Increasing focus has been added toward self-collected samples as a means to increase the participation in the screening programs for cervical cancer. Urine samples have also been tested for this purpose, but the knowledge on performance is still quite sparse.
In this study, the high-risk HPV status of a self-collected (SC) vaginal sample and a urine sample is being compared to a concomitantly physician-taken liquid-based cytology (LBC) sample. The results of the three HPV tests are being compared to the histological diagnosis of biopsies taken in parallel.
Women referred to colposcopy at the gynecological departments at Lillebaelt Hospital and Odense University Hospital is being in-rolled in the study. A total of 300 women will be included.
A urine sample and a SC vaginal sample using the Evalyn Brush are performed by the women after a short instruction and just before the medical examination. At the colposcopy an LBC sample and biopsies are taken by the gynecologist. The urine, SC vaginal sample and LBC sample are analyzed for the presence of high-risk HPV using the Cobas HPV test, Roche. The biopsies is diagnosed by a pathologist and used as the gold standard.
At this point 70 women have been included in the study.
Urine and a SC vaginal sample are available from all women, while data from LBC and biopsy samples are currently available from 60 and 58 women, respectively. The results of the HPV test are distributed as follows:
For urine and the SC vaginal samples, 46 were identified as positive in both sample types and 19 as negative. The concordance was 93 % (65/70). For SC vaginal samples and LBC samples the agreement was 87% (52/60), while for urine and LBC a concordance of 80% (48/60) was found.
The sensitivity of the three sample types was the same (94% and 100% for CIN2+ and CIN3+, respectively). For the specificity differences was observed. At CIN2+ the specificity was 38%, 43% and 57% for SC vaginal samples, urine and LBC, respectively. For CIN3+ the specificity was 33%, 38% and 53% for SC vaginal samples, urine and LBC, respectively. The differences are not statistically significant at this point.
Among the women one adenocarcinoma was identified and for this patient all three sample types were HPV positive.
These initial data indicate that the sensitivity of a urine sample is equally good as a SC vaginal sample and comparably to the physician-taken LBC samples. Further data are needed in order to determine whether the specificity of a urine sample and vaginal samples is significantly lower than a LBC sample taken by the physician. The updated data will be presented.