In our recent newsletter, from December 2022, we highlight cancer screening and discuss a new study that touts the life saving magic of a little-used screening device for lung cancer, annual CT scans. In fact, this study is part of a long-term follow-up a prospective randomized trial that divides smokers and ex-smokers under age 75 into two groups: one receiving usual care, one getting lung CT scans every year. After twenty years, the small cohort who received screening did better than the placebo group. In fact, according to the new study, annual scans allowed potentially lethal cancers to be picked up early, and the death rate from lung cancer in the screened group was 20%, compared to 85% in the unscreened group. Impressive results, but what do they really mean?
As we have discussed, and as several of our videos elucidate, any statistics presented in percentage numbers are inherently deceptive. For instance, let’s say a wonder drug can reduce your chance of dying of cancer by 50%. That sounds amazing, but we have to ask, 50% of what? Let’s say that out of 1000 people who don’t take the wonder drug 3 people die after 10 years, but in people who take the wonder drug only 2 people die. That is a 50% reduction in death, but in fact only 1/1000 lives are saved after 10 years of treatment. We also would need to know if anyone suffered harmful side effects from the wonder drug to determine if it’s effective. By simply stating that the drug reduces death by 50% obscures the true risks and benefits of the drug.
So too is the case with lung cancer screening. When the initial screening trial (NLST) came out, similar exaggerated claims about its efficacy were reported, all by using percentage numbers. In fact, as I show in an AAFP article I wrote, and an NPR story in which I participated, the benefit of lung cancer screening is small. After 5 years of screening, 2.5/1000 lung cancer deaths were prevented, and these numbers held up after 20 years. Thus there is a benefit, but much smaller than the 80% survival statistic would let us believe.
How did the authors arrive at 80% survival? They don’t tell us, but it’s actually fairly simple to derive. It turns out that about 25% of lung cancers resolve on their own, and at least another 25% remain stagnant and cause no harm. Those cancers are all picked up by screening, are treated aggressively, and are cured. But they are non-lethal cancers, and thus “curing” them will inflate the survival percentage without leading to a clinically meaningful benefit, since if we never found those cancers they would not have killed anyone. Finding non-lethal cancers and claiming cure will certainly inflate the cure rate but will not add to clinical benefit.
In fact, finding and fixing non-lethal cancers in such a large numbers exposes people to unnecessary risk. That’s the main problem with lung cancer screening, and it’s a downside of all screening that the authors of this study fail to recognize. We have shown in our article that screening does find a small number of lethal cancers early and allow cure, again 2.5/1000 people benefit, which makes this screening test more effective than mammograms (1/1000 cancer deaths averted after a lifetime of screening), colonoscopies (1/1000 cancer deaths averted), and PSAs (no lives saved with screening). But how many people are harmed with lung cancer screening?
To start, the 50% of non-lethal cancers that are found and “fixed” means that a huge number of people are victimized by treatments they don’t need and that achieve no benefit and ample harm. While these people are told that a cancer was found and treated, that is fallacious and deceptive. Many such people receive surgeries, radiation, and chemotherapy for something that would have been harmless if left alone, and a measurable number of these people are killed and maimed by the treatment without deriving any benefit from it.
But that’s only the tip of the iceberg when it comes to lung cancer screening. Of people screened for 5 years, 750/1000 of them will be told that they may have a cancer even when they don’t have one, as these tests have huge numbers of false positives. This could lead to further radiologic tests with high radiation exposure, lung biopsies/surgeries that carry a 1% risk of death, and significant anxiety among those told they might have a cancer. In fact, because of the astronomical false positive rate, and the dangerous ramifications of telling people they might have cancer and they need more tests, most people stop getting lung cancer screening after they experience this, and most doctors stop ordering the test. Only about 5% of doctors will offer their patients lung cancer screening for reasons that should seem apparent.
And thus does this screening test help elucidate the problems with screening in general. Here are a few issues that are especially true of lung cancer screening, but which haunt all screening tests:
The studies conducted to assess the efficacy of screening are done in controlled university settings, utilize a highly selected group of patients not likely to be harmed by interventions, and are highly scrutinized. Once screening is introduced in the community setting, typically the false positive rates increase, the harms are higher, and the benefits smaller. We have no community studies of lung cancer screening even though Medicare has promised they are forthcoming.
·In all screening we will find and “fix” non-lethal cancers, thus harming a large number of people whose cancers are harmless. For lung cancer screening and mammograms, approximately 50% of cancers found would not have harmed the person screened, but they will likely still be treated and people will be harmed. In prostate screening with PSAs, 99% of cancers found are nonlethal if left alone but will lead to dangerous and potentially deadly treatments if found.
·All screening tests have false positives. For mammograms, the false positive rate is approximately 4%, for PSA’s it’s closer to 20%, and for lung cancer screening it is 75%. False positives lead to more testing, potentially to interventions and treatment that are harmful, and to severe anxiety. The false positive rate of lung cancer screening is medically unacceptable.
·The numbers used to tout the value of these tests are almost always exaggerated. Thus is the 20% reduction in breast cancer deaths from lifetime mammograms utterly deceptive, since as our NPR video shows, lifetime screening saves 1/1000 people and leads to a large amount of unnecessary testing and treatments that may kill more people than the test saves. The 80% benefit of lung cancer screening translates to 2.5/1000 people saved over two decades of annual testing and leads to a huge amount of harm in people who are both treated for non-lethal cancers or who are told they have cancer when they don’t.
·Screening becomes even less effective and more harmful as people age. That is why even groups that exaggerate the benefits of screening acknowledge that after age 75 the risks of screening drastically outweigh the benefits and, in elders, screening likely kills more people than it saves.
With all screening tests, we can find cancers and treat them, but that’s not the metric on which we should be focused. Rather, screening benefit must be presented in absolute terms, how many people out of 1000 are saved from the screen compared to people not screened and over how many years of screening is that benefit derived. Contrarily, we also must know how many people out of 1000 are harmed from screening or receive a false-positive result with its concomitant adverse impact. It is also crucial to know if the studies done on screening enroll patients and doctors who are reflective of the community at large and the patients we are screening particularly. And finally, how do these numbers change based on underlying risk of cancer death, on personal health habits, and on age.
Screening seems to make sense in an abstract way, but in most cases the benefits of screening are tiny and the risks could be large, especially as we get older. We may think our lives are saved with a screen, but most likely we would have been just fine without the screen and we are exposed to unnecessary risk and anxiety from the screen. Lung cancer screening is both the most beneficial and most harmful screen we offer, but to say that it is effective at curing 80% of cancers is irresponsible and deceptive and could lead many victims down a precarious road that they don’t fully understand until it's too late.