The purpose of this study was to test the hypothesis that when ultrasound (US) guidance is not feasible, abdominal biopsies can be performed safely and accurately under magnetic resonance imaging (MRI) guidance in a low-field environment. MRI-guided abdominal biopsy was performed on 31 consecutive patients, in whom US-guided abdominal biopsy was not possible because the lesion was not visualized in US (n=27) or an US-guided procedure was not considered safe (n=4). The locations of the lesions were liver (n=14), pancreas (n=6), lymph node (n=4), retroperitoneal mass (n=3), adrenal gland (n=3) and spleen (n=1). The average size of the lesion was 2.2 cm (range 1-4 cm) in maximum diameter. All procedures were done by using a 0.23-T open-configuration C-arm-shaped MRI scanner with interventional optical tracking equipment and software. Fine-needle aspiration (FNA) biopsy was performed on all 31 patients; 18 patients underwent both FNA biopsy and cutting needle core biopsy. Procedures were evaluated for diagnostic sensitivity, specificity and accuracy as well as procedure time and complications. The FNA biopsy specimens were adequate for interpretation in 27 (87%) of 31 cases. Two of these proved to be false-negative findings during follow-up or subsequent biopsy. The final diagnosis was malignant in 15 and benign in 16 patients. The sensitivity, specificity and accuracy of FNA biopsy were 71, 100 and 81%, respectively. Of the 18 core-needle biopsies, one was determined false-negative owing to nonrepresentativeness. The sensitivity, specificity and accuracy of histological samples were 90, 100 and 94%, respectively. The needle time was 19 min on average and the mean room time was 1 h 48 min. No immediate or late complications occurred. MRI-guided abdominal biopsy can be performed safely and accurately in a low-field environment in patients for whom an US-guided procedure is not feasible.