[Optimizing biopsy procedures during colposcopy and detection of high-grade cervical lesions]

Objective: To investigate the value of biopsies on diagnosing cervical intraepithelial neoplasia of grade 2 or worse (CIN Ⅱ⁺), and optimize biopsy procedures of risk-based colposcopy practice. Methods: A prospective study was performed on 346 women referred to colposcopy following abnormal cervical cancer screening results at the First Affiliated Hospital of Nanjing Medical University, from April 2017 to December 2019. Up to 4 cervical biopsies were taken during colposcopy and each biopsy specimen was evaluated separately in histology. CIN Ⅱ⁺ identified by any biopsy was the reference standard of disease used to evaluate the diagnostic value of targeted biopsy and random biopsy, and to quantify the improved detection of CIN Ⅱ⁺ by taking multiple biopsies. Cervical cytology, HPV genotyping, and colposcopic impression were used to establish different risk strata to select various multiple biopsies procedures during colposcopy to improve accuracy and efficiency of CIN Ⅱ⁺ detection. Results: In total 346 women, 190 (54.9%, 190/346) cases of them were diagnosed as CIN Ⅱ⁺. (1) In total 346 women, 96.8% (184/190) CIN Ⅱ⁺ were detected by targeted biopsies, 27.9% (53/190) CIN Ⅱ⁺ were detected in biopsies targeted grade 1 abnormal colposcopic findings (G1) on the cervix, and 68.9% (131/190) CIN Ⅱ⁺ were detected in biopsies targetrd grade 2 abnormal colposcopic findings (G2) on the cervix. Colposcopy had a sensitivity of 68.9% when the biopsy threshold was G2, sensitivity increased to 96.8% when the biopsy threshold was defined to be G1. Among women with G1, adding 2 targeted biopsies to the first biopsy were sufficient to detect all CIN Ⅱ⁺, among women with G2, adding 1 targeted biopsy was sufficient. (2) Among 270 women, random biopsies targeted normal colposcopic findings on the cervix were performed in addition to targeted biopsies and in total 3.2% (6/190) additional CIN Ⅱ⁺ were detected. As the number of cervical quadrants involved by abnormal colposcopic images increased, random biopsy detected fewer CIN Ⅱ⁺ that would have otherwise been missed by targeted biopsies (P=0.010). (3) Women with atypical squamous cells,cannot exclude high grade squamous intraepithelial lesion (ASC-H), high grade squamous intraepithelial lesion (HSIL) or atypical glandular cell (AGC) referral cytology, HPV 16-positive, G2 were more likely to have CIN Ⅱ⁺(P<0.01); for those meeting only one category, the yield of CIN Ⅱ⁺ increased from 34.0% for one biopsy to 51.0% for two biopsies, the absolute increase in CIN Ⅱ⁺ yield increased from the first to the second biopsy was 17.0%, two biopsies were sufficient to detect all CIN Ⅱ⁺; for those meeting at least two categories, the yield of CIN Ⅱ⁺ increased from 90.7% for one biopsy to 92.6% for two biopsies, the absolute increase in CIN Ⅱ⁺ yield increased from the first to the second biopsy was 1.9%, two biopsies were sufficient to detect all CIN Ⅱ⁺; for those not meeting any category, the yield of CIN Ⅱ⁺ increased from 8.8% for one biopsy to 17.6% for two biopsies, to 23.5% for three biopsies, the absolute increase in CIN Ⅱ⁺ yield increased from the first to the second biopsy, from the second to the third biopsy was 8.8%, 5.9%, three biopsies were sufficient to detect all CIN Ⅱ⁺. Conclusions: Performing multiple targeted biopsies could improve efficiency of CIN Ⅱ⁺ detection. Adding random biopsies to multiple targeted biopsies showed very limited additional benefit for detection of CIN Ⅱ⁺. The biopsy procedures undertaken during the colposcopy visit could be modified based on various colposcopic impressions and reasons for referral.