumors with multiple raised markers
measured prior to definitive therapy, the marker showing An important interfering factor to be considered before
highest elevation should be used for follow-up.[7,10] If in any interpretation is presence of a hook effect.[30] This
a given case, tumor markers were not evaluated in the is especially true if the value of a tumor marker does
pretreatment setting, it is advisable to use multiple markers not correlate to the clinical situation. Hook effect
for monitoring in the post-therapy setup.[5,7,10] is an inherent flaw of certain methods of detection
Table 8: Selected examples of serologic tumor markers and malignant diseases associated with
each[27]
Tumor marker Associated malignancy
Primary Other malignancies
Oncofetal antigens
AFP Primary HCC Teratoblastomas of the ovary and testes
CEA Colorectal carcinoma Various carcinomas
Hormones
b-hCG Choriocarcinoma Testicular cancers (non-seminomatous), trophoblastic tumors
Calcitonin Medullary carcinoma Cancer of the thyroid, liver cancer, renal cancer
Metanephrines Pheochromocytoma Neuroblastoma, ganglioneuromas
Chromogranin A Pheochromocytoma, neuroblastoma MEN, small-cell lung cancer, carcinoid tumors
IGF- 1 Pituitary cancer Insulinoma
Glycoproteins
CA 15-3 Breast cancer Various carcinomas
CA 19-9 Pancreatic and gastric carcinomas Various carcinomas
CA 72-4 Gastric carcinoma Various carcinomas
CA 125 Ovarian carcinoma Various carcinomas
Isoenzymes
PSA Prostate cancer
NSE Small-cell lung carcinoma Neuroblastoma, kidney tumors
Cellular components/products
LASA-P Various carcinomas, leukemia, lymphoma, Hodgkin’s disease
SCC-A Squamous cell carcinoma of the uterus, cervix, lung, and head and neck
TAG 72 Gastric carcinoma Colorectal, lung, pancreatic and ovarian cancers
Immunoglobulins Multiple myeloma Gammopathies
AFP: Alfa fetoprotein; b-hCG: Beta human chorionic gonadotropin; CA: Carbohydrate antigen; CEA: Carcinoembryonic antigen; HCC: Hepatocellular carcinoma; LASA-P: Lipid-
associated sialic acid P; MEN: Multiple endocrine neoplasia; NSE: Neuron-specific enolase; PSA: Prostate-specific antigen; SCC-A: Squamous cell carcinoma antigen
machinery, and the cell becomes autonomous. Autonomic AFP: Alfa fetoprotein; b2M: Beta 2 microglobulin; WD: Well differentiated
expression of various unrelated genes (for the given tumor
tissue type) is the reason for the expression of ectopic Physiological influences that need to be considered in
tumor markers in advanced malignant diseases [Table 9]. interpreting the results include effects of aging and
Ectopic tumor markers denote dedifferentiation by menopause; metabolism and route-of-elimination kinetics
indicating activation of unrelated genes and, in other words, of the tumor marker; coexisting disease, like renal or
are associated with poorer prognosis or metastases. For liver failure; hormonal imbalances, like hyperthyroidism/
example, elevated levels of AFP may be detected in patients hypothyroidism; etc.[20,21] Life-style influences on tumor
with gastrointestinal tract malignancy with metastases to marker values include states like smoking (increases
liver, although the liver function tests may be normal. levels of CEA, AFP, etc.), alcoholism (altered liver and
renal parameters), obesity (hormonal imbalances, altered
Interpretation steroidal metabolism in peripheral fat), etc.
However, if no therapy has been given in the recent past, during With the explosion in the pool of knowledge, clinical
monitoring for a malignancy, a linear rise in three consecutive application of tumor markers in the field of oncology
samples (i.e., over a two time intervals) on a log scale should represents a classical example of “losing sight of the forest
be noted before a recurrence can be established. [37,38] Sampling for leaves of a tree.” It is indeed unfortunate that the
for tumor marker levels should ideally be repeated after 5-6 emphasis (in medical education, research and literature) has
half-lives of the marker in question (or the marker with the been on more and more extensive and in depth knowledge
longest half-life if multiple markers are being considered); of individual tumor markers, their pathophysiology, genetic
but if found elevated, the next sampling after 2-4 weeks, for origin, etc., rather than basic broad understanding and its
additional evidence, may be justified.[36] application in routine clinical practice.
Judicious application of tumor markers to clinical 17. Einhorn LH, Lowitz BB. Testicular Cancer. In: Casciato DA,
Lowitz BB, editors. Manual of clinical oncology. 3rd ed.
practice needs a thorough understanding of the basics of Boston: Little, Brown; 1995. p. 228-36.
pathophysiology, the techniques of identification or testing, 18. Testis: TNM classification of malignant tumors. In: Sobin LH,
reasons (in cases of both benign and malignant tumors) Wittekind C, editor. 5th ed. New York: Wiley-Liss; 1997.
for out-of-range levels of tumor markers, as well as the p.174-9.
19. The international non-Hodgkin’s lymphoma prognostic factors
knowledge of evidence of their role in any given malignancy. project: A predictive model for aggressive non-Hodgkin’s
lymphoma. N Engl J Med 1993;329:987-94.
20. Fateh-Moghadam A, Stieber P. Sensible use of tumour
REFERENCES markers, 2nd ed. Marloffstein-Rathsberg: Hartmann Verlag;
1993. p. 11-31.
1. Sturgeon C, Hammond E, Chang SL, Sölétormos G, Hayes DF. 21. Basuyau JP, Leroy M, Brunelle P. Determination of tumor
NACB: Practice guidelines and recommendations for use of markers in serum: Pitfalls and good practice. Clin Chem Lab
tumor markers in the clinic: Quality requirements [Section 2]. Med 2001;39:1227-33.
Draft Guidelines 2006. Available from: [Link] 22. Fearson ER. Progressing from gene mutations to cancer.
rdonlyres/3CAF1DC0-2E83-4BB1-8517-9300F58DF1DB/0/ In Abeloff: Clinical Oncology, 3rd ed. In: Abeloff MD,
chp2_qreqs.pdf. [accessed on 2008 Mar 20]. Armitage JO, Niederhuber JE, Kastan MB, McKenna WG,
2. Waxman J. Tumor markers. Quart J Med 1995;88:233-41. editors. Pennsylvania: Elsevier Churchill Livingston; 2004.
3. Gold P, Freeman SO. Demonstration of tumor specific antigens 23. Dalen AV. The basis of tumor marker determinations-
in human colonic carcinomata by immunological tolerance Recommendations of the EGTM. Kuwait Med J 2001;33:1-2.
and absorptions techniques. J Exp Med 1965;121:439-62. 24. Henry NL, Hayes DF. Uses and abuses of tumor markers
4. Diamandis EP. Tumor markers: Past, present, and future. in the diagnosis, monitoring, and treatment of primary and
Tumor markers: Physiology, pathobiology, technology, and metastatic breast cancer. Oncologist 2006;11:541-52.
clinical applications. In: Diamandis EP, Fritsche H Jr, Lilja H, 25. Bigbee W, Herberman RB. Tumor markers and
Chan D, Schwartz M, editors. Washington DC: AACC Press; immunodiagnosis. In: Bast RC Jr, Kufe DW, Pollock RE, et al.,
2002. p. 3–8. editors. Cancer Medicine. 6th ed. Hamilton, Ontario, Canada:
5. Sokoll LJ, Chan DW. Clinical chemistry: Tumor markers. In BC Decker Inc.; 2003.
Abeloff: Clinical Oncology, 3rd ed. In: Abeloff MD, Armitage 26. Sturgeon C, Aronsson AC, Duffy MJ, Hansson LO, Klapdor R,
JO, Niederhuber JE, Kastan MB, McKenna WG, editors. Dalen A van. European group on tumour markers: Consensus
Pennsylvania: Elsevier Churchill Livingston; 2004. recommendations. Anticancer Res 1999;19:2789.
6. Schrohl AS, Anderson MH, Sweep F, Schmitt M, Harbeck N, 27. Recommendations of European Group on Tumor Markers
Foekens J, et al. Tumor markers: From laboratory to clinical (EGTM). Anticancer Res 1999;19(4A):2791-819.
utility. Mol Cell Proteomics 2003;2:378-87. 28. Lee P, Pincus MR, McPherson RA. Diagnosis and management
7. Goedegebuure PS, Liyanange U, Eberlein TJ. Tumor biology of Cancer using serological tumor markers. In: Henry’s
and tumor markers. In: Townsend: Sabiston textbook of clinical diagnosis and management by laboratory methods.
surgery, 17th ed. In: Townsend CM, Beauchamp RD, Evers In: McPherson RA, Pincus MR, editors. 21st ed. Philadelphia:
BM, Mattox KL, editors. Philadelphia: Elsevier Saunders; 2004. Elsevier Saunders; 2007. p. 1353-66.
8. Johnson PJ. A framework for the molecular classification of 29. Wu JT, Nakmura R. Human circulating tumor markers.
circulating tumor markers. Ann N Y Acad Sci 2001;945:8-21. Chicago: American Association of Clinical Pathologist; 1997.
9. Lindblom A, Liljegren A. Tumor markers in malignancies. Br 30. Wu JT, Christensen SE. Effect of different test designs of
Med J 2000;320:424-7. immunoassays on ‘‘hook effect’’ of CA 19-9 measurement.
10. Cooper DL. Tumor markers. In: Cecil textbook of medicine. In: J Clin Lab Anal 1991;5:228.
Goldman, et al., editors. 22nd ed. Philadelphia: WB Saunders 31. Wu JT, Knight JA. Monoclonal immunoassays for tumor
Company; 2004. markers. Clin Chem 1988;28:1.
11. Wu JT. Diagnosis and management of cancer using serological 32. Bast RC Jr, Knauf S, Epenetos A, Dhokia B, Daly L, Tanner M,
tumor markers. In: Henry’s Clinical Diagnosis and Management et al. Coordinate elevation of serum markers in ovarian cancer
by Laboratory Methods. In: McPherson RA, Pincus MR, editors. but not in benign disease. Cancer 1991;68:1758.
21st ed. Philadelphia: Elsevier Saunders 2007. 33. Wu JT. Expression of monoclonal antibody defined tumor
12. Locker GY, Hamilton S, Harris J, Jessup JM, Kemeny N, markers in four carcinomas. Ann Clin Lab Sci 1989;19:17.
Macdonald JS, et al. ASCO 2006 update of recommendations 34. Katopodis N. Lipid associated sialic acid test for the detection
for the use of tumor markers in gastrointestinal cancer. J Clin of human cancer. Cancer Res 1982;42:5270.
Oncol 2006;24:5313-27. 35. Nahm MH, Hoffmann JW: Heteroantibody: Phantom of the
13. Harris L, Fritsche H, Mennel R, Norton L, Ravdin P, Taube S, immunoassay. Clin Chem 1990;36:829.
et al. American society of clinical oncology 2007 update of 36. Sturgeon C. Practice Guidelines for tumor marker use in the
recommendations for the use of tumor markers in breast clinic. Clin Chem 2002;48:1151-9.
cancer. J Clin Oncol 2007;25:5287-312. 37. Fleisher M, Dnistrian AM, Sturgeon CM, Lamerz R,
14. Canadian Task Force on the Periodic Health Examination. Whittliff Jl. Practice guidelines and recommendations for use
Canadian Guide to Clinical Preventive Health. Ottawa: of tumor markers in the clinic. In Tumor markers: Physiology,
Canada Communication Group; 1994. p. 797-809. pathobiology, technology, and clinical applications. In:
15. Cong WM, Wu MC, Zhang XH, Chen H, Yuan JY. Primary Diamandis EP, Fritsche H Jr, Lilja H, Chan D, Schwartz M,
hepatocellular carcinoma in women of mainland China: editors. Washington, DC: AACC Press; 2002.
A clinicopathologic analysis of 104 patients. Cancer 38. Bonfrer JM. Working group on Tumor marker criteria. Tumor
1993;71:2941-5. Biol 1990;11:287-8.
16. Bast RC Jr, Urban N, Shridhar V, Smith D, Zhang Z, Skates S,
et al. Early detection of ovarian cancer: Promise and reality. Source of Support: Nil, Conflict of Interest: None declared.
Cancer Treat Res 2002;107:61-97.