Soft tissue sarcomas often present as large and histopathologically heterogenous tumors. Proliferation has repeatedly been identified as a prognostic factor and immunostaining for Ki-67 represents the most commonly used proliferation marker. There is, however, a lack of consensus on how to evaluate Ki-67 staining regarding optimal cut-off levels, selection of tumor areas, and the number of tumor cells to evaluate. We assessed the impact of targeting peripheral versus central tumor areas using tissue microarray-based staining for Ki-67 throughout the tumor diameter in 25 leiomyosarcomas. In 18/25 tumors, Ki-67 expression was higher in the tumor periphery. If 10% staining tumor nuclei was used as cut-off and the maximal Ki-67 staining section in the tumor periphery was considered, 21/25 tumors would have been classified as highly proliferative compared to 14/25 if the tumor center had been analyzed. Similar results were obtained also when higher cut-off levels were used and if the mean expression rather than the maximal expression was considered and the differences were neither caused by necrosis nor by hypoxia (assessed as HIF-1alpha expression). Our findings suggest that the determination of proliferation in soft tissue sarcomas should be standardized for clinical application of Ki-67 as a prognostic marker.