Response to Comments: Multimarker Serum Tests Related to Ovarian Cancer Testing

Comments 1-42 were submitted by Vermillion.

Critique of Proposed NGS OVA1 LCD (DL38371) by Vermillion has been completed and the following gaps have been noted to bring to the attention of NGS:

1. In NGS’ proposed LCD on OVA1, NGS notes that, “It is estimated that only about 33% to 60% of ovarian cancers are treated by a gynecologist/oncologist.⁵”

1. Gap: 40% to 67% of ovarian cancer patients are not getting to a Gyn Oncologist and receiving NCCN guideline-adherent care is a major healthcare problem for a cancer as deadly as ovarian cancer.
2. NCCN guidelines recommend that all patients with ovarian cancer undergo surgery by a gynecologic oncologist (Morgan et al, 2008).
3. Contemporary population-based studies have shown that the proportion of women with ovarian cancer that undergo surgery by these gynecologic oncologists is less than 50% (Carney et al, 2002; Olaitan et al, 2001; Elit et al, 2002; Vernooj et al, 2008; Earle et al, 2006).
4. Following diagnosis, less than 40% of patients receive care that adheres to National Comprehensive Cancer Network (NCCN) guidelines (treatment by a gynecologic oncologist) although studies have shown adherence to these guidelines is associated with improved oncologic outcomes (Bristow et al, 2013c; Morgan et al, 2013).
5. Early involvement of specialists is associated with improved survival (Carney et al, 2002; Engelen et al, 2006; Eisenkop et al, 1992; Chan et al, 2007).
6. Individual reports and meta-analysis have reported that referral to a gynecologic oncologist improves outcomes for ovarian cancer patients with higher adherence to guidelines, a higher fraction of optimal cytoreduction, optimal chemotherapy and improved overall survival (du Bois et al, 2009; Vernooiji et al, 2007; Engelen et al, 2006; Mercado et al, 2010).
   

The draft has been amended to include some of the above information as well as supporting references. The studies by Olaitan and Vernooij (2009) were not included as they were not conducted in the United States.

Actual clinical practice referral rates are 60%. If MVIA1/OV1/MVIA2 resulted in a referral for every positive test, referral rates would be anywhere from 42%-67% (depending on study referencing). So even the test would not increase referrals.

Also, Bristow suggests that the test wouldn’t help in cases of geographical scarcity of gynecologic oncologists - almost half of the incident ovarian cancer cases in the U.S. occur in a county without a gynecologic oncologist.

b. NGS also notes in this proposed LCD that the estimated 5-year survival rate for ovarian cancer is 92.5% when ovarian cancer is diagnosed in Stage I. But only 15% of ovarian cancer cases are diagnosed in Stage I.

1. Gap: The majority (over 65%) of ovarian cancer cases are diagnosed in late stages when the 5-year survival is <30% (SEER data).
2. This is the status quo for ovarian cancer: 5-year survival of <30% on over 65% of all cases; and at least 40% of all ovarian cancer patients not receiving NCCN guideline-adherent care. This status quo is unacceptable and must change.

NGS agrees with the above statements. However, there are no data demonstrating how OVA-1 will direct patients to receive NCCN guideline-adherent care.

c. In this proposed LCD, NGS states that 15 studies were identified using Hayes search criteria, but only six of those studies met the inclusion criteria.

1. Gap: If this is a non-biased review, we should understand which nine studies were excluded? What criteria were used to exclude them?
2. In the interest of patient care and transparency, the search criteria and exclusion criteria should be disclosed publically, as well as which studies were excluded for consideration and why.
3. How can this proposed LCD be commented on comprehensively unless those important details are disclosed?

The proposed LCD has been amended to reflect that a PubMed search was performed. Hayes was utilized as an ancillary tool.

The following studies were excluded for the following reasons:

OVA1 was not the index test:

1. Amonkar et al., Development and preliminary evaluation of a multivariate index assay for ovarian cancer. PLoS One. 2009;4(2). 2009
2. Yip et al., Comprehensive serum profiling for the discovery of epithelial ovarian cancer biomarkers. PLoS One. 2011;6(12) 2011
3. Coleman et al., Validation of a second-generation multivariate index assay for malignancy risk of adnexal masses. Am J Obstet Gynecol. 2016 Jul;215(1). 2016

OVA1 performance assessed as an individual test; test not performed with the intended FDA-cleared intended usage as an adjunct test

4. Grenache et al. Clinical performance of two multi-marker blood tests for predicting malignancy in women with an adnexal mass. Clin Chim Acta. 2015 Jan 1;438:358-63. 2015

The primary endpoint of the study was to examine how pelvic imaging influenced the OVA1 score, not the performance of the OVA1 test

5. Goodrich et al. The effect of ovarian imaging on the clinical interpretation of a multivariate index assay. Am J Obstet Gynecol. 2014 Jul;211(1):65. 2014

Cost-evaluation studies

6. Kim et al. Management of complex pelvic masses using a multivariate index assay: a decision analysis. Gynecol Oncol. 2012 Sep;126(3):364-8. 2012
7. Havrilesky et al., 2015 Costs, effectiveness, and workload impact of management strategies for women with an adnexal mass. J Natl Cancer Inst. 2014 Dec 16;107(1):322
8. Forde et al., Cost-effectiveness analysis of a multivariate index assay compared to modified American College of Obstetricians and Gynecologists criteria and CA-125 in the triage of women with adnexal masses. Curr Med Res Opin. 2016;32(2):321-9 2016

Use of OVA1 for indications other than adnexal mass assessment

9. Macuks et al. An ovarian cancer malignancy risk index composed of HE4, CA125, ultrasonographic score, and menopausal status: use in differentiation of ovarian cancers and benign lesions. Tumour Biol. 2012 Oct;33(5):1811-7. 2012

NGS also states that “most” of the included OVA1 studies had overlapping populations. That is likely due to the nine studies excluded from consideration as well as studies missed using Hayes search criteria. Since NGS has not disclosed which studies were excluded and why, it is not possible to clearly identify which ones were missed and not considered.

Please see above comments.

Ueland (2011): “Patients were enrolled from 27 sites throughout the United States, including women’s health clinics, obstetrics and gynecology groups, gynecologic oncology practices, community and university hospitals, and health maintenance organizations" (see the Appendix,available online at http://links.lww.com/AOG/A240). This article, published in Obstetrics and Gynecology (the official publication of ACOG) rated this as Level III evidence.

Bristow (2013) “Consecutive patients who met inclusion criteria were prospectively enrolled at 27 sites throughout the United States.” This study did not provide a weblink or any reference as to where exactly participants came from but it can be inferred, perhaps erroneously that these were the same 27 sites? The authors noted “independent of the original dataset reported by Ueland et al.” but it is not clear what this means.

Bristow (2013) “From the combined clinical trial populations, the subset of patients enrolled by nongynecologic oncologist providers was selected for further study…A total of 1110 subjects were prospectively enrolled at 44 sites across the United States,”. It appears this study combined at least some of the patients in the prior two studies.

Ware Miller (2011) et al utilized the same patients “This multi-institutional trial enrolled patients from 27 primary care and specialty sites across the United States (see the Appendix, available online at http://links.lww.com/AOG/A243).” This article, also published in Obstetrics and Gynecology, rated this as Level III evidence.

Lastly, Longoria (2014) “Female subjects with a planned surgical procedure for removal of an adnexal mass were enrolled in 2 prospective OVA1 studies at 44 sites across the United States”. It appears this was same population as the second Bristow study.

d. In the Ueland et al 6, study, NGS states:“…The study results showed that 269 of the 524 patients included in the analysis portion of the study were enrolled by non-gynecologic oncologists meaning 321 were enrolled by gynecologic oncologists. This most likely skewed the assessment/conclusion that OVA1® demonstrated a high sensitivity and positive predictive value (PPV). Another limitation was the sensitivity appeared to vary depending on menopausal status and not all results were reported separately for patients enrolled by nongynecologic oncologists as specified by the FDA clearance.

Gaps: The test is indicated for women planned for surgery and this intervention occurred before the gynecologic oncologist performed the surgery. ii. The enrollment by gynecologic oncologists reflects the necessity to have an adequate number of cancers to reach conclusions. It does not “skew” the data.

1. What was the designated “adequate number of cancers to reach conclusions” and how was it calculated? The word “skewed’ has been removed and this paragraph has been amended for clarity. However, it was used to indicate that if the test is being ordered by a gynecologic oncologist, this indicates that the patient is already being treated by the type of provider to whom you are attempting to refer the patient.

This study demonstrates that 61% were operated on by a gynecologic oncologist as the study was working with NCCN centers.

It was not clear how exactly patients were recruited. The Ueland and Longoria studies referenced 33 sites http://links.lww.com/AOG/A240 and http://links.lww.com/AOG/A243. There is no indication that these sites were all NCCN centers.

iii. Since this study (Ueland) compared OVA1® against other methodologies, to include CA 125, why would this study “skew” the assessment/conclusion in favor or OVA1® only?

The word skew has been explained above and it has been removed.

Based upon what criteria is it a “limitation” if the sensitivity appeared to vary depending on menopausal status? ¹. OVA1® is FDA cleared using a different cut-off for pre- and postmenopausal patients.

The test itself per FDA label indication is “Intended use:
The OVA1^TM Test is a qualitative serum test that combines the results of five immunoassays into a single numerical score. It is indicated for women who meet the following criteria: over age 18; ovarian adnexal mass present for which surgery is planned, and not yet referred to an oncologist “

There is nothing in OVA1®’s FDA clearance that excludes use of OVA1® by gynecologic oncologists (it does specifically state that OVA1® is not intended as a screening or stand-alone diagnostic assay), or prohibit use of gynecologic oncologist patients in peer reviewed published studies.

But it should be noted here that the way NGS implies support of CA 125 use in the “Analysis of Evidence” section of this proposed LCD is a non-FDA cleared use of CA 125, and a use of CA 125 that is specifically not covered by CMS per NCD (190.28).

So at the same time that this proposed LCD incorrectly critiques OVA1® based upon its FDA clearance, the LCD also implies support of non-FDA cleared use of CA 125 that is not covered by CMS as the reason (“in light of this”) to deem OVA1® to be considered not medically reasonable or necessary?

Thank you for the information.

This review incorrectly states that the Ueland study, “… skewed the assessment/conclusion that OVA1® demonstrated a high sensitivity and positive predictive value (PPV)” (our highlight added).

1. Gaps: This study did not conclude that OVA1® demonstrated a high PPV.

This has been corrected.

The (Ueland) study concluded that OVA1® demonstrated a high negative predictive value (NPV) of 98.1% when combined with clinical assessment

Given that ovarian cancer is relatively uncommon, one would expect the negative predictive value to be high. The positive predictive value is however low with the attendant risks of over referral and false positives with the subsequent potential for increased patient stress/anxiety and unnecessary procedures.

Per OVA1®’s FDA clearance, OVA1® should not be used without an independent clinical/radiological evaluation

The assessment of an adnexal mass includes a radiologic evaluation regardless of biomarker evaluation, OVA1 or otherwise.

There is a big difference between PPV and NPV.

This was a typo and this has been acknowledged.

If the NGS review confused the PPV and NPV results of the Ueland study, it does raise the question of what else was missed in this review overall, especially if nine unknown studies were excluded from consideration for undisclosed criteria.

These studies have been detailed as above.

Additionally, across all studies, the PPV for MVIA is 31%-41%. The PPV for physician assessment is 60-70%. The PPV for CA-125 is 56-75%. So, the false positives are increasing. It stands to reason that since evidence supports that physicians will refer even if they don't think the patient has a malignancy, that the test itself shows it will not increase referrals over current clinical practice, and the immediate direct cost of the test is so much higher than CA-125/ultrasounds/what the provider community already has in their toolset, perhaps a better solution is to educate providers to refer to a gynecologic oncologist in all cases of a tumor.

Other Gaps: Important findings from Ueland’s study that were missed by NGS in this proposed LCD are:

1. As a stand-alone test, OVA1® (MIA) detected 97% of malignancies in all stages and menopausal status, compared to just 78% identified by CA 125-II.
2. OVA1® detected 82% (33/40) of malignancies missed on physical assessment alone.

This is misleading. You’re highlighting the “missed” malignancies while skimming over the fact that the false positive rate is increasing.

OVA1® detected 76.0% (38/50) of malignancies missed by CA 125.

Adding OVA1® (MIA) to the physician assessment correctly identified 70.0% (14 of 20) malignancies that had been missed by non-gynecologic oncologists and 95.0% (19 of 20) malignancies missed by gynecologic oncologists.

Adding the physician assessment to the findings with OVA1® identified 86.0% (43/50) of malignancies missed by CA 125, including all advanced cancers.

The Ueland study has been referenced. And of note, the study was noted to be of Level III evidence (per Obstetrics and Gynecology). We are concerned that the NPV comes at the expense of a lower PPV and a lower specificity leading to patients who do not have the malignancy (but who tested positive for a malignancy) being referred to a gynecologic oncologist. Therefore, it is not clear how OVA1 facilitates getting the patient to the right provider at the right time.

In Bristow, Hodeib, Smith et al13 and Bristow, Smith and Zhang et al11 NGS states that: “…both clinical impression and CA 125-II were more accurate in identifying benign disease. As in the previous study, although sensitivity improved with the OVA1® test, specificity declined and was reported as 53.5% with OVA1® alone and 50.7% with OVA1® combined with clinical impression compared to 92.5% with clinical impression and 86.1% and 94.5% for CA 125-II (using two different cut-off values) in predicting disease”

Gap: NGS decides to focus on Specificity in this study as a weakness of OVA1®. But in touting the higher specificity of CA 125, NGS completely misses the much more important false negative rate of CA 125.

Simply put CA 125 misses an unacceptable level of ovarian cancers in both pre- and postmenopausal patients.

This paragraph has been amended.

NGS also notes in the Bristow study that: “However, the authors noted that both the OVA1® and OVA500 trials were designed to measure accuracy in prediction of malignancy rather than the test’s effect on patient referral for sub-specialty care.”

1. Gap: Comparison to pathology which is the gold standard is a direct reflection of the sensitivity as well as an indicator as to what patients should be referred to a gynecologic oncologist. The accuracy of the risk assessment directly impacts decisions on patient referral for sub-specialty care, as per OVA1®’s FDA clearance.

This is acknowledged. Given the very low specificity and PPV, it is possible that referral patterns will not change or might even increase, including benign referrals. Given that gyn oncologists were included in the original trials, this is a confounder. There are no prospective studies that we are aware of demonstrating that referral patterns will change, or that patient quality of life will improve or that mortality will decrease as a result of OVA 1.

Important findings from the Bristow 2013 studies that were missed by NGS in this proposed LCD are:

Early Stage Disease: correctly predicted ovarian malignancy in 91.4% (95%CI=77.6–97.0), compared to 65.7% (95%CI.=49.2–79.2) for CA 125.

This has been added

Missed Malignancies: correctly identified 83.3% malignancies missed by clinical impression and 70.8% cases missed by CA 125-II.

Predicting absence of an ovarian malignancy: OVA1® was superior in predicting the absence of malignancy, with a negative predictive value of 98.1% (95%CI=95.2–99.2).

Please see previous comments.

The Bristow studies clearly show the superiority of OVA1® to identify early stage compared to CA 125, and the improvement of the diagnosis of early stage ovarian cancer is exactly what is needed for ovarian cancer outcomes to improve.

Neither of the Bristow studies directly demonstrated that OVA1 would improve ovarian cancer outcomes. The ability of OVA1 to improve outcomes in women with ovarian cancer has yet to be determined.

Twelve to 40% of women referred to a gynecologic oncologist have a confirmed ovarian malignancy at surgery, meaning more than 60% are referred for benign disease. In this multivariate index assay trial, the specificity of physician assessment alone (nongynecologic oncologists) was 83%, yet the number of nonmalignant tumors referred to the gynecologic oncologist was still high. Surprisingly, 72% of all benign ovarian tumors were referred to a gynecologic oncologist for surgery, including 45% of patients referred despite a negative physician assessment. So the calculated specificity and predictive values of physician assessment do not correlate with observed patient referrals.

So physicians are referring even if they don't think that the patient has a malignancy

1. The studies also show that CA 125 misses ovarian cancers that OVA1® detects.
2. Lastly, the studies show that when combined with clinical impression (as per OVA1®’s FDA clearance), the NPV is as high as 99%, as stated in the ACOG guidelines PB 174 , when both the OVA1® and imaging results are Low Risk. a. This means that OVA1® allows these truly low risk patients to have “peace of mind” and do not need to see a specialist, but rather can be treated by their general gynecologist.

As noted by Bristow (2013) “Despite a low risk indication of ovarian cancer according to clinical assessment, CA125, and modified-ACOG guidelines, approximately 50% of patients were nevertheless referred to a gynecologic oncologist for surgery. This observation reflects the fact that, in real-world clinical practice, there are many variables that could trigger referral including nuanced interpretation of serum biomarker tests, imaging results, and clinical triage algorithms, either individually or in combination”, p 581.e6.

In regards to Ware Miller et al10 NGS states: “Ware Miller et al10 used the same study population as the Ueland et al study6 and like the studies above, patients evaluated by gynecologic oncologists were included in the study contrary to FDA clearance label.”

1. Ware Miller was evaluating a different endpoint. The study evaluated the impact of replacing CA125 with OVA1 in conjunction with ACOG guidelines.
2. OVA1®’s FDA clearance Indications for Use does state that OVA1® “…is indicated for women who meet the following criteria: … and not yet referred to an oncologist.”

Gap: Label does not state that OVA1® cannot be used by a gynecologist-oncologist. It specifically says that OVA1® “… is not intended as a screening or stand-alone diagnostic assay…”

The documentation from the FDA is as follows:

Intended use:

The OVA1™ Test is a qualitative serum test that combines the results of five immunoassays into a single numerical score. It is indicated for women who meet the following criteria: over age 18; ovarian adnexal mass present for which surgery is planned, and not yet referred to an oncologist. The OVA1 Test is an aid to further assess the likelihood that malignancy is present when the physician’s independent clinical and radiological evaluation does not indicate malignancy. The test is not intended as a screening or stand-alone diagnostic assay.

To eliminate any potential confusion, “and not yet referred to an oncologist“ has been italicized in the manuscript.

The label does not exclude use of the test for patients enrolled by gynecologic oncologist patients in peer review published studies.

Aside from the fact that one cannot enroll in a peer review published study, the label clearly states that this is for patients who have not yet been referred to an oncologist. Therefore, the oncologist should not be involved prior to the test. The above statements are acknowledged; the manuscript has been amended.

Does NGS have any questions regarding the study’s conclusions?

1. “CONCLUSION: Replacing CA 125 with the multivariate index assay improves the sensitivity and negative predictive value of the College referral guidelines while decreasing specificity and positive predictive value. The high sensitivity is maintained in premenopausal women and early-stage disease.” DOI: 10.1097/AOG.0b013e31821b1d80

No questions. The author’s conclusions do not substantiate the clinical utility OVA1.

NGS summarizes these studies above as follows: “Taken together, these studies provide a low quality evidence for the clinical validity of the OVA1 ® test.

1. Vermillion and the researchers who published this peer reviewed study do not agree that this is “… a low quality of evidence…”?

2. By what standard/criteria is this subjective opinion being based?

The Ueland and Ware-Miller studies were both designated as “level III evidence by Obsteterics and Gynecolgy (the official publication of ACOG). The Eskander study was a retrospective review (also Level III). The first Bristow study (2012) included gyn oncologists and the second Bristow study (2013) utilized “combined datasets” from the two prior studies and none of the studies include how the study was powered or a sample size calculation. Furthermore, although some of the studies were prospective, aspects of the study design are unclear.

1. How is the inclusion of emergency room physicians (Bristow 2012) appropriate to evaluate a test in women who are to undergo surgery for an adnexal mass.
2. It is not clear exactly how patient were recruited. Are Clinical Research Consultants, Women's Health Research, Precision Trials and North Coast Women's Care Research a primary care women’s health clinic, general obstetrics and gynecology group practice, a gynecologic oncology practice, a community and university based hospital, or a health maintenance organizations? How does a patient get recruited from an HMO?

There does not appear to be a single level I study in the entire series of articles.

In regards to Eskander et al5 NGS states that: “The study showed that 80% of the patients with elevated OVA1® had been referred to gynecologic oncologists prior to surgery with equal referral rates in pre and postmenopausal women. The study had several limitations which included small patient population (n=136), retrospective study design, physician response and recall has the potential to introduce selection bias, which could impact the rates of malignancy and referral outcomes data analyzed in the study. Furthermore, patient outcomes, such as morbidity and mortality or other long term outcomes were not assessed.”

Reiterating passages from various studies does not connote validity.

Gaps: This is NOT a small number of patients for an ovarian cancer study, due to the low prevalence of the disease: 136 patients with 122 having surgery with diagnosis of 78 cancers.

1. 65 primary ovarian cancers identified via OVA1 with 48% at early stage.
2. While a retrospective study has limitations compared to prospective trials, it does not invalidate conclusions.
3. This was a chart review study and therefore physician response and recall is not an issue with “potential” selection bias.

There was no justification of sample size in the study. Chart review studies are particularly prone to bias. A chart review is level III evidence.

These findings greatly DIFFER from the NGS noted 33% to 60% who get gynecologic oncologist treatment for their ovarian cancer.

1. 89% of women with an adnexal mass presumed to be malignant prior to surgery either had the surgery performed by a gynecologic oncologist (80%) or had an evaluation with a gynecologic oncologist prior to surgery (9%)
2. 100% of patients with primary ovarian cancer had gynecologic oncologist involvement prior to their first surgery.
3. 94% had their first surgery performed by a gynecologic oncologist, with the remaining 4 (6%) having gynecologic oncologist involvement prior to their first surgery.

As did patients with benign disease. It is not clear “which findings” are being referred to. If the purpose of this test is to streamline referrals, it is not clear that is the case

In its Analysis of Evidence (Rationale for Determination), NGS quotes from ACOG PB #174 (2016) that: “The combination of an elevated CA 125 level and a pelvic mass in a post-menopausal woman is highly suspicious for malignancy, and patients with these findings should be referred to or treated in consultation with a gynecologic oncologist”. In light of this, ordering/performing any of the lab tests reviewed in this document would be considered not medically reasonable or necessary for post-menopausal women and therefore is noncovered.

Gaps: So after removing several OVA1® studies (based on undisclosed criteria) that were “not performed as specified by OVA1’s FDA clearance” in this proposed LCD, NGS cites CA 125 as adequate for evaluation despite CA 125 not being FDA cleared for this use.

1. CA 125 is not supported by CMS via NCD or LCD (to include NGS) for this use.
2. Thus NGS is instructing CMS providers in the NGS service area that CA 125 is adequate for evaluation such that OVA1 isn’t medically reasonable or necessary, but at the same time CMS DOES NOT COVER CA 125 either and CA 125 is not FDA cleared for this use, whereas OVA1® is.

References to CA-125 have been revised and are abstracted directly from society publications.

So what realistic choice do CMS providers have in the NGS service area in terms of CMS covered ACOG Level B recommended tumor marker use on indeterminate pelvic masses? 1. It appears that none are covered despite the availability of FDA cleared and ACOG recommended OVA1®. 

The purpose of this LCD is not to provide choices; the purpose is to review the scientific evidence with regards to Multimarker Serum Tests for ovarian cancer

NGS should note that CA 125 typically elevates late and misses early stage disease as opposed to OVA1® (Longoria et al, 2014).

1. While OVA1® may have a lower PPV and specificity compared to CA 125, the clinical issue is false negatives CA125 and lack of appropriate referral. 

To quote NGS from earlier in this policy:

1. Ovarian cancer is staged surgically and the prognosis of ovarian cancer is closely related to the stage of the tumor at the time of diagnosis. If malignancy is detected while still localized in the ovary, the estimated five-year survival rate has been found to be about 92.5%1. Conversely, five-year survival rate among women diagnosed with advanced stage ovarian cancer range from 20% to 40%1.2,3. Overall, early-stage diagnosis is correlated with higher five-year survival rate establishing a need for early detection intervention.

1. If this is true (and it is), how can NGS cite use of non-FDA cleared CA 125 as being adequate as the main rationale for denying OVA1® coverage, if CA 125 has such poor sensitivity for early stage disease (both in pre- and postmenopausal populations) and is not FDA cleared for this use?

References to CA-125 have been revised and are abstracted directly from society publications.

Where are the peer reviewed published studies showing CA 125 performance having superior sensitivity and NPV compared to FDA cleared OVA1®?

3. Where are the clinical utility studies showing how use of CA 125 improves patient outcomes to support the tacit (but non-covered) approval of NGS in this proposed LCD?
4. In essence, NGS is disregarding pertinent evidence in recommending CA125.

References to CA-125 have been revised and are abstracted directed from society publications.

Critique of Omissions from Proposed NGS OVA1 LCD (DL38371) by Vermillion after review of the new proposed LCD; the following gaps have been noted to bring to the attention of NGS:

Two OVA1® studies were listed in the proposed LCD reference list, but no mention was made of them in the proposed LCD’s Summary of Evidence. We believe the following evidence should be considered:

Longoria (2014) showed adding OVA1® to clinical assessment:

1. Results in higher sensitivity for early stage cancer compared to clinical assessment alone, CA 125-II and modified ACOG guidelines;
2. Successfully identifies 95% of early-stage cancers;
3. Results in a reduction of missed early-stage malignancies:
4. 85% (23/27) reduction missed by clinical assessment alone
5. 78% (25/32) reduction missed by CA 125 II
6. 65% (13/20) reduction missed by the modified ACOG guidelines.

The Longoria paper has been explicitly cited with a paragraph dedicated to this study alone. As this is a summary/review, we cannot include every piece of data from every study.

Goodrich (2014): 1. Showed serum biomarkers and imaging are a complementary set of clinical tools:

1. When the MIA score is further stratified by imaging risk and menopausal status, there is a better understanding of the clinical risk of ovarian malignancy;
2. When performing parallel testing, there’s high sensitivity making it very unlikely to miss an ovarian malignancy;
3. Interpretation in series obviously improves test specificity;
4. When testing in series, there are fewer false-positive outcomes which increases test specificity;
5. Only 1.6% of tumors were malignant when both tests (imaging and OVA1®) indicated low risk.
6. Understanding the specific statistical impact of test combinations may be useful when discussing the urgency of surgery, type of procedure planned, the specific surgical approach, and/or patient counseling and consent.

The primary endpoint of this study was to examine how pelvic imaging influenced the OVA1 score, not the performance of the OVA1 test. This study was excluded.

Per OVA1®’s FDA clearance, OVA1®, “… is intended to serve as one of the diagnostic tools available to the primary clinicians the patient is most likely to see first for evaluation of non-specific signs and symptoms suspicious for ovarian cancer …”

1. Goodrich (2014) shows how use of ACOG recommended Level B OVA1® complements and improves the use of ACOG recommended Level A TVUS on indeterminate pelvic masses.

Please see above.

One study listed in the proposed LCD’s reference list is from Hayes: OVA1 (ASPiRA Labs) https://evidence.hayesinc.com/report/gte.ova206027

While OVA1® may have a lower PPV and specificity compared to

1. This link simply brings us to a very brief summary that is overly general about ovarian cancer and the OVA1® test.
2. To get more information on what is actually in the publication one must “Request Access” from Hayes.
3. In the past this publication has not been publically available and must be purchased from Hayes for $7,000.
4. In the interest of transparency, Vermillion believes NGS should not consider the opinions of publications that are not open to the public, such as the restricted Hayes opinion that must be purchased to be viewed.
5. This study cited in NGS’ policy should either be disclosed to the public for review and comments, or excluded from consideration.

Unfortunately, many publications, including the results of NIH funded trials are not made available to the public. For the sake of the requested transparency, the summary of Hayes findings has been included in the manuscript.

The following studies and policies were not included in the NGS proposed LCD references but should be considered by NGS:

Dunton et al (2019a) Ova vs CA 125 in African American Women showed:

1. OVA1® (MIA) has a higher sensitivity for detection of malignancy in African American women compared to CA 125 and application of the ACOG CA 125 recommendations.
2. OVA1® (MIA) will result in a higher rate of referral to the gynecologic oncologist for evaluation and management as per guideline recommendations.
3. Application of the ACOG cutoff of CA 125 >200 U/mL has the lowest sensitivity which means it will miss the highest number of malignancies in African American women.

Dunton et al (2019b) MIA & CA 125 and HE4 Testing in African American Women showed:

1. ROMA in African American women with adnexal masses has lower sensitivity for the detection of malignancy than does OVA1® (MIA).
2.  Implementation of OVA1® (MIA) in the evaluation of adnexal masses will increase the sensitivity of the detection of malignancy compared with ROMA, with the most marked results in African American women.

Noted. Dunton et al provided a retrospective review including patients in the Ueland and Bristow (2013) studies which included 23 African America women with an ovarian malignancy; this is a small sample.

eviCore has had positive coverage of OVA1® since the start of 2018: eviCore OVA1®, MOL.TS.260 A v2.0 2019

Of note, eviCore has since changed it’s coverage to now consider OVA1®, investigational and/or experimental effective 1/1/2020: eviCore OVA1®, MOL.TS.260. B v1.0.2020

Forde (2016) showed:

1. Use of OVA1® (MIA) is associated with decreased overall health care costs based on the increased referral to gynecologic oncologists with subsequent improved overall survival, costs and QALYs and incremental cost-effectiveness ratio (ICER).
2. Use of OVA1® (MIA) was cost-effective, resulting in fewer reoperations and pre-treatment CT scans.
3. When compared to modified ACOG criteria, OVA1® (MIA) resulted in 1930 fewer reoperations (1.9% of patients) and 2412 fewer CT scans (2.3% of patients) which would have been required due to more patients with false negative results.
4. A ‘refer all’ strategy is very difficult to implement outside of research setting.
5. Approximately 80% of masses will be benign, referring all cases may be inadvertently inefficient: i. Reduce access to gynecologic oncologists for women with true gynecologic malignancies
6. Extend waiting time for high-risk women
7. Refer a woman to a higher cost hospital and staff when there are less costly options available that can provide the level of care needed for a benign mass.
8. An additional potential consequence of the ‘refer all’ strategy is a gradual reduction in surgical volume for obstetrician gynecologists, which may over time lead to a decrease in surgical proficiencies and inability of non-urban areas to generate sufficient patient care to sustain local obstetrician-gynecologists.

The study by Forde was a cost-evaluation study and was therefore not included in this review.

Novitas LCD L35396 has had positive coverage of OVA1® (MIA) since October 2015: https://www.cms.gov/medicare-coverage-database/details/lcd-details.aspx?LCDId=35396

Novitas based it’s coverage decision on the FDA label and not review of literature as shown in their LCD.

Summary: NGS should grant coverage to OVA1® (MIA) now:

1. OVA1® has by far demonstrated the best sensitivity for early stage ovarian cancer detection for both pre- and postmenopausal women as well as the best NPV when combined w/imaging and physician assessment.
2. OVA1® has by far demonstrated the best sensitivity for ovarian cancer detection in African American women.

Please see previous discussion. The cohort of AA women in the study was too small to make this conclusion.

The choice NGS has with this LCD is either keeping the status quo (which its own proposed LCD essentially paints as a picture that is unacceptable), or

d. Approving coverage of FDA cleared and ACOG recommended OVA1® that clearly is best available ACOG recommended Level B choice available today to help change that status quo.

The option we had at NGS was to analyze the evidence which we believe we have done in this LCD.

Several comments were received discussing the literature on the ROMA test

We appreciate the literature which was submitted. These were reviewed and are reflected in the LCD.

A consulting company noted that NGS is the only MAC not covering ROMA. It is inappropriate for NGS to limit coverage for Medicare Beneficiaries when the other MACs cover this test.

They also referenced these studies:

DL38371 reference # 18, the meta-analysis by Dayyani failed to mention the following key findings:

1. ROMA (0.921 [0.855-0.960]) had a numerically greater diagnostic performance than CA125 (0.883 [0.771-0.950]) and HE4 (0.899 [0.835-0.943]). This was also observed in each of the subgroup populations, in particular the postmenopausal patients and patients with early OC.
2. The sensitivity and specificity (95% confidence interval) results showed ROMA (sensitivity 0.873 [0.752-0.940]; specificity, 0.855 [0.719-0.932]) to be numerically superior to CA125 (sensitivity 0.796 [0.663-0.885]; specificity 0.825 [0.662-0.919]) in all patients and for the early- and late-stage OC subgroups
3. The results presented support the use of ROMA to improve clinical decision making, most notably in patients with early OC.

ROMA’s better performance than CA125 in early stage EOC was also shown in DL38371 reference # 17, the meta-analysis by Wang and Moore et al, Obstetrics and Gynecology, 2011 (118) 280-288

This is not so. Apart from the Novitas LCD which has been addressed in a prior response, The LCDs referenced are from the Moldx program which is titled general Molecular Diagnostic Tests (MDT) and gives general instructions on how to apply for unique test identifiers and do not discuss coverage of any biomarkers. This LCD gives a range of codes for which this is applicable which happens to contain the codes addressed in our LCD.

Both studies referenced here compare ROMA with OVA 1 which is not the essence of this policy. Of note, OVA1 has also been reviewed and is considered non covered as well.

Comment from Dr. Robert Bristow:

I am writing to request that NGS reconsider the negative coverage determination for OVA1 testing. I understand that the panel was concerned about studies funded by the maker of OVA1. I can assure you that our peer reviewed study was performed under IRB approval and conducted with scientific integrity. The conclusions are valid. From the paper of which I was the lead author, “For all ovarian malignancies, the sensitivity of the multivariate index assay was 95.7% (95%CI=89.3–98.3) when combined with clinical impression. The multivariate index assay correctly predicted ovarian malignancy in 91.4% (95%CI=77.6–97.0) of cases of early-stage disease, compared to 65.7% (95%CI=49.2–79.2) for CA125-II.”

Importantly, OVA1 performed better that CA125 in early stage disease as demonstrated in our publication. Our paper also showed better detection in uncommon ovarian histologies. Additionally, I am aware of 2 recent publications on OVA1’s superior performance in African American women. African American women have been shown to have lower CA125 and more commonly have the less common ovarian histology.

See prior answers under in comments 1-42.