pathtalk.org is a weblog about pathology.

Lung Cancer Screening in the News

The National Cancer Institute has commissioned the National Lung Cancer Screening Trial, to the tune of 50,000 participants and $200 million. This huge and expensive trial comparing patients screened by CT scan with those screened by X-ray hopes to definitively answer the question of whether lung cancer screening is beneficial. The US Preventative Services Task Force currently rates the evidence for (or against) screening as “insufficient“. The study is to be completed in 2009.

However, motivated critics are not waiting to discredit the effort. The Lung Cancer Alliance has cried foul because two of the academics on the study have previously testified for tobacco companies regarding the lack of evidence for lung cancer screening:

Since late last year, the Lung Cancer Alliance, a Washington, D.C., nonprofit that supports screening, has asserted in letters to the NCI and its parent, the National Institutes of Health, that two of the study’s key researchers have conflicts of interest because they have accepted money from tobacco companies to be expert defense witnesses in lawsuits. The suits sought to force the companies to pay for annual CT screening.

The Alliance… also charged the study has design flaws that could bias its outcome against screening.

A few thoughts, in no particluar order:

• Nevermind that the LCA is partially funded themselves by GE, a maker of CT scanners. I always found it funny how groups lobbying in the “public interest” always get a pass on their funding sources, all the while pointing out conflicts of interest. Their money doesn’t appear out of thin air.
• Aside from monetary motivations, their ethics are often given a pass as well. I could make a good argument that it is unequivocally unethical for a group to advocate for a potentially harmful intervention that has not been shown to be effective.
• Their objections to the study design are laughable as well:
  

  The Lung Cancer Alliance and others also complain about the trial’s design, in part because it compares patients receiving CT scans with those receiving X-rays. If abnormalities are detected by X-ray and a CT confirms cancer, the X-ray, not the CT, will be credited with the discovery. Critics also say that 50,000 patients are too few to detect a benefit.

  50,000 subjects of a population at something like 7% lifetime risk of lung cancer is too small? The first studies of mammography were of similar size. The objection to the manner of comparison may be even more ridiculous - I didn’t know control groups were a bad thing, did you?

Having said all that, I can think of few better victories for preventative medicine if the results comes out in favor of screening. However, given the history of lung cancer screening, all the risks associated with lung surgery, and the many potential confounding lung nodules walking around out there, I wouldn’t put any money it. And true patient advocates shouldn’t either.

More here and here.

One of the objections I have heard to CT screening for lung cancer is that asymptomatic patients with early-stage tumors like the ones that would be detected by CT screening might not actually derive any mortality benefit from having them removed. I suppose the thinking there is that these small tumors may behave in an indolent fashion and death from other causes is likely to occur first.

I did a quick Pubmed search about this and found an interesting article published in Chest in 2003, in which the authors undertook a retrospective study of observation-only management of early stage unresectable lung cancers. Of the 128 patients they identified, 49 received no therapy whatsoever. Survival was poor (14.2 months), but interestingly to me, the cause of death was cancer in only 53% of this subset.

What it all means to me is still unclear. And of course, when you add the cost of millions of CT scans and the morbidity and mortality associated with additional diagnostic and therapeutic interventions to the equation, things get even more complicated.

There is no doubt that stage 1 lung cancer patints have a better prognosis than stage 2 and up with an approximately 70% long term survival. The 30% of patients that recur inevitably die. So what is different about those 30% of patients/tumors? Is it possible that a bad tumor (one that will kill you) is bad whether or not it is detected when it is big or when it is small? Will screening just decrease the survival rate for stage 1 cancers by detecting intrinsically bad tumors earlier? Gene expression studies showing that stage 1 tumors that recur have gene expression patterns similar to tumors that presented as stage 3 or 4 suggest that this may be the case. I will be curious to see the results, politics aside.

After a scan of the 2003 Chest article indicated above, it looks like that 53% is a little misleading. It looks like 53% died of cancer — but several were still alive, meaning (by my calculation) 65% of those who died had cancer as the cause of death.

It also raises the question of how cause of death was determined for those patients for the purposes of that study. At my institution I find clinicians signing death certificates heavily weigh their determination of cause of death towards final mechanism, while autopsy/forensic pathologists point more at the underlying “causative” condition. (For example, a pulmonary embolus versus complications of carcinoma.)

Either way, the data this study should be able to provide is likely to go a long way towards answering a lot of questions.