Precision Medicine Can Help Reduce "Overdiagnosis" Problem

Maurie Markman, MD

It would be difficult to name a more vexing oncology problem than overdiagnosis—that is, poorly defined conditions that are initially labeled as “malignant” but ultimately do not become clinically relevant cancers.¹ The intense and never-ending debates regarding the potential benefits versus harms associated with screening strategies for breast and prostate malignancies are testimony to this complex issue.

In patients with common solid tumors, it is well-established—and not debatable—that those with early-stage cancers have a statistically defined superior prognosis compared with those who present with and are required to be treated at later stages of the disease. Therefore, it would appear intuitively obvious that an earlier diagnosis in the natural history of a malignancy in an individual patient should result in a more favorable prognosis.

Unfortunately, the biology of cancer is not that simple and as our increasingly sensitive diagnostic strategies, such as imaging technology and tumor marker algorithms, are widely employed, it is evident that not all conditions labeled as a “cancer” will progress to cause disease during an individual’s lifetime or that they may be diagnosed at a later time without a change in the outcome. Simply stated, the problem is determining which cancers fall into which category.

There are several dimensions to this complex issue, not all necessarily relevant to particular audiences or a given person. For example, the relative risk a specific individual may be willing to accept for the potential “harm,” such as the anxiety associated with what may turn out to be a false-positive test result, for the potential benefit of earlier detection of a malignancy (eg, stage I versus stage II or III) is likely to be highly variable.

Further, an individual with good health insurance may be far less concerned about the cost of what might be labeled by some as “excessive screening” than a large employer responsible for paying the bills for their employees (and dependents) for screening procedures. The payment issue is particularly relevant for federal government-supported efforts where taxpayers as a group will be responsible for the associated costs of screening.

In addition, individuals with a personal experience (family member, healthcare provider) caring for patients who “presented with an advanced cancer” may have an understandably strong desire to want to do whatever is reasonably possible to optimize the chances for an earlier diagnosis.

Testing for Risk Factors

Unfortunately, the debate regarding overdiagnosis of cancer today appears to have become less about pragmatic solutions and more about somewhat arcane discussions about the interpretation of large population-based screening trials often initiated more than a decade ago that used diagnostic approaches and biomarker algorithms not currently employed in routine practice.How might the rapidly evolving and increasingly clinically relevant role of precision medicine favorably influence the development of diagnostic paradigms that may reduce the risk of the overdiagnosis of malignant conditions? Today, the major focus of molecular efforts in oncology has been on optimal treatment strategies following the diagnosis but there is no reason to conclude this basic approach cannot in certain well-defined situations be effectively used earlier in the cancer management continuum.

A relevant example of the value of precision medicine is the use of molecular testing to discover the presence of a germline BRCA mutation, which may lead (after counseling) to the performance of a prophylactic surgical procedure (bilateral salpingo-oophorectomy) that substantially reduces the risk of the development of epithelial ovarian cancer or primary carcinoma of the peritoneum. Evidence of a genetically defined Lynch syndrome will also result in a management strategy for both screening and potentially prophylactic surgical interventions for several malignant entities.

Although not specifically related to overdiagnosis issues, other uses of precision medicine that currently are not as established as BRCA or Lynch syndrome testing may serve as a paradigm for how such efforts might be used in the future. Investigators have begun to examine molecular profiles of patients with myeloproliferative conditions to help define the statistical likelihood (risk) of encountering negative outcomes.²

Such prognostic data might be helpful in optimizing disease management. For example, a genetic biomarker has been associated with the development of arterial thrombosis in patients with essential thrombocythemia, which has a median life expectancy of 20 years.² Although this is a relatively rare medical condition with a long life expectancy, the development of a serious but uncommon event can have a devastating impact on the clinical outcome.

Knowledge of a heightened risk for this adverse event may lead to a rational screening strategy that may favorably influence the ultimate course of the illness. Although such an approach might not be cost effective over several decades for all patients presenting with essential thrombocythemia, the presence of this molecular factor (plus additional clinical characteristics) might result in a focused strategy that meaningfully alters the natural history of this condition for many individuals.

Similarly, when discussing concerns about the overdiagnosis of cancer, the presence of specific germline variants or unique somatic molecular events within a biopsied mass could serve as a clinically useful prognostic marker for the statistical likelihood that a lesion discovered through screening will become an invasive and progressive cancer.³

Of course, such testing will almost certainly not be definitive as to whether a given abnormality will become a “true” cancer, but it is reasonable to speculate that future research in this area will find such knowledge to be of merit in the subsequent clinical decision-making process.

References

1. Welch HG, Fisher ES. Income and cancer overdiagnosis—when too much care is harmful. N Engl J Med. 2017;376(23):2208-2209. doi: 10.1056/NEJMp1615069.
2. Tefferi A, Vannucchi AM. Genetic risk assessment in myeloproliferative neoplasms. Mayo Clin Proc. 2017;92(8):1283-1290. doi: 10.1016/j.mayocp.2017.06.002.
3. Evans JP, Powell BC, Berg JS. Finding the rare pathogenic variants in a human genome. JAMA. 2017;317(18)1904-1905. doi: 10.1001/jama.2017.0432.