With the advent of mass screening by Pap smear, cervical cancer incidence and death rates declined by more than 60 percent in the U.S. between 1955 and 1992. It was a triumphant demonstration of the value of early detection. But the model has never worked so well for other common cancers.
Figure: Vaginal cells drawn by George Papanicolaou, inventor of the pap smear (Wellcome Images)
Mammography, for example, fails to detect one in four tumors in younger women while also delivering many false-positive results. Less than 5 percent of positive initial findings prove to be cancer. And believe it or not, we still don’t know for sure whether any cancer screening test saves lives. The evidence in favor of screening is based on fewer deaths due to the target cancer, not reductions in overall mortality, according to a critique published in the British Medical Journal calling for higher standards of evidence for cancer screening.
“We believe it is both scientifically inaccurate and misleading to say that screening tests save lives,” says Vinay Prasad, a specialist in blood cancers for the OHSU Knight Cancer Institute. He co-authored the critique with journalist Jeanne Lenzer and David Newman, a professor at the Icahn School of Medicine at Mount Sinai in New York.
“Saving lives is reducing overall mortality and we don’t know if screening decreases dying for any reason. It’s quite possible that it doesn’t,” Prasad says. “I don’t think patients understand that.”
How screening can falter
Looking at clinical trials of screening tests, he and co-authors found that most had discrepancies between disease specific and overall mortality. A systematic review published earlier this year by Stanford researcher John Ioannidis and colleagues found no evidence for reductions in overall mortality for any cancer screening test in meta-analyses of multiple trials.
“Screening may still be highly effective (and thus justifiable) for a variety of other clinical outcomes, besides mortality. However, our overview suggests that expectations of major benefits in mortality from screening need to be cautiously tempered,” Ioannidis and co-authors concluded.
There are two explanations why screening tests that prevent cancer deaths don’t appear to reduce overall mortality. It could be that existing studies are simply underpowered to detect the overall benefit. But it’s also possible that the reduction in cancer deaths has been offset by increases in deaths that result from adverse, unintended consequences of screening.
During the first decades of mass screening by Pap smear, doctors often overdiagnosed and overtreated women with suspicious findings. During the 1950s some clinicians routinely performed radical hysterectomies and even radiation therapy for pre-cancerous lesions: “…preinvasive epithelia carcinomas were treated as aggressively as the ‘most locally advanced but still operable cases’ – that is, with extensive elimination not only of the uterus, but also surrounding tissues and lymph nodes,” science historian Ilana Löwy has noted.
More recently, the PSA test for prostate cancer has come to exemplify the potential for harm. Less than one in three men with an elevated PSA level turn out to have prostate cancer at biopsy. Mass screening contributes to over one million prostate biopsies annually, biopsies that are associated with serious risks, including hospitalization and death, Prasad and co-authors note. “Moreover, men diagnosed with prostate cancer are more likely to have a heart attack or commit suicide in the year after diagnosis or to die of complications of treatment for cancers that may never have caused symptoms.”
Screening tests inevitably detect some tumors that do not need to be diagnosed. By definition, an overdiagnosed cancer is one that was destined to never cause harm because it is slow-growing, for instance, or unable to metastasize. Studies of cancer screening tests typically ignore overdiagnosis, Prasad and co-authors noted. Out of 57 studies, only 7 percent quantified overdiagnosis.
They argue that screening tests should be judged by their impact on overall mortality. That’s easier said than done. In the case of colorectal cancer, one group estimated that a clinical trial would need 4 million participants to demonstrate a reduction in overall death. The expense would approach $1 billion.
But conducting randomized trials within existing large national registries could dramatically reduce the cost, Prasad and co-authors say. And national screening programs could introduce randomized trials when they adopt new methods or make other changes. Australia and The Netherlands have shown that it is feasible and acceptable to people undergoing screening.
“We’ve assumed that cancer screening improves longevity. We haven’t proved it,” Prasad says. And without knowing the mortality benefits of screening, doctors can’t give people the information they need to make an informed choice. “We must be honest about this uncertainty,” Prasad says.
Why cancer screening has never been shown to “save lives”—and what we can do about it, by Vinay Prasad, Jeanne Lenzer, and David H. Newman. BMJ (2015)
Does screening for disease save lives in asymptomatic adults? Systematic review of meta-analyses and randomized trials, by Nazmus Saquib, Juliann Saquib, and John PA Ioannidis. Int J Epidemiol (2015)
Cancer, women, and public health: the history of screening for cervical cancer, by Ilana Löwy. História, Ciências, Saúde-Manguinhos (2010)
Assessment of changes to screening programmes: why randomisation is important, by Katy J. L. Bell, Patrick Bossuyt, Paul Glasziou, and Les Irwig. BMJ (2015)