BACKGROUND: The liver surgeon's decision to operate is based on imaging studies. However, no clear practical guidelines are available enabling surgeons to safely predict tumor-free margins after a partial hepatectomy. The aim of this retrospective study is to provide surgeons with simple and easily applicable practical guidelines. METHODS: We retrospectively stringently selected 42 anatomical right or left hepatectomies whose main characteristic was to pass along the median hepatic vein, which was preserved. This vein is an easily visualized anatomical landmark on preoperative imaging and is never transgressed by the surgeon. We compared the minimum distance between the tumor and this vein measured on preoperative imaging, and the minimum tumor-free excision margin measured on the specimen by the pathologist. RESULTS: The median tumor-free excision margin was 5 mm at pathological analysis, significantly different (P < .0001) from the tumor-free margin measured on preoperative imaging (15 mm). The mean difference between these two measurements was 10 +/- 4 mm (median, 9 mm). This difference was partly the result of the transection and partly the result of technical deviations in relation to the ideal resection line. CONCLUSIONS: The liver surgeon must consider that roughly a 5 to 8 mm tumor-free margin will disappear during hepatectomy when comparing measurements on the basis of preoperative imaging versus tumor-free specimen margins. If the histologically assessed minimum 2-mm tumor-free margin is added, the surgeon must plan to have a 7 to 10 mm tumor-free margin on preoperative imaging. However, few technical solutions exist that would enable the surgeon to increase the safety margin in borderline cases.