We previously reported that the rate of ovarian malignancy for a large population-based cohort of women over age 50 years, with a “complex” adnexal mass <6 cm and no other signs of malignancy, was low at 1.3% (95% confidence interval, 0.8–2.1%) and that all cases of epithelial cancer or borderline tumor demonstrated growth by 7 months of ultrasound observation. Our findings suggest that initial short-term observation of small masses is safe and that ongoing monitoring of stable masses beyond 7 months is likely of limited value. The American Congress of Obstetrics and Gynecology Practice Bulletin on Management of Adnexal Masses states “Repeat imaging is recommended if there is uncertainty regarding a diagnosis…. The frequency of repeat imaging has not been determined.” One of the aims of our study was to evaluate the time necessary for ascertainment of malignancy, given this lack of previous data. In our practice, we support limiting observation of stable masses without solid components to 1 year and stable masses with solid components to 2 years. This strategy is similar to that described in UpToDate.

In a recent “Viewpoint,” Ormsby et al argue that indefinite, potentially lifelong monitoring of stable adnexal masses is “cost effective and low risk.” However, no studies have actually demonstrated this to be true. It is important to differentiate between initial short-term serial ultrasound observation and potentially lifelong monitoring of stable masses. Short-term ultrasound observation of a discovered adnexal mass can aid detection of malignancy–in our study, all epithelial cancers demonstrated growth by 7 months of observation. Ormsby et al state that “Malignancy has been found in apparently stable masses which eventually enlarged and increased in morphologic complexity in up to 3 years after initial detection,” citing data from the United Kingdom Collaborative Trial of Ovarian Cancer Screening (UKCTOCS). A closer look at the UKCTOCS data, however, does not support this claim, as only 3 cancers were diagnosed >1 year following initial detection of an abnormality, and in all 3 cases, the initial abnormalities were small simple cysts that did not progress and appear entirely unrelated to the later presentation of cancer. An ultrasound finding of a benign mass, like a normal ultrasound, does not preclude subsequent development of cancer during follow-up.

Ormsby et al state that we underestimated the risk of cancer by omitting women with masses whose reports did not include the word “complex,” instead of focusing on specific morphologic features of the masses. The study aim was not to establish a morphologic index for malignancy, but rather to describe the experience of a cohort of older women with masses other than simple cysts, since, from the point of view of the clinician, this represents broadly the cohort of women for whom there is typically some concern for early malignancy. The discussion presented by Ormsby et al regarding “morphologic complexities” of adnexal masses does not relate to the question of whether long-term monitoring of stable masses is beneficial. We recommend immediate removal of masses with highly worrisome ultrasound characteristics, such as large papillary solid components or enhanced vascularity. The overwhelming majority of masses have an indeterminate or benign appearance and are candidates for observation. However, the potential benefit of observation in terms of cancer detection wanes with time.

Ormsby et al also criticize our inclusion of women with symptoms, stating “women with symptoms have been found to have a higher prevalence of ovarian cancer than that reported by screening trials.” Indeed, had we excluded symptomatic women from our cohort, we would have potentially underestimated the risk of cancer, so we view this as a strength of the study.

In arguing for indefinite monitoring, Ormsby et al cite, as a guiding principle, “Pascal’s wager that it is better to do something that proves unnecessary than to do nothing and find that it would have saved or lengthened life.” However, this fails to consider potential harms. In our view, it is not better to do something that is likely both unnecessary and harmful when there is no evidence that it saves or lengthens life.

The question of whether recognition of distinct type I and type II pathways for ovarian cancer can be translated into effective screening is an open one, but this has not been demonstrated. Type 2 cancers, which represent the majority of ovarian epithelial malignancies, arise primarily from fallopian tube precursors, which helps explain the failure of screening to detect these cancers at early stage. Type 1 cancers may have ovarian precursors but, as noted by the Society of Radiologists in Ultrasound, “although ovarian cystadenoma and cystadenofibroma may be precursor lesions for borderline (low malignant potential) tumors and low-grade carcinoma, the rate of transformation is exceedingly slow, and these lesions can be considered benign.” Furthermore, while a more indolent growth pattern may be observed for type 1 cancers, any prediction of benefit from prolonged monitoring must take into consideration the fact that benefit is realized only if the stage at diagnosis is earlier than would otherwise occur. Since this subset of cancers come to clinical attention more often at early stage, such benefit is less likely. In our study, among the 1463 women we studied, of the 7 cancers found, 3 were type 1 and all demonstrated growth on follow-up ultrasound within 7 months with no additional cancer diagnoses within 24 months of follow-up. Similarly, in UKCTOCS, all of the type 1 cancers found were diagnosed within the first year.

As gynecologic oncologists, we are invested in identifying early ovarian cancer. However, we have a responsibility to ask ourselves whether what we are doing is actually serving that goal and whether we are overall helping or harming patients at the end of the day. In a recent survey study using vignettes, 1 of every 3 physicians indicated that they engage in ovarian cancer screening of low-risk women. This belief that screening works despite the substantial evidence to the contrary, concern for “missing something,” inability to manage patient anxiety, and even possible financial incentives likely contribute to this behavior. Also contributing to “overtesting” is the detection of masses as incidental findings. In our study, approximately 20% of masses were initially detected by computed tomography or magnetic resonance imaging, and potentially many more were incidental findings on ultrasound if it was ordered for another concern. In his recent magazine article Overkill: America’s Epidemic of Unnecessary Care , Atul Gawande observed “It has been hard for patients and doctors to recognize that tests and scans can be harmful. Why not take a look and see if anything is abnormal? People are discovering why not. The United States is a country of three hundred million people who annually undergo around fifteen million nuclear medicine scans, a hundred million computed tomography and magnetic resonance imaging scans, and almost ten billion laboratory tests. Often, these are fishing expeditions, and since no one is perfectly normal you tend to find a lot of fish.”

The “fish” here is the benign adnexal mass. There is no clear benefit of removal of an asymptomatic benign adnexal mass. An autopsy study showed that 56% of postmenopausal women who died from nongynecologic causes were found to have benign ovarian cystic or solid lesions. Ormsby et al also minimize the potential harm, citing a 2% complication rate for surgery in the UKCTOCS. However, in the Prostate, Lung, Colorectal, and Ovarian Cancer Screening trial, 15% of women having surgery had at least 1 serious complication. While the actual rate of surgical complication clearly varies depending on patient and surgeon factors, and may differ in the trial setting compared to the real world, it is not a potential harm that should be discounted. Why are any complications acceptable if the masses in question did not need to be removed? The harm of anxiety that can accompany monitoring for a mass should also not be discounted. In our experience, many women with stable benign masses elect surgery during the course of prolonged observation due to either cumulative anxiety or because a follow-up report raises concerns for progression due to variability in technique and reporting styles, even though the mass is unchanged. Equally likely is that a new incidental finding is noted on a follow-up scan leading to additional testing. Few practicing clinicians will be unfamiliar with this scenario.

In the conclusion of his essay, Gawande describes operating on a woman with a small microcarcinoma of the thyroid that would have never caused a problem in her lifetime. The woman had a postoperative bleed requiring reoperation. Nevertheless, she thanked him for relieving her anxiety. He notes, “I couldn’t help reflect on how that anxiety had been created. The medical system had done what it so often does: performed tests, unnecessarily, to reveal problems that aren’t quite problems, to then be fixed, unnecessarily, at great expense and no little risk.” For the management of adnexal masses, we need to consider to what degree this represents the current state of affairs and what we can do to limit potential harm from overtesting and overdiagnosis.