Interobserver variability in the diagnosis of minimal and mild endometriosis
Olaf Buchweitz1, Pia Wülfing, Eduard Malik
1Department of Obstetrics and Gynecology, University of Münster, Albert Schweitzer Str. 33, D-48149 Münster, Germany. buchweo@mednet.uni-muenster.de
Study Objective:
To evaluate the observer's influence on the diagnosis and classification of endometriosis according to the rASRM classification.
Design:
: Prospective analysis.
Setting:
University hospital.
Patients:
Digital videotapes of laparoscopies in three patients with typical endometriotic disorders (rASRM I, II and no endometriosis).
Intervention:
One hundred and eight gynecologic surgeons were asked to indicate the endometriotic lesions on a prepared surgical sketch and to classify the site according to the rASRM classification.
Measurements And Main Results:
Total number, location and morphology of endometriotic lesions, rASRM classification. The interobserver correlation concerning the number of lesions ranged between 18% (rASRM II) and 30%. (rASRM I). There was marginal correlation regarding the location of endometriotic lesions. Kendall W coefficient ranged from 0.14 (rASRM II) to 0.44 (rASRM I) (p < 0.001). Only 13% (rASRM II) to 22% (rASRM I) of observers used the correct endometriotic classification. None of the participants specified the morphological characteristics of endometriotic lesions according to the rASRM classification.
Conclusion:
Visual assessment of an operative situs with minimal and mild endometriosis is subject to a considerable interindividual variability. One and the same lesion is assessed quite differently by different observers. Histopathological verification seems to be necessary to objectify the diagnosis of endometriosis.