Urine spectroscopy and cervical images in one test — an AI registered as a diagnostic test
A study developing and validating a system for cervical cancer screening in low and middle income countries, detecting HPV by infrared spectroscopy of urine and reading cervical images with AI. The registered intervention is a diagnostic test.
Trial overview (primary data)
- StatusRecruiting
- ConditionsCervical Cancer, Screening
- InterventionsDIAGNOSTIC_TEST: Screening with HPV test
- SponsorInternational Agency for Research on Cancer
- Target enrollment3,200 participants
- Period2023-12-09 〜 2027-12-31
Key points
- The AI analyses infrared spectroscopic signals from urine to judge whether high-risk human papillomavirus is present.
- For women screening positive it interprets cervical images from a dedicated camera to detect high-grade precancers and cancers and to determine the transformation zone type.
- Part one, over the initial two years, analyses urine from 1,100 women to improve the technology.
- Part two, over the following three years, screens 2,100 women and compares sensitivity and specificity against a validated HPV test, with a cost analysis.
- The sponsor is the International Agency for Research on Cancer, target enrolment 3,200, with completion expected 31 December 2027.
1When the test itself is the AI
The AI in the previous article analysed records from outside the course of treatment. Here the AI is built into the test. Reading the spectroscopic signal and interpreting the cervical images are both its work, and so the intervention field says diagnostic test.
2Two stages of testing in one system
- 1First screenUrine from unscreened women goes to infrared spectroscopy, where the AI judges whether high-risk HPV is present
- 2TriageFor women screening positive, cervical images are captured with a dedicated camera
- 3InterpretationThe AI detects high-grade precancers and cancers from those images
- 4Treatment decisionIt also determines the type of transformation zone, which helps in deciding treatment
Screening, triage and the treatment decision are carried by one AI system. That is what the phrase one stop in the registered title points at: a form affordable enough and self-contained enough to work where specialist staff are few.
3A study in two parts
Development and validation sit one after the other inside the same registration. Build in part one, confirm in part two, for a combined target of 3,200.
4The diagnostic test box
The diagnostic test box holds 91 of 776 (11.7%). An AI reading images can be registered as a device or as a diagnostic test alike. The next article takes up an approved AI registered as a device.
Why it matters
Where an AI is built into the process of testing itself, the registered intervention becomes a diagnostic test. Carrying screening, triage and the treatment decision in one system is a design aimed at settings where specialist staff are scarce. Placing development and validation one after the other within a single registration is likewise characteristic of research where the technology has not yet settled.
FAQ
What is an automated visual exam?
Why use urine?
Why determine the transformation zone type?
Sources (primary)
Source: ClinicalTrials.gov (U.S. NIH/NLM, public domain). This site does not provide medical advice. Verify the latest and exact details with the official source. This site is not endorsed or certified by the NIH/NLM.
- ClinicalTrials.gov (study record, original)
- NCT ID: NCT06042543