Proteomics-Based Testing Related to Ovarian Cancer
OVA1 and ROMA tests are combinations of several separate lab tests and involve a proprietary algorithm for determining risk (i.e., they are what the American Medical Association’s CPT calls “Multianalyte Assays with Algorithmic Analyses” [MAAAs]).
CPT instructs that these codes cannot be reported with the component tests (i.e., codes 86304 and 86305 cannot be reported with 81500, and codes 82172, 82232, 83695, 83700, 84134, 84466, and 86304 cannot be reported with 81503).
A variety of gene-based biomarkers have been studied in association with ovarian cancer. Of particular interest have been tests that integrate results from multiple analytes into a risk score to predict the presence of disease. Two tests based on this principle (Ova1™ test and ROMA™ test) have now been cleared by U.S. Food and Drug Administration (FDA)for use in women with adnexal masses as an aid to further assess the likelihood that malignancy is present.
The OVA1 and ROMA tests have both been analytically validated and clinical performance has been reported in prospective multicenter clinical studies. Changes in the observed sensitivity and negative predictive value of testing compared with clinical assessment has been small and of uncertain diagnostic value. Studies on the diagnostic accuracy of these tests compared with other diagnostic tools have had mixed findings, but do not report that ROMA is superior to other risk prediction tools that use standard clinical information or single markers. No studies have been performed that directly evaluate the impact on patient management e.g., referral patterns, and no studies have evaluated the impact on health outcomes. Clinical input from academic medical centers and specialty societies did not show consensus that this test improved outcomes when used as a tool to triage patients with adnexal masses. As a result of the evidence and clinical input, these tests are considered investigational pending more information about their performance and impact on outcomes.
In 2009, more than 21,000 women in the U.S. are diagnosed each year with ovarian cancer and more than 14,000 died of this disease. (1) The mortality rate depends on three variables: 1) characteristics of the patient; 2) the biology of the tumor (grade, stage, and type); and 3) the quality of treatment (nature of staging, surgery and chemotherapy used). (2) In particular, comprehensive staging and completeness of tumor resection appear to have a positive impact on patient outcome.
In 1997, the Society of Surgical Oncology recommended ovarian cancer surgery and follow-up treatment be performed by physicians with ovarian cancer disease expertise. (3) To date, dozens of articles have been published relevant to this recommendation looking at long-term outcomes, short-term outcomes, and process measures (e.g. types of treatment such as complete staging or tumor debulking). At least two meta-analyses have concluded that outcomes are better in patients with ovarian cancer when they are treated by gynecologic oncologists. (4, 5) Data have been most convincing for patients with advanced stage disease.
Adult women presenting with an adnexal mass have an estimated 68% likelihood of having a benign lesion. (6) About 6% have borderline tumors, 22%, invasive malignant lesions, and 3%, metastatic disease. Clinicians generally agree that women with masses that have a high likelihood of malignancy should undergo surgical staging by gynecologic oncologists. However, women with clearly benign masses do not require referral to a specialist. Criteria and tests that help differentiate benign from malignant pelvic masses are thus desirable.
In 2005, the American College of Obstetricians and Gynecologists (ACOG) and the Society of Gynecologic Oncologists (SGO) jointly released referral guidelines that address criteria for referring women with pelvic masses that are suspicious for ovarian cancer to gynecologic oncologists. (7) Separate criteria were developed for pre-menopausal and post-menopausal women. In premenopausal women, referral criteria included at least one of the following: elevated CA 125 (greater than 200 unites per mL), ascites, evidence of abdominal or distant metastasis, or a positive family history. The referral criteria in postmenopausal women were similar, except that a lower threshold for an elevated CA-125 test was used (35 units per ML) and nodular or fixed pelvic mass was an additional criterion.
Two proteomic tests have now been cleared by the U.S. Food and Drug Administration (FDA) with the intended use to triage patients with adnexal masses. A suggested use of the test is to identify women with a positive test who have a higher likelihood of malignant disease and may benefit from referral to a gynecologic-oncology specialist. Patients with positive results may be considered candidates for referral to a gynecologic oncologist for treatment.
On July 16, 2009, the OVA1TM test (Vermillion, Inc. Fremont, CA) was cleared for market by the FDA as a 510(k) submission. On September 1, 2011, the Risk of Ovarian Malignancy Algorithm (ROMATM test, Fujirebio Diagnostics, Inc., Malvern, PA) was cleared by the FDA as a 510(k) submission. Because the OVA1 test had been found to be a class II medical device by virtue of the July 2009 clearance, ROMA was found to be substantially equivalent to that predicate device. Intended use of OVA1 is as an aid to further assess the likelihood that malignancy is present when the physician’s independent clinical and radiological evaluation does not indicate malignancy. Intended use of ROMA is as an aid in assessing whether a premenopausal or postmenopausal woman who presents with an ovarian adnexal mass is at high or low likelihood of finding malignancy on surgery. Neither test is FDA-cleared as a screening or stand-alone diagnostic assay.
FDA product code: ONX
Black Box Warning
On December 10, 2011, the FDA published an amendment to the regulation for classifying ovarian adnexal mass assessment score test systems to restrict these devices so that a prescribed warning statement that addresses off-label risks be highlighted by a black box warning. (8) The warning is intended to mitigate the risk to health associated with off-label use as a screening test, stand-alone diagnostic test, or as a test to determine whether or not to proceed with surgery.
Medical policies are systematically developed guidelines that serve as a resource for Company staff when determining coverage for specific medical procedures, drugs or devices. Coverage for medical services is subject to the limits and conditions of the member benefit plan. Members and their providers should consult the member benefit booklet or contact a customer service representative to determine whether there are any benefit limitations applicable to this service or supply.
This policy was originally created in 2010 and was updated regularly with searches of the MEDLINE database. Most recently, the literature was reviewed through September 25, 2014. The literature review focuses on the following three issues related to evaluation of diagnostic tests:
Assessment of a diagnostic technology typically focuses on 3 parameters: 1) technical performance; 2) diagnostic performance (sensitivity, specificity, and positive and negative predictive value in appropriate populations of patients; and 3) demonstration that the diagnostic information can be used to improve patient outcomes (clinical utility).
Technical performance is typically assessed with 2 types of studies, those that compare test measurements with a gold standard and those that compare results taken with the same device on different occasions (test-retest).
Diagnostic performance is evaluated by the ability of a test to accurately diagnose a clinical condition in comparison with the gold standard. The sensitivity of a test is the ability to detect a disease when the condition is present (true -positive), while specificity indicates the ability to detect patients who are suspected of disease but who do not have the condition (true -negative). Evaluation of diagnostic performance, therefore, requires independent assessment by the 2 methods in a population of patients who are suspected of disease but who do not all have the disease.
Clinical utility involves assessing the data linking use of a test to improvement in patient management and/or health outcomes. While in some cases, tests can be evaluated adequately using technical and diagnostic performance. When a test identifies a new or different group of patients with a disease; randomized trials are needed to demonstrate impact of the test on the net health outcome.
Descriptions of the developmental process for the OVA1 test have been published in FDA documents and in a perspective by Fung in 2010. (9-11) candidate biomarkers were selected based on initial studies using mass spectroscopy but were converted to standard immunoassays to improve analytical performance. Seven final markers were evaluated, none of which individually appeared to be highly specific for malignant ovarian disease. However, the choice of 5 of these (CA 125, prealbumin, apolipoprotein A-1, beta 2 microglobulin, and transferrin) produced a composite profile that did appear to have discriminatory ability. The test, as cleared by FDA, is performed on a blood sample, which is to be sent to a reference laboratory for testing using the 5 immunoassays described above. Results of the 5 determinations are entered manually into an Excel® spreadsheet used by the OvaCalc software. This software contains an algorithm which combines the 5 discrete values into a single unitless numerical score from 0.0 to 10.0.
Details of the algorithm appear proprietary, but development is described as an empiric process, based on use of banked samples from academic partners, on a small prospective study of samples from Europe and using a designated subset of samples from the clinical study used to support submission to the FDA. It appears at an undisclosed point in the developmental process as a result of interaction with FDA; separate cutpoints were developed for premenopausal and postmenopausal women.
A similar developmental process was described for ROMA by Moore et al. (12) They studied 9 biomarkers and chose human epididymis secretory protein 4 (HE4) and CA 125 because these markers in tandem produced the best performance. The algorithm developed was subsequently modified to include menopausal status and was independently validated. (13) Again, Separate cut-offs were used for premenopausal and postmenopausal women.
The OVA1 is a qualitative serum test that combines immunoassay results for the 5 analytes described above (CA 125, prealbumin, apolipoprotein A-1, beta 2 microglobulin, and transferrin) into a single numerical score. Analytical performance for the test demonstrated good test precision (coefficient of variation (CV) ranging from 1% to 7.4%, depending on the sample levels studied) and good reproducibility (CV from 2.8% to 8.9%). The test appears linear, reagent and samples stable, and there was no observed interference evaluating common endogenous substances (e.g., hemoglobin, bilirubin)
The ROMA test is also a qualitative serum test that combines 2 analytes HE4 EIA and the ARCHITECT CA 125, along with menopausal status into a numerical score. Analytical performance for the ROMA also exhibited good precision with a total CV ranging from 0.49% to 7.72%, depending on both sample values and menopausal status. The reproducibility of the test was acceptable, with a CV that ranged from 0.98 to 25.9%, with highest values observed in patients with low scores, as expected. The reagents are variably stable, and users are instructed to follow package inserts for stability on each analyte used. The test was unaffected by interference with hemoglobin, bilirubin, lipids, or human anti-mouse antibodies (HAMA). However, high levels of rheumatoid factor (more than 500 IU/mL) did appear to cause elevations in test values, and testing in patients with elevated rheumatoid factor is not recommended.
Diagnostic performance of the OVA1 test was evaluated in a single prospective, double-blind clinical study using 27enrollment sites. (11) The study was supported by the commercial sponsor of the test. Patients underwent a complete clinical evaluation before surgical intervention, and only patients with adnexal masses who had a planned surgical intervention were included. The study enrolled a total of 743 patients, with 146 subjects used in the training set and 516 in the testing set. Seventy-four patients were excluded because of missing information or samples. The final prevalence of cancer in the population was 27%.
Using pathologic diagnosis as the gold standard, test performance, when combined with presurgical assessment for benign disease, was as follows in the hands of non-gynecological oncologists (see Table 1).
Table 1. Diagnostic Performance of OVA1
* Confidence intervals not provided.
No additional studies evaluating the diagnostic performance of the OVA1 test were identified in literature updates.
In 2014, Wang et al. published a meta-analysis of studies evaluating the diagnostic accuracy of the ROMA algorithm and comparing it to the performance of single markers HE4 and CA125.14 To be included in the meta-analysis, studies had to investigate both HE4 and CA125 or calculate ROMA, include women with ovarian cancer and benign gynecologic disease, use pathology diagnosis as the reference standard, and collect blood samples before treatment was initiated. A total of 32 studies met these inclusion criteria; 6 of these were conducted in the United States. Findings of the overall pooled analysis of diagnostic accuracy are presented in Table 2.
Table 2. Diagnostic Performance of ROMA compared with HE4 and CA125: Meta-analysis findings
Findings were similar when diagnostic performance in premenopausal women and postmenopausal women were evaluated separately. ROMA had similar or higher sensitivity than HE4 and CA125, and HE4 had the highest specificity.
The study with the largest sample size was published by Karlsen et al. in 2012. (15) The study included 1218 women presenting with pelvic masses. Prior to diagnosis, HE4 and CA125 levels were obtained, and ROMA and the Risk of Malignancy Index (RMI; an index consisting of ultrasound findings, menopausal status and CA125 levels) were calculated. At a fixed sensitivity of 94.4%, the specificity of ROMA was 76.5% and the specificity of RMI was 81.5%. At a fixed specificity of 75%, the sensitivity of ROMA was 94.8% and the sensitivity of RMI was 96.0%. Accuracy of ROMA and RMI were not compared statistically, but appeared to be similar. In another study, Kaijser and colleagues evaluated 360 women with pelvic masses who were scheduled for surgery. (15) The study compared the diagnostic accuracy of ROMA and an ultrasound-based prediction model (LR2) developed by the International Ovarian Tumor Analysis Study (IOTA). Histology was used as the reference standard. The overall performance of LR2 (94% sensitivity and 82% specificity) was significantly better than ROMA (84% sensitivity and 80% specificity).
The ideal study design to evaluate clinical utility of proteomics-based testing would be a randomized controlled trial comparing patient management decisions (e.g., referral patterns) and/or health outcomes (e.g., mortality) in patients managed with the tests with those managed according to best current clinical practices. No randomized or nonrandomized studies with these comparisons were identified.
A TEC Assessment was completed in 2012 on “Multi-analyte testing for the evaluation of adnexal masses.” (17) The Assessment included evaluation of both the OVA1 and ROMA tests in regards to their impact on health outcomes. The following conclusions were made:
“The evidence regarding the effect of OVA1 and ROMA and effects on health outcomes is indirect and based on studies of diagnostic performance of the tests in patients undergoing surgery for adnexal masses. Although the studies show improvements in sensitivity and worsening of specificity with the use of the tests in conjunction with clinical assessment, there are problems in concluding that this results in improved health outcomes. The clinical assessment performed in the studies is not well characterized. Although OVA1 improves sensitivity, specificity declines so much that most patients test positive. ROMA does not improve the sensitivity of testing to a great extent. Underlying these issues is uncertainty regarding whether there would be actual health benefits based on altering patient referral based on these tests…Whether use of OVA1 or ROMA improves the net health outcome or is as beneficial as other diagnostic strategies has not been demonstrated. Referring all patients to a physician with expertise in staging and debulking ovarian cancer is a reasonable clinical alternative with no harm.”
In 2014, Kaijser et al. published a study that did not directly evaluate clinical utility but that provides relevant information. (18) It was a retrospective cohort study that included 101 newly diagnosed cases of biopsy-proven invasive ovarian cancer. Blood samples obtained before treatment were analyzed; HE4 and CA125 levels were measured and the ROMA algorithm was calculated. Median overall survival in the study cohort was 3.7 years. In a multivariate analysis controlling for confounding variables, neither HE4 levels nor ROMA were independently associated with progression-free survival (PFS) or disease-specific survival (DSS). For example, for ROMA and the outcome of PFS, the adjusted hazard ratio (HR) for each 10% increase in risk was 0.98, 95% confidence interval (CI), 0.88 to 1.11. Patients were not prospectively managed according to their HE4 levels or ROMA score and thus the actual impact of these tests on PFS and DSS cannot be determined from this study.
Clinical Input Received through Physician Specialty Societies and Academic Medical Centers
While the various physician specialty societies and academic medical centers may collaborate with and make recommendations during this process, through the provision of appropriate reviewers, input received does not represent an endorsement or position statement by the physician specialty societies or academic medical centers, unless otherwise noted. Results of clinical input in 2012 revealed mixed support for the use of this test as a tool for triaging patients with an adnexal mass. Reviewers agreed that the evidence was insufficient to determine the impact of these tests on referral patterns. For indications other than triaging patients with an adnexal mass, there was a lack of support for use of these tests.
Summary of Evidence
The OVA1 and ROMA tests have both been analytically validated and clinical performance has been reported in prospective multi-center clinical studies. Changes in the observed sensitivity and negative predictive value of testing compared to clinical assessment has been small and of uncertain diagnostic value. Studies on the diagnostic accuracy of these tests compared to other diagnostic tools have had mixed findings, but do not report that ROMA is superior to other risk prediction tools that use standard clinical information or single markers. No studies have been performed that directly evaluated the impact on patient management e.g., referral patterns, and no studies have evaluated the impact on health outcomes. Clinical input from academic medical centers and specialty societies did not show consensus that this test improved outcomes when used as a tool to triage patients with adnexal masses. As a result of the evidence and clinical input, these tests are considered investigational pending more information about its performance and impact on outcomes.
Clinical Practice Guidelines and Position Statements
The American Congress of Obstetricians and Gynecologists addressed the use of the OVA1 test in their guidelines on the role of the obstetrician-gynecologist in the early detection of epithelial ovarian cancer, last updated in 2011. (19) This document made the following statements:
The National Institute for Health and Clinical Excellence issued guidance in 2011 on the recognition and management of ovarian cancer. (21) These guidelines made the following recommendations:
The National Comprehensive Cancer Network (NCCN) guideline on ovarian cancer (v3.2014) includes The following statement: (22)
“It has been suggested that specific biomarkers (serum HE4 and CA-125) along with an algorithm (Risk of Ovarian Malignancy Algorithm [ROMA]) may be useful for determining whether a pelvic mass is malignant or benign
The FDA has approved the use of HE4 and CA-125 for estimating the risk of ovarian cancer in women with a pelvic mass. Currently, the NCCN panel does not recommend the use of these biomarkers for determining the status of an undiagnosed pelvic mass.”
U.S. Preventive Services Task Force Recommendations
In 2012, The U.S. Preventive Services Task Force recommended against screening women for ovarian cancer (D recommendation). (23) The task force has not addressed proteomics-based testing related to ovarian cancer.
Medicare National Coverage
There is no national coverage determination (NCD). In the absence of an NCD, coverage decisions are left to the discretion of local Medicare carriers.
Unlisted chemistry procedure
Oncology (ovarian), biochemical assays of two proteins (CA-125 and HE4), utilizing serum, with menopausal status, algorithm reported as a risk score
Oncology (ovarian), biochemical assays of five proteins (CA-125, apoliproprotein A1, beta-2 microglobulin, transferrin and pre-albumin), utilizing serum, algorithm reported as a risk score
Type of Service
Place of Service
Add to Pathology/Laboratory Section - Reviewed by OAP on May 12, 2011. New medical policy.
Replace policy. Policy updated to change the use of the OVA1™ and ROMA tests from medically necessary to investigational for all indications. Rationale updated based on a literature review through September 2012, results of TEC Assessment, and results of clinical vetting. References 7, 13, 17-28 added; others renumbered or removed. New CPT codes added. Policy statement changed as noted.
Update Related Policies. Add 2.03.501.
Update Related Policies. Add 12.04.66 and remove policy 2.04.34 as it was archived.
Replace policy. Policy updated with literature search through September 30, 2013. References 14, 15 and 20 added; other references renumbered or removed. No change to policy statement. Title changed to Proteomic-based Testing Related to Ovarian Cancer. Clarification note added that this policy is only to be used when HE4 is included in the ROMA combination test. When HE4 is billed as an individual test, 12.04.66 – Serum Biomarker Human Epididymis Protein 4 (HE4) should be used.
Annual review. Policy updated with literature review through September 25, 2014. References 1, 14, 18 and 23 added. Policy statement unchanged. Policy title changed to “Proteomics-Based Testing Related to Ovarian Cancer”. ICD-9 and ICD-10 diagnosis and procedure codes removed; these are not utilized in adjudication of the policy.
Disclaimer: This medical policy is a guide in evaluating the medical necessity of a particular service or treatment. The Company adopts policies after careful review of published peer-reviewed scientific literature, national guidelines and local standards of practice. Since medical technology is constantly changing, the Company reserves the right to review and update policies as appropriate. Member contracts differ in their benefits. Always consult the member benefit booklet or contact a member service representative to determine coverage for a specific medical service or supply. CPT codes, descriptions and materials are copyrighted by the American Medical Association (AMA).