Questions About Cancer? 1-800-4-CANCER

Levels of Evidence for Cancer Genetics Studies (PDQ®)

Health Professional Version
Last Modified: 07/25/2012

Clinical Validity

Clinical validity refers to the predictive value of a test for a given clinical outcome (e.g., the likelihood that cancer will develop in someone with a positive test). It is primarily determined by the sensitivity and specificity with which a test identifies people with a defined clinical condition within a given population. Sensitivity of a test refers to the proportion of people who test positive for a clinical condition among those who actually have the clinical condition; specificity refers to the proportion of people who test negative for a clinical condition among those who do not have the clinical condition. In the case of genetic susceptibility to cancer, clinical validity can be considered at two levels:

  1. Does a positive test identify a person as having an increased risk of cancer?
  2. If so, how high is the cancer risk associated with a positive test?

Thus, the clinical validity of a genetic test is the likelihood that cancer will develop in someone with a positive test result. This likelihood is affected not only by the presence of the gene mutation itself but also by any other modifying factors that might affect the penetrance of the mutation (e.g., the mutation carrier's environmental exposures or personal behaviors) or by the presence or absence of mutations in other genes. For this reason, the clinical validity of a genetic test for a specific mutation may vary in different populations. If the cancer risk associated with a given mutation is unknown or variable, a test for the mutation will have uncertain clinical validity. A summary of definitions of concepts relevant to understanding clinical validity and other aspects of cancer genetics testing has been published.[1] The test should be evaluated in the population in which the test will be used.

Clues to whether a particular familial cancer syndrome has a genetic basis can be derived informally, by inspecting the pattern of affected and unaffected people in a series of families; or formally, using an analytic technique known as segregation analysis. Segregation analysis provides quantitative data in support of, or against, the likelihood that a particular genetic mode of inheritance might explain the patterns observed in the study families.

Evidence that a particular gene might explain a specific cancer predisposition syndrome often derives initially from linkage studies that use collections of families meeting stringent clinical criteria for a specific cancer susceptibility syndrome. The demonstration of strong linkage of cancer susceptibility to a gene or genetic region in a pattern consistent with autosomal dominant inheritance provides evidence in support of both the mode of inheritance and the particular gene that might underlie the risk. Once linkage is established, a strong case for association between the genetic trait and disease can be made, even though the families used in the study may not be representative of the general population. The genetic trait measured in linkage studies is not always the causal factor itself but may be a genetic trait closely linked to it. Additional molecular studies are required to identify the specific gene associated with inherited risk, after linkage studies have determined its general chromosomal location.

Linkage studies, however, provide only limited evidence concerning either the range of cancer types associated with a mutation or the magnitude of risk and lifetime probability of cancer conferred by a mutation in less selected populations. In addressing these questions, the best information for clinical decisions comes from naturally occurring populations in which people with all degrees of risk are represented, similar to those in which clinical or public health decisions must be made. Thus, observations about cancer risk in families having multiple members with early breast cancer are applicable only to other families meeting those same clinical criteria. Ideally, the families tested should also have similar exposures to factors that can modify the expression of the gene(s) being studied. The mutation-associated risk in other populations, such as families with less dramatic cancer aggregation, or in the general population can best be assessed by direct study of those populations.

References
  1. Grann VR, Jacobson JS: Population screening for cancer-related germline gene mutations. Lancet Oncol 3 (6): 341-8, 2002.  [PUBMED Abstract]