The use of computed tomography (CT) has seen enormous growth over the past decade.
In the US, approximately 63 million examinations were performed in 2005 (Niagara
Health Quality Coalition 2004) compared to 35 million in 2000. The increased number
of clinical applications (e.g., in emergency and trauma, paediatric, cardiac, and vascular
disorders) made possible by the fast scanning capabilities of multidetector CT
(MDCT) will drive even greater growth.
CT is already the main cause of radiation dose to the US population (Wiest et al.
2002; Mettler et al. 2000), and this will surely increase as the number of examinations
per patient increases. This is a serious concern with which the radiology community is
now confronted. The signifi cant uncertainty associated with radiation risk estimates,
long delays between exposure and cancer manifestation, and the fact that carcinogenesis
is proved by statistical inference rather than by direct observation tend to reduce
the perceived urgency to reduce radiation dose delivered by CT. However, the radiology
community needs to be made aware that the small but acceptable risk–benefi t decisions
made at the individual patient level are amplifi ed by the huge number of CT procedures
performed each year. In a recent report on the biological effects of ionizing radiation
(Beir 2005), the overall probability of death due to a solid tumor induced by a single
10-mGy CT examination is estimated to be approximately 0.00041. This apparently
very low risk – multiplied by the 63 million CT examinations performed each year
– suggests in fact that 25,420 fatal cancers are induced by CT every year.
This calculation, however, has a number of major fl aws. The most important fl aw is
the fact that the risk factors were derived for generally healthy individuals in the population
of Japanese A-bomb survivors, whereas patients who undergo CT are usually
older and have a lower life expectancy than those in the general population. Moreover,
the health benefi t of CT-derived diagnostic information is immediate, whereas the risk
of induced cancer is decades away. Nevertheless, this mathematical calculation was
meant to underscore the importance of restraint in the use of MDCT.
Given these fi gures, what are radiologists supposed to do? Should they refuse to perform
CT examinations on the patients referred to them? Conservative estimates of the
benefi t-to-risk ratios for CT are 100:1 and even higher. This discussion does suggest,
however, that CT should not be performed for dubious or trivial clinical indications.
Appropriateness criteria need to be vigilantly applied for all patients referred for a CT
examination. Appropriate medical training in radiation risk management would be
helpful in reducing the number of inappropriate requests for CT examinations. Academic
radiologists should push for this training and organize dedicated lectures in
medical schools. In training hospitals, CT examinations requested by young residents
should be approved by senior physicians.