Maurie Markman, MD, explains the concerns surrounding liquid biopsies in cancer screening, and why regulatory agencies should be cautious and not approve any proposed molecularly based cancer diagnostic screening strategy until the tool is shown to effect cancer-specific survival compared with a control population not undergoing such testing.
In a provocative article, Siddhartha Mukherjee, MD, a cancer laboratory scientist and 2011 Pulitzer Prize winner for The Emperor of All Maladies, discussed future developments that are likely to occur with increasingly sensitive blood testing strategies for the early detection and diagnosis of malignant disease.1 Mukherjee appropriately distinguishes how the role of early detection with liquid biopsy will play in 2 scenarios: (a) individuals with known cancer where the goal is to discover documented evidence of recurrence as soon as possible; and (b) individuals without prior knowledge of the presence of a malignancy.
Although one may argue the actual technical approach (eg, examination for a unique molecular marker or markers to confirm the presence of cancer) may be similar in the 2 described settings, the implications for subsequent evaluation, the costs of such management, and the emotional effect of a positive test result may be strikingly different. It is critical to emphasize that this commentary focuses on a discussion of cancer detection employing novel strategies in the absence of any signs (eg, abnormal imaging study) or symptoms (eg, abdominal pain) that might suggest the presence of a cancer.
Imaging studies and blood tests employed to detect early-stage cancer and recurrent disease have been the standard of care in cancer treatment for decades. It is understandable that patients with cancer who are in remission may seek reassurance that the disease has not recurred, whereas others without a malignancy may desire a form of comfort that they do not have one of the multiple types of this large group of diseases.
It would be difficult to argue with the statement that recurrence of early-stage germ cell tumors is the poster child for the clinical utility of routine blood testing, where the availability of 2 simple tests—β-human chorionic gonadotropin (HCG) and α fetoprotein—have made it possible to reliably document the presence of disease progression in the absence of signs or symptoms allowing for the initiation of effective curative therapy on the basis of these findings. Unfortunately, in most other settings, although it may be possible to demonstrate early recurrence based solely on results of imaging studies or blood-based testing—for example, rising CA-125 in patients with ovarian cancer2—it is uncertain if the natural history of the illness and survival has been favorably influenced by somewhat speedier intervention.
It is essential to note the sophistication and costs of imaging and laboratory-based testing in recent times and that such studies may be employed independent of any formal evaluation in controlled clinical trials as to their effect on objective clinical outcomes, specifically, survival.
It is also relevant to acknowledge that investigative efforts that focus solely on a population-based survival effect may miss opportunities to provide critical justification for why many patients desire routine, disease-specific follow-up imaging or blood tests, such as the reassurance and peace of mind a negative finding may provide. As a clinician who has cared for women with ovarian cancer over a period of several decades it is my experience that it is the rare patient who elects to discontinue CA-125 monitoring at the completion of chemotherapy, even when informed of the data from randomized trials that routine testing has not been demonstrated to improve survival compared with a strategy where the blood test is obtained only in the presence of any signs or symptoms suggestive of disease recurrence (eg, abdominal discomfort or bloating).2
At least in the case of disease recurrence, it is possible to focus on the recognized natural history of the specific malignancy, know the most likely metastatic sites for the primary cancer, employ imaging and increasingly relevant molecular diagnostic strategies demonstrated to be of value in the population of patients with the cancer, and have a reasonably well-defined approach to subsequent disease management if recurrent cancer is discovered.
For many individuals without a known cancer, considerable debate continues regarding the magnitude of the benefits of specific forms of screening, including the optimal age to begin or end the process as well as the frequency of testing. It is reasonable to suggest that there is a general consensus within the medical community that a number of strategies have been documented to discover early stages of cancer or even premalignant abnormalities where intervention can reduce cancer-related mortality.
However, in contrast to a search for recurrent cancer, in the asymptomatic individual without any evidence for the presence of a malignancy, the potential to produce far more harm than good is a major issue, as has been clearly documented in the multidecade, and essentially fruitless, effort to develop an ovarian cancer screening test.3 There are several concerns that must be confronted by a company or academic institution desiring to develop such a screening strategy.
First, as only a small number of individuals within a large population will be anticipated to receive a diagnosis of specific malignancy, there is an absolute requirement for both a highly sensitive and specific test to ensure a very low risk of either a false negative or false positive result. One may reasonably argue that the concern for false positive results is the more serious issue as this outcome may lead to additional costly testing, unwarranted anxiety for patients, and invasive procedures associated with serious morbidity or even mortality.3
At least with a false negative result, the individual will not be in a worse position than if they had never undergone the test, except in a situation where they have mistakenly elected to pursue the widely advertised promise of the new technology rather than undergoing the older, but well-established approach to screening for a particular condition.
A second concern is for the risk of lead-time bias associated with any approach to finding a cancer at what is believed to be an earlier point in its development where therapy may be assumed to be more effective and less aggressive. In fact, this discovery may have no effect on the actual natural history beyond what would have been achieved with a later diagnosis. However, survival outcomes may appear to be longer simply because the disease was discovered to be present at an earlier point in time.
Therefore, at least in the opinion of this commentator, regulatory agencies should be cautious and not approve any proposed molecularly based cancer diagnostic screening strategy until the tool is shown to effect cancer-specific survival compared with a control population not undergoing such testing. Unfortunately, it is not difficult to suggest the potential for serious harm rather than clinical benefit in the absence of definitive comparative trials.