Interventional magnetic resonance (IMR) machines have produced unique opportunity for image-guided surgery. The open configuration design and fast pulse sequence allow virtual real time intraoperative scanning to monitor the progress of a procedure, with new images produced every 1.5 sec. This may give greater appreciation of anatomy, especially deep to the 2-dimensional laparoscopic image, and hence increase safety, reduce procedure magnitude and increase confidence in tumour resection surgery. The aim of this paper was to investigate the feasibility of performing IMR-image-guided general surgery, especially in neoplastic and laparoscopic field, reporting a single center -- St. Mary's Hospital (London, UK) -- experience. Procedures were carried out in a Signa 0.5 T General Elettric SP10 Interventional MR (General Electric Medical Systems, Milwaukee, WI, USA) with magnet-compatible instruments (titanium alloy instruments, plastic retractors and ultrasonic driven scalpel) and under general anesthesia. There were performed 10 excision biopsies of palpable benign breast tumors (on female patients), 3 excisions of skin sarcoma (dermatofibrosarcoma protuberans), 1 right hemicolectomy and 2 laparoscopic cholecystectomies. The breast lesions were localized with pre- and postcontrast (intravenous gadolinium DPTA) sagittal and axial fast multiplanar spoiled gradient recalled conventional Signa sequences; preoperative real time fast gradient recalled sequences were also obtained using the flashpoint tracking device. During right hemicolectomy intraoperative single shot fast spin echo (SSFSE) and fast spoiled gradient recalled (FSPGR) imaging of right colon were performed after installation of 150 cc of water or 1% gadolinium solution, respectively, through a Foley catheter; imaging was also obtained in an attempt to identify mesenteric lymph nodes intraoperatively. Concerning laparoscopic procedures, magnetic devices (insufflator, light source) were positioned outside scan room, the tubing and light head being passed through penetration panels. Intraoperative MR-cholangiography was performed using fast spin echo (SSFSE) techniques with minimal intensity projection 3-dimensional reconstruction. About skin sarcomas, 2 of them were skin recurrences of previously surgically treated sarcomas (all of them received preoperative biopsy) and the extent of the lesion was then determined using short tau inversion recovery (STIR) sequence. The skin was closed in each case without need for any plastic reconstruction. The breast lesions were visualized with both Signa and real-time imaging and all enhanced with contrast: 2 (20%) were visualized only after contrast enhancement; intraoperative real time imaging clearly demonstrated a resection margin in all cases. Maximum dimensions of breast specimens (range 8-50 mm, median 24.5 mm) were not significantly different from those measured by Signa (p>0.17, Student's paired t-test) or real time images (p>0.4): also there was no significant difference in lesion size between Signa and real time images (p>0.25). All postprocedure scans clearly demonstrated complete excision. The extent of the tumor at MR imaging was greater in each case than suggested by clinical examination. Adequate resection margins were planned using STIR sequences. Histological examination confirmed clear surgical margins of at least 1 cm in each case. During right hemicolectomy, both intraoperative SSFSE and FSPGR contrast imaging revealed the lesion and details of the colonic surface; imaging of the lymph node draining right colon was only partially successful, due to movement artifact. Concerning laparoscopic procedures, both FSE and SSFSE techniques produced reasonable images of the gallbladder and intrahepatic ducts, but the FSE imaging was of poor quality due to respiration artifact; however, SSFSE allowed visualization of the gallbladder and part of the common bile duct. About skin sarcomas, the extent of the tumor at MR imaging was greater in each case than suggested by clinical examination and in each case the complete tumor excision was confirmed. Histological examination confirmed clear surgical margins of at least 1 cm in each case. Intraoperative MR scanning reliably identifies palpable breast tumours and skin sarcomas and is sufficiently accurate to guide their surgical excision. Further work may be done to develop laparoscopic and open abdominal surgery as well.