No abstract

No abstract

In their recent article in Cancer Cytopathology, Blatt et al 1 present surprising sensitivity estimates for Papanicolaou testing, human papillomavirus (HPV) DNA testing, and cotesting for cervical intraepithelial neoplasia (CIN) and cancer. They conducted a follow-back study that started from the results of all colposcopy-guided biopsies and assessed the most recent HPV and cytology results (up to 1 year before colposcopy).The American Cancer Society recommends immediate colposcopy for women who have low-grade squamous intraepithelial lesions or worse, independently from the result of human papillomavirus test, and for those who have atypical squamous cell of undetermined significance and test positive for human papillomavirus (Fig. 1).2 For HPV-positive/cytology-negative women, the American Cancer Society recommends cotesting after 1 year with referral to colposcopy if cytology or HPV findings are positive. 2Because of the management of HPV-positive/ cytology-negative women, this study design gave rise to 2 biases. First, a biopsy specimen would be available only if a woman complied with 1-year retesting and colposcopy. In a setting with a low rate of completion for follow-up (approximately 25%), the probability of colposcopy would be lower for HPV-positive/cytologynegative women versus cytology-positive women. Second, it is likely that a substantial proportion of HPV-positive women with high-grade CIN and normal cytology at the baseline developed an abnormal cytology after 1 year. This was, for instance, the case for 9 of 10 CIN21 patients found among HPV-positive/ cytology-negative women in the New Technologies for Cervical Cancer trial.3 With the outlined follow-back design, they would be (mis)classified as HPV-positive/ cytology-positive or even as cytology-positive/HPVnegative if during follow-up they became negative for HPV.These 2 biases resulted in an overestimation of the cytology sensitivity. Indeed, a 91% sensitivity of cytology for CIN21 is really surprising: in studies with a follow-up design conducted with a screening population, the sensitivity has ranged from 60% to 70%. 4 Moreover, a number of women may have undergone colposcopy because of symptoms, and this makes the observed predictive values not meaningful for screening: approximately 1 of 1000 HPV-negative/cytology-negative women in the study had an invasive cancer.1 If this is taken to represent the general HPV-negative/cytologynegative population, we should abandon both tests as screening methods and simply periodically perform random cervical biopsies in the general population.Despite these flaws, this study confirms that the gain in cross-sectional sensitivity of HPV versus cytology for invasive cancer is much smaller, if any, than the gain in sensitivity for CIN. 4 Actually, screening can prevent new invasive cancers, and with this aim in mind, strong evidence indicates that HPV testing is more effective than cytology 5 and that cotesting may not reduce the risk of future cancer more than HPV alone but will generate substantially more ...

No abstract

No abstract