Evaluation of new and current preventive strategies for cervical and vaginal cancer

Abstract

This thesis addresses clinical research questions on primary and secondary prevention of cervical and vaginal cancer by targeting four groups of women in the organized cervical screening program; women entering the screening program, women exiting the screening program, non-attenders in the screening program, and women previously treated for precancerous lesions.

In Study 1, we evaluated the colposcopic and histopathologic findings in women with 1 and 3 year human papillomavirus (HPV) persistence, respectively, in women exiting the cervical screening program at age 56-64. The design was a randomized health care policy offering HPV screening to 50% of resident women exiting the screening program in the Stockholm-Gotland region of Sweden during January 2012 through May 2014. Women with HPV persistence were referred to colposcopy. Biopsies as well as an endocervical sample analyzed for cytology and HPV was taken at the examination. HPV was analyzed for 14 different possibly oncogenic genotypes. We found that HPV prevalence was 5.5% (405 women of 7325 attending). After one year, 91 of 176 (52%) were persistently HPV positive and after three years 55 of 137 (40%) (p=0.042). Histopathologically confirmed cervical intraepithelial lesions grade 2 or worse (CIN2+) was found in 19 of 82 women (23%) and 9 of 45 women (20%) in the 1 and 3 year groups, respectively, indicating that it was safe to postpone repeat HPV testing for three years in women exiting the cervical screening program. There was a high risk for CIN2+ at follow-up and noteworthy yields from HPV genotyping as well as endocervical sampling and random biopsies in the absence of visible colposcopic lesions.

In Study 2, we compared the colposcopic evaluation in vaccinated and unvaccinated women entering the organized cervical screening program. Women in the 1994 and 1995 birth cohorts who entered the cervical screening program at age 23 in one region in Sweden were identified. Colposcopy was performed within two to four months after a positive screening result and evaluated with two different tests; Swedescore and Colposcopic impression, respectively. Endocervical sampling with analysis for cytology and 14 possibly oncogenic HVP genotypes, as well as punch biopsies for histopathologic confirmation were taken in all women. Out of 165 women with positive screening tests, 160 (98%) attended colposcopy, of which 90 (56%) were vaccinated and 70 (44%) were unvaccinated. Only 7 out of 90 (5%) women in the vaccinated group were HPV 16/18-positive, compared with 23 out of 70 (33%) in the unvaccinated group (P<.001). The sensitivity was slightly higher for the detection of CIN2+ in unvaccinated women using both colposcopic tests (Swede score, 0.67 vs 0.75; Colposcopic impression, 0.67 vs 0.68), but the difference was not statistically significant. The results indicate that colposcopic examination is still a useful tool in vaccinated women entering the organized cervical screening program.

In Study 3, we identified all women aged 33–62 who had not participated in the organized cervical screening program for at least 10 years in one Swedish county. HPV self-sampling kits were sent to all eligible women. A randomized written reminder was sent to non-responders after 8 weeks. HPV-positive women were referred directly to colposcopy without prior triage. Biopsies for histopathologic confirmation were taken in all women. Among eligible women, 150/741 (20.2%) returned the self-sample kit or attended routine screening; 11 of the responders returned the kit after the reminder. In total, 23/150 (15.3%) of returned kits were HPV positive. Out of the 23 HPV-positive women, 17 (74%) attended colposcopy; 11/17 (65%) had a histopathological high-grade intraepithelial lesions (HSIL) or an invasive cervical cancer. The direct send kit strategy and referral of all HPV-positive women to colposcopy without prior triage appears to be feasible if resources are available and should be prioritized given the high prevalence of HSIL lesions and cancer among non- attenders.

In Study 4, we identified women resident in Sweden between 1 Jan 1999 and 31 Dec 2018 and diagnosed with HSIL/adenocarcinoma in situ (AIS) at age 22 or above and their outcome of interest (cervical and vaginal cancer) through the National Cancer Register (NCR). Women were linked to the Swedish National Cervical Screening Registry (NKCx) to retrieve information on cervical screening history including cytology and HPV tests. After various exclusions and censoring for hysterectomy, there were 67693 women at baseline out of which 213 were identified as having a subsequent cervical or vaginal cancer. Testing was divided into four different states defined by test of cure vs no test of cure stratified by follow-up testing: 1) no follow-up, 2) irregular follow-up, 3) regular follow-up, and 4) the time period after being eligible for the test of cure and before initiating follow-up. For each state, we found an excess risk of invasive cervical or vaginal cancer for women without a test of cure: no follow- up (HR 13.9, CI 1.6-121.1), irregular follow-up (HR 2.3, CI 1.1-4.7), regular follow-up (HR 1.4, CI 0.8-2.2), the time period after being eligible for the test of cure and before initiating follow-up (HR 4.8, CI 1.8-13.1), though the excess was modest and statistically not significant for regular follow-up. The results indicate that efforts to increase participation in follow-up during the first three years after treatment of HSIL/AIS should be prioritized.