Advertisement

Outcomes from ultrasound follow-up of small complex adnexal masses in women over 50

      Objective

      The discovery of a complex adnexal mass in an older woman often raises concern for cancer. We evaluate outcomes for a large population-based cohort of women older than age 50 years with a small complex adnexal mass reported on ultrasound, without elevated CA125 or other evidence of malignancy, including time to detection of malignancy and stage at diagnosis for those initially observed.

      Study Design

      Women older than age 50 years who had an ultrasound during 2007-2011 reporting a complex adnexal mass 1-6 cm in size were identified. Previous or subsequent pelvic ultrasounds were reviewed to determine when the mass was first identified and whether there was change over time. Women with concurrent elevated CA125, evidence of metastatic disease, or less than 24 months of clinical follow-up were excluded. Surgical pathology from removal and diagnoses of ovarian cancer within 24 months of follow-up were identified.

      Results

      Among 1363 complex masses identified, 18 cancers or borderline tumors (1.3%; 95% confidence interval, 0.8–2.1%) were found. Six cases were diagnosed among 204 women who had immediate surgery after initial ultrasound (15%), and 12 additional cases were found among 994 women with at least 1 repeat ultrasound (73%). Growth was apparent on ultrasound by 7 months for all borderline and epithelial ovarian cancers. Of the 12 cases diagnosed during follow-up, 10 were found to be stage 1 at surgery.

      Conclusion

      Among isolated adnexal masses reported as complex and 1-6 cm on pelvic ultrasound in women older than 50 years, the overall risk of malignancy is low. All cases of epithelial cancer and borderline tumor demonstrated growth by 7 months of observation.

      Key words

      Ovarian cancer is the fifth leading cause of cancer death among women in the United States, with approximately 22,000 cases/year. There is no effective screening test
      • Reade C.J.
      • Riva J.J.
      • Busse J.W.
      • Goldsmith C.H.
      • Elit L.
      Risks and benefits of screening asymptomatic women for ovarian cancer: a systematic review and meta-analysis.
      and early warning signs are non-specific, leading to 75% of cases being detected at late stage.
      US Cancer Statistics Working Group
      United States Cancer Statistics: 1999–2010 Incidence and mortality web-based report.
      When an ovarian mass is discovered in an asymptomatic postmenopausal woman and there is no evidence of metastatic disease, clinicians most often rely on the ultrasound characteristics of the mass as well as the CA125 blood test
      • McDonald J.M.
      • Doran S.
      • DeSimone C.P.
      • et al.
      Predicting risk of malignancy in adnexal masses.
      • Myers E.R.
      • Bastian L.A.
      • Havrilesky L.J.
      • et al.
      Management of adnexal mass.
      • Alcázar J.L.
      • Errasti T.
      • Zornoza A.
      • MÃnguez J.A.
      • Galán M.J.
      Transvaginal color Doppler ultrasonography and CA-125 in suspicious adnexal masses.
      • Varras M.
      Benefits and limitations of ultrasonographic evaluation of uterine adnexal lesions in early detection of ovarian cancer.
      to assess the risk of cancer. Ultrasound imaging, however, is often nonspecific, interpretation varies, depending on the experience of the reading radiologist,
      • Van Calster B.
      • Timmerman D.
      • Bourne T.
      • et al.
      Discrimination between benign and malignant adnexal masses by specialist ultrasound examination versus serum CA-125.
      and CA125 blood levels are frequently elevated in advanced disease but are elevated in only 50% of early-stage ovarian cancers,
      • Einhorn N.
      • Sjövall K.
      • Knapp R.C.
      • et al.
      Prospective evaluation of serum CA 125 levels for early detection of ovarian cancer.
      • O’Connell G.J.
      • Ryan E.
      • Murphy K.J.
      • Prefontaine M.
      Predictive value of CA 125 for ovarian carcinoma in patients presenting with pelvic masses.
      making neither test highly accurate in detecting early disease. Screening using these modalities has been shown to lead to unnecessary surgery without improvement in ovarian cancer mortality.
      • Reade C.J.
      • Riva J.J.
      • Busse J.W.
      • Goldsmith C.H.
      • Elit L.
      Risks and benefits of screening asymptomatic women for ovarian cancer: a systematic review and meta-analysis.
      • Buys S.S.
      • Partridge E.
      • Black A.
      • et al.
      PLCO Project Team
      Effect of screening on ovarian cancer mortality: the Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Randomized Controlled Trial.
      When confronted with a woman with a complex adnexal mass, even in the absence of elevated CA125 or other signs of cancer, the clinician must balance concern for cancer with surgical risks.
      • Medeiros L.R.
      • Rosa D.D.
      • Bozzetti M.C.
      • et al.
      Laparoscopy versus laparotomy for benign ovarian tumour.
      This decision is hindered by uncertainty regarding the potential risk of observation. Moreover, if surgery is deferred and the mass is found to be stable but persistent on follow-up, there are currently no evidence-based guidelines for how long follow-up is needed to exclude ovarian cancer. As a result, the clinical management of these women is characterized by wide variability in surgical intervention as well as the schedule and duration of follow-up, with time and cost incurred by both the patient and the health care system despite unclear benefit. We therefore sought to evaluate clinical outcomes from a large population-based cohort of women older than age 50 years who had a small (1–6 cm) complex mass reported on pelvic ultrasound to assess both the risk associated with observation and the duration of observation required for the ascertainment of malignancy.

      Materials and Methods

      Kaiser Permanente Northern California is a closed integrated health care delivery system, providing care through 19 medical centers to approximately 35% of the population in the geographic areas served by the health plan. All formal ultrasound studies, laboratory tests, pathology reports, and diagnoses are captured in electronic databases and are linked to electronic medical records.
      The study population was defined as women older than 50 years of age who had a pelvic ultrasound from 2007 through 2011 in which a mass described as complex and 1–6 cm in size was reported. Ultrasound reports were derived from all 19 medical centers within the system, representing 408 radiologists. Women with a history of ovarian cancer, a diagnosis of metastatic cancer of any kind, or with evidence of metastatic disease on imaging or an abnormal CA125 blood test (>35 U/mL) drawn within a month of the ultrasound were excluded.
      We chose these selection criteria as those that typically define women for whom ovarian cancer is a significant concern but for whom observation would be potentially considered. Women were also excluded if they did not maintain active membership in the health plan for at least 24 months after the initial detection of the mass for any reason other than death.
      The study was designed to have a sample size that could provide reasonable precision around an estimated 1% risk of subsequent diagnosis of ovarian cancer, which was based on the known incidence of ovarian cancer in the population and the volume of gynecological pelvic ultrasound done for an adnexal abnormality annually within the system. With a sample size of 900, the 95% confidence interval around an observed incidence of 1% is 0.3-1.7%. The study period of 2007-2011 was determined based on preliminary data that indicated that 5 years of data would be needed to ensure a sample size of at least 1000 cases.
      Following approval from the Kaiser Permanente Northern California Institutional Review Board for Health Services, women meeting the selection criteria were identified using an electronic database that houses all radiology reports done within the health care system. The reports from these studies were searched to identify text strings that included the word complex and any of the following: ovary, ovarian, adnexa, adnexal, cyst, neoplasm, or mass. Chart review of the electronic medical record was then done to review the entirety of the ultrasound report to identify those cases in which a complex adnexal mass was found whose greatest dimension was 1-6 cm and to screen for exclusion criteria.
      Any additional pelvic ultrasounds done prior to or following the identified ultrasound were reviewed to identify the study when the complex mass was first described. The mass was considered to be the same from one ultrasound to the next if the laterality of the mass was the same and there was no intervening radiological evidence of resolution. Repeat ultrasound studies done for follow-up of the mass were categorized as showing the mass to be increased, decreased, resolved, or stable.
      Given the inherently variable nature of ultrasound measurements secondary to differences in technique and equipment and to capture changes in complexity as well as size, change was assessed by the following rules: if the radiologist’s stated impression was that the mass was unequivocally increased in size and/or there were new solid features not previously seen, the mass was considered increased, regardless of measurements given. If the radiologist’s stated impression was that the mass was unequivocally decreased in size or if the mass was judged to be simple rather than complex on a follow-up study, it was considered decreased. If the mass was no longer seen, it was considered resolved. If the radiologist reported the mass as only slightly or minimally changed or if the radiologist equivocated on whether there was any change, it was considered stable, regardless of measurements given. If there was no comment on the report comparing the mass to the previous studies, the mass was considered changed if the greatest dimension of the measurement differed by more than 25% from the prior study.
      Age at the time of the initial ultrasound and self-reported race/ethnicity were determined from health plan databases. All surgical pathology records were reviewed to determine whether the mass was removed and the histological findings. Diagnoses of cancer made in the 24 months following the initial ultrasound were identified by chart review and tumor registry data. We defined malignancy to include epithelial ovarian cancer and fallopian tube and primary peritoneal cancers as well as stromal tumors with low malignant potential and borderline ovarian tumors.
      The exact confidence intervals around the incidence of malignancy were calculated using binomial proportions. We also compared age, race/ethnicity, and mass size at index ultrasound by subsequent management. We divided the cohort into 4 groups: (A) repeat ultrasound and eventual removal (n = 218), (B) surgery only (n = 204), (C) no repeat ultrasound or surgery (n = 165), and (D) repeat ultrasound only (n = 776).
      We used a χ2 test for categorical variables and an analysis of variance and a Wilcoxon rank-sum test for continuous variables. The Tukey-Kramer method was used for adjusting P values in multiple comparisons. All analyses were performed using SAS version 9.3 (SAS Institute Inc, Cary, NC).

      Results

      We identified 1363 women who met the inclusion criteria, 18 of whom (1.3%; 95% confidence interval, 0.8–2.1%) were subsequently found to harbor either an ovarian cancer or borderline tumor. In all cases the diagnosis was made at the surgical removal of the mass.
      Two hundred four women (15%) underwent surgical removal without repeat ultrasound. Of these, 6 (2.9%) were malignant (2 epithelial ovarian cancers and 4 borderline tumors). The majority of women, 994 of 1363 (73%) were followed up with 1 or more ultrasounds (mean, 2.2; range, 1–9). Of these women, 218 of 994 (22%) ultimately underwent surgery, with 6 borderline tumors and 6 cancers identified, including 5 epithelial cancers (4 of 5 high grade) and 1 granulosa cell tumor. There were 169 women who had neither formal repeat ultrasound nor surgical removal (12%). However, they were followed up clinically for at least 24 months as active health plan members with no diagnoses of ovarian, fallopian tube, or peritoneal cancer during that time.
      For 261 women (19%), the initial ultrasound was done to follow up an incidental finding on computed tomography (n = 240), magnetic resonance imaging (n = 15), PET scan (n = 3), renal ultrasound (n = 2) or plain film (n = 1). Age, demographic comparison, and mass size for women with and without repeat ultrasound and/or subsequent surgical removal are shown in Table 1.
      Table 1Characteristics of women aged older than 50 years with isolated complex adnexal mass of ≤6 cm by subsequent management
      VariableTotal (n =1363)Group A: At least 1 repeat sonogram, eventual removal (n = 218)Group B: Surgical removal, no repeat sonogram (n = 204)Group C: No repeat sonogram, no surgical removal (n = 165)Group D: Repeat sonogram, no surgical removal (n = 776)
      Age, y
      Mean (SD)61.2 (10.1)59.5 (8.1)60.8 (7.8)61.5 (11.9)61.8 (10.7)
       50-59750 (55.0)129 (59.2)101 (49.3)98 (59.4)422 (54.4)
       60-69319 (23.4)56 (25.7)72 (35.1)29 (17.6)162 (20.9)
       70-79203 (14.9)31 (14.2)30 (14.6)18 (10.9)124 (16.0)
       80+92 (6.7)2 (0.9)2 (1.0)20 (12.1)68 (8.8)
      Race/ethnicity, n, %
       White874 (64.1)135 (61.9)133 (64.9)108 (65.5)498 (64.2)
       Black76 (5.6)8 (3.7)13 (6.3)12 (7.3)43 (5.5)
       Hispanic184 (13.5)36 (16.5)31 (15.6)14 (8.5)102 (13.1)
       Asian133 (9.8)26 (11.9)12 (5.9)17 (10.3)78 (10.1)
       Other97 (7.1)13 (6.0)15 (7.3)14 (8.5)55 (7.1)
      Mass size, mean/median (SD), cm3.6/3.6 (1.4)4.1/4.2 (1.3)4.2/4.4 (1.3)3.2/3.0 (1.3)3.4/3.3 (1.3)
       1-1.915712 (7.6)15 (9.6)26 (16.6)104 (66.2)
       2-2.930336 (11.9)23 (7.6)46 (15.2)198 (65.4)
       3-3.931144 (14.2)39 (12.5)44 (14.2)184 (59.2)
       4-4.929556 (19)49 (16.6)29 (9.8)161 (54.6)
       5-5.921751 (23.5)61 (28.1)12 (5.5)93 (42.9)
       68019 (23.7)17 (21.2)8 (10)36 (45)
      Suh-Burgmann. Small complex adnexal masses in women older than 50 years. Am J Obstet Gynecol 2014.
      Of the 994 masses that were followed up with repeat ultrasound, 155 (15.6%) were found to regress, and 160 (16.1%) were found to resolve. When outcomes were examined based on size, although the vast majority of masses in each size category were initially observed, surgical removal generally correlated with increasing mass size, with 45% of 6 cm masses removed compared with 17% of masses of 1–1.9 cm.
      For the 12 women who elected initial observation and whose masses were later found to be malignant or borderline, the time interval to the first repeat ultrasound varied from 2 to 7 months. For the 6 masses later diagnosed as borderline tumors, 3 demonstrated growth on their initial ultrasound done at 2, 6, and 7 months, whereas 3 were observed to be initially stable at 2 months. One mass that was stable on a 2 month ultrasound was removed at that point, whereas 2 demonstrated growth on follow-up ultrasounds done at 3 months and 6 months from the initial ultrasound.
      For the 5 masses later diagnosed as an epithelial ovarian cancer, in all cases, the mass demonstrated growth on the initial repeat ultrasound at 2-7 months. Only 1 mass that was categorized as malignant remained stable for more than 7 months and was subsequently removed at 13 months, revealing an adult granulosa cell tumor. The clinical characteristics and management for malignant and borderline cases are detailed in Table 2.
      Table 2Clinical course and characteristics of malignant masses
      CharacteristicAgeMass size
      Greatest dimension.
      Initial CA125, U/mLManagementSurgical pathology
      Epithelial cancer
       1664.616Immediate removal1a high grade serous (FS)
       2545.428Immediate removal1c high grade (FS)
       3532.6none6 mo, increased, removed1a mucinous (FS), concurrent 1a endometrial cancer
       46058 initial, 10 at 2 mo2 mo, increased, removed1c high grade (FS)
       5794.5No initial, 59 at 7 mo7 mo, increased, removed3c high grade serous (FS)
       6513.1No initial, 8 at 7 months7 mo, increased, removed1a clear cell (FS)
       7532.1No initial, 10 at 5 mo, 10 at 9 mo5 mo, increased; 9 mo increased, removed1a clear cell in endometrioma (CS)
      Borderline
       1625.216Immediate removal1a serous (FS)
       254511Immediate removal1a serous (CS)
       371427Immediate removal1a mucinous and serous (CS)
       460529Immediate removal3c serous (FS)
       5514.3162 mo, stable; 6 mo, increased, removed1a serous (CS)
       6844.452 mo, stable, removed1a mucinous (CS)
       7662.319 initial, 25 at 6 mo6 mo increased, removed3c serous (FS)
       8634.122 mo, increased, removed1a mucinous (CS)
       9644.182 mo, stable; 3 mo increased, removed1a serous (CS)
       10783.99 initial; 7 at 17 mo6 mo, increased; 17 mo increased; 24 mo stable, removed1a serous (CS)
      Stromal tumor
       1572NoneStable at 5 mo; 8 mo, 13 mo, removed1a adult granulosa cell (CS)
      CS, clinically staged; FS, fully staged.
      Suh-Burgmann. Small complex adnexal masses in women older than 50 years. Am J Obstet Gynecol 2014.
      a Greatest dimension.
      Four of the 10 borderline tumors had staging procedures done at the time of surgical removal that included the evaluation of retroperitoneal nodes and omentum, and the remainder were unstaged but were clinically stage 1 based on gross surgical findings. Six of the 7 epithelial ovarian cancers were fully staged, with one case determined to be stage 3c and the remainder stage 1. No staging was done for the granulosa cell tumor, but it was presumed to be stage 1a.
      Overall, 422 of the 1363 women (31%) ultimately underwent surgical removal, with 404 (96%) demonstrating benign adnexal pathology, most commonly cystadenofibroma, serous cystadenoma, and nonneoplastic cyst (Table 3). Among the women who did not have surgical removal, there were no additional clinical diagnoses of ovarian cancer within 24 months follow-up. A flow diagram illustrating the distribution of the groups and outcomes is shown in the Figure.
      Table 3Pathology findings from surgical removal
      Pathology findingSurgical removal, no repeat sonogram (n = 204)Surgical removal after at least 1 follow-up sonogram (n = 218)
      Borderline42.0%62.8%
      Epithelial cancer21.0%52.3%
      Granulosa cell tumor0010.5%
      Brenner62.9%52.3%
      Cystadenofibroma/adenofibroma/fibroma3919.1%2913.3%
      Endometrioma42.0%146.4%
      Fibrothecoma10.5%00
      Leiomyoma10.5%00
      Mucinous cystadenoma104.9%146.4%
      Nonneoplastic cyst
      Cysts are not considered to be pathologically significant including functional cysts and benign cystic phenomena such as cortical inclusion cysts.
      3215.7%3917.9%
      Nonspecific (no specific pathological abnormality)3919.1%2812.8%
      Nongynecological42.0%31.4%
      Other31.5%83.7%
      Paratubal cyst62.9%10.5%
      Serous cystadenoma3818.6%4420.2%
      Mixed serous/mucinous cystadenoma0020.9%
      Struma ovary II10.5%10.5%
      Mature teratoma136.4%188.3%
      Tuboovarian abscess10.5%00
      Suh-Burgmann. Small complex adnexal masses in women older than 50 years. Am J Obstet Gynecol 2014.
      a Cysts are not considered to be pathologically significant including functional cysts and benign cystic phenomena such as cortical inclusion cysts.
      Figure thumbnail gr1
      FigureOutcomes following identification of complex adnexal masses 1-6 cm
      Flow diagram showing management and outcomes for women over age 50 with isolated finding of complex adnexal mass 1-6 cm in size.
      Suh-Burgmann. Small complex adnexal masses in women older than 50 years. Am J Obstet Gynecol 2014.

      Comment

      We found that the risk of malignancy among women older than age 50 years who had a reported ultrasound finding of a complex adnexal mass 1-6 cm in size and who did not have an elevated CA125 or other evidence of cancer was low at 1.3%. If one includes only epithelial cancer cases and excludes borderline tumors and stromal tumors of low malignant potential, the risk was 0.5% or 1 in 200. All epithelial cancers and borderline tumors that were observed demonstrated growth by 7 months with no additional diagnoses during 24 months of clinical follow-up. Of the 7 total cases of epithelial cancer, 6 were fully staged and 5 were determined to be stage 1.
      Recently published data from the Kentucky ovarian cancer screening trial
      • Pavlik E.J.
      • Ueland F.R.
      • Miller R.W.
      • et al.
      Frequency and disposition of ovarian abnormalities followed with serial transvaginal ultrasonography.
      • Saunders B.A.
      • Podzielinski I.
      • Ware R.A.
      • et al.
      Risk of malignancy in sonographically confirmed septated cystic ovarian tumors.
      demonstrated that for asymptomatic complex masses detected in the course of the trial, the risk of malignancy was lower and the rate of resolution higher than previously assumed, suggesting that greater consideration should be given to the initial observation to reduce unnecessary surgery.
      When observation is elected, however, uncertainty around risk and the duration of observation needed has led to recommendations for ongoing, indefinitely repeated annual or biannual monitoring of stable but persistent lesions,
      • Levine D.
      • Brown D.L.
      • Andreotti R.F.
      • et al.
      Management of asymptomatic ovarian and other adnexal cysts imaged at US: Society of Radiologists in Ultrasound Consensus Conference Statement.
      despite the fact that: (1) no benefit has ever been demonstrated from long-term monitoring, (2) stability over time argues strongly against malignancy, (3) benign lesions are not generally precursors of malignant lesions, and (4) indefinitely repeated ultrasound monitoring exposes women to many of the same risks seen with ovarian cancer screening and as such may actually result in harm.
      • Reade C.J.
      • Riva J.J.
      • Busse J.W.
      • Goldsmith C.H.
      • Elit L.
      Risks and benefits of screening asymptomatic women for ovarian cancer: a systematic review and meta-analysis.
      • Buys S.S.
      • Partridge E.
      • Black A.
      • et al.
      PLCO Project Team
      Effect of screening on ovarian cancer mortality: the Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Randomized Controlled Trial.
      The main potential benefit of ultrasound observation is to identify those malignancies for which detection and intervention at an earlier stage improve survival and to avoid unnecessary surgery. Whether benefit is derived from repeated sonographic follow-up studies for lesions that are known to be benign and for how long are questions not addressed by this study. However, if ongoing surveillance of asymptomatic benign adnexal masses is beneficial, then ultrasound screening to detect these lesions should be indicated because the prevalence of these abnormalities is high.
      • Dørum A.
      • Blom G.P.
      • Ekerhovd E.
      • Granberg S.
      Prevalence and histologic diagnosis of adnexal cysts in postmenopausal women: an autopsy study.
      • Greenlee R.T.
      • Kessel B.
      • Williams C.R.
      • et al.
      Prevalence, incidence, and natural history of simple ovarian cysts among women >55 years old in a large cancer screening trial.
      • McDonald J.M.
      • Modesitt S.C.
      The incidental postmenopausal adnexal mass.
      However, outside specific indications such as infertility evaluation, such screening is not advised.
      We want to emphasize that the aim of the study was to evaluate real-world clinical outcomes for women for whom the decision between initial observation and surgical removal of an adnexal mass is difficult: older women with an isolated small complex mass. We specifically were interested in evaluating the time to detection of malignancy for those masses initially observed because of the lack of previous data published specifically addressing this question.
      This study was not intended to and does not establish which complex masses should and should not be observed. We therefore did not attempt to dissect the individual ultrasound characteristics that led to the designation of the mass as complex on the ultrasound report. It is understood that this term will encompass a range of ultrasound appearances. However, we have found that the very use of the term complex when describing an adnexal mass in a postmenopausal woman is enough to trigger concern for cancer.
      The study also does not establish the utility of factors such as symptoms, physical examination, and family history for predicting malignancy. Various strategies and tools have been proposed, with ongoing debate, to predict malignancy among adnexal masses, including the risk of malignancy index, which incorporates elevated CA125 levels and clinical evidence of metastatic disease
      • Karlsen M.A.
      • Sandhu N.
      • Hogdall C.
      • et al.
      Evaluation of HE4, Ca125, risk of ovarian malignancy algorithm (ROMA) and risk of malignancy index (RMI) as diagnostic tools of epithelial ovarian cancer in patients with a pelvic mass.
      • Van Gorp T.
      • Veldman J.
      • Van Calster B.
      • et al.
      Subjective assessment by ultrasound is superior to the risk of malignancy index (RMI) or the risk of ovarian malignancy algorithm (ROMA) in discriminating benign from malignant adnexal masses.
      ; risk of ovarian malignancy algorithm, which combines CA125 and HE4 levels
      • Van Gorp T.
      • Veldman J.
      • Van Calster B.
      • et al.
      Subjective assessment by ultrasound is superior to the risk of malignancy index (RMI) or the risk of ovarian malignancy algorithm (ROMA) in discriminating benign from malignant adnexal masses.
      ; ultrasound-based rules
      • Kaijser J.
      • Bourne T.
      • Valentin L.
      • et al.
      Improving strategies for diagnosing ovarian cancer: a summary of the International Ovarian Tumor Analysis (IOTA) studies.
      ; and use of the multivariate index assay in combination with clinical criteria.
      • Longoria T.C.
      • Ueland F.R.
      • Zhang Z.
      • et al.
      Clinical performance of a multivariate index assay for detecting early-stage ovarian cancer.
      In our study, we specifically excluded women with elevated CA125 or other evidence of metastatic disease from our cohort.
      The study has additional limitations related to its retrospective observational design. First, the determination of 7 months as the time period required to observe growth of all epithelial cancers and borderline tumors is subject to detection bias. All cases subsequently found to be epithelial cancer demonstrated growth on the initial follow-up ultrasound; however, the interval to the first follow-up study varied from 2 to 7 months. Growth may have been apparent prior to that time.
      Second, we do not know what the growth on ultrasound for the masses that were not observed would have been. As we describe, of the women who underwent immediate removal without observation, 3% were found to have a malignancy compared with 1% for the women who were observed, indicating that additional factors appropriately influenced the decision for surgery in some cases. However, the immediate removal of more clinically worrisome masses would be expected to bias the results toward a longer, not shorter, observation time for assessing malignancy in the remaining masses. Also, because benign ovarian neoplasms are not anticipated to be precursors of ovarian malignancy, neither the decision for observation nor the length of observation would be expected to influence the ultimate determination of whether an adnexal mass represents a malignancy. Therefore, the outcome of malignancy itself should not be biased by the management chosen.
      Third, because the study was limited to masses of 6 cm or smaller, we do not know whether the findings apply to larger masses.
      Fourth, for the 12% of the women in the cohort who had neither surgery nor a repeat formal ultrasound, we were not able to assess why repeat ultrasound was not done for each case. Given that the masses were on average smaller, less concern on the patient and/or provider’s part likely played a role. In some cases, a repeat study was ordered but was declined. It is also possible that an informal sonogram demonstrating resolution was performed in the office, which was not documented in the medical record. Although these women had neither surgery nor a repeat formal ultrasound, they were clinically followed as active health plan members for at least 24 months; therefore, diagnoses of ovarian cancer during this time would have been detected, even if diagnosed nonsurgically.
      Fifth, staging was clinical for 6 of 10 borderline tumors, 1 of 7 epithelial cancers, and the granulosa cell tumor. Although the stages assigned represent the most likely outcomes had full staging been done, we cannot prove that they are correct. Although we acknowledge that conclusions about stage distribution among women electing observation who subsequently proved to have epithelial ovarian cancer are limited by the small number of cancers detected, our finding that 4 of 5 of those fully staged were stage 1 (80%) is consistent with findings reported from Van Nagell et al
      • van Nagell Jr., J.R.
      • DePriest P.D.
      • Ueland F.R.
      • et al.
      Ovarian cancer screening with annual transvaginal sonography.
      in which 70% of women detected by screening ultrasound were found to have early stage disease.
      And sixth, we assume that if a mass is an early cancer, it would become clinically apparent within 24 months of first detection on ultrasound. In ovarian cancer screening trials, screen failures are typically defined as cancers diagnosed within 12 months of a negative screen. However, some tumors may be too indolent to be detected within 24 months. Indeed, we did find a single case of a 2 cm adult granulosa cell tumor that remained stable for 13 months prior to removal. Granulosa cell tumor is generally indolent and carries a favorable prognosis, with the vast majority of cases being cured by surgery alone.
      • Malmstrom H.
      • Hogberg T.
      • Risberg B.
      • Simonsen E.
      Granulosa cell tumors of the ovary: prognostic factors and outcome.
      • Segal R.
      • DePetrillo A.D.
      • Thomas G.
      Clinical review of adult granulosa cell tumors of the ovary.
      Although early detection of growth leading to removal seems likely to confer a survival advantage for small epithelial cancers, it is unclear whether the same can be said for these types of indolent tumors. Indeed, the Prostate, Lung, Colorectal, and Ovarian cancer screening trial did not consider these tumors or borderline tumors to be cancers.
      • Buys S.S.
      • Partridge E.
      • Black A.
      • et al.
      PLCO Project Team
      Effect of screening on ovarian cancer mortality: the Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Randomized Controlled Trial.
      A model has been proposed to categorize ovarian cancers into type 1 and type 2 tumors based on the hypothesis that type 1 tumors (eg, low-grade carcinomas, mucinous, endometrioid, and clear cell subtypes) are slow growing and arise through precursor lesions, whereas type 2 tumors are fast growing with early metastases.
      • Kurman R.J.
      • Visvanathan K.
      • Roden R.
      • Wu T.C.
      • Shih IeM.
      Early detection and treatment of ovarian cancer: shifting from early stage to minimal volume of disease based on a new model of carcinogenesis.
      The model would suggest that ultrasound observation would be more likely to lead to the detection of early stage type 1 tumors but not type 2 tumors.
      A full discussion and debate of the potential implications of the model on adnexal mass follow-up and detection of ovarian cancer is beyond the scope of this study. However, it is interesting to note that of the 7 epithelial cancers, 3 cancers were type 1 tumors and all were early stage, and 4 cancers were type 2 tumors, 3 of which were early stage.
      The strengths of this study include the large size of the cohort and the fact that it is population based. All women meeting the selection criteria were included in the analysis, regardless of clinical indication. We were interested in studying this entire population because many assumptions guiding management of adnexal masses in older women were established during an era when masses typically came to attention because of either symptoms such as pain or from being palpable on examination, and it is unclear whether these assumptions of risk apply to the current population of women found to have a mass on pelvic ultrasound, given the increased utilization of ultrasound and the increased detection of masses as incidental findings.
      Although much has been learned about the natural history of adnexal masses from data from ovarian cancer screening trials,
      • Pavlik E.J.
      • Ueland F.R.
      • Miller R.W.
      • et al.
      Frequency and disposition of ovarian abnormalities followed with serial transvaginal ultrasonography.
      • Saunders B.A.
      • Podzielinski I.
      • Ware R.A.
      • et al.
      Risk of malignancy in sonographically confirmed septated cystic ovarian tumors.
      • Greenlee R.T.
      • Kessel B.
      • Williams C.R.
      • et al.
      Prevalence, incidence, and natural history of simple ovarian cysts among women >55 years old in a large cancer screening trial.
      • van Nagell Jr., J.R.
      • DePriest P.D.
      • Ueland F.R.
      • et al.
      Ovarian cancer screening with annual transvaginal sonography.
      • Sharma A.
      • Gentry-Maharaj A.
      • Burnell M.
      • et al.
      UK Collaborative Trial of Ovarian Cancer Screening (UKCTOCS)
      Assessing the malignant potential of ovarian inclusion cysts in postmenopausal women within the UK Collaborative Trial of Ovarian Cancer Screening (UKCTOCS): a prospective cohort study.
      we sought to evaluate outcomes from care as it is actually provided, not as assigned, determined, or recommended by a protocol. As a result, the outcomes may be more representative of what is achieved in practice since they are based on management by a large group of community-based women’s health providers and reflect generally accepted standards of practice. All ovarian malignancies were captured, regardless of the management selected or observation time, with the only potential exception arising from malignancies whose behavior was too indolent to become clinically apparent within 24 months of their initial detection.
      In summary, we found that among women older than age 50 years with complex adnexal masses 1-6 cm in size and no other evidence of malignancy on imaging or by CA125, the risk the mass represented a malignancy or borderline tumor was low at 1.3% and that all epithelial cancers and borderline tumors that were observed demonstrated growth on ultrasound by 7 months. Among malignancies that were eventually identified in women who were observed, 10 of 12 cases were stage 1, indicating that delay of surgery to permit initial observation does not necessarily adversely affect the stage of the cancer identified at surgery. These findings lend support for short-term initial observation of isolated small complex masses and suggest that ongoing monitoring of stable masses beyond 7 months for the purpose of excluding malignant etiology is of limited value.

      References

        • Reade C.J.
        • Riva J.J.
        • Busse J.W.
        • Goldsmith C.H.
        • Elit L.
        Risks and benefits of screening asymptomatic women for ovarian cancer: a systematic review and meta-analysis.
        Gynecol Oncol. 2013; 130: 674-681
        • US Cancer Statistics Working Group
        United States Cancer Statistics: 1999–2010 Incidence and mortality web-based report.
        Department of Health and Human Services, Centers for Disease Control and Prevention, and National Cancer Institute, Atlanta, GA2013 (Available at:) (Accessed April 1, 2014)
        • McDonald J.M.
        • Doran S.
        • DeSimone C.P.
        • et al.
        Predicting risk of malignancy in adnexal masses.
        Obstet Gynecol. 2010; 115: 687-694
        • Myers E.R.
        • Bastian L.A.
        • Havrilesky L.J.
        • et al.
        Management of adnexal mass.
        Evid Rep Technol Assess (Full Rep). 2006; 130: 1-145
        • Alcázar J.L.
        • Errasti T.
        • Zornoza A.
        • MÃnguez J.A.
        • Galán M.J.
        Transvaginal color Doppler ultrasonography and CA-125 in suspicious adnexal masses.
        Int J Gynaecol Obstet. 1999; 66: 255-261
        • Varras M.
        Benefits and limitations of ultrasonographic evaluation of uterine adnexal lesions in early detection of ovarian cancer.
        Clin Exp Obstet Gynecol. 2004; 31: 85-98
        • Van Calster B.
        • Timmerman D.
        • Bourne T.
        • et al.
        Discrimination between benign and malignant adnexal masses by specialist ultrasound examination versus serum CA-125.
        J Natl Cancer Inst. 2007; 99: 1706-1714
        • Einhorn N.
        • Sjövall K.
        • Knapp R.C.
        • et al.
        Prospective evaluation of serum CA 125 levels for early detection of ovarian cancer.
        Obstet Gynecol. 1992; 80: 14-18
        • O’Connell G.J.
        • Ryan E.
        • Murphy K.J.
        • Prefontaine M.
        Predictive value of CA 125 for ovarian carcinoma in patients presenting with pelvic masses.
        Obstet Gynecol. 1987; 70: 930-932
        • Buys S.S.
        • Partridge E.
        • Black A.
        • et al.
        • PLCO Project Team
        Effect of screening on ovarian cancer mortality: the Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Randomized Controlled Trial.
        JAMA. 2011; 305: 2295-2303
        • Medeiros L.R.
        • Rosa D.D.
        • Bozzetti M.C.
        • et al.
        Laparoscopy versus laparotomy for benign ovarian tumour.
        Cochrane Database Syst Rev. 2009; : CD004751
        • Pavlik E.J.
        • Ueland F.R.
        • Miller R.W.
        • et al.
        Frequency and disposition of ovarian abnormalities followed with serial transvaginal ultrasonography.
        Obstet Gynecol. 2013; 122: 210-217
        • Saunders B.A.
        • Podzielinski I.
        • Ware R.A.
        • et al.
        Risk of malignancy in sonographically confirmed septated cystic ovarian tumors.
        Gynecol Oncol. 2010; 118: 278-282
        • Levine D.
        • Brown D.L.
        • Andreotti R.F.
        • et al.
        Management of asymptomatic ovarian and other adnexal cysts imaged at US: Society of Radiologists in Ultrasound Consensus Conference Statement.
        Radiology. 2010; 256: 943-954
        • Dørum A.
        • Blom G.P.
        • Ekerhovd E.
        • Granberg S.
        Prevalence and histologic diagnosis of adnexal cysts in postmenopausal women: an autopsy study.
        Am J Obstet Gynecol. 2005; 192: 48-54
        • Greenlee R.T.
        • Kessel B.
        • Williams C.R.
        • et al.
        Prevalence, incidence, and natural history of simple ovarian cysts among women >55 years old in a large cancer screening trial.
        Am J Obstet Gynecol. 2010; 202: 373.e1-373.e9
        • McDonald J.M.
        • Modesitt S.C.
        The incidental postmenopausal adnexal mass.
        Clin Obstet Gynecol. 2006; 49: 506-516
        • Karlsen M.A.
        • Sandhu N.
        • Hogdall C.
        • et al.
        Evaluation of HE4, Ca125, risk of ovarian malignancy algorithm (ROMA) and risk of malignancy index (RMI) as diagnostic tools of epithelial ovarian cancer in patients with a pelvic mass.
        Gynecol Oncol. 2012; 127: 379-383
        • Van Gorp T.
        • Veldman J.
        • Van Calster B.
        • et al.
        Subjective assessment by ultrasound is superior to the risk of malignancy index (RMI) or the risk of ovarian malignancy algorithm (ROMA) in discriminating benign from malignant adnexal masses.
        Eur J Cancer. 2012; 48: 1649-1656
        • Kaijser J.
        • Bourne T.
        • Valentin L.
        • et al.
        Improving strategies for diagnosing ovarian cancer: a summary of the International Ovarian Tumor Analysis (IOTA) studies.
        Ultrasound Obstet Gynecol. 2013; 41: 9-20
        • Longoria T.C.
        • Ueland F.R.
        • Zhang Z.
        • et al.
        Clinical performance of a multivariate index assay for detecting early-stage ovarian cancer.
        Am J Obstet Gynecol. 2014; 210: 87.e1-87.e9
        • van Nagell Jr., J.R.
        • DePriest P.D.
        • Ueland F.R.
        • et al.
        Ovarian cancer screening with annual transvaginal sonography.
        Cancer. 2007; 109: 1887-1896
        • Malmstrom H.
        • Hogberg T.
        • Risberg B.
        • Simonsen E.
        Granulosa cell tumors of the ovary: prognostic factors and outcome.
        Gynecol Oncol. 1994; 52: 50-55
        • Segal R.
        • DePetrillo A.D.
        • Thomas G.
        Clinical review of adult granulosa cell tumors of the ovary.
        Gynecol Oncol. 1995; 56: 338-344
        • Kurman R.J.
        • Visvanathan K.
        • Roden R.
        • Wu T.C.
        • Shih IeM.
        Early detection and treatment of ovarian cancer: shifting from early stage to minimal volume of disease based on a new model of carcinogenesis.
        Am J Obstet Gynecol. 2008; 198: 351-356
        • Sharma A.
        • Gentry-Maharaj A.
        • Burnell M.
        • et al.
        • UK Collaborative Trial of Ovarian Cancer Screening (UKCTOCS)
        Assessing the malignant potential of ovarian inclusion cysts in postmenopausal women within the UK Collaborative Trial of Ovarian Cancer Screening (UKCTOCS): a prospective cohort study.
        BJOG. 2012; 119: 207-219