Materials and methods

We compared conventional ultrasonography and tissue harmonic imaging (THI) to evaluate the utility of THI in the study of patients who have undergone cholecystectomy. We consecutively examined 76 cholecystectomised patients (53 females, 23 males, mean age 56.9 (14.5, age range 20-81 years); 71% of the patients had undergone laparotomic surgery and 20% laparoscopic surgery. All the patients were examined after fasting for at least eight hours. The patients were first examined with conventional US and then with THI. Imaging was performed with a commercially available vectorial 2V4 transducer and a 512 Sequoia ultrasound unit (Acuson, Mountain View, California).

For each examination the operator optimised the scanning parameters for the two methods taking care to use the same focussing area. The harmonic images were acquired at a transmission frequency of 2 MHz and a reception frequency of 4 MHz. The conventional US image was obtained at a mean frequency of 3.35 MHz. Both the conventional and harmonic imaging examinations were recorded as clips on magneto-optical disks to allow blinded evaluation by the operator performing the examination and two other radiologists. Images were evaluated based on identification of the distal third of the common bile duct, better evaluation of choledocus content; resolution of anatomical detail to establish relationships between the choledocus, head of the pancreas, adjacent vascular structures and duodenal wall and, in the presence of bile duct lithiasis, the ability to identify the site, size and number of stones within the choledocus (table I). For each criterion a score from 1 to 4 was assigned depending on the degree of diagnostic confidence and conspicuity of the different structures. The results obtained by the three operators were analysed using the Wilcoxon sign rank tests. Interobserver agreement was evaluated using the weighted Kappa test: Cohen’s Kw [20-22].