In our opinion, the article “Predicting the risk of malignancy in adnexal masses based on the Simple Rules from the International Ovarian Tumor Analysis group” once again addresses the problem of ultrasound and ovarian cancer diagnosis from the wrong perspective; that is, by thinking first in the probability of cancer. This perspective results in a higher number of surgeries performed in the United States than are warranted by the number of actual cancer cases.

The first publications in this field showed that ultrasound had a high sensitivity for the diagnosis of ovarian cancer. The problem was–and still is–that the results have a low level of specificity. This leads to major surgeries being performed for minor adnexal pathologies, or in healthy women, with a number of unnecessary complications.

An adnexal mass should be evaluated by an expert, considering it carries the risk of something as rare and devastating as an ovarian cancer. This assessment, although subjective and operator dependent, is the most appropriate way to reach a specific diagnosis and is, currently, still the best way to interpret diagnostic images. Through this procedure, a conclusive diagnosis can be reached in approximately ≥90% of cases. These findings include hemorrhagic corpus luteum, dermoid cysts, endometrioid cysts, hydrosalpinx, cystadenomas, Tarlov cysts, several other benign entities, and finally cancer. International Ovarian Tumor Analysis instantaneous descriptors are only 46% accurate, and only in relation to 4 of these entities. This is, in our opinion, the explanation for the low specificity obtained.

We think that the proper diagnosis of ovarian cancer starts with first attempting a precise diagnosis of the most frequent functional and benign adnexal entities. If no precise diagnosis is made, then, and only then, should malignancy markers be evaluated. Finally a structured and objective report should be issued, following the current trend in most areas of diagnostic imaging. This allows for proper communication and adequate management of the case, based on the specific diagnosis or the relative risk of malignancy. We have published a system and methodology that has the highest accuracy rates and that was validated under a variety of different conditions.

In summary: to adequately identify ovarian cancer through gynecological ultrasound, we should first identify and tag the various functional and benign images. Only once we have done this with accuracy should we even start thinking of a possible ovarian cancer. We believe that this schedule answers the needs posed by Dr Goldstein in the last paragraph of his editorial in the April 2016 issue of the American Journal of Obstetrics and Gynecology .