Maurie Markman, MD
One of the basic tenets of science is the concept that all accepted “truths” are subject to objective testing and, through this process, can be shown to be false. No matter how solid the existing evidence, at any time new data or a novel analysis may prove the current view to be incorrect. Further, any claim that cannot be challenged—“a personal belief”—is not within the domain of science and scientific thought.
During their training, physicians learn much about how we have progressed to our current level of understanding of anatomy, physiology, and the multiple biological mechanisms through which the normal transitions to the abnormal. They also learn that what is considered “truth” today in the biological foundation of modern medicine may be substantially altered in the coming years or decades as a result of changes in our understanding of the molecular basis of health and illness.
Now, widely anticipated paradigm-changing approaches to population-based Big Data hold the promise of adding an essential element of objectivity to the current—often distressing—state of affairs where clinicians are frequently required to “guess” the appropriate standard of care for real-world patients based on the results of clinical trials whose participants may only be representative of their patients by being either male or female.
Examples of Faulty Beliefs
Unfortunately, it is not difficult to find rather striking examples of why we must always remember that “what we think we know” in the realm of medical science may ultimately prove to be incorrect or at least questionable.
In the increasingly clinically relevant and highly complex genomic arena, a recent report has noted that an unknown but distressing number of African Americans have been classified as having a heightened risk for the development of hypertrophic cardiomyopathy due to misclassification of genetic variants associated with the condition.1
This potentially serious error is believed to have occurred because an inadequate number of members of this population were included in the control arms of older studies that attempted to define the genetic risk. This experience is highly relevant for ongoing and future studies of genetic risk for other human illnesses, including cancer. Since an increased risk is defined in comparison to a baseline, it is critical that the control population in any epidemiological analysis include the normal germline variants observed within a particular ethnic group or a geographically isolated population.
Closer to the oncology domain, a recent report has challenged the decades-old dogma regarding the relationship between exposure to external non-naturally occurring radiation and the subsequent development of cancer.
There has been a strong presumption that radiation exposure resulting from the atomic bombs that fell over Hiroshima and Nagasaki near the end of World War II would result in a substantial increase in cancer, a decrease in life expectancy, and a major increase in the mutations observed in the offspring of the survivors.2
Beginning in 1947 after the war ended, extensive follow-up was initiated among a group of 120,000 survivors and 77,000 of their offspring. In fact, while the risk of cancer was increased (42%) among this surviving population, there was a limited overall decrease in life expectancy found (approximately 1 year) and, to date, no increase has been observed either in abnormalities or the mutation rate among their offspring.2
Of relevance to both individuals working within the nuclear industry and potentially those exposed to medical radiation such as health care workers and patients, the established safety requirements for radiation have been based on assumptions of risk that are possibly overstated based on the extensive and long-term follow-up data now available from the large survivor population in Japan.
Chemotherapy Advice Questioned
As a final example, we turn to the arena of long-established routine recommendations regarding particular strategies designed to reduce the risk among patients with cancer who have an already recognized heightened probability for experiencing a serious infectious episode.
In a provocative analysis, the objective utility was reported for several such approaches widely employed by oncologists and cancer programs for children with acute myeloid leukemia undergoing intensive therapy.3
In a carefully conducted review of 339 patients managed at 37 different institutions, the investigators were unable to find any benefit for several widely utilized non-pharmacologic strategies or policies, including the requirement for adherence to specific dietary restrictions (neutropenic diet) and for limitations on social contact or exposure to pets.