Abbreviations: M1, Triage model 1; M1A, Triage model 1A; M2, Triage model 2; DS, p16/Ki67 dual staining test; HSIL/CIN2+, histologic high-grade squamous intraepithelial lesion with a quantification of cervical intraepithelial neoplasia in grade 2 or worse; PPV, positive predictive value; NPV, negative predictive value; CI, confidence interval.
Discussion
The landscape of cervical cancer prevention has evolved significantly with the approach of primary HPV-based screening and limited genotyping. Our critical examination of the necessity of obligatory colposcopy referral for all major screening abnormalities detected through this approach highlights several key points. Overreliance on colposcopy for all major screening abnormalities detected through primary HPV-based screening with limited genotyping may lead to overtreatment. In our population-based non-interventional analysis the PPV for detection of HSIL/CIN2+ in both triage models with reflex DS (M1A and M2) was significantly higher than for model with reflex cytology alone (M1), 44.2% vs. 28.3% (p < 0.0001) and 45.7% vs. 28.3% (p < 0.0001), respectively. Also, both M1A and M2 triage models were associated with an approximately 40% reduction in number of colposcopies required compared to M1 (95/92 vs. 152). The difference in PPV between triage models with DS incorporation, M1A and M2, was insignificant and almost imperceptible (44.2% vs. 45.7%), as well as difference in number of colposcopies needed to detect one HSIL/CIN2+ case (2,26 vs. 2,19) and number of missed HSIL/CIN2+ cases (1 case per each triage model). The highest PPV was achieved in HPV 16/18+ DS-positive cases in M2 (45.7%). NPVs of both models with DS triage were very high (98.3%, no statistically significant difference), highlighting patients’ good protection. All analyzed models had a PPV levels above 10%, meeting the criterion for the acceptable triage strategy in cervical cancer screening [38].
The results we obtained for selected combinations of screening tests results were similar to many other studies, including data presented by Wright et al [39]. The PPV in HPV 16/18+ cases with DS triage published in that paper was slightly lower (35.1% vs. 45.7% in our study) with nearly identical NPV for the same group (96.5% vs. 98.3%, respectively). Also, data presented by FDA was in similar range, although analyzed separately for HPV 16+ and HPV 18+, both with DS triage (PPV: 44.0%/23.9% and NPV: 93.8/95.3%, HPV 16+/HPV 18+ respectively) [18]. Our study showed lower PPV and higher NPV in the HPV 16/18+ DS-positive cases compared to Øvestad et al. (PPV: 45.7% vs. 70.3%, NPV: 98.3% vs. 84.4%) [40]. The PPV in HPV 16/18+ cases presented by Naucler et al. was nearly identical to the result in our analysis (25.1% vs. 28.3%) and the value obtained by El-Zein et al. was around 15% higher than in our study (43.8% vs. 28.3%) [41,42]. In turn, PPV for group analyzed in M1A model (NILM/ASC-US/LSIL DS-positive and ASC-H+ cases), with a triage strategy using cytology combined with DS, we have not found a reliable comparator to our research. This is the first study analyzing the following approach. Due to no full-fledged studies on the subject available, a complete in detail comparison of all PPV levels obtained in HPV 16/18-positive patients with major screening abnormalities with other studies was not possible.
In our analysis the highest PPV was obtained in model with DS-alone triage (M2) with statistically insignificant difference between this triage model and model with cytology combined with DS (M1A). Also, both triage models allowed just an effective reduction in number of colposcopies required in each model comparing to M1 and missed the same number of HSIL/CIN2+ cases – only 1 case per model, ensuring high safety. That confirms the use of DS as a triage test in primary HPV-based cervical cancer screening strategy for women with selected major screening abnormalities, allowing accuracy improvement and reduction in number of invasive procedures (e.g., colposcopy). The important difference between these two triaging models is the possibility to refer patients directly to expedited treatment, not only colposcopy. In M1A due to reflex cytology, for selected screening tests results colposcopy/expedited treatment or expedited treatment is recommended. In M2 triage, it does not bring such potential, DS-positive cases need to be referred to colposcopy and DS-negative cases should go to 1 year follow-up.
In our study AGC results were excluded from the evaluated models and were managed based on separate ASCCP 2019 recommendations algorithm [14,15]. European Consensus Statement on Expert Colposcopy 2023 also recommends distinct management for all AGC cases, compared to squamous epithelial lesions [43].
There are several strengths of the study. The non-interventional analysis of the largest number of LBS results in Poland and Central Eastern Europe, the insight into the results of screening tests in private-based opportunistic cervical cancer screening and the wide age range of participants are some of most important. Joint analysis of all locally approved cervical cancer screening tests (liquid-based cytology, HRHPV and DS), which allowed us to compare their accuracy and assess potential of these tests to be used in a triage strategy for women with major screening abnormalities in primary HPV-based screening has also been a strength. Moreover, the study introduces one of the largest investigations on cytologic-virologic-immunocytochemical-histologic correlations in cervical cancer screening. A qualified gynecologic cytopathologist evaluated all LBC and DS. However, the study has limitations, such as being a retrospective analysis and not all patients with abnormal screening results had a colposcopy with biopsy at the Center. The results of colposcopic biopsies carried out outside the facility were not included in the study due to different colposcopic protocols and/or histologic terminology and/or lack of p16 stain in cervical histologic specimens. Our study was conducted in a single private funds-based Center and the results may not be generalized to other populations. The sample size was another limitation and could affect the accuracy of the PPV estimates. To confirm our results further studies with the multicenter design and greater sample sizes are required.
Conclusions
In conclusion, the evolution of cervical cancer secondary prevention strategies calls for a comprehensive assessment of clinical practices. As we move toward more accurate and personalized approaches, such as primary HPV-based screening with limited genotyping, the necessity of obligatory colposcopy referral for all major screening abnormalities warrants careful consideration. The use of DS as a triage test allows reduction in number of colposcopy referrals while ensuring safety for the HPV 16/18-positive patients. The strategy incorporating cytology as the first triage testing might also improve referrals to expedited treatment in selected HPV-positive cases.
Author Contributions: Conceptualization, Karolina Mazurec, Maciej Mazurec and Martyna Trzeszcz; Data curation: Maciej Mazurec and Karolina Mazurec; Analysis of data: Karolina Mazurec; Interpretation of data: Karolina Mazurec and Maciej Mazurec; Investigation: Maciej Mazurec, Martyna Trzeszcz and Karolina Mazurec; Methodology: Karolina Mazurec, Maciej Mazurec and Martyna Trzeszcz; Writing draft: Karolina Mazurec, Maciej Mazurec and Martyna Trzeszcz; Draft editing: Karolina Mazurec, Maciej Mazurec and Martyna Trzeszcz; Supervision and review: Robert Jach; Review: Agnieszka Halon and Joanna Streb. All authors have read and agreed to the published version of the manuscript.
Data Availability Statement : The data presented that support the findings are available from the corresponding author upon reasonable request.