Tumor markers are substances that are made by neoplastic cells and released into the bloodstream and detectable by measurement with laboratory testing on blood samples from the patient. Tumor markers may or may not have specificity for certain cancers.For example:
carcinoembryonic antigen (CEA) can be elaborated by a variety of cancers such as those arising in the gastrointestinal tract.
Prostate specific antigen (PSA) is, clearly, highly specific for prostatic adenocarcinomas.
Some markers are not entirely specific for neoplastic cells, such as human chorionic gonadotropin (HCG) and testing may detect small amounts that circulate, or may detect increased levels with a normal physiologic event (pregnancy).
Another problem is the amount of marker that is present. An ideal tumor marker would be detectable, even if the neoplasm were still quite small, allowing early detection and treatment with a better prognosis. However, in reality, the amount of marker can be related to tumor size (stage) and differentiation (grade). The larger the size of the tumor, the more marker will be present. The more differentiated the tumor, the more likely the marker will be elaborated. A tumor with a very low stage may not make enough marker to be reliably measured with routine laboratory testing. A poorly differentiated tumor may also not elaborate the marker in large amounts, or may not elaborate it at all.
An additional problem is predictive value of the test. In order for any positive laboratory test to have a high predictive value (i.e., is the disease really there or not there) the test must have a high specificity (can you exclude persons without the disease) and sensitivity (can you find persons with the disease). Few laboratory tests approach 99% sensitivity and specificity, but even at that high level, if the prevalence of the disease (how many people in the population have it) is low, then the predictive value of a positive test will be low. The predictive value of a negative test is generally better for any test.
Screening with tumor markers for a general population, only a few of whom have the disease, tends to yield many more false positives than true positives. This is not cost effective use of resources in medicine. You can increase the predictive value by narrowing down the population to be screened.
Another problem occurs when non-neoplastic conditions increase the level of a tumor marker and make the separation of patients with cancer more difficult. For example, the levels of CEA are increased in persons who are smokers. Persons with prostatitis or nodular hyperplasia of the prostate have increased PSA levels (generally up to double the normal upper limit).
Return to the main menu