Serum levels of CA 125 and CA 15-3 were measured in 148 patients with primary endometrial carcinoma. CA 125-positive levels were found in 33 and 22% of the cases using a cutoff of 35 and 65 U/ml, respectively. Thirty-five (24%) and 12 (8%) patients had CA 15-3 levels higher than 30 and 50 U/ml, respectively. Among 144 patients with clinical stage I-II, 17 (12%) had extrauterine disease. CA 125 (> 65 U/ml) and CA 15-3 titers (> 30 U/ml) were found in 59 and 47% of occult stage III with respect to 16 and 18% in stage I-II of disease, respectively (P = 0.0001 and P = 0.01). The combined use of CA 125 and CA 15-3 resulted in a reduction of false-positive results of CA 125 with an acceptable sensitivity of 41%. Low-risk patients (G1 and M0-M1 tumors) showed a CA 125 positivity (> 35 U/ml) of 10% with respect to 37% of high-risk patients (G2-G3 and M2 tumors) (P = 0.0026). CA 125 positivity (> 65 U/ml) was 22% in patients without metastatic lymph node involvement, compared to 58% of cases with histologically positive lymph nodes (P = 0.022). A similar trend, although not statistically significant, was found for CA 15-3. A good correlation was found between CA 125 and CA 15-3 serum levels and clinical course of disease during chemotherapy. A statistically significant relationship was demonstrated between CA 125 (> 65 U/ml) (P = 0.0027) and CA 15-3 positivity (CA 15-3 > 30 and 50 U/ml) (P = 0.0004 and P = 0.00025) and a shorter survival. Our data show that CA 125 and CA 15-3 may be used as predictors of extrauterine spread and in monitoring of chemotherapy response in endometrial cancer. Moreover, the presence of elevated levels of these antigens may identify a subset of patients with a particularly poor prognosis.