Background: Retrospective studies have shown that a significant number of ovarian borderline tumors, ovarian metastases, and nonepithelial tumors were erroneously diagnosed as ovarian carcinomas. This may lead to unnecessary morbidity, suboptimal therapeutic modalities, and unintended bias in clinical trials. The aim of this study was to investigate the frequency and clinical significance of such diagnostic discrepancies.
Material and methods: Original histological slides from patients with ovarian carcinomas included in phase III chemotherapy trials of the Working Group on Gynecological Oncology (AGO) were reviewed by at least two specialized pathologists. Diagnostic discrepancies were classified as being either clinically relevant (major) or clinically not relevant (minor).
Results: A total of 454 out of 533 patients from the AGO OVAR11 (ICON7) trial gave consent to the second opinion on the pathology results. All of the 104 institutes of pathology responsible for the original diagnoses contributed to the study. The first diagnosis and the second opinion pathology review were identical in 295 out of 454 (65%) cases. In 128 cases (28.2%) a minor discrepancy was found and 31 cases (6.8%) were shown to have a major discrepancy.
Conclusion: The assumption of a significant number or erroneous diagnoses in chemotherapy trials of ovarian carcinomas was confirmed. A pathology review seems therefore desirable and may help to reduce unnecessary morbidity and optimize therapeutic strategies. Moreover, improvement of quality in therapy trials may become possible. In another study a new concept allowing a rapid pathology review before randomization of patients has now been successfully tested and it may well have potential to form the basis for modern networking consultation pathology.