Correlation of findings on transvaginal sonography with serum CA 125 levels

This study was performed to correlate findings on transvaginal sonography with differential levels of serum CA 125.

The study population included 144 patients who underwent transvaginal sonography for suspected pelvic mass and in whom a serum CA 125 level was established as part of the diagnostic workup. Levels of less than 36 U/ml were considered nonelevated; levels equal to or greater than 36 U/ml but less than 50 U/ml were considered mildly elevated; levels of 50-100 U/ml, moderately elevated: and levels greater than 100 U/ml, markedly elevated. Seventy-eight patients had pathologically proven gynecologic disease. Sixty-six patients were followed up with serial sonograms; benign gynecologic disease was diagnosed in 39 of these patients. In the 5-year period included in the study, the patients followed up with serial sonograms were imaged an average of 2.6 times during 2.1 years.

Transvaginal sonography had a sensitivity of 97%, a specificity of 97%, and an overall accuracy of 94% for correctly revealing pathologically proven gynecologic disease. Interpreters of the sonograms misdiagnosed no carcinomas, even in patients with nonelevated levels of CA 125. Likewise, no benign lesions were incorrectly interpreted as malignant, even in patients with markedly elevated levels of CA 125. When interpreters used CA 125 alone to evaluate for malignancy, those interpreters achieved a sensitivity of 85% and a specificity of 29% in all patients with levels of CA 125 greater than or equal to 36 U/ml. Also, the use of CA 125 alone to evaluate for malignancy had an even lower specificity (17%) in postmenopausal patients who had levels of CA 125 that exceeded 100 U/ml.

Transvaginal sonography is a sensitive tool for diagnosing gynecologic disease in spite of misleadingly high or low levels of CA 125. Evaluation of levels of CA 125 should be discouraged before sonograms are obtained.