Avaliação da Acurácia Diagnóstica do exame de ultrassom transvaginal baseado em diagnósticos histeroscópicos para lesões intrauterinas

Abstract

Objective: Evaluate transvaginal accuracy for intra cavitary diagnosis, based on hysteroscopic findings, as a gold standard method. Methods: This is an observational prospective study in 307 women from 388 patients evaluated in the hysteroscopy diagnostics service in Hospital das Clínicas da Universidade Federal de Minas Gerais, a quaternary health care center in Brazil. The 307 women selected to participate in this study have been evaluated in the quaternary hospital by an hysteroscopic exam accomplished by two doctors who have experience in office hysteroscopy, using the same method and material for examination. All those patients brought a transvaginal ultrasound, previously done in a public or complementary heath care, exactly like it happens in our routine women health care in most places in Brazil. We evaluated sensibility, specificity and accuracy for the results obtained from transvaginal ultrasound in comparison with the diagnosis obtained from hysteroscopy. We even evaluate the agreement between the diagnosis from the two exams for polyps, myoma, endometrial high thickness, atrophy and other findings , a group formed by anatomic lesions inside the uterus. Results: Our study evaluated a total of 307 women, aged 56,55 years ( 12,3 years). The results for polyp were: sensitivity 39.8%, specificity 72.7% and accuracy of 52.8%. A concordance between the two tests was poor, with a kappa index 0.11 and p = 0.025. For myoma, we obtained sensitivity 46.7%, specificity 95.0% and accuracy of 87.9%. The agreement was moderate with respect to kappa 0.46 and p <0.001. For stunning endometrial thickening, sensitivity was 68.7%, 41.7% specificity and accuracy of 47.6%. There was no agreement between the two exams, as p= 0.126, with a kappa index of 0.06. For atrophy we found sensitivity of 6.7%, specificity of 99.3% and accuracy of 90.2%. The agreement between the tests was weak, with a kappa index 0.10 and p = 0.006. For "other findings", sensitivity was 15.6%, specificity was 99.6% and accuracy was 87.3%. The concordance between the results was mild with a 0.23 kappa index and p <0.001. Conclusion: The present study demonstrated that there is weak to light concordance between the diagnostic results of transvaginal ultrasound and hysteroscopy for polyp and "other findings, poor agreement for endometrial atrophy, moderate concordance for myoma, and that there is no agreement between the two tests for diagnosis of endometrial thickening.