Medical devices & diagnostics · Ovarian cancer
AOA Dx
A Denver company reading lipids, gangliosides and proteins out of a tube of blood to find ovarian cancer in women who already have symptoms. Three peer reviewed papers are on the record. The test is still not on the market.
Nine months
Ingrid Kolstoe was already at stage 4 when she learned what was wrong. “Ovarian cancer has traditionally been called a silent killer because it really whispers,” she told CBS Colorado in 2024. “In the beginning, the symptoms are very subtle. Many women are just told they have IBS.”
In the United States it takes about nine months on average to get from first symptom to an ovarian cancer diagnosis. High grade serous tumors, the most common and most aggressive subtype, double in volume every 2.2 to 4 months. That arithmetic is the whole business case for AOA Dx.
A marker from 1987
The blood marker clinicians reach for, CA125, was introduced in 1987. Screening healthy women with it has been tried and failed: in the UK Collaborative Trial of Ovarian Cancer Screening, CA125 plus transvaginal ultrasound hit 89.5% sensitivity and 99.8% specificity and still produced no significant reduction in deaths. Guidelines now advise against screening women without symptoms.
So AOA Dx aimed narrower, at the woman already sitting in a clinic with bloating, pelvic pain, early satiety or urinary changes. In the company's own published cohorts, CA125 at the standard 35 U/mL cutoff caught 76.2% of all ovarian cancers and 65.2% of early stage ones. What it misses is what matters: five year survival runs above 90% at stage I and near 30% once the disease has spread.
Anna, Oriana, Alex
The company name is the founders' initials. Anna Jeter, Oriana Papin-Zoghbi and Alex Fisher had worked together for about a decade at two earlier women's health diagnostics startups, both of which the company says ended in strategic exits, the second in 2018. Then they went looking for something to build themselves.
“None of us on the founding team were originally scientists... We needed to really create a partnership,” Papin-Zoghbi told the Angel Invest Boston podcast in 2021. They spent roughly a year screening technologies before meeting Professor H. Uri Saragovi of McGill University, who had been quantifying tumor marker gangliosides, a class of glycolipid, in blood. AOA Dx signed an exclusive license to that work in August 2021, weeks after Y Combinator's Summer 2021 batch and a Form D for $2.5 million. Surbhi Sarna, founder of nVision Medical, joined the board in February 2022.
From one marker class to four
The original pitch was gangliosides. The published science grew wider. A 2023 Frontiers in Oncology paper from the Saragovi group reported GD2 and GD3 gangliosides as candidate markers across stages and subtypes. By the time the company opened its own Denver laboratory in 2024, the test had become a multi-omic model: untargeted lipidomics by mass spectrometry, protein immunoassays, later metabolomics, and machine learning stitched across them.
The samples came from two partners. CU Anschutz supplied a retrospective cohort through its Gynecologic Tissue and Fluid Bank; the University of Manchester supplied a prospectively collected subset of its DETECT study. Train on one country, test on the other. That structure is what turned company claims into something a journal would print, and in August 2025 Reuters and the BBC both covered the result. At ASCO in June 2026 the company said it had moved those signals into a targeted, clinical grade assay at 92% early stage sensitivity, and it is now preparing an early access program.
What is proven, and what is still claimed
| Evidence | What the record shows | Source type |
|---|---|---|
| Peer reviewed model | Cancer Research Communications, Sep 2025. Cohorts of N = 433 and N = 399. A model of fewer than 20 lipid and protein features, trained on cohort 1 and tested on cohort 2, reached AUC 92% (95% CI 87% to 95%) for ovarian cancer and 88% (83% to 93%) for early stage disease. The authors call it a proof of concept and say analytic and clinical validation is still required. | Public record |
| Second paper | Diagnostics, Sep 2025. 509 specimens. Sensitivity is reported against fixed specificity: 96.1% for early stage at 70% specificity, 94.8% at 80%. The headline 98.7% is early stage against healthy and gastrointestinal controls only, at 70% specificity. | Public record |
| Third paper | Diagnostics, Jul 2026. 503 participants. A biology paper on coordinated lipid, metabolite and protein signatures. No new diagnostic performance claim. | Public record |
| Sensitivity headlines | Releases and trade coverage repeat 94.8% early stage and 94.4% all stage sensitivity without saying both sit at a fixed 80% specificity, which calls one control in five positive. | Context omitted |
| Sample counts | The April 2025 release describes “Approximately 1,000 patient samples”; the August 2025 release says “two large, independent cohorts, totalling >950 samples”. The paper they announce reports 433 plus 399, or 832. | Differs from paper |
| Conference versus journal | The April 2025 AACR release gave early stage AUCs of 92% in cohort 1 and 89% in cohort 2. The published paper reports 91% and 88%. | Differs from paper |
| OVERT and APEX | OVERT first patient announced Jan 2023, described as a multi-center national trial; APEX first participant May 2026. Neither appears in a ClinicalTrials.gov search as of Sep 22, 2026, and neither has a published readout. | Not found |
| Regulatory status | No FDA 510(k) or PMA record. In 2021 the CEO described a 510(k) De Novo path and said the test would be on the US market by 2025. | Public record |
| Federal grants | No NIH or NSF award to AOA Dx found. The only research grant announced is CAD 650,000 to Professor Saragovi from the Canadian Institutes of Health Research in 2021. | Not found |
| Partners | CU Anschutz and the University of Manchester supplied cohorts and co-authored. Sonrai Analytics signed a data partnership in 2025. Labcorp Venture Fund is on the cap table. | Company-stated |
Read plainly: the biology has cleared peer review three times, with academic co-authors and an independent test cohort, which is more than most diagnostics startups can show this early. What has not happened is the part that decides whether the test works in practice. Every number in print comes from banked samples analyzed by the company that sells the test, the strongest result is labelled proof of concept by its own authors, and the prospective studies that would settle it have produced no readout and no public registration.
What to watch
- A readout from OVERT or APEX, and whether either is registered on a public trial registry.
- Whether the launch comes as a laboratory developed test or as an FDA submission, five years after the CEO described a 510(k) De Novo path.
- Whether 92% early stage sensitivity holds when the matching specificity is stated beside it.
- A price, a launch date for the early access program, and any payer coverage decision.
- Sales of the remaining $22.5 million of the offering registered in October 2025.
In their words
“None of us on the founding team were originally scientists... We needed to really create a partnership.”
Oriana Papin-Zoghbi, CEO and co-founder, Angel Invest Boston, 2021 · Interview
“We expect to be on the market in the US by 2025. Another four years out from here.”
Oriana Papin-Zoghbi, Angel Invest Boston, 2021 · Interview
“The standard of care today is still largely based on technology that became available almost 40 years ago.”
Anna Jeter, co-founder and chief regulatory officer, Inside Precision Medicine, 2026 · Interview
“When a disease that you're studying is rare, the tests have to be extremely sensitive and specific for you to not have a high rate of false positives.”
Kian Behbakht, MD, University of Colorado Cancer Center, 2025 · Independent
“are there confounding factors that can shift these biomarkers so you get false positives?”
Benjamin Bitler, PhD, University of Colorado Cancer Center, 2025 · Independent
“less than 2% of all funding goes to diseases that affect women or predominantly affect women”
Oriana Papin-Zoghbi, Denver7, 2024 · Independent
“This is the moment where our science becomes a product.”
Oriana Papin-Zoghbi, ASCO release, June 2026 · Company release
“AOA Dx's platform shows significant promise for ovarian cancer early detection, offering a practical solution for symptomatic women.”
Professor Emma Crosbie, University of Manchester, 2025 · Quoted in release
Related companies
Sources
- Public recordSerum lipidomics with protein biomarkers and machine learning for early detection of ovarian cancer
- Public recordClinical evaluation of a multi-omic diagnostic model for early-stage ovarian cancer detection
- Public recordMulti-omics profiling in a symptomatic cohort, ovarian cancer serum
- Public recordGD2 and GD3 gangliosides as diagnostic biomarkers in epithelial ovarian cancer
- Public recordFive Form D filings, AOA DX Inc., CIK 0001869920
- Public recordSearch for AOA Dx, AKRIVIS and OVERT, no matching studies
- Public record510(k) and PMA databases, no records for AOA Dx
- IndependentHealth Rounds: Blood test finds early ovarian cancer
- IndependentNew blood test could detect ovarian cancer early, researchers say
- IndependentPutting Early Ovarian Cancer Detection to the Test
- IndependentDenver lab, CU researchers trial early ovarian cancer screening method
- IndependentDenver lab developing first-of-its-kind tool for early detection
- IndependentLipid-protein biomarker makes clinical debut
- InterviewOriana Papin-Zoghbi, AOA Dx, episode transcript
- IndependentBoston biotech AOA Dx has raised new funding and moved to Colorado
- IndependentAOA Dx closes $7M funding round for ovarian cancer test
- CompanyAOA Dx news archive and press releases
- CompanyAKRIVIS GD product page and team page
Profile researched and written by Healthcare Discovery. Last updated September 29, 2026.
