Wednesday, September 05, 2007

Early screening for lung cancer

A major risk that occurs in smokers and ex smokers is lung cancer. Are there steps that smokers or ex smokers can take to offset the effects of this terrible disease? The difficulty is that a disease like lung cancer can emerge in otherwise healthy people without early warning signs at all. Once it is diagnosed on the basis of symptoms the five-year survival rate is only 14% - it is a terrible prognosis and the leading cancer killer by a wide margin.

When a lung cancer is detected the tumour is ordinarily the size of a small orange and the cancer has usually already spread. An x-ray can reveal a cancer the size of a grape but a CAT scan can reveal a cancer tumour no bigger than a grain of rice. When such small cancers are detected the five-year survival rate jumps from 14 to 70%. Henschke et al. in an article in The Lancet in 1999 report on checks of 1000 smokers and ex smokers aged 60+ who showed no cancer symptoms using CAT scans and found 27 tumours of which 23 were early cancers. X-rays identified only 4 of these cancers.

This type of early diagnosis seems a possible way of significantly reducing mortality among those who have ever smoked. I would be interested in comments from those with a medical background - I don't have one - on this idea and particularly on the economics of such screening. Some useful websites I found are here and here. An excellent article with an extensive list of technical medical followup studies is here.

The literature is mixed with some sources claiming no beneficial effects and others stating that much can be achieved. Recent studies have suggested that 5-year survival rates of those identified using CAT scan procedures as suffering from early-identified cancers that are then surgically removed increases to 92% with the cost of a scan being between $200-$300US. There was one cancer victim detected for every 65 individuals tested in this sample of adults aged 40+ so the cost per cancer victim identified is over $16,000 although this could be reduced if the sample could be more effectively pre-screened by age and perhaps smoking history but even without this qualification the cost does not seem prohibitive.

The case for seeking to early screen for lung cancer is controversial. One claim is that there is over-diagnosis of cancers (many false positives) as a result of their procedures. In defence of the screening procedures however almost all of those identified as having small cancers who did not have them removed subsequently died within 5 years.

I have been trying to learn a bit about lung cancer. A useful guide to those suffering from this affliction – and those interested in dispelling myths - is C.I. Henschke, P. McCarthy and S. Wernick’s, Lung Cancer. It is full of interesting facts. It explains how lung cancers occur and how sufferers should seek treatment and perhaps how they should deal with the prospect of imminent death. It is a direct though very compassionate book.

I was interested in the argument that low tar cigarettes might be causing the adenocarcinomas in the smaller airways and alveoli in the lungs because people draw on these cigarettes more deeply.

To answer the question I posed initially – if you do smoke how should you act to reduce your risk of dying of lung cancer? The authors suggest that you make every attempt to quit smoking but otherwise take a low dose CAT scan annually.

What about ex smokers? The authors themselves only screen patients who are aged 40+ with at least a 10 pack-year smoking history. (Your pack-year history is the average number of packets you have smoked * number of years you have smoked). Quitting at any age increases your longevity – it is never too late to quit - but if you have ever smoked your risk of eventually suffering from lung cancer is permanently raised. About half of those suffering lung cancers in the US are ex smokers so the ideal response to the lung cancer issue is simple: If you don’t smoke don’t start and if you do smoke quit!

Update: The references Francis Xavier Holden provides in his comment, and the summary comments he makes, suggest it is too early to conclude that CAT scans are useful for the early detection of lung cancers. This seems a fair summary to me. The scans often identify slow-growing tumors not associated with cancer deaths. Faster growing malignancies are likely to occur between scans. Further research underway using very large scale surveys will report results in 2010. Interestingly, regular x-rays seem to increase lung cancer risks presumably through adverse effects of the x-rays themselves. Ex-smokers and current smokers in their 50s might not reduce their mortality by having regular CAT scans and regular, high-frequency x-rays may prove counterproductive.


Anonymous said...

The simple solution to start with would be people to pay for this test themselves -- it seems like a good investment if you smoked. Somehow I doubt that its in human nature however (especially for men if you look at health seeking behavior of males).

hc said...

Conrad, I agree. This is a 'screening' procedure rather than a treatment so I would expect it not to be covered by health insurance. In most cases it is not covered in the US unless someone has cancer and the CAT scans are being used to test the status of the disease.

Anonymous said...

I notice one report claims that it is about the same cost as breast cancer screening with equal or better results. But I can't find anything to back that claim up.

I somehow doubt it.

The cost quoted, $200-$300US, looks to me as if it might be the marginal cost of one CT scan, not the per person cost of a proper formal public screening program. A screening program is much more than the cost of a CT scan.

I'd like to see an epidemiologist and say a health economist have a go at costings/ benefits for a formal screening program. Theres skills /workforce /availability issues if you are goin to,say, run 50% of the adult population through the screening.

Its not only imaging for lung cancer that has a lot of false positives. Imaging in general tends towards false positives especially, paradoxically, as it gets better and clearer pictures of smaller areas.

Control studies of "healthy" or non symptomatic individual's images in lower back pain, as an example, show the same "abnormalities" as those with back pain.

I'd like to see studies of imaging of lungs in large numbers of "healthy" non smokers and also "healthy" ex smokers.

That said: I'd be inclined to say something like this:
If you are over 50 yo (or even 40)
Have at some period been an everyday smoker

You might find it a worthwhile investment in early intervention to get a CT lung scan.


If you have some breathing problems or coughing or URTI

Its probably a good idea to have a CT lung scan instead of a x-ray.


Be aware that there is a large chance of false positives and be prepared for further tests to confirm or negate the presence of malignancy.

All of the above is off the top of my head after scanning your links. I haven't looked at Cochrane or anything meta .

hc said...

fxh, I hear what you are saying but have a query. Is a one-off test much use? If you have cancer and it is pea-sized and can be removed with a fair chance of recovery then that's great. But isn't ther probability of that low - more likely you will observe a large cancer that has metastasized or no cancer at all. After all don't these cancers grow quickly?

Hence if you are an ex heavy smoker don't you need regular - say annual - testing?

The point about false positives is important - I understand this occurs with many other cancer tests.

Anonymous said...

harry as you probably know the whole to screen or not to screen decision is a million times more complicated than most people think. And its based ona lot of highlevel stats /probabilities which in turn have to be based on large number studies over a wide range of cohorts and combined with elapsed time. Not to mention controls etc and thats even before you get to the economics for populations.

But yes - I'd say in very general terms if the scan picks up a pea sized malingnancy (and assuming it can be safely and timely excised - not neccessarily a given)then it matters how long that lump takes to be "scannable".

Say it only takes a month of growth to get to see-able size, then in theory all smokers and ex smokers, say 40+, would need to be screened every month. However, it is extrememly unlikely that the surgery could be done within a month - even in a "perfect" health system.

So - a sensible achievable screening would have to decide what size is optimum for intervention, how long it takes to grow to that size, what the % risk is of leaving it till that size or larger, the time to surgery etc etc.

All very complicated.

Further. The risk of re-occurance even after surgery is probably reduced by ongoing treatment so that will be increased costs to be allocated to the screening.

Possibly theres still a greater value in QALYs but its certainly not straight forward for a population.

Its even more complicated than it might seem for an individual who can afford a CT scan every so often. But maybe that every so often is once a month?

Anonymous said...

Cochrane; but its only up to 2003 .

Not enough evidence to support regular screening for lung cancer

Lung cancer is the most common cause of cancer related death in the western world. It takes about 20 years to develop and cigarette smoking is a known cause. Most lung cancers are not found until they are advanced but regular screening is offered to those considered at high risk of the disease. The review of trials found early detection methods such as chest x-ray, testing sputum or CT scan do not appear to have much impact on either treatment or number of deaths from lung cancer. The review found frequent chest x-ray may cause harm. More research is needed.

JAMA MArch 2007

Conclusions Screening for lung cancer with low-dose CT may increase the rate of lung cancer diagnosis and treatment, but may not meaningfully reduce the risk of advanced lung cancer or death from lung cancer. Until more conclusive data are available, asymptomatic individuals should not be screened outside of clinical research studies that have a reasonable likelihood of further clarifying the potential benefits and risks.

Anonymous said...

Harry - this is the best summary both in plain terms and scientific I can find:

Dr. William C. Black, professor of radiology at Dartmouth Medical School explains why early detection of lung cancer may not be beneficial:

Anonymous said...

One of the significant points is that especially in Cancer studies survival statistics are often misleading. populist calls for screening will always ignore these issues.

1. Lead time bias. Cancers may be found earlier, but the patient doesn't live longer than he/she would have without the screening diagnosis.

2. Length bias. Screening tends to detect slowly growing tumors. Rapidly progressing tumors tend to show themselves by causing symptoms in between screening rounds or after the patient has stopped getting screening CT scans.

3. Overdiagnosis bias. Screening detects tumors that would not have become clinically significant before the patient dies of other causes.

4. Survival statistics may only pertain to the small minority of screened individuals who are diagnosed with the cancer through screening, not those diagnosed outside of screening programs.

Anonymous said...

I think you need to look at the average time tumours start spreading -- this differs an enormous amount between types of cancers. This like testicular cancer, for example, only take about 70 days, but other types of cancer are much slower (like prostate).

hc said...

fxh, I have now had a chance to go through the links you have provided. Thanks they were very good.

My own reading of the evidence is that CAT scans tend to pick up slow-growing tumors that are not associated with cancer deaths. The fast growing ones are likely to develop between scans - as I suggested above.

Note that one study suggests frequent x-rays increase cancer risks because of adverse effects of the x-rays.

This has been a very productive post for me at least - I learnt a lot.

Anonymous said...


I was under the impression that you can have a tumor, and it can sit there for quite some time and grow very slowy (your point 2). It can then become aggressive later on and start growing faster and spread to other organs (metastasis, what's basically benign today isn't neccesarily so tommorow, which is why you get the benign ones chopped out if you can) -- this is why picking up slow growing tumors is not a bad thing and is is what you are mainly screening for, the cancer before it becomes fast growing. Only some types of cancers basically start off as nasty or have very fast growth trajectories (like testicular cancer, may sure your squeeze them each month :).

THere are also pre-cursors to cancer you can pick up that hang around for ages in some cases. Pollups and bowel cancer are the obvious case which people don't get screened for enough (probably because the procedure is aweful and because there is only some chance they turn into cancer). I think it all depends on the type of cancer and what trajectories they typically follow.

hc said...

conrad, I am right out my depth here but my understanding is you may be right. Some grow slowly then take off. Some grow slowly or not at all and never take off. Inspecting tumors by a CAT scan and operating to remove small tumors will help to reduce mortality if the first type of tumor is identified but not the second.

One of FXH's hyperlinks looks at evidence of success by comparing the mortality of a population who have had the procedure with the mortality you would expect were the procedure not applied. There is no improvement.

It would be nioce if we could get an onctologist surgeon to comment.

Anonymous said...

Actually, now I remember, there is some person who is visiting our department for a few months -- she was talking about genetic markers for lung cancer susceptibility and their interaction with various food types (eat brocolli! -- although I'm not sure if this was used just as an example). In case I am feeling extroverted, I might ask her about it (although I think she is into the ethics of the issue, not the issue -- but she might know the answer you want).

Anonymous said...

It would be nioce if we could get an onctologist surgeon to comment.

Unfortunately this isn't necessarily so at all. Oncologists, even physicians, have an inherant bias towards action. Surgeons have an increased bias toward action and that action is almost always surgery.

Like everyone else oncologists are flawed and make mistakes like ..."jumping to conclusions, seeing what you expect to find, availability/non-availability heuristics, bias toward action rather than inaction, confirmation bias, diagnosis momentum, fundamental attribution error, gambler’s fallacy, omission bias, order effects, overconfidence bias, representativeness restraint, Ockham’s razor error, triage cueing and counter transference."

The most rational interpretation would come from epidiemiologists working with public health, wait for it, economists.

Anonymous said...

conrad and hc

Some malignancies (or any illness really) can start small and take ages to grow and become harmful and the lead time is years. Others start small and quickly become big and deathly. [Note here that BIG doesn't always mean deadly just as SMALL doesn't always mean not deadly]

Other things hibernate so to speak for ages then accelerate to nastiness.

With regard to lung cancer I don't know. (I haven't looked)