Indeterminate findings on imaging-guided biopsy: should additional intervention be pursued?

OBJECTIVE: Our purpose was to determine the outcome of patients in whom the results of imaging-guided biopsies were indeterminate and to examine factors that may affect outcome. MATERIALS AND METHODS: During a 25-month period, 619 consecutive imaging-guided biopsies (CT, n = 268 [43%]; sonography, n = 351 [57%]) were performed on abdominal lesions. Of these biopsies, findings from 454 biopsies (73%) were positive for malignancy, findings from 21 biopsies (3%) yielded a benign diagnosis, and findings from 15 biopsies (2%) were nondiagnostic. Findings from the remaining 129 biopsies (21%) were considered indeterminate because the pathologic diagnosis revealed benign tissue, inflammation, stromal elements, or atypical cells. The frequency of malignancy in lesions of patients in this indeterminate group was determined. RESULTS: Thirty-eight of the 129 patients were lost to follow-up. Of the remaining 91 patients, 55 (60%) had true-negative (benign) results, and 36 (40%) had false-negative (malignant) results. A history of malignancy had no significant effect on the true- or false-negative rate (p = .799). However, the guidance technique had a statistically significant effect: CT and sonographic guidance yielded 25 (49%) and 11 (28%) false-negative results, respectively (p = .037). The number of needle passes, needle type and gauge, and biopsy site did not have a statistically significant correlation with the false-negative rate. The presence of atypical cells in the cytologic aspirate was highly predictive, with a 71% false-negative rate (p = .008). CONCLUSION: We found a high incidence of malignancy (40%) in lesions deemed indeterminate at the time of biopsy. Thus, close follow-up or additional intervention should be pursued in such cases. Sonographic guidance resulted in a statistically significant decrease in the false-negative rate when compared with CT guidance; however, a history of malignancy, the biopsy site, the needle gauge and type, and the number of passes did not have a significant effect on the false-negative rate.

Evaluation of a breast biopsy phantom for learning freehand ultrasound-guided biopsy of the liver.

RATIONALE AND OBJECTIVES: We evaluated whether a breast biopsy phantom device would aid in the development of skills in freehand ultrasound liver biopsy. METHODS: Three radiologists who were inexperienced in freehand biopsy of the liver were observed. Each radiologist was timed and scored during attempts to biopsy lesions created in a beef liver. The time required for biopsy and the success of each pass was recorded. A commercially available breast biopsy phantom was then used by each of these radiologists during two 20-min practice sessions. Posttraining testing on the beef liver was performed in the same manner as pretraining testing. RESULTS: Freehand biopsy practice using the breast biopsy phantom resulted in a reduction in the mean time required for biopsy from 32 to 17 sec. Each of the three subjects reduced the mean time required for successful biopsy after training using the breast biopsy phantom. The total number of lesions missed was reduced from 14 of 43 to 0 of 45. CONCLUSION: Practice using the ultrasound breast biopsy phantom improves performance in freehand ultrasound biopsy of the liver.