This retrospective trial was carried out at a center for cervical cancer prevention in Germany in a population of women referred to this institution for colposcopy. The outpatient consultation has a supraregional catchment area, and up to 2,000 women present themselves with a genital precancerous lesion every year. The center has been certified by the German Cancer Society since 2013.
We included women over the age of 35 with abnormal screening results according to the new guidelines for organized cervical cancer screening (shown in Fig. 1). We excluded women under the age of 35. Furthermore, women who were pregnant or had a history of cervical cancer disease, previous cervical treatments (in particular conization) in the past 3 years, or a known cervical precancerous lesion were excluded.
Approval for this study was obtained from the institutional review board. The trial was registered in the German clinical trials register.
Two gynecologists performed the examinations. Both are experts in cervical cancer screening and have broad practical experience in colposcopy.
The examination was carried out with a video colposcope (Vidan® 2 full-HD Video Colposcope, Schmitz und Söhne, Wickede an der Ruhr, Germany) using a × 30 magnification.
The colposcopic examination procedure was the same for each patient. First, detailed inspection of the external anogenital area was performed. After the introduction of a self-retaining speculum, the vagina and especially the cervix uteri were assessed in the native state. After the application of 5% acetic acid solution, the cervix uteri and adjacent vagina were then assessed for suspicious changes at 7-15x magnification. Lugol's solution was not applied as a standard procedure.
Biopsies were performed the area of maximum changes (maximum 3 biopsies). In addition, endocervical curettage was performed if an endocervical lesion was suspected and, where possible, if the transformation zone was not fully visible. If the squamocolumnar junction was completely visible (transformation zone type 1 or 2) and if normal colposcopic findings were present, no biopsy was taken. Similarly, if the squamocolumnar junction was not completely visible (transformation zone type 3) and concomitant cervical canal stenosis and unremarkable ectocervix were present, no biopsy was taken. We defined cervical stenosis as the inability to insert a 4.5 mm curette into the cervical canal.
The examination results were documented according to the 2011 terminology of the International Federation for Cervical Pathology and Colposcopy [11].
First and secondary endpoints
The primary outcome parameter of this retrospective study was the posttest probability of a CIN3 + lesion on the cervix uteri depending on the initial cytological (PAP test) and virological (HPV test) findings in women aged 35 years and older. The secondary outcome parameter was the posttest probability of a CIN2 + lesion on the cervix uteri depending on the initial cytological (PAP test) and virological (HPV test) findings in women aged 35 years and older.
In addition, the influence of HPV subtype on the primary and secondary outcome parameters was investigated. For HPV types, differentiation between types 16, 18, and 12 of other HPV types in the high-risk group was performed as a total result (31, 33, 35, 39, 45, 51, 52, 56, 58, 59, 66, and 68) using polymerase chain reaction (PCR) technology with the Cobas system (Roche Diagnostics GmbH, Mannheim, Germany).
Data collection and statistical analysis
Patient data from colposcopic examinations were entered into a study database. All relevant information from the digital patient records was collected. These were abnormal co-test results, which led to the patient's presentation at the consultation, as well as complete documentation of the clarification colposcopy. All histological results were included in the evaluation for the presence of precancerous lesions (biopsy as part of the clarification colposcopy and histological processing of the specimen removed in the case of a performed conization or hysterectomy).
The database was created using the Excel program (Microsoft Corporation, Redmond, Washington, USA). The statistical analysis was performed using Microsoft Excel and SAS 9.4 (SAS Institute, North Carolina, USA). Mean, standard deviation, median, and extrema were used to describe the metric data. For categorical data, absolute and relative frequencies were determined. To examine the association between two categorical variables, the chi-square test was performed.