OBJECTIVE Successful surgical treatment of patients with retroperitoneal soft tissue sarcoma (RSTS) is based on pre-operative planning that starts with a correct pre-operative diagnosis. In a population-based study, we determined which patients were initially treated for assumed other conditions. The effect of an erroneous diagnosis on the installed treatment was analysed. METHODS With the help of the Dutch Network and National Database for Pathology (PALGA), data were collected on 143 patients in the Netherlands in whom a primary RSTS was found and confirmed histologically between 1 January 1989 and 1 January 1994. Satisfactory clinical information was obtained on 138 patients, 64 males and 74 females (54%). The median age was 60 (range 18-88) years. RESULTS At the time of actual treatment 37% of the patients with RSTS were assumed to have another disorder (group 1 n=51), whereas 87 patients were diagnosed as having RSTS (group 2). In group 1, an acute presentation was more common (18 vs 2% P=0.002), and the tumour was less often palpable at physical examination (43 vs 69% P=0.004), while clinical work-up less frequently included CT-imaging (57 vs 89% P<0.001) and a biopsy (29 vs 77% P<0.001). Although tumours in group 1 were smaller (median diameter 13 vs 19 cm P<0.05), this was not reflected in a better operative result: less patients underwent complete tumour resection (51 vs 57%) and more patients underwent surgery for tumours that proved to be irresectable (14 vs 1% P=0.004). CONCLUSIONS (1) More than one-third of patients with RSTS are misdiagnosed and inappropriately treated; and (2) biopsies and cross-sectional imaging improve diagnosis.