Intracervical sonographic-pathologic correlation: preliminary results.

OBJECTIVE

To determine whether current-generation endoluminal ultrasonic transducer technology could visualize dysplastic and malignant cervical lesions.

METHODS

Inclusion criteria for patients enlisted in the study were abnormal Papanicolaou test results, an abnormality seen at colposcopy, and consent for a cone biopsy. In addition, we included 4 women who were undergoing hysterectomy but had no evidence of cervical abnormalities at the time of surgery, for a total of 28 women. We used a 20-MHz annular array intravascular transducer. We correlated all sonographic and pathologic results with regard to being normal or abnormal using the Pearson product moment correlation coefficient, and we evaluated interobserver variation by having 5 blinded sonologists interpret each examination and calculating kappa statistics.

RESULTS

We performed intracervical sonography on 24 women with abnormal Papanicolaou test results and on 4 women without abnormalities. Nineteen of the 24 women with proven disease had abnormal sonographic images. Five women with pathologically proven cervical intraepithelial neoplasia I had normal sonographic images. Lesions ranged in size from 0.3 to 7 mm. Two carcinomas (5 and 7 mm diameter), multiple nabothian cysts, cervicitis, and 2 giant cell foreign body reactions were visible on sonography. The correlation coefficient for abnormal sonographic and pathologic findings was 0.87 (P < .005). We obtained a kappa value of 0.75 for interobserver variation.

CONCLUSIONS

Intracervical sonography was able to visualize dysplastic lesions and carcinomas as hypoechoic defects with surrounding echogenic areas of glandular cervical mucosa. Lesions were shown with sufficient resolution and interobserver variation to suggest that the technique may be clinically applicable, particularly for planning surgical therapy. However, design of a transducer that can be easily sterilized between examinations, similar to current endovaginal probes, would be necessary to make intracervical sonography a clinically feasible examination.