LEARNING CURVE OF EUS-GUIDED TISSUE ACQUISITION, A TIMELAPSE STUDY AT A NEWLY ESTABLISHED EUS SERVICE

Aims EUS is gaining ground in todays diagnostic routine due to its ability to get dynamic, accurate representation of peri-gastro-duodenal structures, but mostly due to its capability of obtaining tissue samples amenable to histopathologic studies.

Methods The accuracy of sampling pancreatic malignant tissue through EUS-FNA of a novice trainee at a tertiary center was retrospectively assessed over the first 150 procedures, spread over 25 months. Three tertiles of 50 patients each were used to define the measured parameters.

Results We detected increasing accuracy of tissue acquisition associated with growing number of procedures per month. This proportionality was assessed by the parameters of diagnostic accuracy with slight increase in the second and third groups as compared with the first one (from 82 % to 90% and 90% respectively), number of passes declining from an average of 3 in the first group to about 2 in the second and third tertiles, better evaluation of invasiveness of the primary tumor or metastatic spread as compared with CECT or MRI imaging. A total of 45 patients in the second and third groups were correctly diagnosed by both EUS and CT/MRI, up from 41 in the first group. Four patients in the third group were misdiagnosed by CT or MRI, up from 1 in the second and none in the first groups, but with adenocarcinoma found at EUS. Four patients in the first group and only one in the third were misdiagnosed by all imagistic methods, but interative imaging ultimately led to surgery or definitive diagnostic. Altogether, a total of 40 patients were eventually sent to surgery, 15 of which were offered curative surgery by means of Whipple operation, without any relation to diagnostic accuracy.

Conclusions We found that the EUS operator reaches pretty readily a plateau when it comes to acurrately obtaining malignant tissue.