It was only a couple of months ago that an article was published in the New England Journal of Medicine that claimed spiral CT scans for the early detection of lung cancer could significantly decrease deaths from this disease.
Now, a study reported in this week's Journal of the American Medical Association says that there is no evidence that screening for lung cancer with chest CT scans does anything to reduce deaths from lung cancer.
It does substantially increase the number of cancers detected and the number of surgeries performed, and exposes patients to significant risks for no net gain in survival.
Why all the confusion?
This past October, the New England Journal study indicated spiral CT scans were able to reduce deaths from lung cancer by finding nodules early. The survival rates of the patients with lung cancers found had been outstanding, according to the report --much better than had ever been seen previously.
Although hailed by many experts and advocates as the answer to our prayers to reduce the suffering and loss of life from this terrible disease, there were those at the time who felt the study did not provide sufficient evidence to recommend universally that current and former smokers rush out to get screened.
Left unanswered in that study was the possibility that the investigators were diagnosing lung cancers that never would have made a difference to the patient.
That's hard to believe, but it is a real concern.
More Screening, More Surgery
This week's JAMA report adds some credibility to that argument, claiming that the evidence does not support the conclusion that lung cancer screening with CT scans reduces deaths from lung cancer.
In this study, the authors reported on the results of spiral CT scanning performed on over 3,200 patients screened at three academic medical centers, two in the United States and one in Italy.
They carefully examined the number of lung cancers diagnosed, the number of surgeries performed, the stage of the lung cancer at diagnosis and the rates of death from lung cancer in the patients who were screened. They then compared these rates with a set of data based on previous lung cancer trials to find out how many surgeries and deaths were observed compared to how many were expected.
The results were astonishing, to say the least.
In patients screened with spiral CT scans for the early detection of lung cancer, lung cancers were diagnosed three times more than would have been expected.
The number of surgeries to remove a lung cancer was 10 times (yes, 10 times) greater than what was expected.
Even with all of those lung cancers that were diagnosed and all those surgeries that were performed, there was no evidence that fewer people were diagnosed with advanced lung cancers than were expected.
Most important, the number of deaths from lung cancer was almost exactly the number of deaths predicted.
Scan Detects Nondeadly Tumors
We have always thought of lung cancer as an almost uniformly fatal disease. That's because chest X-rays don't find these small nodules early, and most patients aren't diagnosed with lung cancer until they develop symptoms. When that happens, the disease is usually advanced and beyond the hope of cure with currently available treatments.
The spiral CT scan gave new hope to those who thought that finding lung cancers early would improve the outlook for patients with this disease. These scans can find small lesions in the lung earlier than chest X-rays. The thinking has been we should be able to do for lung cancer what we've done for breast cancer using mammography.
But here's the rub: More recent evidence suggests that there may be lung cancers present that do not cause a person significant problems through his lifetime. This is not unlike the situation that occurs with prostate cancer, where many of the cancers we diagnose through screening -- and treat -- may never cause a man harm or death.
The problem is that we can end up treating many more people than really necessary to help a few.
In fact, as pointed out in an editorial in the same issue of JAMA, the cancers diagnosed by CT scans are indeed slow growing. That may explain why they are less likely to cause symptoms and harm, perhaps even death.
If CT scans were catching tumors that were precursors of significant, even deadly cancer, then we would have expected to see the risk of dying from lung cancer diminish over time after CT scans started being used, when an initial rise in the number of lung cancers was detected. That simply did not happen in this study.
The editorial concluded that we need to be cautious before we wholeheartedly and enthusiastically embrace CT screening for the early detection of lung cancer.
If this study is correct, the practical implication is that we would see many more people diagnosed with lung cancer.
Already the American Cancer Society predicts there will be 213,380 people diagnosed with lung cancer in the United States in 2007, and 160,390 deaths.
If you apply the findings of this study, we theoretically could diagnose more than 500,000 people with lung cancer (assuming every smoker and former smoker was screened, which is not likely to happen), and still see the same number of deaths.
As a result, we would have many more lung cancer survivors -- and many more people grateful that we were able to save their lives.
But the fact is we would have done nothing to reduce the number of people who died from the disease, and that is really where we need to focus our efforts.
A Tale of Two Studies
I suspect that, in the end, neither the optimism of the first report nor the pessimism of the second report will prevail. There are other trials under way that I believe will give us a better insight into whether or not screening has a benefit in decreasing deaths from lung cancer.
In the meantime, if you are a heavy smoker or a former heavy smoker and want to get a chest CT scan to look for early lung cancer, it is important that you know the facts as they exist. And, you need to have a very careful and complete discussion with your health care professional regarding the potential risks and benefits of getting screened.
But for the nation as a whole, it is simply too premature to make a blanket recommendation that everyone at risk be screened.
We look forward to more information from the ongoing clinical trials to provide us with the guidance that we really need to answer this paradox and resolve the dilemma these two studies have created.
Dr. Len Lichtenfeld is deputy chief medical officer for the American Cancer Society. You can view the full blog by clicking here.