INTRODUCTION — Use of serum biomarkers for the diagnosis of epithelial ovarian cancer is an active area of investigation. Biomarkers or panels of biomarkers are generally used in combination with each other or with other findings (eg, ultrasound). Several serum biomarkers are commercially available. Cancer antigen 125 (CA 125) has been the focus of most investigations and has been used most extensively.

High-grade serous epithelial ovarian, fallopian tube, and peritoneal carcinomas should be considered as a single clinical entity due to their shared clinical behavior and treatment and possible shared pathogenesis [1]. No serum biomarker can distinguish between these types of carcinoma. For these reasons, we will use the term epithelial ovarian cancer to refer to this group of malignancies in the discussion that follows.

Use of serum biomarkers for the diagnosis of epithelial ovarian cancer in patients who have an adnexal mass is reviewed here. Related topics are discussed separately, including:

●Ovarian cancer screening in patients who are asymptomatic and have no adnexal mass (see "Screening for ovarian cancer")

●General principles of the evaluation of an adnexal mass (see "Approach to the patient with an adnexal mass" and "Ultrasound categorization of adnexal masses")

●Ovarian cancer diagnosis (see "Epithelial carcinoma of the ovary, fallopian tube, and peritoneum: Clinical features and diagnosis")

●Epithelial ovarian cancer treatment and posttreatment surveillance (see "Epithelial carcinoma of the ovary, fallopian tube, and peritoneum: Surgical staging" and "First-line chemotherapy for advanced (stage III or IV) epithelial ovarian, fallopian tube, and peritoneal cancer")

●Nonepithelial ovarian cancer histologies (germ cell tumors, sex cord-stromal tumors) (see "Ovarian germ cell tumors: Pathology, epidemiology, clinical manifestations, and diagnosis" and "Sex cord-stromal tumors of the ovary: Epidemiology, clinical features, and diagnosis in adults")

TERMINOLOGY — The general definition of a biomarker is any indicator of disease. A tumor maker is a biomarker specific for malignancy. The United States National Institutes of Health defines a biomarker as "a characteristic that is objectively measured and evaluated as an indicator of normal biologic processes, pathogenic processes, or pharmacologic response to a therapeutic intervention." For epithelial ovarian cancer, this may include patient symptoms, physical examination, imaging studies, or laboratory tests. Biomarkers for epithelial ovarian cancer have been identified in serum, urine, ovarian cyst fluids, ascites, and other body fluids. In this topic review, the term biomarkers will refer specifically to serum biomarkers.

BACKGROUND — Epithelial ovarian cancer (EOC) presents several diagnostic challenges. EOC is most commonly diagnosed at a late stage, which contributes to its poor prognosis (table 1). In addition, it is difficult to differentiate between benign and malignant adnexal masses, which results in many surgical procedures for masses that are ultimately determined to be benign.

Symptoms of ovarian cancer are usually present, but are often nonspecific and do not usually result in diagnosis at an early stage. (See "Early detection of epithelial ovarian cancer: Role of symptom recognition".)

Definitive diagnosis of ovarian cancer requires histopathologic examination of excised tissue, and many procedures performed to evaluate a pelvic mass ultimately identify a benign process. As an example, in a large ovarian cancer screening randomized trial, among 570 patients who underwent surgical evaluation for suspected ovarian cancer, 20 cases of malignancy were found (3.5 percent) [2]. Thus, improving the methods for distinguishing malignant from benign masses would avoid unnecessary surgery.

Currently, no serum biomarker is both highly sensitive and specific for the diagnosis of EOC. Thus, investigators have combined serum markers or used multimodality strategies to improve the detection of malignancy.

BIOMARKERS

Cancer antigen 125 — Use of cancer antigen 125 (CA 125) as a biomarker for EOC was first described in 1983 [3]. CA 125 is the most widely used biomarker for EOC, and it is approved by the US Food and Drug Administration (FDA) for monitoring response to therapy in patients with known EOC. CA 125 is often used off-label for evaluation of an adnexal mass alone or combined with other serum biomarkers and/or pelvic ultrasound [4]. (See 'Diagnostic performance' below.)

The gene for CA 125 was cloned in 2001 and is called MUC16 [5]. The CA 125 antigen is a large transmembrane glycoprotein derived from both coelomic (pericardium, pleura, peritoneum) and müllerian (fallopian tube, endometrial, endocervical) epithelia [6,7]. The antigen complex contains two major domains, A and B. A portion of the extracellular domain includes repeat sequences that bind the OC125 and M11 monoclonal antibodies.

The original CA 125 test reacts with OC125, and the newer CA 125 II test utilizes both the OC125 and M11 moieties [8]. Both tests are commonly used in clinical practice. While CA 125 II may be more specific, there are no data to support the superiority of one test over the other [9]. The normal values for the two assays are:

●CA 125 – ≤35 units/mL

●CA 125 II – <20 units/mL

CA 125 and the second generation test (CA 125 II) are both in regular use. In fact, there are dozens of CA 125 assays in use today, and each of them has different test performance. For practical purposes, though, they are considered clinically equivalent [10-12]. The caveat is that clinicians should avoid comparing values from different assays when making clinical decisions. When monitoring CA 125 or CA 125 II, serial tests should be done with the same assay, which usually means using the same laboratory.

Absolute CA 125 cutoffs remain clinically arbitrary, particularly for premenopausal patients. Some experts have suggested that normal values may range from 20 to 200 units/mL; these variations take into consideration elevations in serum due to benign indications common in premenopausal patients, such as menses, fibroids, endometriosis, or ovulation.

In terms of timing of measurement of CA 125, the CA 125 likely does vary (though slightly) with the menstrual cycle [13]. CA 125 values may increase during the menses, but this increase is likely to be clinically insignificant. Ideally, the blood for testing should be obtained during the first half of the cycle to minimize the fluctuation, but as long as subsequent testing is done at the same time of the cycle, serial monitoring will not be affected regardless of when the blood is drawn.

Diagnostic performance — Serum CA 125 is the most commonly used laboratory test for the evaluation of adnexal masses for EOC, but the test has several shortcomings. The diagnostic performance of CA 125 is limited, particularly for early-stage disease. In addition, it is mainly useful in postmenopausal patients.

CA 125 testing alone has a low sensitivity, particularly for early-stage ovarian cancer, and a low overall specificity. The specificity is particularly low in premenopausal patients. The diagnostic performance of CA 125 for preoperative identification of the adnexal mass suspicious for ovarian cancer was illustrated in a meta-analysis of 77 studies [14]. An elevated CA 125 (defined as >35 units/mL) had a sensitivity of 78 percent and a specificity of 78 percent. One factor that lowers sensitivity is that CA 125 is not consistently produced by some histologic types of EOC, including mucinous, clear cell, and mixed müllerian ovarian tumors [6,15]. For early-stage ovarian cancers, some studies have reported a wide range of sensitivities, as low as 25 percent for stage I (range 25 to 75 percent) and as low as 61 percent for stage II (61 to 96 percent) [16,17].

The performance of CA 125 in postmenopausal patients with an adnexal mass compared with premenopausal patients with an adnexal mass was demonstrated in another meta-analysis [18]. Among six studies that defined an elevated CA 125 as >35 units/mL, in postmenopausal patients, the sensitivity for ovarian cancer was 69 to 87 percent, and the specificity was 81 to 93 percent. For premenopausal patients, the sensitivity was 50 to 74 percent, and the specificity was 69 to 78 percent (one study reported a specificity of 92 percent). The low specificity in premenopausal patients is because an elevated CA 125 is also associated with many conditions other than EOC, and many of these are found in reproductive-age patients (table 2 and figure 1).

In addition to the association of an elevated CA 125 in conditions other than EOC, serum CA 125 levels may vary with other factors. As an example, in the Prostate, Lung, Colorectal, and Ovarian Cancer (PLCO) Screening Trial of over 25,000 women, in those without ovarian cancer, an elevated CA 125 level (defined as ≥35 units/mL) was significantly more likely in patients age 60 years or older compared with ages 55 to 59 years and in those with a history of cigarette smoking [19]. In addition, the level was significantly less likely to be elevated in patients who were obese (body mass index ≥30 kg/m²) or had a previous hysterectomy.

Based upon the poor diagnostic performance of CA 125 in premenopausal patients with an adnexal mass, there has been some discussion of using a higher CA 125 level as the threshold for suspicion of malignancy. Serum CA 125 >200 units/mL had been used as a criterion for referral to a gynecologic oncologist for premenopausal patients in 2002 guidelines from the American College of Obstetricians and Gynecologists (ACOG) and the Society of Gynecologic Oncologists (SOGC) (table 3) [20]. A retrospective study of patients with pelvic masses at tertiary care centers found that, for the diagnosis of ovarian cancer in patients >50 years old, a CA 125 of >200 units/mL had a positive predictive value of 70 percent and a negative predictive value of 85 percent; the combination of all the criteria in the guidelines had a sensitivity of 70 percent [21]. However, this guideline was replaced by a 2016 ACOG practice bulletin that recommended referral of postmenopausal patients with an adnexal mass and CA 125 levels >35 units/mL and referral of premenopausal patients with an adnexal mass based on the judgment of the clinician after considering both the CA 125 level and other clinical factors [22]. ACOG noted that the previous CA 125 threshold of >200 units/mL for referral of premenopausal patients to a gynecologic oncologist was based on expert opinion and that no evidence-based threshold is currently available.

The use of CA 125 for screening for ovarian cancer in asymptomatic patients is reviewed separately. (See "Screening for ovarian cancer", section on 'Cancer antigen 125 (CA 125)'.)

Human epididymis protein 4 — The human epididymis protein 4 (HE4) assay was approved by the FDA in 2008 for monitoring for recurrent or progressive disease in patients with EOC [23]. It is also used for the evaluation of an adnexal mass as a component of the Risk of Malignancy Algorithm (ROMA) and serum Overa tests.

HE4 is an antigen derived from human epididymis protein, a product of the WFDC2 gene that is overexpressed in patients with serous and endometrioid ovarian carcinoma [24,25].

The laboratory reference range of HE4 is ≤150 pM. This value represents the upper 95^th percentile for both premenopausal and postmenopausal patients and is reported in the FDA package insert for the HE4 testing kit.

Serum HE4 may be affected by pregnancy and age. In the absence of ovarian malignancy, median HE4 levels have been reported to be significantly lower in pregnant compared with nonpregnant patients and in premenopausal compared with postmenopausal patients [26].

In patients with an ovarian malignancy, HE4 may have prognostic implications. One study identified that a preoperative HE4 >277 pM was significantly associated with ovarian cancer mortality, particularly in patients with serous histology [27]. However, one cohort study of postmenopausal patients with adnexal masses suggested that HE4 provided no additional value when used alongside both ultrasound and CA 125 [28].

HE4 may also be useful in detecting recurrence in ovarian cancer patients with a nonelevated CA 125 at time of diagnosis. In a preliminary study of eight ovarian cancer patients with a normal CA 125 at diagnosis, recurrences were detected with HE4 elevation alone [29].

Carcinoembryonic antigen — Carcinoembryonic antigen (CEA) is a protein normally found in embryonic or fetal tissue. Serum levels disappear almost completely after birth, but small amounts may be present in the colon.

In adults, CEA may be elevated in malignancies that produce the protein, particularly mucinous cancers associated with the gastrointestinal tract or ovary. CEA may also be elevated in the following malignancies [30]:

●Breast

●Pancreas

●Thyroid

●Lung

Benign conditions that have been associated with an elevated CEA include [30]:

●Cigarette smoking

●Mucinous cystadenoma of the ovary or appendix

●Cholecystitis

●Liver cirrhosis

●Diverticulitis

●Inflammatory bowel disease

●Pancreatitis

●Pulmonary infections

The ratio of CA 125 to CEA has diagnostic implications. Among patients undergoing pelvic mass evaluations, when CEA levels were >5 ng/mL, 68 percent were found to have nonovarian malignancies. In those with a CA 125/CEA ratio >25, a primary ovarian cancer was found in 82 percent [31].

Individual laboratory assays vary slightly, but the typical upper limit of normal for CEA in nonsmokers is 3.8 micrograms per liter (mcg/L). For smokers, the upper limit of normal is 5.5 mcg/L [32].

CEA testing may also be clinically useful in monitoring patients with pseudomyxoma peritonei, whether the cause is of a benign or malignant etiology [30].

Cancer antigen 19-9 — Cancer antigen 19-9 (CA 19-9) is a mucin protein that may be elevated in ovarian cancer but is used sparingly in ovarian cancer management [33]. CA 19-9 is used primarily to monitor disease response to therapy or detect cancer recurrence in patients with a documented gastric cancer, pancreatic cancer, gallbladder cancer, cholangiocarcinoma, or adenocarcinoma of the ampulla of Vater.

Other — Many biomarkers for EOC are under investigation [34]. The following serum markers have been reported to be potentially useful: osteopontin [35,36], mesothelin [37], lysophosphatidic acid (LPA) [38], haptoglobin [39], transthyretin [40], apolipoprotein A1 [40], serum C-reactive protein [41], and OVX1 [42].

BIOMARKER PANELS AND MULTIMODAL TESTS

OVA1 — OVA1 (also referred to as the multivariate index assay) is a test that includes five serum biomarkers. It was approved by the US Food and Drug Administration (FDA) in 2009 to further assess the likelihood of malignancy in patients who are planning to have surgery for an adnexal mass [43]. OVA1 is performed exclusively through Quest Diagnostics, which is available in 130 countries worldwide.

The OVA1 test incorporates five proteins that are variably expressed in ovarian cancer. Two are up-regulated (cancer antigen [CA] 125 II, beta 2 microglobulin) and three down-regulated (transferrin, transthyretin, apolipoprotein A1) [3,6,44-46].

The proprietary OvaCalc software combines the values for each assay and uses the OVA1 algorithm to generate an ovarian malignancy risk index score. The numeric result ranges from 0.0 to 10.0, and is interpreted as follows [47]:

●Premenopausal patients:

•Low probability of malignancy: OVA1 <5.0

•High probability of malignancy: OVA1 ≥5.0

●Postmenopausal patients:

•Low probability of malignancy: OVA1 <4.4

•High probability of malignancy: OVA1 ≥4.4

Menopausal status is determined either by clinician report or some clinicians also test serum follicle stimulating hormone to determine menopausal status. (See "Clinical manifestations and diagnosis of menopause".)

As with many immunoassays, there may be interference if triglyceride levels exceed 4.5 g/L or rheumatoid factor levels ≥250 international units/mL. This information is provided primarily to assist laboratories with assay interpretation only. There is no need to routinely test for interferents, though if these levels are known to be elevated, OVA1 should not be ordered.

Diagnostic performance — The OVA1 test is a panel of five serum biomarkers that was introduced in 2009 (see 'OVA1' above). It is intended for use to further assess the likelihood of malignancy in patients who are planning to have surgery for an adnexal mass. It has not been studied for ovarian cancer screening.

Prospective studies of the performance of OVA1 include:

●A prospective series that included 524 patients scheduled for surgery for an ovarian tumor included those enrolled by both general gynecologists and gynecologic oncologists [48]. The study compared the diagnostic performance of OVA1 with CA 125 II alone (the thresholds values used were >200 units/mL for premenopausal patients and >35 units/mL for postmenopausal patients) and clinician assessment (based upon history, physical examination, imaging, and CA 125). For a diagnosis of an ovarian malignancy, OVA1 had a better sensitivity than other measures but a lower specificity than CA 125 alone. The results were as follows: OVA1 sensitivity 93 percent, specificity 43 percent; CA 125 sensitivity 69 percent, specificity 84 percent; and clinician assessment sensitivity 75 percent, specificity 79 percent. The combination of OVA1 and clinician assessment had a sensitivity of 96 percent and specificity of 35 percent.

Results were also stratified by menopausal status and stage for OVA1 compared with CA 125. By menopausal status, the sensitivities for EOC were as follows: premenopausal (OVA1: 96 percent; CA 125: 56 percent) and postmenopausal (OVA1: 100 percent; CA 125: 92 percent). In addition, OVA1 had a higher sensitivity than CA 125 for the diagnosis of a primary ovarian cancer in patients with stage I or II disease, particularly in premenopausal patients (82 versus 29 percent, respectively) [48]. The specificity of OVA1 for ovarian malignancy was 52 percent for premenopausal and 33 percent for postmenopausal patients [49].

●A prospective series of 494 patients scheduled for surgery for an adnexal mass and enrolled by general gynecologists reported the following performance measures: OVA1 (sensitivity 92 percent; specificity 54 percent); CA 125 II, using threshold values of >200 units/mL for premenopausal patients and >35 units/mL for postmenopausal patients (74 percent; 95 percent) [50].

Another study evaluated use of OVA1 compared with CA 125 as part of the American College of Obstetricians and Gynecologists (ACOG) referral criteria [51]. Use of OVA1 rather than CA 125 was associated with an increase in the sensitivity of detecting ovarian cancer (from 77 to 94 percent) but a decrease in the specificity (68 to 35 percent). For premenopausal patients, the test performance for the ACOG referral criteria with use of OVA1 had a sensitivity of 91 percent and specificity of 43 percent compared with a sensitivity of 58 percent and specificity of 71 percent for the original ACOG criteria.

In early-stage ovarian cancer, based on pooled data from two prospective OVA1 studies, the sensitivity for detecting ovarian malignancy for OVA1 compared with CA 125 was the following: all ovarian malignancies (92 versus 71 percent); stage I (88 versus 57 percent); and premenopausal patients with stage I and II disease (86 versus 36 percent). OVA1 detected 78 percent of early-stage malignancies that were not detected by CA 125 and 68 percent of those not detected by the ACOG criteria [52].

An analysis of 1110 patients with an adnexal mass concluded that the risk of malignancy increased with rising OVA1 scores. In addition, the likelihood of malignancy was higher for a high-risk versus low-risk ovarian imaging score. The combination of OVA1 and imaging helped to further stratify malignant risk of an ovarian tumor [53].

Overa — In 2016, the FDA cleared Overa as a second-generation multivariate index assay. Initially known as OVA2, this novel panel test utilizes different biomarkers to improve the relatively low specificity of OVA1. Overa combines CA 125 II, human epididymis protein 4 (HE4), apolipoprotein A1, follicle-stimulating hormone, and transferrin [54]. Indications for use are the same as for OVA1. Because follicle-stimulating hormone is part of the panel, determining menopausal status is not required. The Overa score ranges from 0.0 to 10.0 and has the following clinical interpretation:

●Low risk of malignancy <5.0

●High risk of malignancy ≥5.0

Diagnostic performance — Overa is a second-generation multivariate index assay that was redesigned and validated using the same study population as OVA1 [55]. Like OVA1, it is not a screening test but is intended for use to assess the likelihood of malignancy in a patient who is scheduled for surgery for a known ovarian tumor to determine if referral to a specialist is indicated. The test result is calculated using proprietary software named OvaCalc (version 4.0.0). Since OVA1 and Overa were evaluated using the same data set, performance comparisons between the two tests are valid.

●A prospective series of 493 banked serum patient samples from the OVA500 trial were used to validate the new multivariate index assay [55]. Test sensitivity is similar between Overa (91 percent) and OVA1 (94 percent), while test specificity is significantly higher for Overa (69 percent) than for OVA1 (54 percent). The negative predictive values for both Overa and OVA1 are 97 percent, while the positive predictive value for Overa is 40 percent versus 31 percent for OVA1.

Using Overa, the sensitivity for early-stage malignancies (stage I, II) is 86 percent for all subjects and 89 percent for premenopausal patients, similar to OVA1 (91 percent for both) [50,55].

Overa detected 75 percent of malignancies that were missed by clinician assessment alone, compared with 83 percent for OVA1 [50,55].

Serum biomarkers and imaging are complementary tests, and, as for OVA1, the combination of imaging and Overa improves the preoperative assessment of an adnexal mass [56].

Risk of Malignancy Algorithm — The Risk of Malignancy Algorithm (ROMA) includes CA 125 and HE4. It was approved by the FDA in 2011 to further assess the likelihood of malignancy in patients who are planning to have surgery for an adnexal mass [57]. ROMA is available internationally.

ROMA uses CA 125 and HE4 testing and interprets the results using two separate logistic regression algorithms, depending on menopausal status [58]. The algorithms are not proprietary and may be determined with a calculator, through websites, or on smartphone applications. A ROMA score is then reported:

●Premenopausal patients – High risk of malignancy ≥13.1 percent

●Postmenopausal patients – High risk of malignancy ≥27.7 percent

Menopausal status is determined by clinician report.

Diagnostic performance — The Risk of Malignancy Algorithm (ROMA) consists of serum CA 125 and HE4 values that are computed in an algorithm that includes menopausal status. (See 'Risk of Malignancy Algorithm' above.)

The results of a prospective multi-institutional study of 531 patients considered high risk for ovarian cancer (incidence was 24 percent) initially validated this model and reported a sensitivity for ovarian cancer of 92 percent for postmenopausal patients and 76 percent for premenopausal patients [58]. To validate this model in a lower risk population, the ROMA investigators performed a second prospective multi-institutional study of 472 patients (ovarian cancer incidence 10 percent) [59]. In this low-risk cohort, the sensitivity of ROMA was 92 percent and the specificity 76 percent in postmenopausal patients, and 100 percent and 74 percent in premenopausal patients.

A meta-analysis of 11 studies of patients with a pelvic mass found that the overall sensitivity of ROMA for EOC was 89 percent and specificity was 83 percent [60]. For diagnosis of early-stage EOC, the sensitivity was 81 percent and specificity was 76 percent. For premenopausal patients, the sensitivity was 82 percent and specificity was 82 percent. For postmenopausal patients, the sensitivity was 93 percent and specificity was 79 percent.

When ROMA, HE4, and CA 125 were compared, ROMA was the most sensitive (ROMA: 86 percent; HE4: 80 percent; CA 125: 84 percent), and HE4 was the most specific (ROMA: 84 percent; HE4: 94 percent; CA 125: 78 percent); however, these differences were not statistically significant [60]. This analysis did not include a study of 389 patients that found that the combination of HE4 and CA 125 (ROMA) did not improve the performance over the individual biomarkers alone [61].

Risk of Malignancy Index — The risk of malignancy index (RMI) was originally developed in 1990 and is a multimodality approach that combines serum CA 125, pelvic ultrasound, and menopausal status into an index score to predict the risk of ovarian cancer in patients with an adnexal mass [62].

The RMI is primarily used in the United Kingdom, and the calculation for RMI I is included in the United Kingdom National Institute for Health and Clinical Excellence (NICE) guidelines [63,64]:

●RMI I is a product of the ultrasound scan score (U), menopausal status (M), and serum CA 125 level (RMI I = U x M x CA 125). The NICE guidelines advise that all patients with an RMI I score of ≥200 should be referred to a specialist.

●The ultrasound result is scored 1 point for each of the following characteristics: multilocular cyst, solid areas, metastases, ascites, and bilateral masses. U = 0 for an ultrasound score of 0 points, U = 1 for an ultrasound score of 1 point, and U = 3 for an ultrasound score of 2 to 5 points.

●Menopausal status is scored as 1 = premenopausal and 3 = postmenopausal. "Postmenopausal" is defined as no period for more than one year or a patient over 50 years of age who has had a hysterectomy.

●Serum CA 125 is measured in international units/mL.

A number of modifications to the RMI have since been proposed (the versions are referred to as RMI I through IV), which involve assigning different values to each variable [62,65-69]. As an example, in the RMI II, the score for menopause is 1 for premenopausal status and 4 for postmenopausal status, and the ultrasound score is expressed as 1 or 4.

Diagnostic performance — The RMI is a multimodality score that combines serum CA 125, pelvic ultrasound, and menopausal status. (See 'Risk of Malignancy Index' above.)

A systematic review of 109 studies evaluated RMI I through IV compared with 79 other ovarian cancer prediction scoring systems [69]. RMI I and II (using a cutoff score of 200) had the best diagnostic performance: RMI I (sensitivity: 78 percent; specificity: 87 percent) and RMI II (sensitivity: 79 percent; specificity: 81 percent). Similar results were reported in a meta-analysis of 66 studies, with the only notable difference being that the specificity of RMI II was 89 percent [18].

The advantage of the RMI system over other serum biomarker tests is that it combines the three most important clinical elements in predicting the malignant risk of an ovarian tumor. Conversely, the challenge for the RMI is that the formula includes the actual CA 125 value (rather than assigning a point value to different ranges of CA 125 levels). This can be problematic in early-stage cancers and in premenopausal patients, in whom the CA 125 level is often misleading.

ADNEX model — The assessment of different neoplasias in the adnexa (ADNEX) model was first reported in 2014 [70]. It is designed for use in patients with an adnexal mass planned for surgery. The unique aspect of this model is that the results are intended to predict not only whether the mass is malignant, but multiple outcomes, including: benign, borderline, stage I invasive, stage II to IV invasive, and secondary metastatic adnexal tumors. This study has not been validated outside of this European research collaborative group.

It is a computerized model that combines three clinical and six ultrasound predictors:

●Age

●Serum CA 125

●Type of center (oncology centers, defined as tertiary referral centers with a specific gynecologic oncology unit, versus other hospitals)

●Ultrasound features:

•Maximum diameter of lesion

•Proportion of solid tissue (the maximum diameter of the largest solid component divided by the maximum diameter of the lesion)

•>10 cyst locules

•Number of papillary projections (0, 1, 2, 3, >3)

•Acoustic shadows

•Ascites

The results are reported as a percent risk of malignancy.

The model is available at www.iotagroup.org/adnexmodel/.

Diagnostic performance — The ADNEX model has been evaluated in one prospective study [70]. The study included patients from 24 centers in 10 countries, from data from the International Ovarian Tumour Analysis (IOTA) group; data were collected and analyzed as a development set (n = 3506) and a validation set (n = 2403) and a combined pooled data set (n = 5909). In the study, the serum CA 125 was missing in 31 percent of patients, and the value was imputed. Based upon the validation data, for a threshold result of a 10 percent risk of malignancy, distinguishing a benign from a malignant mass had a sensitivity of 97 percent and specificity of 71 percent. The area under the receiver operating characteristic curves (AUC) was calculated based upon the pooled data set; for the discrimination between benign and malignant tumors, the AUC was 0.94. The AUCs for other predictions were benign versus borderline (0.85), benign versus stage I ovarian cancer (0.92), benign versus stage II to IV cancer (0.99), benign versus secondary metastatic (0.95), borderline versus stage I cancer (0.75), and stage II to IV versus secondary metastatic (0.82).

The study was reported by a group of British researchers, and the model was compared with the RMI since this is a standard ovarian cancer prediction tool in the United Kingdom. Based upon the validation data set, for discrimination between benign and malignant tumors, the RMI showed a decrease in AUC (0.88) and sensitivity at a 10 percent malignancy risk threshold (67 percent sensitivity) but an increase in specificity (91 percent) [70].

The ability to distinguish between a benign mass versus an early-stage ovarian cancer appears promising, but the low specificity may result in unnecessary surgery for benign masses. This specificity decreases if a lower malignancy risk threshold is used (for a malignancy risk of 5 percent, specificity was 59 percent [70]), and many experts would advise surgery for patients with a risk at that level or lower. Also, cancer prevalence and inclusion of the type of center (oncology versus other hospital) may impact the results. This depends upon the health care system, the number of oncology centers, the threshold for referral, and whether a center with a specific gynecologic oncology unit also includes general gynecology services. As with other models, this is applied only to patients with an adnexal mass planned for surgery. Thus, a clinician has already made the judgment that surgery is required. In this model, the judgment regarding referral to an oncology center is also included. Thus, the model may be less helpful for clinicians in the key decision of whether to refer to an oncologist. The ADNEX model requires further study for validation and for comparison with other prediction tools.

Other — Other serum biomarkers panels have been investigated for use for ovarian cancer screening but not for the evaluation of patients with an adnexal mass or symptoms associated with ovarian cancer. Use of biomarkers for ovarian cancer screening is discussed in detail separately. (See "Screening for ovarian cancer", section on 'Other tests'.)

CLINICAL APPROACH — Serum biomarkers are used as part of the evaluation for malignancy in patients with an adnexal mass. This evaluation also includes a medical history, physical examination, and imaging studies. The clinician uses this information to arrive at the degree of clinical suspicion that a mass is malignant. Based upon this, the clinician decides whether the patient should undergo diagnostic surgery or continued surveillance. For masses that are highly suspicious for ovarian cancer, referral to a gynecologic oncologist is appropriate (table 3).

General principles of the evaluation and management of an adnexal mass are discussed in detail separately. (See "Approach to the patient with an adnexal mass".)

The evaluation of patients with symptoms associated with epithelial ovarian cancer, but with no adnexal mass, has not been well studied. There may be a role for serum markers in the assessment of these patients.

Preoperative measurement of biomarkers in patients with possible ovarian cancer has other potential functions. A baseline level is established for use for further monitoring during and after treatment. In addition, biomarkers may play a role in predicting whether optimal cytoreduction is feasible. (See "Cancer of the ovary, fallopian tube, and peritoneum: Surgical cytoreduction", section on 'Predicting feasibility of cytoreduction'.)

Patients with an adnexal mass — The degree of suspicion for malignancy in an adnexal mass is based upon all the available clinical information, including: appearance of the mass on imaging, evidence of metastatic disease, menopausal status, risk factors, biomarker testing, and symptoms. The clinician must determine whether to proceed with surgical exploration and whether referral to a gynecologic oncologist is appropriate.

Use of the combination of these factors to determine the management of an adnexal mass is discussed in detail separately. (See "Approach to the patient with an adnexal mass".)

Deciding to proceed with surgery — The decision whether to proceed with surgical evaluation for a patient with an adnexal mass depends mostly upon the appearance of the mass on imaging and other factors, rather than on a biomarker test.

The main biomarker that has been studied for use in the initial evaluation of an adnexal mass is cancer antigen 125 (CA 125), although this is not its US Food and Drug Administration-approved indication. OVA1, Overa, and the Risk of Malignancy Algorithm (ROMA) have only been studied in patients for whom surgery has already been planned and thus likely have a higher prevalence of ovarian cancer than the general population of patients with an ovarian tumor. In the absence of data regarding the use of OVA1, Overa, and ROMA in the initial evaluation of an adnexal mass, we do not recommend use of these tests alone to decide whether to proceed with surgical exploration for an adnexal mass.

CA 125 adds limited information to the decision to proceed with surgery. It has a low sensitivity overall, and particularly for detection of early-stage disease. In addition, it has a low specificity, particularly in premenopausal patients, as discussed above. (See 'Diagnostic performance' above.)

In premenopausal patients, the decision whether to proceed with surgical evaluation of an adnexal mass is a diagnostic dilemma. Pelvic ultrasound can identify ovarian masses with characteristics of malignancy, but many benign tumors in younger patients also have complex ovarian morphology. For some patients, the decision to proceed with surgery remains uncertain, even after consideration of the ultrasound findings combined with risk factors, symptoms, and physical examination. In such cases, many clinicians draw a serum CA 125. However, the result is informative only if it is very elevated. The sensitivity and specificity of CA 125 is low in general, and particularly in premenopausal patients [18]. Many benign conditions found in this population are associated with a slightly elevated CA 125 (table 2). The definition of a very elevated CA 125 is not well established; a level of >200 units/mL has been proposed, but is supported by few data [21]. Ultimately, referral of a premenopausal patient with an adnexal mass to a gynecologic oncologist is based on the judgment of the clinician after considering both the CA 125 level and other clinical factors [22] or in patients in whom there remains diagnostic uncertainty. Thus, a CA 125 level alone without other information is not useful for determining which patients should proceed to surgery. If a CA 125 is drawn, the patient should be counseled about its limitations.

The risk of ovarian cancer is significantly higher in postmenopausal than premenopausal patients. Surgical exploration is typically performed for most complex ovarian masses in postmenopausal patients. Exceptions to this are apparently benign masses that are stable in size and appearance (eg, a mass with an appearance consistent with an endometrioma that was documented prior to menopause). Ovarian cysts with a simple sonographic appearance (unilocular, thin walls, anechoic fluid) that are <10 cm in diameter are unlikely to be malignant [71]. In our practice, for postmenopausal patients with a simple cyst, we draw a serum CA 125, and if the result is <35 units/mL and no symptoms or risk factors associated with ovarian cancer are present, we follow with serial ultrasound. Based on a United Kingdom Collaborative Trial of Ovarian Cancer Screening (UKCTOCS) nested cohort study, the use of human epididymis protein 4 in addition to CA 125 and transvaginal ultrasound provides little additional value in the differential diagnosis of an adnexal mass in postmenopausal patients and is not recommended [28].

Prior to surgical exploration, another indication for drawing a CA 125 is as a baseline to monitor response to therapy, if ovarian cancer is ultimately diagnosed. (See "First-line chemotherapy for advanced (stage III or IV) epithelial ovarian, fallopian tube, and peritoneal cancer", section on 'Evaluation after adjuvant chemotherapy'.)

Referral to a gynecologic oncologist — When the decision to proceed with surgical evaluation of an adnexal mass has been made, referral to a gynecologic oncologist is appropriate for patients with a high likelihood of ovarian cancer.

The American College of Obstetricians and Gynecologists (ACOG) has published guidelines regarding referral of patients with an adnexal mass to a gynecologic oncologist (table 3) [72]. As noted above, one study found that the use of OVA1 with these criteria had a higher sensitivity for ovarian cancer than the referral criteria alone [51].

Given the available evidence, we suggest OVA1, Overa, or ROMA rather than CA 125 to decide whether a patient who is planned for surgical evaluation for an adnexal mass should be referred to a gynecologic oncologist. Test availability and expense may factor into which test is most appropriate for each clinical situation.

A recent report evaluated the impact of the OVA1 test on patient referral. The cancer prevalence for the study population was 21 percent and the observed clinical practice referral rate was 60 percent. The authors compared test sensitivity with the predicted percentage of patients referred to a specialist. The results were: clinical assessment (73; 24 percent), CA 125 (68; 20 percent), modified ACOG criteria (79; 56 percent), and OVA1 (90; 56 percent). They concluded that observed referral rates were higher than intuitively anticipated, and that the high sensitivity of OVA1 did not lead to an increase in the number of patients referred to a gynecologic oncologist, but enriched the referred cohort by including 90 percent of the cancers [73].

The ACOG referral criteria are discussed in detail separately. (See "Approach to the patient with an adnexal mass", section on 'When to refer to a gynecologic oncologist'.)

Patients with ovarian cancer symptoms — Some patients develop symptoms of epithelial ovarian cancer (EOC), but no adnexal mass is found on evaluation. The role of serum markers in this patient population is not well defined.

The usual indication for a surgical evaluation for ovarian or fallopian tube carcinoma is the finding of an adnexal mass that is suspicious for malignancy. In general, other features (eg, symptoms, risk factors, laboratory results) may contribute to the clinical suspicion of malignancy, but are usually not a sufficient indication for surgery. One exception to this is that some experts perform diagnostic laparoscopy for patients who have both EOC-associated symptoms (table 4) and an elevated serum biomarker.

Symptoms associated with EOC are discussed in detail separately. (See "Early detection of epithelial ovarian cancer: Role of symptom recognition".)

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Ovarian and fallopian tube disease".)

SUMMARY AND RECOMMENDATIONS

●Ovarian cancer is the second most common gynecologic malignancy and the most common cause of gynecologic cancer death in the United States. Epithelial ovarian carcinoma (EOC) is the most common histologic type of ovarian cancer. It is typically diagnosed at an advanced stage. (See 'Introduction' above.)

●Use of serum biomarkers for the diagnosis of EOC is an active area of investigation. Biomarkers or panels of biomarkers are generally used in combination with each other or with imaging. (See 'Introduction' above and 'Biomarkers' above.)

●Several biomarkers are commercially available, including cancer antigen 125 (CA 125), carcinoembryonic antigen (CEA), human epididymis protein 4 (HE4), the OVA1 panel, the Overa panel, the Risk of Malignancy Algorithm (ROMA), and the Risk of Malignancy Index (RMI). (See 'Biomarkers' above.)

●Serum CA 125 is the most commonly used laboratory test for the evaluation of adnexal masses for EOC, but the test has several shortcomings. CA 125 testing alone has a low sensitivity, particularly for early-stage ovarian cancer, and a low overall specificity. The specificity is particularly low in premenopausal patients. (See 'Diagnostic performance' above.)

●The OVA1 test is a panel of five serum biomarkers that was introduced in 2009. ROMA consists of serum CA 125 and HE4 values that are computed in an algorithm that includes menopausal status that was introduced in 2011. Overa includes CA 125, HE4, and three additional markers. Each of these biomarker tests is intended for use to further assess the likelihood of malignancy in patients who are planning to have surgery for an adnexal mass. (See 'OVA1' above and 'Risk of Malignancy Algorithm' above and 'Overa' above.)

●The performance of OVA1 or Overa in predicting malignancy is improved when combined with imaging. (See 'Diagnostic performance' above and 'Diagnostic performance' above.)

●The decision on whether to proceed with surgical evaluation for a patient with an asymptomatic adnexal mass depends on several factors, but mostly on the appearance of the mass on ultrasound or other imaging test rather than a specific serum biomarker result. The only biomarker that has been studied for use in the initial evaluation of an adnexal mass is CA 125, although this is not its approved indication. OVA1, Overa, and ROMA have been studied only in patients who have already been evaluated and for whom surgery is planned, and we do not recommend use of these tests alone to decide whether to proceed with surgical exploration for an adnexal mass. (See 'Deciding to proceed with surgery' above.)

●For making the decision whether to refer a patient with an adnexal mass to a gynecologic oncologist, we suggest OVA1, Overa, or ROMA rather than CA 125. Test availability and expense may impact which test is most appropriate for each clinical situation. (See 'Referral to a gynecologic oncologist' above.)