A meaningful evolution has occurred over the past 30 years in the evaluation of ovarian tumors. In the 1980s, any palpable ovarian tumor was recommended for surgical removal.1 In the early 2000s, studies showed that unilocular cysts were at very low risk for malignancy, and surveillance was recommended.2 In the following decade, septate cysts were added to the list of ovarian tumors unlikely to be malignant, and nonsurgical therapy was suggested.3 It is estimated that 10% of women will undergo surgery for an adnexal mass in their lifetime, despite the fact that only 1 in 6 (13%–21%) of these masses is found to be malignant.4,5
A comprehensive, morphology-based pelvic ultrasonography is the first and most important step in evaluating an ovarian tumor’s risk of malignancy to determine whether surgery or surveillance is required.
Ovarian cancer continues to be the leading cause of gynecologic cancer death. Despite achieving superior surgical and cancer outcomes, a gynecologic oncologist performs only 40% of the initial ovarian cancer surgeries.6 Premenopausal and menopausal ovarian tumors are different in cause and consequence. Only 15% of premenopausal tumors are malignant, most commonly germ cell tumors, borderline ovarian tumors, and epithelial ovarian cancers. Tumors in menopausal women are less common but are more likely to be malignant. In actuality, up to 50% of tumors in this population are malignant. The most common of these malignancies are epithelial ovarian cancers, cancers metastatic to the ovary, and malignant stromal tumors.
Effective and evidence-based preoperative evaluations are available to help the clinician estimate a tumor’s risk of malignancy and determine which tumors are appropriate for referral to a specialist for surgery.
The actual incidence and prevalence of ovarian tumors are not known. From a review of almost 40,000 ultrasonography scans performed in the University of Kentucky Ovarian Cancer Screening Program, the estimated incidence and prevalence of ovarian abnormalities are 8.2 per 100 women annuallyand 17%, respectively.7 Seventy percent of these abnormalities have a unilocular or simple septate morphology and are at low risk for malignancy.7 The remaining 30% of abnormalities are high risk, although this represents only 9% of the total population evaluated. Since the vast majority of these abnormalities are expected to be asymptomatic, most will go unrecognized in the general population. For women who have an ovarian abnormality on ultrasonography, the majority will be at low risk for malignancy and will not require surgery.
Ovarian ultrasonography plus morphologic scoring comprise a comprehensive approach
The recently published recommendations of the First International Consensus Conference report on adnexal masses are summarized in TABLE 1.8 The expert panel reviewed the evidence and concluded that effective ultrasonography strategies exist and are well validated, and that low-risk asymptomatic ovarian cysts do not require surgical removal.
While no single ultrasonographic findingcan differentiate a benign from a malignant mass, morphologic scoring systems improve our ability to estimate a tumor’s malignant potential. In the United States, most practitioners in women’s health have ready access to gynecologic ultrasonography, but individual training and proficiency vary. Since not everyone is an expert sonographer, it is useful to employ an objective strategy when evaluating an ovarian tumor. The focus of a comprehensive ovarian ultrasonography is to recognize morphologic patterns that reflect a tumor’s malignant potential. While tumor volume is useful, tumor morphology is the most prognostic feature.
International Ovarian Tumor Analysis group
The International Ovarian Tumor Analysis (IOTA) group has published extensively on sonographic definitions and patterns that categorize tumors based on appearance.9 Simple rules and the ADNEX risk model are 2 of the group’s approaches (FIGURE 1).10,11 Both methods have been validated as effective for differentiating benign from malignant ovarian tumors, but neither has been used to study serial changes in ovarian morphology.
Regardless of the strategy employed, 25% of ovarian ultrasonography evaluations will be interpreted as “indeterminate” or “risk unknown.”10 The IOTA strategies have been successfully used in Europe for years, but they have not yet been studied or adopted in the United States.
Kentucky morphology index
The morphology index (MI) from the University of Kentucky is an ultrasonography-based scoring system that combines tumor volume and tumor structure into a simple and effective index with a score ranging from 0 to 10 (FIGURE 2).12 A rising Kentucky MI score has a linear and predictable increase in the risk of ovarian malignancy. In a review of almost 40,000 sonograms, 85% of the malignancies had an MI score of 5 or greater (TABLE 2).12 Using this as a cutoff, the sensitivity and specificity for predicting malignancy was 86% and 98%, respectively.12
When comparing the ADNEX risk model with the Kentucky MI, investigators reviewed 45,000 ultrasound results and found that the majority of cancers were categorized by the ADNEX model in the lowest 4 of the 10 risk-of-malignancy groups, compared with only 15% for the MI.13 This clustering or skew is potentially problematic, since we expect higher scores to be more predictive of cancer than lower scores. It also infers that the ADNEX model may not be useful in serial surveillance strategies. Moreover, the ADNEX model identified only 30% of early stage cancers compared with identification of 80% with use of the MI.13
Serial ultrasonography is a concept similar to any longitudinal biomarker evaluation. In the United Kingdom Collaborative Trial of Ovarian Cancer Screening (UKCTOCS) program, the Risk of Ovarian Cancer Algorithm (ROCA) employs serial measurements of cancer antigen 125 (CA 125) to improve cancer detection. Serial ultrasonography similarly can be applied to better characterize a tumor’s physiology as well as its morphology. Over time, malignant ovarian tumors grow naturally in volume and complexity, and they do so at a rate faster than nonmalignant tumors. If this physical change can be measured objectively with ultrasonography, then serial sonography becomes a valuable diagnostic aid.
In comparing serial MI scores with clinical outcomes, studies have shown that malignant tumors exhibit a rapid increase, nonmalignant tumors have a stable or gradual rise, and resolving cysts show a decrease in MI score over time (FIGURE 3).12 An increase in the MI score of 1 or more per month (≥1 per month) is concerning for malignancy, and surgical removal should be considered. If the MI score of an asymptomatic ovarian tumor does not increase by 1 per month, it can be surveilled with intermittent ultrasonography.