Cervical biopsy strategy under colposcopy: evidence from a retrospective study

Abstract Objective To study the factors affecting the detection of cervical precancer and cancer under colposcopy, and to provide evidence for optimizing colposcopy biopsy strategy. Methods This was a retrospective study that included 1007 female patients who underwent colposcopy and cervical biopsy. Factors related to the detection rate of cervical precancer and cancer were analyzed. The accuracy of colposcopy diagnosis and possible influencing factors were also analyzed. Then subgroup analysis was performed in different types of cervical transformation zones to compare the efficiency of different biopsy methods. Results Abnormal cytology and positive detection of high-risk human papillomavirus (HPV) were risk factors for precancer and cancer, while positive detection of non-high-risk HPV was a protection factor. The consistency rate between colposcopy diagnosis and pathology diagnosis was 55.5%. The accuracy of colposcopy was affected by the type of cervical transformation zone rather than the seniority of the examiners. For type I transformation zone, the detection rate of high-grade lesions using 4-quadrant random biopsy (31.4% vs 21.2%, P = 0.038) and multi-point targeted biopsy (28.4% vs 12.5%, P = 0.001) was higher than corresponding control group. For type II transformation zone, the detection rate of high-grade lesions was higher when taking multi-point biopsy at ≥ 3 sites (21.7% vs 12.2%, P = 0.021). For type III transformation zone, the detection rate of lesions was higher when performing endocervical curettage (53.3% vs 35.5%, P = 0.061), but there was no statistical significance. Conclusions This study provides a basis for biopsy strategy during colposcopy. Biopsy is recommended for patients with both high-risk HPV infection and abnormal cytology. Multi-point targeted biopsy of ≥ 3 sites is recommended in type I and type II transformation zone. ECC may increase the detection rate of lesions in type III transformation zone.