Abstract
Purpose: To compare 3 different methods of pancreatic tissue sampling. Methods: All pancreatic mass bxs performed at our institution since 3/91 were reviewed retrospectively. We compared 3 diagnostic methods: EUS-guided FNA (22 ga via Pentax 32-UA echoendoscope), CT-guided needle bx/FNA (18-23 gal and operative transduodenal (Tru-cut) needle bx. Final dx was determined by resection, open bx, positive needle bx (presuming no false positives) or survival >1 yr without progression. Specificity of cytology was 100%. Results: Overall Results Biopsy Method Sensitivity NPV Accuracy EUS 83% (48/58) 38% (6/16) 83% (54/65) Radiological 76% (60/79) 24% (6/25) 78% (66/85) Surgical 72% (13/18) 38% (3/8) 81% (21/26) Combined 78% (121/155) 31% (15/49) 80% (141/176) Size <3 cm Biopsy Method Sensitivity NPV Accuracy EUS 65% (17/26) 31% (4/13) 70% (21/30) Radiological 72% (13/18) 0% (0/5) 72% (13/18) Surgical 63% (5/8) 0% (1/4) 67% (6/9) Combined 67% (35/52) 23% (5/22) 70% (40/57) Size ≥3 cm Biopsy Method Sensitivity NPV Accuracy EUS 97% (28/29) 0% (0/1) 97% (28/29)1 Radiological 78% (43/55) 25% (4/16) 80% (47/59) Surgical 80% (8/10) 50% (2/4) 83% (10/12) Combined 84% (79/94) 29% (6/21) 85% (85/100)2 1 p = 0.035 for comparison with CT. 2 p = 0.027 for comparison with <3 cm Conclusions: Needle sampling of pancreatic masses was useful for the confirmation of malignancy but not for it's exclusion (due to the low NPV). Methods were significantly more accurate with larger masses and, for these masses. EUS was significantly more accurate than radiologic sampling.
| Original language | English (US) |
|---|---|
| Pages (from-to) | AB149 |
| Journal | Gastrointestinal endoscopy |
| Volume | 47 |
| Issue number | 4 |
| State | Published - 1998 |