To evaluate the technical success and diagnostic accuracy of image-guided percutaneous biopsy of the gastrointestinal (GI) tract. An interventional radiology database was used to retrospectively identify patients who underwent image-guided percutaneous biopsy of the GI tract. The medical and imaging records were reviewed to assess biopsy results and complications. Twenty-nine patients (M:F = 19:10, mean age 65 years) were identified who underwent image-guided (CT/US) percutaneous biopsies of the GI tract. All biopsies were performed using coaxial technique with 18 g core and 22 g FNA needles. Twenty-two biopsies (76 %) were for circumferential or segmental wall thickening; the remaining were wall thickening with exophytic mass. Mean tumor diameter was 7.4 cm (SD = 3.4 cm), average wall thickening was 2.1 cm (SD 7.2), and exophytic mass was 10.0 cm (SD = 4.2). Tumor locations included stomach (n = 5), small bowel (n = 8), and colon (n = 16). Malignancy was confirmed in 22 patients, 12 of whom underwent excision, with 3 false positive samples. Benign disease was diagnosed in 7 patients, including 3 with pathology confirmation, with one false negative sample. The overall sensitivity was 83 % and accuracy was 84 %. There was one complication presenting as bleeding at the biopsy site, treated conservatively. Percutaneous biopsy of the GI tract is an uncommon procedure. The results of this study suggest that it is a safe and sensitive procedure that may be considered for small bowel lesions in which endoscopy is not feasible, for submucosal lesions, or in the setting of patients with previously negative endoscopic biopsies.