The Search for Cancer Diagnostic Markers: A Race Against Time

September 1, 2005
K. John Morrow, Jr., PhD

K. John Morrow, Jr., PhD, is president of Newport Biotechnology Consultants, 625 Washington Avenue, Newport, KY 41071,

BioPharm International, BioPharm International-09-01-2005, Volume 18, Issue 9

At this time there are no effective diagnostic tests for cancer that are rapid, economical, highly specific, and highly sensitive. This deficiency means that many cases of malignancy go undetected until long past the time of effective treatment. Much cancer research, including investigations carried out by both the academic and private sector, is focused on combining therapy and diagnostics as "theragnostics."

At this time there are no effective diagnostic tests for cancer that are rapid, economical, highly specific, and highly sensitive. This deficiency means that many cases of malignancy go undetected until long past the time of effective treatment. Much cancer research, including investigations carried out by both the academic and private sector, is focused on combining therapy and diagnostics as "theragnostics."

Three approaches to diagnostics development include (1) a search for individual markers of specific malignancies; (2) a screening of a number of markers simultaneously in malignancies, to establish a diagnostic profile; (3) exploiting the antisera of patients in order to identify new antigens on cancer cells.

Classical markers of malignancy include prostate-specific antigen (PSA), carcinoembryonic antigen (CEA), and alpha fetoprotein, (AFP). Although widely utilized as diagnostic indicators, all have significant shortcomings.

Because of the complexity of the malignant phenotype and the fact that oncogenes are derived from normal cellular functions, designing a valid cancer test based upon a single marker may be extremely challenging, or even impossible. For this reason an evaluation of malignant cells based upon simultaneous profiling of a number of functions may be the most fruitful approach.

In the last decade the search for better identification of states of cancer has focused on new immunodiagnostic recognition systems for malignant disorders. This is one facet of the field of pharmacodiagnostics — the joining of cancer therapy to the outcome of a test measurement from a patient biopsy. Whereas the realization of effective cancer treatment requires successful cancer therapeutics, these markers may be exploited in the short run as cancer detection systems.

Detection of states of malignancy using antibodies has an extended history. The knowledge that the best outcome of the patient depends on the earliest possible detection and surgical removal of a tumor has long driven the search for effective diagnostics. It has been recognized since the origins of modern cancer treatment in the early 20th century that by the time a cancer produces clinical symptoms, the window of curative possibility has already passed.

Until recently, cancer screening was based on the use of low-tech approaches, such as the digital rectal examination for prostate cancer or X-rays for detecting breast cancer. This situation began to change in the 1970s with the development of tests based on the use of cancer-associated antigens using histochemical, immunofluorescent, and radioisotopic markers. These antigens have been localized using radiological, microscopic, and serum-based assay procedures. However, despite decades of use and exhaustive clinical evaluation, in most cases the results are highly controverted, and few tests are recognized as universally beneficial.

To be truly effective on a wide basis, a diagnostic test for cancer must be relatively non-invasive and economical. Thus biopsy of tissues (liver, kidneys, testicles) and the use of expensive medical equipment (MRI and CAT scans) do not lend themselves to population screening, or screening of the same individual on multiple occasions. For this reason, research has focused on markers present in the circulatory system because they can be detected in blood and/or urine. Table 1 outlines some of the most important diagnostic approaches currently in use.

Table 1. Diagnostic Approaches Currently in Use


Bogen and Sompuram1 propose three major approaches to the exploitation of cancer markers. The first is the traditional approach of characterization of a single cancer-related marker and its validation as an immunodiagnostic test. This strategy, which was advanced in the early days of hybridoma work, involved screening antibodies produced from mice immunized with mixtures of cancer cells. Labor intensive, thousands and thousands of hybridoma clones had to be screened, tested, and retested to generate a single candidate marker. With the technical improvement of recombinant DNA technology, routine manipulation of gene sequences allowed the cloning of recombinant antibodies and obviated the need for laborious hybridoma screening.

The second approach has evolved with sophisticated instrumentation for multiplexing and has become more widely available. In this permutation, a number of markers are monitored simultaneously with the expectation that the combined statistical power of numerous markers will yield an assay with a much higher specificity than any one marker counted singularly. In most cases the identity of the markers is not known, but an empirical profile of the particular cancer is developed. This approach has been described by Espina et al.2 who argue that each patient's cancer has a unique profile of genetic alterations. They have used genomic and proteomic tools to categorize the molecular derangements of individual tumors. The concept that measuring multiple analytes will yield superior diagnostic results is an appealing one. Carpelan-Holmstrom et al.3 employed a logistic regression algorithm combining the serum markers CEA, CA72-4, and CA 19-9 to improve diagnostic performance for gastrointestinal malignancy. A comparable study by Louhimo et al.4 investigated the combined performance of HGCß, CEA, CA 242, CA 72-4, and CA19-9, also with improved accuracy. While these results are encouraging, they have not yet resulted in a practical clinical assay.

A number of investigators are pursuing a search for cancer markers by screening the serum proteome using SELDI-TOF, 2-D gel analysis, and other approaches.5,6 However, this is an extremely challenging area, as there are many difficult problems to be resolved. Some of these technologies allow for multiplexed measurement of specific proteins in a rapid, low-cost format, which generates a tremendous amount of data from a single experiment. At this time protein microarray applications have been largely confined to basic research problems and such screening has not yet generated useful cancer markers.1

Yet a third approach is a more indirect way of identifying tumor-specific antibodies in the serum of affected individuals. In many cases when the antigens were identified, they have been found not to be tumor-specific.


Kits for the detection of disease states through antibody interactions have been widely used for many years. Not only the PSA-based kits, but also CEA, AFP, and many other markers have been employed in both diagnostic and therapeutic products.

While one of the most widely used markers for cancer screening is the prostate specific antigen (PSA) test, its reliability has been widely questioned.7 There are many kits available (such as the Biosafe PSA test) but despite years of screening and intensive investigation, Vicini et al. assert "the overall benefit of monitoring serum PSA after treatment for prostate cancer remains controversial."8 These researchers point out that the negative consequences of incorrect diagnosis in terms of cost, patient risk, and psychological anxiety are so substantial that much more investigation is warranted to define an appropriate application of therapy. Debate concerning the sensitivity, specificity, and positive predictive value for clinical response is widespread, and no pattern of PSA kinetics after treatment has conclusively been associated with a specific outcome. Indeed, the authors assert that 5 to 25 percent of patients ultimately experience failure and disease recurrence (beyond five years) even among those whose PSA levels predict the most optimal consequence.

Recently "PSA velocity," the rate of increased PSA levels over time, has been evaluated as a marker of malignancy risk.9 These authors determined that individuals with a velocity significantly greater than 2 ng/year had an increased risk of death from prostate cancer, despite radical prostatectomy. However, the size of the group examined was small, and additional studies will be required to establish the utility of this diagnostic approach. These questions regarding the PSA test illustrate the urgent need for new markers of malignancy.

Ekström et al.10 have explored an interesting means of protein discovery. They have developed a microplatform for analyzing samples of seminal plasma in a search for new protein markers of prostatic disease. They characterized proteins coisolated in affinity chromatography runs (using anti-PSA monoclonals) of prostate-specific antigen, and identified a protein known as prolactin-inducible protein, which may play a role in both tumor progression and fertilization. Proteins that copurify with PSA might serve as defining markers of degree or type of malignancy.

Ovarian cancer is another area that exemplifies the difficulties and pitfalls of establishing accurate and sensitive diagnostic markers. One of the most common causes of cancer-related deaths in women, this disease claimed 13,900 lives in 2001. Because of the difficulty of diagnosing this disorder, Cathro and Stolar11 have examined several markers including calretinin, inhibitin, and the Wilm's tumor gene. They found that in a group of 111 primary ovarian tumors, none of the antigens served to unambiguously distinguish normal from diseased tissue. For example in ovarian sex cord stromal tumors, calretinin detected 76 percent of tumors similar to those identified by inhibitin. Moreover, these markers were evaluated by immunohistochemistry and may be satisfactory for a serum-based test.

Table 2. Site-Specific Diagnostic Tests for Cancer

Another marker of ovarian cancer is bikunin, a glycoprotein that mediates suppression of metastasis in cancer cells.12 Higher levels of expression in ovarian cancer cells are correlated with lower levels of invasiveness, apparently by downregulating the plasminogen activator and its receptor. Bikunin is produced in the liver, and the authors surmise that different individuals respond with differing levels of synthetic capacity. This would be reflected in the level of suppression of tumor invasiveness. While there is a marked and statistically significant difference in mean serum levels in patients with aggressive disease as compared to disease-free controls and patients with indolent disease, the standard deviation is quite large. Thus while bikunin levels may eventually have predictive value for disease outcome, it is unlikely that this marker will be of value in immunodiagnostic screening programs.

In addition to PSA there is only one FDA-approved test for a cancer related antigen, nuclear matrix protein 22 (NMP22). This is a lateral flow enzyme immunoassay cleared for general screening of urine. However, pyuria and hematuria significantly interfere with the test, and its overall sensitivity was only 66 percent. Numerous other biomarkers have been identified for bladder cancer and are under investigation.1

Hepatocellular carcinoma (HCC) surveillance with alpha-fetoprotein (AFP) has been recommended for persons with cirrhosis, but AFP level lacks the requisite sensitivity for the early detection of HCC.13 Other promising biomarkers, such as des-gamma carboxyprothrombin, lens culinaris agglutinin-reactive AFP, human hepatocyte growth factor, and insulin-like growth factor-1, have not been validated for clinical use. According to the National Cancer Institute, there is an urgent need for new biomarkers for early HCC.


It is hardly surprising that the search for effective diagnostic cancer markers that would lend themselves to simple, economical, and effective detection of early cancer has been so fruitless. If the last 30 years of brilliant discovery in molecular biology has proven anything, it is that oncogenes are a normal component of cellular machinery, which have gone awry to a state that is overexpressed, underexpressed, expressed inappropriately, in the wrong place at the wrong time at the wrong level, too much, or too little. If we accept this premise, perhaps we have the basis for an understanding of their nature and how to detect and analyze it. Because clearly a cancer cell is qualitatively different from a normal cell, and cancer cells most to be feared are malignant cells, which are qualitatively different on yet another level. Thus in principle at least, we should be able to distinguish them by biochemical means.

It is an obvious and inescapable fact that there is NO normal cell, or benign cancer cell, that forms masses that break apart, moves through the circulation, establishes itself in totally inappropriate regions of the body, abandons all (or almost all) of its normal functions and then embarks upon a self-destructive rampage of multiplication.

If we accept this picture of the cancer cell, the logical conclusion is that the overall differences between a normal cell and a malignant cell are so great that the two can be easily distinguished. Yet a system based on a single marker such as PSA will never be satisfactory, and an effective assay system will have to be based upon the simultaneous assessment of multiple markers. It is not clear how such a system might be configured, but I would suggest that immunoassays that combine measurements for several markers simultaneously might offer the possibility of an effective diagnostic. One appealing way in which such platforms could be designed would entail the use of single-chain antibodies joined together as bivalent antibodies or "tetrabodies."14

It is a depressing reality that the tremendous strides in our understanding of the molecular laws governing living creatures, which have fostered so much accomplishment and understanding in recent years, have not been accompanied by a concomitant introduction of new therapies. In fact, the number of new drugs introduced into the market place has slowed, and the situation is even worse if one looks at the number of truly new drugs introduced per year, versus copycat drugs and trivial modifications of successful, already existent drugs (such as chiral products). While there are a number of explanations for this unpleasant fact, certainly a large share of the problem rests with the conclusion that the classical approach of targeting a single molecular entity with a single agent does not meet the challenge of conditions such as cancer, obesity, aging, and cardiovascular disease, in which a whole network of functions has been profoundly altered.

This challenge can be met, but it may require a whole new level of understanding and a new paradigm of treatment.

K. John Morrow, Jr., Ph.D. , BioPharm International Editorial Advisory Board and president, Newport Biotechnology Consultants, 625 Washington Ave., Newport, KY 41071, 513.237.3303, Fax: 513.271.0744,


1. Bogen SA, Sompuram SR. Recent trends and advances in immunodiagnostics of solid tumors. BioDrugs 2004;18(6):387-98.

2. Espina V, Geho D, Mehta AI, Petricoin EF 3rd, Liotta LA, Rosenblatt KP. Pathology of the future: molecular profiling for targeted therapy. Cancer Invest 2005; 23(1):36-46.

3. Carpelan-Holmstrom M, Louhimo J, Stenman UH, Alfthan H, Jarvinen H, Haglund C. Estimating the probability of cancer with several tumor markers in patients with colorectal disease. Oncology 2004; 66(4):296-302.

4. Louhimo J, Carpelan-Holmstrom M, Alfthan H, Stenman UH, Jarvinen HJ, Haglund C. Serum HCG beta, CA 72-4 and CEA are independent prognostic factors in colorectal cancer. Int J Cancer 2002;101(6):545-8.

5. Anderson NL, Anderson NG. The human plasma proteome: history, character, and diagnostic prospects. Mol Cell Proteomics 2002; 1(11):845-67.

6. Petricoin EF, Liotta LA. SELDI-TOF-based serum proteomic pattern diagnostics for early detection of cancer. Curr Opin Biotechnol 2004; 15(1):24-30.

7. Hernandez J, Canby-Hagino E, Thompson IM. Biomarkers for the detection and prognosis of prostate cancer. Curr Urol Rep. 2005; 6(3):171-6.

8. Vicini FA, Vargas C, Abner A, Kestin L, Horwitz E, Martinez A. Limitations in the use of serum prostate specific antigen levels to monitor patients after treatment for prostate cancer. J Urol. 2005; 173(5):1456-62.

9. D'Amico AV, Chen MH, Roehl KA, Catalona WJ. Preoperative PSA velocity and the risk of death from prostate cancer after radical prostatectomy. New Eng J Med. 2004; 351(2):125-135.

10. Ekstrom S, Wallman L, Malm J, Becker C, Lilja H, Laurell T, Marko-Varga G. Integrated selective enrichment target--a microtechnology platform for matrix-assisted laser desorption/ionization-mass spectrometry applied on protein biomarkers in prostate diseases. Electrophoresis 2004; 25(21-22):3769-77.

11. Cathro HP, Stoler MH. The utility of calretinin, inhibin, and WT1 immunohistochemical staining in the differential diagnosis of ovarian tumors. Hum Pathol. 2005; 36(2):195-201.

12. Matsuzaki H, Kobayashi H, Yagyu T, Wakahara K, Kondo T, Kurita N,et al. Plasma bikunin as a favorable prognostic factor in ovarian cancer. J Clin Oncol. 2005; 23(7):1463-72.

13. Marrero JA, Lok AS. Newer markers for hepatocellular carcinoma. Gastroenterology 2004; 127(5 Suppl 1):S113-9.

14. Todorovska A, Roovers RC, Dolezal O, Kortt AA, Hoogenboom HR, Hudson,PJ. Design and application of diabodies, triabodies and tetrabodies for cancer targeting. J Immunol Methods. 2001; 248(1-2):47-66.

15. Juuti A, Nordling S, Lundin J, Louhimo J, Haglund C. Syndecan-1 Expression - A Novel Prognostic Marker in Pancreatic Cancer. Oncology 2005; 68(2-3):97-106.

16. Seregni E, Coli A, Mazzucca N, Italian Group RIA-IRMA Test, Italian Association of Nuclear Medicine. Circulating tumour markers in breast cancer. Eur J Nucl Med Mol Imaging 2004; 31 (Suppl 1):S15-22.

17. Lai IR, Lee WJ, Huang MT, Lin HH. Comparison of serum CA72-4, CEA, TPA, CA19-9 and CA125 levels in gastric cancer patients and correlation with recurrence. Hepatogastroenterology 2002; 49(46):1157-60.

18. Saad F. UPM3: review of a new molecular diagnostic urine test for prostate cancer. Can J Urol. 2005 Feb;12 Suppl 1:40-3; discussion 99-100.

19. Carlisle MR, Lu C, McDougall IR. The interpretation of 131I scans in the evaluation of thyroid cancer, with an emphasis on false positive findings. Nucl Med Commun. 2003 Jun;24(6):715-35.

20. Mehta K, Haller JO, Legasto AC. Imaging neuroblastoma in children.Crit Rev Comput Tomogr. 2003;44(1):47-61.

21. Ardekani AM, Liotta LA, Petricoin EF 3rd. Clinical potential of proteomics in the diagnosis of ovarian cancer. Expert Rev Mol Diagn. 2002 Jul;2(4):312-20.

22. Gamzu R, Almog B, Levin I, Fainaru O, Niv J, Lessing JB, Bar-Am A. Clinical and economic implications of adding HPV tests to the routine cytology follow-up and management of patients with histologically defined cervical intraepithelial neoplasia grade 1 Gynecol Oncol. 2002 Aug;86(2):129-33.

23. Baum RP, Hellwig D, Mezzetti M. Position of nuclear medicine modalities in the diagnostic workup of cancer patients: lung cancer. Q J Nucl Med Mol Imaging. 2004 Jun;48(2):119-42.

24. Wilkinson NW, Loewen GM, Klippenstein DL, Litwin AM, Anderson TM. The evolution of lung cancer screening. J Surg Oncol. 2003 Dec;84(4):234-8.

25. Xing PX, Young GP, Ho D, Sinatra MA, Hoj PB, McKenzie IF. A new approach to fecal occult blood testing based on the detection of haptoglobin. Cancer. 1996 Jul 1;78(1):48-56.

26. Kocak H, Oner-Iyidogan Y, Kocak T, Oner P. Determination of diagnostic and prognostic values of urinary interleukin-8, tumor necrosis factor-alpha, and leukocyte arylsulfatase-A activity in patients with bladder cancer. Clin Biochem. 2004 Aug;37(8):673-8.

27. Pattari SK, Dey P. Urine: beyond cytology for detection of malignancy. Diagn Cytopathol. 2002 Sep;27(3):139-42.

28. Engelman KK, Ellerbeck EF, Mayo MS, Markello SJ, Ahluwalia JS Mammography facility characteristics and repeat mammography use among Medicare beneficiaries. Prev Med. 2004 Sep;39(3):491-7.

29. Hlawatsch A, Teifke A, Schmidt M, Thelen M. Preoperative assessment of breast cancer: sonography versus MR imaging. R Am J Roentgenol. 2002 Dec;179(6):1493-501.

30. Khandker RK, Dulski JD, Kilpatrick JB, Ellis RP, Mitchell JB, Baine WB. A decision model and cost-effectiveness analysis of colorectal cancer screening and surveillance guidelines for average-risk adults. Int J Technol Assess Health Care. 2000 Summer;16(3):799-810.