Improvement of hysterosalpingographic accuracy in the diagnosis of peritubal adhesions.

Objective: Our purpose was to compare hysterosalpingography with laparoscopy in the diagnosis of peritubal adhesions and to verify whether a combination of radiographic signs improves hysterosalpingographic accuracy.
Subjects and Methods: Thirty candidates for laparoscopy underwent hysterosalpingography before surgery. Two radiologists evaluated the presence or absence and types of radiographic signs of peritubal adhesions (convoluted tubes, vertical tubes, loculation of contrast medium in peritoneum, halo effect, and fixed laterodeviation of the uterus) using two different criteria for normality or abnormality: no sign means a normal result, one or more signs mean an abnormal result (first criterion); no sign or one sign means a normal result, two or more signs mean an abnormal result (second criterion). Interpretation discrepancies were resolved by consensus. Peritubal and periovarian adhesions were evaluated by a single operating surgeon during laparoscopy (recorded on S-VHS videotape) and by a different surgeon reviewing the videotape. The radiographic results obtained using the two criteria in radiologically patent as well as in distally nonpatent tubes were compared with corresponding laparoscopic results by 2 x 2 tables and were statistically analyzed (kappa statistics).
Results: The first criterion displayed poor diagnostic accuracy. The correlation with laparoscopy was not statistically significant in either radiologically patent or distally nonpatent tubes. The second criterion greatly improved the agreement with laparoscopy, but only in patent tubes (kappa = 0.7789; p<0.001).
Conclusion: Hysterosalpingographic accuracy in peritubal adhesion diagnosis can be improved in patent tubes by taking into account more than one of the reported radiographic signs.