Circulating tumor markers

Abstract

The concept of a circulating tumor marker applies to a secreted chemical product of a tumor cell such that the concentration of the chemical in the blood may in some way represent a quantifiable assessment of the tumor burden at that time. The earliest example is the protein produced from myeloma cells discovered by Bence Jones in the mid-19th century. Subsequently a number of oncofetal and other proteins have proved useful and are widely available from antibody (Ab)-based assays. This scene is set to expand dramatically with an increase in our knowledge of the molecular pathology of cancer subtypes and the application of genomic and proteomic analysis techniques. One example is the measurement of serum DNA concentration [1]. The DNA probably derives from necrosis and apoptosis [2]. The specificity can be increased by analyzing tumor DNA, such as that with allelic imbalance of sequences subject in that tumor type to frequent allelic losses [3, 4], or by analyzing the proportion of circulating DNA in long fragments (DNA integrity) [2]. A study of DNA integrity in the serum of patients with breast cancer suggested this test could predict lymph node metastasis [5]. Proteins or protein fragments may be released into the circulation from cancers and detected by surface-enhanced laser desorption-ionization time-of-flight mass spectroscopy (SELDI-TOF) [6]. An early report suggested this may form the basis of a test for ovarian cancer [7], however this has not been reproduced, and bioinformatics methodologies have also been questioned [8]. Despite a number of reports of high sensitivity and specificity [9, 10], at present the reliability of these proteomic profiles in cancer diagnosis remains controversial [11]. Currently, the range of possible tumor markers is broad; however, relatively few tumor markers have been incorporated in routine oncological practice (Table 5-1), possibly because any one marker is expressed in only a few tumors. This problem, however, can be addressed by a broad preliminary screen. For example, we have conducted a study of human chorionic gonadotrophin (hCG), carcinoembryonic antigen (CEA), CA125, and CA19.9 in 74 patients with advanced bladder cancer and found that 43 (58%) had a significant increase of at least one serum marker [12]. This information was useful because of the close correlation between clinical and marker response to chemotherapy. A range of new serum markers is being investigated. YKL-40, a member of the mammalian chitinase-like proteins with some growth-factor activity is found in a number of cancers [13]. In melanoma, it is reported to be increased in 45% of patients and to be an independent prognostic indicator [14]. Serum chromogranin A may add to the assay of neuron-specific enolase in the detection of neuroendocrine differentiation in prostate cancer [15], or other sources. Mesothelin is highly expressed in ovarian cancers as well as mesothelioma, and has been detected in the serum of 67% and 71% of these cancers, respectively [16]. Tissue inhibitor of metalloproteinase-1 (TIMP-1) is over expressed in a range of malignancies including gastric, lung, breast, and colorectal cancers, and increased plasma concentrations have been reported in patients with colorectal cancer, the postoperative concentrations correlating with prognosis [17]. Finally, vascular endothelial 107 growth factor (VEGF) is proangiogenic, and serum VEGF concentrations are prognostic in melanoma, lymphoma, esophageal, small-cell lung (SCLC), ovarian, and colon cancers. The report on a large series of patients with ovarian cancer showed that serum VEGF was an independent indicator of survival, and in particular, patients with stage I disease and serum VEGF concentrations 380 pg/ml had an eight-fold increased risk of mortality [18]. Macrophage inhibiting cytokine-1 (MIC-1) has been found to be better than other markers, including CA19-9, in the diagnosis of pancreatic cancer [19]. The clinical roles of circulating markers might include screening, diagnosis, staging, assessment of prognosis and monitoring of response, remission, and relapse. Additionally, as relatively specific tumor products, marker substances may confer tissue specificity for immunohistochemical (IHC) diagnosis and ligand-targeted techniques for imaging and therapy [20]. To be useful in clinical practice, an ideal marker should be both sensitive and specific. Furthermore, the marker test should reliably indicate the situation to which there is an appropriate therapeutic response. The sensitivity of a test is the probability of the test being positive in patients with the disease. Based on the symbols in Table 5-2, sensitivity equals A/(A + B). The specificity of the test is the probability of a normal test result in patients without the cancer. From Table 5-2 specificity equals D/(C + D). A further concept of value in judging markers is the positive predictive value, which is the probability of a patient having the cancer when the test is positive, i.e., the number of true positive results divided by the total number of positive results (i.e., A/[A + B]). These relatively simple concepts become more complex for marker tests in which no clear cutoff is seen between a normal and abnormal result, e.g., with prostate specific antigen (PSA) as a diagnostic test for prostate cancer. In this setting, higher values of the marker represent a greater probability of the presence of prostate cancer and the appropriate choice of cutoff level for finding cancer may depend upon patient-related factors such as the age [21], or the size of the normal prostate gland. The predictive ability of a test can be represented by the receiver operating curve (ROC), which is a plot of the sensitivity versus (1-specificity) for a binary classification as its discrimination threshold is varied. © 2008 Humana Press Inc.