New diagnostic and prognostic methods to improve the effectiveness of cervical cancer screening

Abstract

This thesis has made use of real-life clinical practice to provide guidance for secondary prevention of cervical cancer. We combined epidemiological, virological and biostatistical investigations using HPV DNA and HPV E6/E7 mRNA tests in women who were identified with abnormal findings through the organised cervical cancerscreening programme. The majority of cytological abnormalities detected through the Swedish cervical cancer-screening programme are minor, i.e. atypical squamous cells of undetermined significance (ASC-US, 3.6 %) or low-grade intraepithelial lesions (LSIL, 2.1 %). Since only a minority of women with minor abnormal cytology harbour underlying cervical disease, or rarely invasive cervical carcinoma (ICC), management of ASC-US/LSIL remains a major challenge to system efficiency. The realisation that infection with high-risk human papillomavirus (HR-HPV) causes cervical cancer and its precursor lesions has led to development of new molecular tests for detection of HRHPV.

Study 1 compared the effectiveness of HR-HPV DNA testing, using Hybrid Capture 2 (HC2), with repeat cytology using Pap smear for detection of underlying high-grade cervical intraepithelial neoplasia (CIN2-3) in an evaluation of 177 women with minor cytological abnormalities, identified through the population-based cervical cancer screening programme. The HPV DNA test was positive in 66% of the women. Sensitivity for detection of CIN2-3 was 82% (95% CI; 80-97) using HR-HPV DNA testing and 61% (95% CI; 45-74) using Pap smear. These results indicated that HRHPV DNA testing found more underlying disease than follow-up with repeat Pap smear and suggest that HR-HPV DNA-positive women should be referred for further examination including colposcopy and directed biopsies.

This study was published in 2005 and since then a large number of other studies have addressed this issue. There is now consistent evidence indicating that HPV triage with HC2 is more accurate than repeat cytology to triage women with equivocal (ASC-US) cytology. HR-HPV DNA testing has also been considered useful as a follow-up test in women treated for high-grade CIN to predict success or failure of treatment. HPV testing identifies residual disease more rapidly, with higher sensitivity and with similar specificity compared with follow-up cytology or histological assessment of section margins. Since expression of the HR-HPV E6/E7 oncogenes, resulting from deregulation of p53 and retinoblastoma protein, is necessary for malignant transformation and persistence in cervical tissue, a diagnostic test that identifies E6/E7 mRNA in 14 HR-HPV types was developed (APTIMA HPV Assay). In study 2 we compared the performance of APTIMA with HR-HPV DNA testing and cytology in 143 women treated for CIN with conisation. These women were followed over a median time of 3.2 years and high-grade residual/recurrent disease was identified in 7 (4.9%) of them. Presence of HR-HPV DNA at the first follow-up visit predicted all (100%; 95% CI; 64.6-100) high-grade disease during follow-up, while the APTIMA failed to identify 3 women with high-grade disease (sensitivity 57.1% 95% CI 25.0-84.2). We concluded that APTIMA sensitivity was too low to be used for follow-up surveillance after conisation.

HPV DNA triage of ASC-US cytology is an established technique, but less effective in women with LSIL due to high prevalence of HPV. HPV DNA testing has high sensitivity, but is also positive in many women without disease (low specificity). Thus, a molecular marker, such as HPV mRNA, which allows more specific identification of a transforming infection, holds promise.

In study 3 we compared the triage effectiveness of APTIMA testing with for HPV16 DNA testing, HPV16/18 DNA testing and repeat cytology, in a cohort of 205 HR HPVpositive women with minor cytological abnormalities. Nine of 25 (36%) women with ASC-US and 64 of 180 (36%) with LSIL developed CIN2+ over a 4-year follow-up period. APTIMA had the highest sensitivity to predict CIN2+ and CIN3+ among patients with ASC-US (77.8% and 100%) and LSIL (78.1 and 75.8%), although specificity was insufficient (<50%). HPV16 DNA testing, HPV16/18 DNA testing and repeat cytology were all less sensitive, but more specific than APTIMA. Our results support the use of APTIMA in triage of women with ASC-US, but not with LSIL. All evaluated tests showed accuracy estimates that indicated poor LSIL triage capability and risk of disease remained even when triage tests were negative. Additional biomarkers need to be evaluated to stratify women with LSIL.

Finally, in study 4 we addressed long-term risk of developing cervical pre-cancer among 314 women with ASC-US or LSIL in relation to age, HPV status and HPV DNA genotype. Median follow-up was 3.8 years. Data for these women were linked to the Swedish National Quality Register for Cervical Cancer Prevention (NKCx) to identify cases of histologically confirmed CIN2+. We showed that HPV status was the most important factor in determining risk for developing pre-cancer. Risk was low among HPV-negative women during the first 4.5 years, suggesting that they could safely return to the regular screening programme. The highest risk of pre-cancer was observed among women positive for HPV16/18, suggesting a need for more aggressive follow-up.

The purpose of our findings is to facilitate risk stratification in women with minor cytological abnormalities and in women treated for CIN, as well as to provide guidance for clinical management. In Sweden the second major peak in incidence of cervical cancer is among women aged 60-85. Older women are clearly a vulnerable risk group for cervical cancer. One approach to reduce CC incidence in Sweden is to extend the screening programme up to at least the WHO-recommended age of 65 years and to complement or replace cytology screening with HPV testing in older women and in those who never participate in screening programmes. Women’s health would greatly benefit from more clinically effective and cost-effective screening strategies to identify more women at risk of developing cervical cancer, thereby reducing costs for society.