Objective To evaluate the diagnostic accuracy of liquid-based cytology versus conventional cytology for cervical neoplasia. Methods Such databases as PubMed, Embase, The Cochrane Library, Cochrane Central Register of Controlled Trials, CNKI and CBM were searched to collect the random control trials (RCTs) about evaluating the diagnostic accuracy of liquid-based cytology versus conventional cytology for cervical neoplasia published before June, 2010. According to the inclusive and exclusive criteria, two assessors independently screened the studies, extracted the data, assessed the quality and conducted meta-analysis by using RevMan 5.0 and Metadisc 1.4 softwares. Results A total of five RCTS were eligible. With the ASCUS regarded as the abnormal critical value, there were significant differences between liquid-based cytology and conventional cytology in specificity for CIN 3+ (RR=0.97, 95%CI 0.97 to 0.97, Plt;0.000 01) and CIN 2+ of high risk population (RR=1.01, 95%CI 1.01 to 1.01, Plt;0.000 01), but no significant differences were found in other outcomes. Conclusion Based on the current evidence of evidence-based medicine, the liquid-based cytology is of neither more specificity nor more sensitivity for detecting high grade CIN than the convention cytology in the regular screening program, but it seems to be of more specificity in high risk group.
Objective To evaluate the diagnostic accuracy of human papillomavirus test for cervical neoplasia. Methods The Cochrane Library (Issue 2, 2010), Cochrane Central Register of Controlled Trials (Issue 2, 2010), and the following databases as CBMdisc, CNKI, MEDLINE, PubMed, and EMbase were all searched from their establishment to June 2010 to get all the randomized control trials (RCTs), and the relevant magazines and the references of the included studies were also searched. The screening, data extraction and quality assessment were conducted in accordance with the inclusion and exclusion criteria by two reviewers independently. The software Metadisc 1.4 was used to perform meta-analyses, and the forest plots and SROC curves were drewn with the RevMan 5.0 software. Results A total of 7 RCTs involving 171 604 subjects were included. The meta-analyses showed, the sensitivity of the HPV test for detecting cervical neoplasia (CIN) was higher than that of the conventional cytology test; the difference in sensitivity for detecting CIN in or above second grade was only found between the test of HPV combined with cervical cytology and the test of conventional cytology (Plt;0.00001), but the HPV test obviously lowered its diagnostic specificity. Among the following three tests for diagnosing CIN, such as, the single test of HPV, the combined test of HPV and cervical cytology, and the test of HPV followed by cytology shunting, the statistic differences compared with the conventional cytology test were found (Plt;0.01) except the last test (P=0.41) which had no difference in diagnosing CIN in or above the first grade. Conclusion The current evidence indicates that, compared with the conventional cytology test, the HPV combined with the cytology test can improve the sensitivity for diagnosing CIN in or above the second grade, but the HPV test cannot improve the specificity for cervical neoplasia. The application of human papillomavirus test for detecting cervical neoplasia needs to be further studied.