Early swab test for endometrial cancer could spare some women a biopsy

What happened

A vaginal swab test for endometrial cancer is moving into a larger trial, and it could one day let some women skip an endometrial biopsy. That is the gist of a report, as first reported by Medical Xpress, which attributes the details to The Washington Post.

The test comes from PinkDx, based in California, and is being studied at 16 U.S. sites. The idea: tumor cells in the uterus shed genetic material that can drift into the vagina, where a clinician-collected swab picks it up for lab analysis.

We could only work from excerpts of the Medical Xpress piece, and we did not retrieve the Post article. Treat the secondary details below accordingly.

What the data actually show

The human evidence is an exploratory company study, PNK001, indexed on PubMed. Women scheduled for hysterectomy had vaginal swabs, ectocervical swabs, endocervical brushings and endometrial tissue collected.

Sequencing data came from 27 participants plus 46 tissue samples from the Cooperative Human Tissue Network. The machine-learning analyses used tissue samples and data from only 20 participants, with surgical pathology as the answer key.

Samples from women with endometrial cancer showed differences in gene expression and more somatic variants across the brush, cervical and vaginal swabs. Cross-validated AUC values ranged from 0.6 to 0.96. An AUC of 0.6 is weak, and this range is not a sensitivity or specificity figure.

The authors concluded that vaginal swabs carried enough signal for RNA-based machine learning to detect cancer, and they picked the vaginal swab for a second study, PNK002, to develop and validate an actual test.

What it does not show

• It does not show accuracy in the people it is meant for. PNK001 enrolled women already scheduled for hysterectomy, not women who came in with abnormal bleeding.
• It is tiny. Wentzensen described this first effort to the Post as very small, and the machine-learning set was 20 women.
• It is not independent. The sponsor, PinkDx, designed and managed the study, generated and analyzed the data, and decided to publish.
• It does not remove biopsy. Women who test positive would likely still need one, according to the report.
• It does not support the word "painless" as a proven benefit. A swab is plainly less invasive, but we found no data on how many women find biopsy painful or skip follow-up because of it. That framing comes from the news item.

The report also says PinkDx plans to run the test only in its own lab, with patient access possible by mid-2027. That timeline comes from secondary coverage, and we could not check it against a primary document.

What it means for you

Nothing changes today. The swab is not on the market, and no evidence we found suggests it is safe to put off an evaluation for abnormal bleeding.

If you have bleeding after menopause, irregular periods, or bleeding while taking tamoxifen, those are standard reasons for an endometrial biopsy. Cleveland Clinic calls the biopsy the most common and accurate test for diagnosing endometrial cancer.

If pain is what keeps you from booking, say so to your clinician. The usual side effect is cramping, with light bleeding or spotting for several days, and StatPearls describes it as a safe, well-accepted office procedure. A hospital leaflet from an NHS trust says most women bleed afterward, and about a quarter bleed for more than a week.

For postmenopausal bleeding, transvaginal ultrasound is often the first test, and UK guidance recommends biopsy when the uterine lining is 4 mm or thicker, per a review in PMC. The same review notes that endometrial sampling is less sensitive than once thought, so bleeding that persists after a negative biopsy needs more evaluation.

Context: vaginal-sample tests are not new

Other groups have tried non-invasive sampling for years, with mixed designs and populations. None of these are PinkDx's method, and they are not head-to-head comparisons.

The small size of that group limits how much its result can be trusted. Wentzensen's NCI work on vaginal and endometrial cell samples is part of the same long-running effort.

The takeaway: the concept is plausible and has been studied for a while, but results vary by method and population. PinkDx's approach, RNA sequencing plus machine learning, has no published diagnostic-accuracy data in symptomatic women yet. Any swab that replaces a biopsy has to beat a standard that is imperfect but well established.