Risk assessment of cervical disease by hrHPV testing and cytology

textabstractAs cervical cancer is an important health problem worldwide with over a half million patients a year and as it is the fourth most common cause of cancer-related death in women, improving the prevention of this disease is a continuing and important process. A major reduction of cancer incidence and mortality has occurred in countries with cervical cancer screening. Because cervical cancer develops through different premalignant stages it can be detected in a premalignant stage, allowing treatment before these stages would be able to develop into cervical cancer. Chapter 1 gives a general introduction about the cervix, human papillomavirus (HPV), the model(s) of cervical carcinogenesis and different measures that are taken to prevent cervical cancer. These measures include screening, triaging of abnormal test results, colposcopic examination, treatment and post-treatment surveillance. In the vast majority of cervical cancers a persistent infection with high-risk HPV types (hrHPV) has been proven to be the causative agent in their carcinogenesis. Besides almost all cervical squamous cell carcinomas, approximately 95% of all cervical adenocarcinomas (ACs) are caused by a transforming infection with a hrHPV type. The remaining ACs are rare and sometimes seem hrHPV-unrelated, which could be caused by detection error or because these tumours are indeed caused by another, not HPV-related carcinogenic mechanism. Chapter 2 describes the attribution of hrHPV in cervical clear-cell adenocarcinoma (CCAC), relatively rare tumours (<<1% of all cervical carcinoma). These tumours have a bimodal age distribution with one peak in the early twenties and another after menopause and are characterised by clear cytoplasm and Hobnail cells. In approximately 60% of the cases this tumour has been associated with intrauterine exposure to diethylstilbestrol (DES), a synthetic oestrogen in the past (falsely) used to prevent miscarriages. In this study of 28 women with CCAC, of whom 15 were DES-exposed in utero, hrHPV was found in 13 (46.4%) tumours. However, after performing immuno-histochemistry with p16INK4a and p53 to distinguish transient hrHPV infections from transforming, carcinogenic infections, only three carcinomas remained in which a causal relation of hrHPV and CCAC was plausible. This demonstrated a very limited role of hrHPV in the carcinogenesis of CCAC. None of the hrHPV-associated tumours were found in women prenatally exposed to DES. In DES-unrelated tumours only a minority (20-25%) seemed hrHPV mediated. In the Dutch population-based screening programme approximately 2.5% of screened women have borderline or mild dysplasia (BMD, PAP2/3a1). These women are retested after 6 months with either cytology of a combination of both cytology and HPV (co-testing), and after 18 months with cytology. If the tests remain abnormal, women are referred for colposcopy. However, not all women with BMD comply with this protocol. Many studies have examined the short-term value of hrHPV-testing in predicting the cumulative risk of CIN3+. In Chapter 3 we have evaluated the long-term cumulative CIN3+ risk in a group of 342 women with an abnormal cytological test result (≥ BMD). These women were followed for a time period of 17 to 19.5 years after detection. Immediate hrHPV-testing clearly stratified the CIN3+ risk; almost all CIN3+ lesions (97.1%) were found in women who tested hrHPV positive. Almost half of all hrHPV-positive women were infected with HPV16; these women had a significantly higher CIN3+ risk than women infected with other hrHPV types. This risk difference between HPV16-positive women and women positive for other hrHPV types, was only found in younger women (<30 years). In older women (≥30 years) the risks in both age groups were similar. The 5-year CIN3+ risk was lower in women who had cleared the virus within 6 months than in women with persistent hrHPV infections (2.2% versus 56.0%), with the highest risks for women with a persistent HPV16 infection (67%). We stratified the CIN3+ risks according to referral cytology and found that both women with BMD and women with >BMD referral cytology had an increased risk of developing CIN3+ within the first 5 years after detection. This risk was twice as high in women with >BMD compared to women with BMD (45% versus 22%). In the subsequent 5 years an increased risk (3.5%) remained for women with >BMD, while for women referred with BMD this risk was with 0.7% similar to that of the general population. Immediate (or delayed, i.e. after 6 months) hrHPV testing clearly stratified the risk in women with BMD; the 5-year risk in hrHPV-negative women was 0.01%, and in hrHPV-positive women 37.5%. Therefore we support the strategy to refer hrHPV-negative women with BMD to routine screening and to refer those who are hrHPV positive for additional testing or colposcopy. When these women do not develop CIN3+ within 5 years, they also may be referred to population-based screening. Additional (baseline) hrHPV-testing in women with >BMD did not result in a group with a risk low enough to refrain from colposcopy, therefore we do not advocate hrHPV testing in this group and advise to refer all these women for colposcopy. As their CIN3+ risk is elevated for at least 10 years, long-term monitoring is required. Chapter 4 focuses on women treated for high-grade cervical disease (CIN2/3). As over 10% of treated women will develop residual/recurrent (post-treatment) high-grade cervical disease, they are closely monitored by cytological testing after treatment. Most published studies concern the risk-assessment of developing post-treatment disease up to a maximum of two years. Currently, treated women in the Netherlands are referred to population-based screening when they have three consecutive negative cytological test results after treatment. This means that it would take at least another three years before women are invited for population-based screening again. In order to evaluate the safety of the current regimen, long-term follow up data is essential. Also because in several other countries yearly follow-up for up to 10 years after treatment. As successful treatment is associated with the elimination of hrHPV, hrHPV testing has been suggested as an improvement in post-treatment surveillance. In Chapter 4.1 a multi-cohort study is described that includes 435 women followed between 5 and 21.5 years after treatment. Different post-treatment test algorithms were analysed; sole cytological testing, sole hrHPV-testing and combined testing with both cytology and hrHPV (co-testing). The overall 5-year CIN2+-risk in this cohort was 16.5%. However, in women who tested consecutively negative for cytology (at 6,12 and 24 months after treatment) this risk was lowered to 2.9% and even to 1.0% in women who tested negative for co-testing at both 6 and 24 months after treatment. The risk of developing CIN3+ in treated women with three consecutive negative cytological test results is similar to the risk of developing high-grade cervical disease in women who test negative for cytology (PAP1) in population-based screening. However, by adding hrHPV-testing to post-treatment surveillance, a better risk-assessment could be reached with even fewer visits. In order to judge the results found in this multi-cohort study, studies which compared different surveillance methods (cytology, hrHPV or co-testing), tested six months after treatment, were systematically reviewed in Chapter 4.2. After a bibliographic database search, relevant studies published between January 2003 and May 2011 were identified by two reviewers with a multi-step process. Then the selected studies were methodological assessed with a modified version of the QUADAS tool (QUality Assessment of Diagnostic Accuracy Studies). Eventually, only eight out of 2410 identified studies remained, incorporating 1513 treated women. The sensitivity of hrHPV testing to predict post-treatment CIN2+ was significantly higher than of cytology (relative sensitivity 1.15; 95%CI 1.06-1.25), while the specificity of these tests was similar (relative specificity 0.95, 95%CI 0.88-1.02). The sensitivity of co-testing was the highest (95%), however this combined test had the lowest specificity (67%). In summary, this review supports the inclusion of hrHPV testing in post-treatment monitoring protocols. The general discussion in Chapter 5 summarizes the findings of this thesis and discusses possible future prospects and clinical consequences