The aim of this study was to compare, prospectively, traditional pervaginal cul-de-sac aspiration cytology with an ultrasonographic-guided aspirate in the detection of residual or recurrent ovarian carcinoma. METHODS Fifty-one patients with ovarian carcinoma were monitored during chemotherapy (21 patients) or follow-up (30 patients) after first-line treatment. All patients underwent both traditional blind pervaginal cul-de-sac aspiration cytology and an ultrasonographic-guided pervaginal aspirate. The samples were classified as class 0 or insufficient when no mesothelial cells were detected in the aspirate. The results of cytological classification of the aspirates were compared with each other according to sampling order. RESULTS Samples were classified as class 0 in 56% when the traditional cul-de-sac aspiration was taken first, and in 73% when ultrasonographic-guided aspiration was taken first (p = 0.249, Fisher's exact test). The number of class 0 samples was smaller among those taken second than among those taken first (22 (44%) vs. 33 (65%), p = 0.046). Four recurrences were detected during the mean follow-up of six months (range 2-11 months) in 30 patients who were followed-up after the first-line treatment. In one case, a positive cul-de-sac cytology was the first and only early indication of recurrence. CONCLUSIONS The use of ultrasonography did not improve the accuracy of the cul-de-sac aspiration. The greater amount of fluid in the cul-de-sac during the second sampling might contribute to achieving a better result.