Experience of the reader.
We have to keep in mind that virtual colonoscopy is not at all an easy technique. There is a considerable learning curve to become used to this new visualization technique of the colon.
The effect of experience has been a mater of debate at many conferences , including the annual International Symposia on Virtual Colonoscopy in Boston.
Recently a study has been conducted unther auspices of ESGAR with respect to the effect of reader experience: it was found that radiologists with specific interest in virtual colonoscopy having read hundreds of cases had better results than radiologists having just a training on a data set of 50 patients, who in turn, had better results than radiologists having no experience at all.
Other evidence of the importance of reader experience can be obtained by comparing results of different studies: there has been a great difference in the reported accuracies between recent major reports, varying from 52 to 92% for the detection of polyps > 1cm (Johnson et al. 2003; Cotton et al. 2004; Pickhardt et al. 2003).
Although many variables could have been detected possibly explaining all these' differences, it is widely believed that the differences in performance were at least partially attributed to the differences in reader experience. This hypothesis is sustained by the report by Spinzi et al., who found an improved sensitivity in 99 patients from 31.8% in the first 25 patients to 91.6% in the last 20 patients (Spinzi et al. 2001). Adequate and widespread access to reader training will therefore become necessary before the technique can gain acceptance as a screening tool..
Expert reading can be obtained by correlation with conventional colonoscopy or by second reading by recognized experts in the field of virtual colonoscopy.
Current recommendations are that radiologists should have expert reading on at least 50 to 75 cases. However, the number of cases will greatly be dependent upon the pathology : radiologists should have gone through different cases with different degrees of difficulties, e.g. normal studies, small , and larger polyps, sessile and pedunculated polyps, diverticular disease, flat lesions. It is therefore a matter of debate if 50 cases is enough.
Since this greatly depends upon the pathology we propose a minimum of 75 reading, guided by Virtual Colonoscopy Reading Centre.
The purpose is to learn how to recognize disease and to learn how to optimize the technique.
