Tumor Markers: From Laboratory To Clinical Utility

Abstract

A very broad definition of a tumor marker is: a tool that enables the clinician to answer clinically relevant questions regarding a cancer disease (1). However, most researchers in this field would probably prefer the following more specific definition of a tumor marker: a molecule, a process, or a substance that is altered quantitatively or qualitatively in precancerous or cancerous conditions, the alteration being detectable by an assay (2). Alterations can be produced either by the tumor itself or by the surrounding normal tissue as a response to tumor cells (2). Regardless of which definition is preferred, the tumor marker itself can be DNA, mRNA, protein, or processes (apoptosis, angiogenesis, proliferation, etc.) measured quantitatively or qualitatively by an appropriate assay. In addition, the types of specimen in which the tumor marker is detected can be different; tissue, blood (plasma/ serum), saliva, urine, etc. are all used. The tumor marker assays can be of very different formats ranging from complex animal models to immunohistochemical test kits. The most commonly used format is probably the immunoassay, which is a well-characterized methodology. However, this field is progressing rapidly, and new and advanced assays such as microarrays and mass spectrometry are becoming established technologies in tumor marker research. The first known tumor marker was described in 1846, when Henry Bence-Jones reported the precipitation of a protein in acidified urine from patients with multiple myeloma. Detection of the monoclonal immunoglobulin light chain in this disease is still in use, and since then numerous potential tumor markers have been reported on in the literature (1). Examples of such markers in clinical use are: alpha-fetoprotein for tumors of the liver, testis, and other germ cell line tumors, CA125 for ovarian cancer, prostate specific antigen (PSA) for prostate cancer, and steroid hormone receptors (estrogen and progesterone receptor) used in management of breast cancer. However, as the field of tumor markers has expanded rapidly over the last two decades with a concomitant increase in published reports, it has become increasingly apparent that a strong need exists for establishment of consensus guidelines for development and use of tumor markers. Such guidelines should be internationally accepted if any of these potential new markers are ever to reach a stage where they will benefit the patients. The guidelines should define the potential specific clinical uses of tumor markers, define specific requirements for the technical development of tumor marker assays, and state specific requirements that are to be fulfilled before clinical implementation of a tumor marker. Suggestions for such guidelines have been made; in 1996, a tumor marker expert panel convened by the American Society of Clinical Oncology proposed a framework to be used for evaluation of tumor marker studies: the tumor marker utility grading system (TMUGS), which also includes a framework for rating published evidence (2). The TMUGS framework is further discussed in “Clinical Testing.” However, work in this field is still ongoing, and some important aspects to consider in the process of designing such guidelines will be covered in this review. The possible clinical uses of tumor markers are manifold, and several categories of markers can be defined. A diagnostic tumor marker is a marker that will aid in detection of malignant disease in an individual. Preferably, the marker should be tissue specific and not be influenced by benign diseases of the particular tissue/organ. Thus, a diagnostic marker should exhibit both high levels of diagnostic sensitivity and specificity (see below) to be of clinical value, especially if the marker is to be used for (mass) screening purposes. A fundamental prerequisite for development of any diagnostic (screening) tumor marker lies in the nature of the disease From the ‡Department of Pharmacology and Pathobiology, Royal Veterinary and Agricultural University, Ridebanevej 9, DK-1870 Frederiksberg C, Copenhagen, Denmark, §Department of Chemical Endocrinology, University Medical Centre Nijmegen, P.O. Box 9101, Geert Groteplein 10, NL-6500 HB Nijmegen, The Netherlands, ¶Clinical Research Unit, Department of Obstetrics & Gynaecology, Technical University of Munich, Ismaninger Strasse 22, D-81675 Munchen, Germany, and Rotterdam Cancer Institute (Daniel der Hoed Klinik), Josephine Nefkens Building, Nr. BE 426, Dr. Molewaterplein 50, NL-3015 GE Rotterdam, The Netherlands Received, June 9, 2003 Published, MCP Papers in Press, June 17, 2003, DOI 10.1074/mcp.R300006-MCP200 1 The abbreviations used are: PSA, prostate specific antigen; TMUGS, tumor marker utility grading system; CEA, carcino-embryonic antigen; uPA, urokinase-type plasminogen activator; PAI-1, plasminogen activator inhibitor type-1; EORTC, European Organisation for Research and Treatment of Cancer; RBG, Receptor and Biomarker Group; ER, estrogen receptor; PgR, progesterone receptor; HCG, human chorionic gonadotropin; QC, quality control; LOE, level of evidence. Review