Lung cancer screening, even with
computed tomography (CT) scanning, won't save any lives,
according to a large study published this month in the
Journal of the American Medical Association.
The finding comes six months after
another major study produced diametrically opposed findings.
The Early Action Lung Cancer Program (I-ELCAP) researchers,
who followed a whopping 31,567 asymptomatic smokers,
predicted last October in the New England Journal
of Medicine that CT screening of asymptomatic smokers
would prevent 80% of lung cancer deaths.
Each side of the screening debate
has brandished the findings that supported their positions,
while denigrating the methodology of the other. But
the truth is that neither study is definitive. We'll
have to wait for the results of a controlled, randomized
trial for that.
MOSTLY
HARMLESS
The JAMA study, led by Drs Peter Bach and Colin
Begg of New York's Memorial Sloan-Kettering Cancer Center,
followed 3,246 asymptomatic current or former heavy
smokers for an average of 3.9 years. It compared the
rates of diagnosis of early and advanced cancers, and
of death, to predictions generated by two validated
models for that risk group.
Dr Bach believes that his team's
findings show the Achilles heel of CT screening: it
detects small tumours that really weren't going anywhere
if left alone. That's why this study led to vastly more
diagnoses and resections without producing any change
in mortality. CT screening was assumed to be preferable
to chest x-rays, which generated too many false negatives.
But the ultra-sensitive CT scanners pick up tumours
that, in clinical terms, might as well be false positives.
There were 144 individuals diagnosed
with lung cancer compared with 44.5 expected cases,
a threefold increase. There was a tenfold increase in
lung resections over what the model predicted for an
unscreened population — 109 resections instead
of a predicted 10.9. But there was no decrease in advanced
cancers detected — 42 cases against an expected
33.4, nor in mortality — 38 deaths compared to
the predicted 38.8 deaths.
"Ours is the first study to ask
whether detecting very small growths in the lung by
CT is the same as intercepting cancers before they spread
and become incurable," said Dr Bach. "We found an answer
and it was No. Early detection and additional treatment
did not save lives but did subject patients to invasive
and possibly unnecessary treatments."
SURVIVAL
vs MORTALITY
So how did I-ELCAP arrive at the conclusion that 80%
of deaths could be prevented? It all comes down to an
overemphasis on survival rates as opposed to mortality.
I-ELCAP looked at survival time after diagnosis, and
found that 10-year survival was 88% among those with
stage I disease who underwent resection. Since 85% of
those whose cancers were detected by CT screening had
stage I disease, which is rarely detected in the normal
course of events, they hypothesised a dramatic improvement
in survival. But they never considered the overall death
rate, nor compared it to what would have been expected
in an unscreened population.
In fact the JAMA study achieved
equally good survival among patients resected for stage
I cancer after detection by screening. "However," the
authors note, "as our study illustrates, excellent survival
of a few individuals does not necessarily equate to
a benefit overall. (...) Despite routine screening,
most of the lung cancers that were ultimately fatal
were not detected until an advanced stage, or until
they caused death."
It must also be said that survival
rates alone are a poor indicator of benefit. Early detection
can increase apparent survival time with no benefit
to the patient: his survival is considered longer because
it is measured from the moment of diagnosis, but he
still dies at the same time.
Drs William Black and John Baron
of Dartmouth Medical Schools put it best in an accompanying
JAMA editorial. "Perhaps the best explanation
for the contrasting results... is the difference in
the primary outcome measures of the two studies: mortality
in the study by Bach et al versus survival in the I-ELCAP
study. While these outcome measures are often mistaken
to be complementary, prolonged survival in cases need
not imply reduced mortality in the population," they
wrote
WAITING
FOR WORD
A lot of specialists have been making the same point
lately, and it's clear that consensus opinion in the
field is happier with the methodology of the JAMA
study. But no one will really be satisfied until a controlled,
two-arm trial takes the question on.
That is now happening, with the
National Cancer Institute's vast 50,000 person, two-arm
trial, due to report in about three years. Nothing is
likely to move in terms of policy until those results
are in.
That goes for Canada too. The Canadian
Task Force on Preventive Health Care's lung cancer screening
recommendations, last updated in 2003, essentially agree
with Bach et al that no clear survival benefit has been
shown. But above all they bemoan the lack of head-to-head
comparisons. Until these are done, all data is potentially
flawed.
|
