Diagnostic accuracy of a systematic approach using Swedescore versus conventional colposcopy for the detection of cervical intraepithelial neoplasia grade 2 or worse: A multicenter, non‐randomized clinical study
Ane‐Kersti Skaarup Knudsen, Line Winther Gustafson, Berit Bargum Booth, Tina Hovgaard Randrup, Ulrik Schiøler Kesmodel, Björn Strander, Ameli Tropé, Pekka Nieminen, Anne Hammer,Abstract
Introduction
Globally, colposcopy is often performed by non‐experts with no certification or routine use of colposcopic scoring tools. A systematic colposcopic scoring tool, the Swedescore, has been developed to enhance the detection rate of cervical intraepithelial neoplasia grade 2 or worse (CIN2+). We aimed to compare the diagnostic accuracy of two colposcopy methods (systematic use of the Swedescore versus conventional colposcopy) for detecting CIN2+.
Material and Methods
This multicenter, non‐randomized clinical study included women (23–65 years) referred for colposcopy and with a visible transformation zone (type 1 and 2). Colposcopy was performed using the systematic scoring tool, Swedescore, or the conventional method. All women had four cervical biopsies collected. Colposcopists collected lesion‐directed biopsies (target biopsies) from abnormal areas with or without non‐lesion‐directed biopsies (random biopsies) from areas that appeared normal. The gold standard was the most severe diagnosis (CIN2+) based on all four biopsies. The primary outcome was CIN2+. We calculated sensitivity (SN), specificity (SP), positive predictive value (PPV), and negative predictive value (NPV) with 95% confidence intervals and compared results between the two methods. Results were stratified by age, body mass index (BMI), human papillomavirus (HPV) vaccination status, screening method, index cytology, index HPV type, and colposcopists. Clinical Trial Registration
Results
Among 929 women included (Swedescore: 534, conventional: 395), the diagnostic accuracy for the detection of CIN2+ was slightly higher with the Swedescore compared to the conventional method (SN: 84.7% vs. 77.8%, SP: 37.3% vs. 30.9%, PPV: 47.6% vs. 43.9%, NPV: 78.3% vs. 66.7%). Sensitivity for CIN2+ was significantly higher with the Swedescore compared to the conventional method in women ≥50 years (92.9% vs. 62.5%), those with high‐grade cytology (92.4% vs. 83.2%), and when colposcopy was performed by residents (91.8% vs. 70.6%). Regardless of HPV vaccination status and screening method, the Swedescore performed slightly better than the conventional method, although the differences were not statistically significant.
Conclusions
Our findings may suggest that, in a non‐expert setting, a systematic colposcopic examination using the Swedescore was associated with a slightly higher diagnostic accuracy for detecting CIN2+ compared to the conventional method. These findings should be interpreted considering baseline imbalances between the two methods.