92 Systematic review of overdiagnosis in cervical cancer screening: how should we define overdiagnosis in cervical cancer screening?

Abstract

Objectives Cervical cancer screening is a common strategy for cancer control worldwide. Although its real target is invasive cervical cancer, the incidence has not been high in developed countries, and precancerous lesions have now become the actual target of cervical cancer screening. Therefore, cervical intraepithelial neoplasia (CIN) 3 has now been generally identified as the actual target for early detection and treatment, while, in some countries, CIN2 has become the treatment target. The definition of overdiagnosis in cervical cancer screening has been unclear. Although most cases of CIN have a high possibility of disappearing, CIN2 and CIN3 lesions have been routinely resected when detected by cervical cancer screening. To clarify the traditional concept of overdiagnosis in cervical cancer screening, a systematic review was performed. Method Medline, Cochrane Central, Embase, and Igaku-Cyuo Zasshi were searched until January 2018. The articles were original articles limited to English-language or Japanese-language publications. Search terms such as ‘cervical cancer’, ‘cancer screening’, ‘cytology’, ‘Pap smear’, ‘HPV testing’, and ‘overdiagnosis’ were used. A modeling approach was also included. Additional references cited in candidate articles were included as needed. To select appropriate articles regarding the concept and frequency of overdiagnosis, a two-stage review process was used: the title and abstract were initially checked and then potential papers were subsequently reviewed. Finally, studies of overdiagnosis in cervical cancer screening were selected. Results One modeling reported from the Netherlands and two articles from a Finnish study which included in a randomized controlled trial for HPV testing were selected. In the modelling approach, the frequencies of overdiagnosis in the screening period were estimated to be 74.8% for CIN1+, 68.0% for CIN2+, and 55.4% for CIN3+. Over the subjects’ lifetime, the frequencies of overdiagnosis were 70.6% for CIN1+, 63.2% for CIN2+, and 50.0% CIN3+. In the first report in the Finnish study, the gap in the cumulative incidence of detected invasive cancers between the Pap smear group and the HPV testing group suggested overdiagnosis of HPV testing. Based on a 4.5 year follow-up from the first screening of this study, the frequency of overdiagnosis was 20.3 (/100,000) for Pap smear and 39.6 (/100,000) for HPV testing. Conclusions In cervical cancer screening, precancerous lesions have been identified as the target of cancer screening. These lesions have been resected, and the adoption of this approach has expanded despite the high possibility of the disappearance of these lesions. Overdiagnosis of cervical cancer screening has not been investigated until recently and the studies regarding overdiagnosis have been few. However, its frequency was high in recent reports. Until recently, overdiagnosis has been ignored in cervical cancer screening and has led to overtreatment of precancerous lesions. In developed countries, the incidence of cervical cancer has decreased and has not become a serious burden. In addition, the natural history of the development of cervical cancer has also been clarified. Although cervical cancer screening has high impact of reeducation from cervical cancer, the balance of benefits and harms including overdiagnosis should be reconsidered.