|Home | About | Journals | Submit | Contact Us | Français|
To measure the occurrence and natural history of simple ovarian cysts in a cohort of older women.
Simple cysts were ascertained among a cohort of 15,735 women from the intervention arm of the Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial, through 4 years of transvaginal ultrasound screening.
Simple cysts were seen in 14% of women the first time their ovaries were visualized. The one-year incidence of new simple cysts was 8%. Among ovaries with one simple cyst at the first screen, 54% retained one simple cyst, and 32% had no cyst one year later. Simple cysts did not increase risk of subsequent invasive ovarian cancer.
Simple ovarian cysts are fairly common among post-menopausal women, and most appear stable or resolve by the next annual exam. These findings support recent recommendations to follow unilocular simple cysts in post-menopausal women without intervention.
With frequent use of transvaginal ultrasound (TVU), clinicians are detecting many simple ovarian cysts among post-menopausal women. Because the natural history of simple ovarian cysts is not fully understood, the proper management of incidental simple cysts in post-menopausal women has been uncertain.1 Aggressive surgical approaches for simple cyst management have given way to recommendations for careful monitoring,2–5 and some have raised the question of whether simple cysts need to be monitored at all.1 We evaluated the prevalence, incidence, and natural history of simple cysts within a cohort of mostly post-menopausal women receiving serial TVU examinations in the intervention arm of the Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening Trial.6
Prorok, et al. have previously described the design and methods of PLCO6, which is the setting and source population for this observational cohort analysis. Briefly, PLCO is a prospective evaluation of whether screening with a posterioranterior chest x-ray, flexible sigmoidoscopy, and CA-125 plus transvaginal ultrasound (TVU) can reduce mortality for lung, colon and ovarian cancers respectively in women. Comparable questions are being investigated for lung, colon, and prostate cancer among men. Women were eligible if they were between ages 55 and 74, and had no previous lung, colon or ovarian cancer diagnosis. Women receiving cancer treatment or participating in another screening or prevention trial were not eligible. Women who had undergone prior oophorectomy were excluded from ovarian cancer screening in the trial.
Recruitment occurred from November 1993 to July 2001. Women were randomly assigned either to the intervention or control arm after stratification by age, gender, and Screening Center (see appendix). Participants responded to a self-administered general risk factor questionnaire at entry.7 Ovarian cancer screening in PLCO included a CA-125 blood test and TVU at baseline, an annual TVU for three additional years, and annual CA-125 tests for five years beyond baseline.6 For TVU screening, qualified sonographers used a 5 to 7.5 MHz transvaginal probe to measure each ovary and describe any observed abnormalities. The examiner spent at least five minutes searching for each ovary, although the examiner could end the search if the iliac vessels were visualized and the ovaries were not observable. Ovaries were measured along major and minor axes of both transverse and longitudinal planes, and the prolate ellipsoid formula (width × height × thickness × 0.523) was used to calculate the volume of each ovary and/or cyst. Quality assurance procedures on a sample of participants included either repeating the screening examination, independent observation of the exam, or independent review of TVU films at a later time.
In the trial, the TVU screening examination was considered positive (abnormal and suspicious for ovarian cancer) when findings included: 1) ovarian volume greater than 10 cubic cm; 2) cyst volume greater than 10 cubic cm; 3) any solid area or papillary projection extending into the cavity of a cystic ovarian tumor of any size; or, 4) any mixed (solid/cystic) component within a cystic ovarian tumor. As part of the trial protocol, women with positive screening exams were referred to regular medical care for follow-up investigation. Ovarian cysts with volume less than 10 cubic cm and no solid areas, septae, or papillary projections noted in the cyst cavity, were not considered a positive screening exam in the trial. These simple cysts are, however, the subject of this cohort analysis. Ovarian cancers were ascertained in PLCO through review of medical records and pathology reports following a positive screen or a report of cancer on annual study surveys.6
This analysis is based on the subset of intervention arm women with both ovaries visualized one or more times during TVU screenings. As in a previous PLCO report, women were classified by the most serious abnormality present.8 Prevalence was the proportion of women having a simple cyst discovered at the screening in which both ovaries were visualized for the first time (prevalent cyst). Most of these were discovered during the baseline screen, but for some women, the first informative screen was in a later round. A one-year incidence rate was the proportion of women who developed a new simple cyst in their second screen with visualized ovaries, after no cysts had been discovered one year before in their first screen with visualized ovaries (incident cyst).
Descriptive evaluation of the natural history of simple cysts was performed by examining the status of all ovaries visualized in 2 consecutive study years (baseline to year 1, year 1 to year 2, or year 2 to year 3). For this particular analysis, each ovary was counted as a separate unit of evaluation. One year change in CA125 levels were compared, in an analysis of variance, between women with simple cysts who had more extensive findings (increased number or complexity) one year later and those who did not. Chi-square or Fisher's exact test, and multivariable logistic regression, were used to evaluate the potential association between detection of simple cysts and subsequent discovery of invasive ovarian cancer, through year 7 after baseline.
Chi-square analysis evaluated possible correlates of simple ovarian cyst occurrence, comparing women with prevalent or incident simple cysts to women without detected cysts at the corresponding screen. The primary objective of this analysis was to evaluate whether known or potential ovarian cancer risk factors correlate with simple cyst occurrence, based on the approach of Hartge.8 Potential predictors in this analysis included age, education, smoking status, number of sisters, first degree family history of ovarian cancer and of breast cancer, number of pregnancies, parity, number of miscarriages, use and duration of oral contraceptives, age at first pregnancy, use of hormone replacement, regular use of non-steroidal anti-inflammatory drugs or aspirin, age at menarche, age at menopause, previous gynecologic surgery, body mass index, history of benign ovarian tumors or cysts, and history of infertility. Multivariable logistic regression models contained factors with P-values less than or equal to 0.2 from the chi-square analyses. All multivariable models also adjusted for race and Screening Center. For dichotomous variables, the ‘no’ category served as the referent, and for categorical variables, the referent was the most common category. P-values below 0.05 were considered statistically significant. The study was approved by the Marshfield Clinic Research Foundation Institutional Review Board.
The trial enrolled 78,237 women and randomized 39,115 to the screening arm (Table 1). Of 4895 screening arm women not eligible for TVU, 4892 had prior oophorectomy, two had died, and one had ovarian cancer prior to the baseline screen. Of 34,220 eligible women, 30,389 (89%) received at least one exam. A total of 15,735 women had both ovaries visualized one or more times during the study. About two-thirds of these women were enrolled between ages 55 and 64, 89% reported their race as Caucasian, and 94% indicated they had received at least a high school diploma (Table 2).
Among the 15,735 women, 2,217 (14.1%) had one or more simple cysts detected at their first fully visualized TVU screening (Table 3). In stratified analysis, prevalence varied by age (p=0.001), with simple cyst detection slightly more common for women ages 55–59 (16%) than for women in older age groups (13%). Among women without a cyst of any kind on their first fully visualized screen, the rate of having a new simple cyst at the second screen one year later was 8.3% (Table 4). The incidence rate of new cyst development did not vary systematically with age, ranging across age groups from 7–9% (p=0.20). When calculating a simple cyst incidence rate for each individual screening year of the trial, the results were 8.4% at year 1 after baseline, 7.4% from year 1 to year 2, and 7.3% from year 2 to year 3.
Among ovaries with a simple cyst at the first visualized screen, 79% had just one cyst, while 15% had two, and 7% had 3 or more. For ovaries with a single simple cyst, 54% retained a single simple cyst in the following year, while 32% no longer had a cyst present (Table 5). Regression to having no cysts at all was less common for ovaries containing 2 or more cysts (23%), although more than half of the time, ovaries with multiple simple cysts had a reduced number of simple cysts the following year.
About 8% of ovaries with a single simple cyst presented the following year with multiple simple cysts, and 6% had developed more complex cysts or solid masses. Ovaries that contained multiple simple cysts were somewhat more likely than ovaries with just one simple cyst to present with complex cysts or solid masses the following year (7% for 2 simple cysts, 11% for 3+ simple cysts). Development of complex cysts or solid masses over the course of a year was rare among ovaries starting with no cyst at all (1%). Average CA-125 level changes over the one-year period were not associated with an increase in the number of simple cysts or progression from simple to more complex cysts.
Women whose ultrasound showed one or more simple cysts at their first fully visualized screening were not at significantly increased risk of subsequently developing invasive ovarian cancer (9 of 2217, or 0.41%) compared to their counterparts with no cysts (55 of 12638, or 0.44%), p=0.85. Women with a new simple cyst at their second fully visualized screening had the same occurrence of subsequent invasive ovarian cancer (1 of 459, or 0.22%) as did women who had no simple cysts on two consecutive screens (11 of 4939, or 0.22%), p>0.99. Findings did not change upon multivariable analysis adjusting for ovarian cancer risk factors: OR (95% CI) comparing those with prevalent cysts to those without, 0.84 (0.41–1.73); comparing those with incident cysts to those without, 0.95 (0.12–7.46).
In multivariable analysis, a modestly reduced risk of simple cysts at the prevalence screen was found in women age 60–64 (OR compared to women age 55–59 0.83, 0.73–0.94), current smokers (OR compared to never smokers 0.75, 0.61–0.91), and women with first pregnancy at age 25–29 (OR 0.85) or age 30+ (OR 0.71). Women with a college (OR 1.32, 1.14–1.53) or post graduate education (OR 1.19, 1.02–1.40) had an elevated odds of a prevalent simple cyst, as did women with previous history of benign ovarian cysts (OR 1.29, 1.09–1.52) and women with previous gynecologic surgery (OR 1.48, 1.33, 1.66). Menopause before age 40 was strongly associated with elevated odds of a simple cyst compared to women whose last menstruation was between ages 50 and 54 (OR 2.09, 1.77–2.46). Current and former smokers were each at about 25% increased risk of having an ovarian cyst at the incidence screen, although this was of borderline significance only for former smokers (OR 1.25, 1.00–1.56). Associations with menopause and gynecologic surgery were similar to those observed for simple cyst prevalence. There was a suggestion that women with post-high school education had a higher incidence, but this was only significant among women with some college (OR 1.36, 1.03–1.79).
In our analysis of four serial TVU examinations within a large cancer screening trial of mostly post-menopausal women, we determined that simple ovarian cysts are a common incidental finding upon ultrasonography. Increases over time in the number or complexity of cysts were infrequent, and many cysts fully resolved between annual exams. Greater numbers of simple cysts correlated with subsequent development of complex cysts, but simple cysts were not associated with development of ovarian cancer, an outcome actively ascertained in the PLCO trial.
Prevalence of simple cysts in this analysis (14%), based on a participant's first screening with fully visualized ovaries, is just slightly lower than previous baseline results from PLCO (15.7%).8 These values fall within, but near the high end, of the range of prevalence estimates from the published literature. Much of the knowledge gained about the occurrence and course of simple ovarian cysts in post-menopausal women comes from smaller case series.9–10 Oyelese11 summarized the results of 5 studies from the late 1980s and 1990s,10, 12–15 with numbers of post-menopausal women examined ranging from 149 to 7705, and simple cyst prevalence ranging from 3.3 to 14.8%. Levine found 17% of 184 women had simple adnexal cysts following first examination with transvaginal and transabdominal ultrasound.16 More recently, in a large ovarian cancer screening study at the University of Kentucky, 18% of 15,106 women had unilocular ovarian cysts discovered during the course of the screening program.17
Estimates of the incidence of simple ovarian cysts are lacking in the published literature. A major strength of the present study was the use of serial TVU examinations to assess the development of new cysts among women whose ovaries had been cyst-free in previous screenings. The incidence of simple cysts was about 8% per year and remained fairly constant over the course of the trial, even among older women in the study. The expectation of continued simple cyst detection even after multiple clear screens, and even as women approach their late 70s, reinforces the importance of a sound strategy for managing simple ovarian cysts when incorporating repeated TVU examinations in clinical practice.
Serial screenings at standard intervals also provided insight into the natural history of simple ovarian cysts. One third of ovaries with simple cysts were cyst-free the next year. Even ovaries with 3 or more simple cysts were cyst-free one year later 25% of the time. Previous studies report resolution rates ranging upward from 23%. The proportion of simple cysts resolving spontaneously over an average of 4.6 years of followup in the University of Kentucky trial was 83%.18 The most common outcome for ovaries with simple cysts in the PLCO study a year later was to have the same number of simple cysts. Finding 54% of simple cysts persisting a year later is comparable to the 49% reported among 134 women followed for a mean of 3 years1 and the 56% persistence over 2 years reported by Castillo19 among 104 women. In our study, only 6% of ovaries progressed from having a single simple cyst to having a more complex cyst a year later, although this was somewhat more common within ovaries with multiple simple cysts. In the University of Kentucky trial, over 22% of cysts developed septae or solid areas upon followup.17
Surgical intervention for abnormal findings in a post-menopausal ovary had been the standard of care for decades.20,21 In the 1980s, the necessity for a surgical approach was questioned based on small study results. A recent evidence-based report was not able to draw definitive conclusions on the sensitivity and specificity of periodic monitoring of simple cysts for cancer detection, but suggested that prolonged followup of simple cysts with interval TVU was likely a safe approach.22 The American College of Obstetricians and Gynecologists subsequently stated that simple cysts found on ultrasound ‘may be safely followed without intervention, even in post-menopausal women’.4 Our study results reinforce the belief that simple cysts are not likely cancer precursors, nor markers of increased risk, and can be managed conservatively. In our study, women with and without simple cysts developed invasive ovarian cancer at an equivalent rate after nearly 8 years of followup. These findings are also consistent with the University of Kentucky report, where only 10 ovarian cancers were detected among 2,700 women with simple cysts after 6 years on average, and none developed among women with an isolated simple cyst.17 Also, while potential correlates of prevalent simple cysts were identified in our study, including higher risk among women that were younger, well-educated, and had early menopause, traditional ovarian cancer risk factors (increasing age, family history of breast or ovarian cancer, nulliparity, infertility, etc.) were not associated with simple cysts, reinforcing a previous baseline report.8
It is a limitation that results were restricted to women with both ovaries visualized during TVU screening. Only 52% of screened women had both ovaries visualized, and occurrence rates of simple cysts may not be the same in women whose ovaries were not visualized. By incorporating simple cysts observed among the 26% of women with only one ovary visualized, and assuming the remaining 22% of women with ovaries never visualized were all cyst free, a lower bound of 11% can be placed on the prevalence among all screened women in the trial. Some ‘persistent’ cysts may reflect cyst resolution and replacement in the same ovary. In a previous study, half of simple cysts resolved within 60 days, and in another, two-thirds of resolving cysts did so within 3 months, suggesting resolution and replacement may have been occurring in some ovaries during the year between screens.15 Also, data for this study do not clearly distinguish whether a complex cyst found following a simple cyst on a previous screen represents morphological progression or the independent appearance of a complex cyst in the same ovary. PLCO quality assessment results do indicate TVU examinations had reasonably good reproducibility. If no cysts were visualized on the TVU, 99% of the time no cysts were seen on the quality assurance examination. If the initial examiner saw a simple cyst it was seen 84% of the time on the quality assurance examination. Finally, the PLCO trial protocol did not include systematic followup of the care received in response to the detection of simple cysts, since simple cysts were not considered positive screens. Therefore we do not have useful data on simple cyst surgeries or pathological interpretation of simple cyst tissue.
In summary, TVU data from participants in a large cancer screening trial confirm that simple ovarian cysts are common incidental findings among women over age 55 upon transvaginal ultrasonography, and remain common after several screening rounds and as women age. Simple cysts frequently resolve or persist without progression. Women with simple ovarian cysts do not appear to be at increased risk of developing invasive ovarian cancer. These findings support recent recommendations to follow unilocular simple cysts in post-menopausal women without intervention.4
The design and conduct of the study; collection, management, analysis, and interpretation of the data; and preparation, review, and approval of the manuscript were supported by grants to participating institutions from the National Cancer Institute, e.g., #N01-CN-25518 to Marshfield Clinic Research Foundation.
Thanks to the PLCO ovarian cancer subcommittee, especially Phil Prorok of the National Cancer Institute for his insightful comments, and Jian-Lun Xu of the National Cancer Institute and John Commins of IMS Inc. for project assistance. Thanks to Deb Multerer of the Marshfield Clinic Research Foundation for assistance with manuscript preparation.
PLCO Screening Centers and their locations include:
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Preliminary results of this study were presented in abstract form at the 2008 annual meeting of the American Society of Clinical Oncology, May 31-June 3, 2008, in Chicago, IL.
Robert T. Greenlee, Epidemiology Research Center, Marshfield Clinic Research Foundation, 1000 North Oak Ave., Mailstop ML2, Marshfield, WI 54449, Phone: 715 389-3537, FAX: 715 389-3880, greenlee.robert/at/mcrf.mfldclin.edu.
Bruce Kessel, Pacific Health Research Institute; John A Burns School of Medicine, University of Hawaii, Honolulu, Hawaii.
Craig R. Williams, Information Management Services, Inc., Rockville, MD.
Thomas L. Riley, Information Management Services, Inc., Rockville, MD.
Lawrence R. Ragard, Westat, Rockville, MD.
Patricia Hartge, Division of Cancer Epidemiology and Genetics, National Cancer Institute, Bethesda, MD.
Saundra S. Buys, University of Utah Health Sciences Center, Salt Lake City, UT.
Edward E. Partridge, University of Alabama at Birmingham, Birmingham, AL.
Douglas J. Reding, Marshfield Clinic, Marshfield, WI.