On the basis of 400 cases of liver masses, the authors have drawn conclusions that can be used to orient etiological investigations: --one or more well demarcated transsonic formations always correspond to a cystic etiology: --a hyperechoic metastatic liver points towards a primary cancer of the digestive tract whereas a hypoechoic metastatic liver is only rarely caused by such a cancer. The authors also propose a strategy for the exploration and surveillance of isolated hyperechoic or transsonic nodules of no more than 3 cm.