Key Takeaways
- A blood RNA signal called Tu-Stroma separated ovarian cancer's microscopic precursor, STIC, from healthy samples with an accuracy score of 0.99, versus 0.59 for the standard CA125
- In a validation group it caught 92.3% of precursor cases and correctly cleared 98.2% of healthy women.
- The study used stored blood from women with known diagnoses; the test is not available and has not been tried prospectively.
The Deadliest Part of Ovarian Cancer Is the Stage Nobody Can Detect
High-grade serous ovarian cancer usually begins as something almost absurdly small: a patch of abnormal cells in the fallopian tube called STIC, invisible to scans and missed by the standard CA125 blood test. By the time this cancer is found, it has usually spread. A study in Nature Medicine now reports a blood signal that appears to betray that hidden first stage.
The signal is Tu-Stroma, an RNA molecule released as cancer cells weaken the healthy tissue around them to make room to grow. In stored blood samples from 506 people — including 43 with isolated STIC and 232 with high-grade serous cancer — Tu-Stroma separated the precursor from healthy controls with an accuracy score (AUC) of 0.99, against 0.59 for CA125, where 1.0 is perfect and 0.5 is chance.
In a separate validation group, the test detected the precursor in 92.3% of women who had it and correctly cleared 98.2% of healthy women. Modelled in a high-risk population, about 72% of positive results would be true alarms.
Why this is a beginning, not a test you can ask for
Every limit of this study matters. The blood came from women whose diagnoses were already known — a design that almost always flatters a new test's accuracy. Tu-Stroma has never been used to screen women followed forward in time, and there is no evidence yet that catching STIC this way changes outcomes or saves lives. The eight-year, international study, led by the Hoag Family Cancer Institute in California, is a proof of principle: prospective trials are the stated next step.
The Twikup Take
Ovarian cancer kills largely because it is found late; a signal that shows up at the precursor stage attacks exactly that problem. The science here is genuinely striking — and the distance between a striking result in stored samples and a screening test in a clinic is measured in years of trials. Both things are true at once.
Image: Illustrative image.