The endometrium in asymptomatic breast cancer patients on tamoxifen: value of transvaginal ultrasonography with saline infusion and Doppler flow.

OBJECTIVE: To define by transvaginal ultrasonography an optimal cutoff for endometrial thickness measurements to be used in screening for endometrial pathologies in asymptomatic breast cancer patients on tamoxifen, and to evaluate the incorporation of saline infusion sonohysterography and Doppler studies into the diagnostic scheme.

METHODS: Sixty tamoxifen-treated women examined by transvaginal ultrasonography with saline infusion were included in this retrospective study. Variables of interest were endometrial thickness and texture, and the presence of intracavitary fluid at ultrasonography, total endometrial thickness, defined as the sum of the two endometrial layers and the presence of polypoid masses at sonohysterography, and uterine artery flow indices at Doppler ultrasonography. The dilatation and curettage performed after the sonographic scan detected pathological endometrial changes in nine cases, including six endometrial polyps, two endometrial hyperplasias, and one endometrial cancer. All parameters evaluated were compared between patients with benign and pathological endometria. Continuous variables that differed significantly between the groups were investigated further by receiver operating characteristics curve analyses and the diagnostic value of combinations of various parameters by binary logistic regression.

RESULTS: The endometrial thickness in patients with proven endometrial pathologies was significantly greater compared with women with benign endometria, both by transvaginal ultrasonography (12.7 +/- 5.5 vs. 7.0 +/- 4.5 mm; P = 0.003) and by sonohysterography (6.3 +/- 2.8 vs. 4.1 +/- 1.7 mm; P = 0.036). While saline infusion sonohysterography also revealed a significantly higher frequency of polypoid masses in the former group (67% vs. 2%; P < 0.001), no other significant differences were defined between the groups in regard to any other sonographic or Doppler parameter evaluated. For the diagnosis of any endometrial pathology, the optimal cutoffs of endometrial thickness at ultrasonography and total endometrial thickness at sonohysterography were 9.5 and 5.5 mm, with sensitivities of 89% and 78% and specificities of 78% and 84%, respectively. A logistic regression model including polypoid lesions (B = -4.935; P < 0.001) and total endometrial thickness at sonohysterography (B = 0.432; P = 0.027) as the only two independent variables had a sensitivity of 100% and specificity of 84%.

CONCLUSION: Saline infusion sonohysterography does, yet Doppler ultrasonography does not, add to the value of endometrial thickness measurements by transvaginal ultrasonography in the screen for endometrial pathologies in asymptomatic breast cancer patients on tamoxifen.