This study explored a novel approach to detecting endometrial cancer using protein analysis of two easily obtained samples: cervico-vaginal fluid (CVF), collected with a simple swab from the vagina, and blood plasma. Rather than looking for DNA mutations, the researchers searched for abnormal protein levels that signal the presence of cancer.
The technique used is called SWATH-MS (Sequential Window Acquisition of All Theoretical Mass Spectra), a highly sensitive form of mass spectrometry that can simultaneously measure thousands of proteins in a single sample. This study enrolled 118 postmenopausal women: 53 with confirmed endometrial cancer and 65 healthy controls.
Cervico-vaginal fluid was chosen because it is in direct contact with the uterine lining where endometrial cancer originates, making it rich in proteins shed by the tumor. Blood plasma was also analyzed as it circulates throughout the body and can carry tumor-derived proteins.
The researchers used SWATH-MS proteomics to generate comprehensive protein profiles from both sample types, then applied statistical analysis to identify proteins that were consistently elevated or reduced in cancer patients compared to controls. Machine learning was then used to identify the smallest panel of proteins that maintained strong diagnostic accuracy.
From the cervico-vaginal fluid analysis, the researchers identified a panel of five proteins that together achieved an AUC of 0.95 for detecting endometrial cancer. The five proteins were HPT (haptoglobin), LG3BP (galectin-3-binding protein), FGA (fibrinogen alpha chain), LY6D (lymphocyte antigen 6D), and IGHM (immunoglobulin M).
An AUC of 0.95 is very close to perfect discrimination, indicating this panel could correctly classify cancer versus non-cancer in the vast majority of cases. The direct contact of CVF with the tumor likely explains why this sample type yielded such strong results.
The blood plasma analysis identified a three-protein panel achieving an AUC of 0.87. The three proteins were APOD (apolipoprotein D), PSMA7 (proteasome subunit alpha type 7), and HPT (haptoglobin). Notably, haptoglobin appeared in both the CVF and plasma panels, suggesting it is a robust marker of systemic inflammation or tumor-related signaling in endometrial cancer.
While 0.87 is slightly lower than the CVF panel, a blood test is simpler and more acceptable to patients than a vaginal swab, making the plasma panel potentially more practical for widespread screening.
Endometrial cancer is most common in postmenopausal women, who typically do not undergo routine gynecological screening after menopause in many health systems. A non-invasive protein-based test from a blood draw or simple swab could be used to screen symptomatic women (for example, those with postmenopausal bleeding) or even as a population-level screening tool.
Currently, diagnosis requires transvaginal ultrasound followed by endometrial biopsy or dilation and curettage. A protein panel that flags high-risk women could streamline this process and catch cancers earlier, when they are most treatable.
While the results are highly encouraging, this was a relatively small study of 118 women. Independent validation in larger, diverse cohorts is essential before either panel can be considered ready for clinical deployment. The study also enrolled postmenopausal women specifically, so the performance in premenopausal women or in mixed populations is not yet known.
The authors also note that combining both CVF and plasma panels could further improve accuracy by capturing both local tumor signals (from CVF) and systemic signals (from plasma). Future studies should also explore whether these proteins can distinguish early-stage from late-stage disease and monitor treatment response.