PinkDx and the art of doing less

There is an awkward little problem with looking for cancer: most of the people being investigated will not have it. That is good news for the patient, eventually. Getting to “eventually” can involve instruments, appointments and a considerable amount of worry. PinkDx is building a company around shortening that journey. Its proposed starting point is a vaginal swab.

To understand the appeal, consider the distinction between finding disease and excluding it. A test that finds a cancer directs attention toward treatment. A reliable negative result might spare someone a procedure undertaken to discover whether treatment is necessary at all. Both decisions matter. Only one tends to make a triumphant headline.

A symptom is an invitation to uncertainty

PinkDx focuses on gynecologic cancer, beginning with cancer of the endometrium, the lining of the uterus. Its website describes a lopsided clinical problem: more than 90% of women diagnosed with endometrial cancer present with abnormal or postmenopausal bleeding, yet only about 5% of women evaluated for bleeding ultimately receive that diagnosis. The figures refer to different groups. They are not mirror images.

That asymmetry explains the company’s interest in ruling cancer out. Bleeding deserves investigation, but the symptom alone does not sort the dangerous cases from the benign ones. The existing workup can involve imaging and an endometrial biopsy. PinkDx’s ambition is to give clinicians a less invasive source of information before some women travel further down that path.

The economic attraction follows the clinical one. If a useful test can reduce unnecessary procedures, its value extends beyond its laboratory result. It could save discomfort, time and resources. Those are the benefits PinkDx seeks to demonstrate; the swab’s existence alone does not establish them.

The sample had to travel less

PinkDx combines real-world clinical cohorts, whole-transcriptome sequencing and machine learning. In plain English, it examines patterns of RNA activity and uses computational classifiers to distinguish disease-associated signals. The clinical records and the specimen are as important as the algorithm: a model needs a trustworthy account of what happened to the person behind each sample.

The company’s approach also includes benign conditions, different cancer stages and diverse patient populations in its research cohorts. This is sensible product design. A diagnostic has to work among the confusing cases that turn up in a clinic, where symptoms overlap and the answer is not supplied in advance.

The first scientific question was unusually concrete: could a signal associated with a cancer inside the uterus be detected outside it? Researchers collected several specimen types from women undergoing hysterectomy, including vaginal samples and tissue. The published exploratory analysis found cancer-associated molecular signatures in the less invasive samples.

The numbers deserve careful handling. The paper’s abstract reports sequencing data from 27 PNK001 participants, alongside 46 tissue samples from a separate network. That is an exploratory analysis, not thousands of women receiving a finished test in routine practice. The researchers selected the vaginal swab for the subsequent development and validation study. The evidence changed the next experiment.

There is a lesson here for anyone building a product: identify the assumption that would make the whole thing impossible, then investigate it directly. PinkDx needed a detectable signal in a tolerable specimen. A beautiful classifier would have been little consolation if the swab contained nothing useful.

The second act brought familiar company

Bonnie Anderson and Giulia Kennedy had worked together at Veracyte, where genomic diagnostics helped reduce unnecessary invasive procedures, including in the evaluation of thyroid nodules. PinkDx transfers that experience to a new clinical question. The previous success supplies expertise, not borrowed validation.

The third cofounder, Timothy McMeekan, is chief financial officer and chief business officer. The current leadership also includes chief medical officer Richard Kloos. It is a company with scientific, clinical and commercial jobs to do, and its leadership reflects that division of labor.

Anderson’s route back to company-building was personal. She had planned to retire and sail with her husband; he died from COVID during her transition out of Veracyte. Conversations with physicians and researchers helped direct her next effort toward women’s health. In a September 2026 interview, she described choosing a vaginal sample close to the suspected disease rather than following enthusiasm for blood-based cancer tests.

“Don’t fall in love with the technology. Fall in love with the problem.”

Bonnie Anderson · September 2026 interview

That principle is easy to print on an office wall. Here it has consequences for the specimen, the study population and the decision the test must support. PinkDx’s public operating philosophy is to map those hurdles early. In diagnostics, commercial planning begins well before there is anything to sell.

$45 million buys a chance to prove it

PinkDx announced a $40 million Series A in April 2024, co-led by Catalio Capital Management and The Production Board, with Mountain Group Partners, Byers Capital and Mayo Clinic participating. A January 2025 second close added $5 million, bringing Blue Venture Fund, Sandbox Clinical Ventures and BEVC into the financing.

That total is capital raised. It is not the price of the test or an accounting of development expenditure. The funding gives PinkDx resources to advance its first diagnostic; commercial success will require a different kind of persuasion.

The intended model is a laboratory-developed test, with samples sent to PinkDx’s California laboratory. Clinicians would use the result, patients would experience the consequences, and payers would help determine access. The immediate alternative is an established workup, rather than a software subscription or a consumer wellness kit.

Insurance therefore belongs inside the story. PinkDx appointed health-economics and outcomes specialist Robert Epstein to its board in August 2024, explicitly connecting his expertise to development and reimbursement. A less invasive test still needs to show that it answers a useful question at a cost the healthcare system will accept.

4,500 is a plan, not a verdict

The next major evidence step is PNK002, the prospective PROACTION study. Its August 2026 registry update lists 16 US locations and estimated enrollment of 4,500 participants, larger than the roughly 3,000 described in earlier coverage. It remains recruiting. Estimated primary completion is October 30, 2026, with longer follow-up extending into 2031.

The study targets women aged 45 and older seeking care for abnormal uterine or postmenopausal bleeding. Its criteria exclude women who have had a hysterectomy or a known history of endometrial cancer, among other conditions. Those boundaries matter: performance in this population would not automatically prove suitability for routine screening or recurrence monitoring.

Columbia describes the investigational workflow as a clinician-collected swab without a speculum, followed by laboratory analysis with results returned in about two weeks. Standard care continues during the study. A trial participant is contributing evidence; she is not purchasing a proven shortcut around clinical evaluation.

For PinkDx, the demanding question is how confidently a negative result can guide a decision. Missed cancers would undermine the purpose of avoiding invasive workups. Equally, a test that flags almost everyone for further investigation would do little to reduce the burden. The useful operating point must be established in the patients the company intends to serve.

Columbia’s July 2026 account puts possible availability as early as mid-2027, if trials succeed. For now, women with symptoms can discuss evaluation and potential research participation with their clinicians. The future product offers a specific promise: a gentler way to gather information that might let a doctor do less. PinkDx has chosen a modest-looking specimen for a decision that is anything but small.