Abstract
Background : Major screening abnormalities in pre-colposcopic stage are tests results that imply direct referral to colposcopy (and/or expedited treatment) without performing additional high-grade squamous intraepithelial lesions or worse (HSIL+) risk selection testing. Currently, both clinically validated HSIL+ risk selection tests, reflex cytology and reflex p16/Ki67 dual staining (DS), are being compared for use in primary HPV-based screening to avoid possible overtreatment, but there is still no sufficient data available for their performance.
Methods : Among 30,066 liquid-based cervical cancer screening tests results, a group of 332 women was selected with available HRHPV tests results with 16/18 limited genotyping, liquid-based cytology, DS, and histology results from standardized colposcopy with biopsy. In HPV 16/18+ cases, three triage approaches were retrospectively analyzed. Predictive values for detection of HSIL+ were calculated and number of colposcopies required in each strategy.
Results : Both triage models with DS used (reflex cytology followed by DS, and reflex DS alone in all cases) had significantly higher PPV for HSIL+ than strategy with reflex cytology alone (44.2%/45.7% vs. 28.3%; p<0.0001). In models with DS, less colposcopies were required (95/92 vs. 152) and less colposcopies were needed per HSIL+ detection (2,26/2,19 vs. 3,54). Only 1 HSIL+ case was missed in both triage models with DS incorporation.
Conclusions : p16/Ki67 dual-stain may be an effective, alone or combined with cytology, triage test to detect HSIL+ in patients with major screening abnormalities in primary HPV-based cervical cancer screening. Performing cytology as the first triage test improves the strategy by enabling referrals to expedited treatment in selected cases.
KEY WORDS: cervical cancer screening; CIN; high-risk HPV; HPV 16 and 18; p16/Ki67 dual staining; triage
Introduction
Cervical cancer still remains a significant global public health burden, particularly in low- and middle-income countries where access to regular screening and preventive interventions is limited [1]. The implementation of cervical cancer screening programs has substantially reduced the incidence and mortality rates of this malignancy in many high-income countries [2]. The emergence of 13-14 high-risk types of human papillomavirus (HRHPV) (family Papillomaviridae , subfamilyFirstpapillomavirinae , genus Alphapapillomavirus , speciesAlphapapillomavirus 9) [3] as a primary etiological agent in the development of cervical cancer has led to revolutionary changes in screening strategies [4-6]. HRHPV testing, with its very high sensitivity, has gained prominence as a cornerstone in early detection efforts [7-9].
In recent years, advancements in genotyping techniques have further refined our understanding of HPV’s role in cervical carcinogenesis [10]. The limited genotyping has opened possibilities to differentiate between HPV genotypes with varying carcinogenic propensities. Incorporating this information into clinical decision-making could enable a targeted approach to colposcopy referral, risk stratification and minimizing unnecessary interventions for individuals at lower risk [11]. Primary HPV-based screening with genotyping, presents a promising approach to enhance the accuracy of cervical cancer secondary prevention, specifically in triaging individuals at highest risk for colposcopy referral [12]. But performing a colposcopy in all major screening abnormalities, with potentially higher risk of histologic high-grade intraepithelial lesions with a quantification of cervical intraepithelial neoplasia in grade 2 or worse (HSIL/CIN2+) implying a direct referral to colposcopy without prior triage testing, identified through primary HPV-based screening with limited genotyping, raises important clinical and resource-related considerations [13]. The cost associated with performing colposcopies in all individuals with major screening abnormalities can strain healthcare systems and lead to resource allocation challenges. Furthermore, it may lead to overtreatment, causing unnecessary stress and discomfort for patients.
The transition to primary HPV-based screening with a limited genotyping new paradigm needs a critical evaluation of the clinical algorithm guiding follow-up steps. HPV 16 and/or 18 positive results (HPV 16/18+) have the highest HSIL/CIN2+ risk among all HRHPV-positive results [14-16]. Currently some of recommendations worldwide, including the 2019 American Society for Colposcopy and Cervical Pathology (ASCCP 2019) Consensus Guidelines, are advising colposcopy for all HPV 16/18+ results without any triage test [14-16]. In the ASCCP 2019 Guidelines, cytology as a triage test for all results is highly valued due to its importance in referring patients to expedited treatment, defined as excisional treatment, without preceding confirmatory colposcopy with biopsy), and not only to colposcopy. It is also recommended that all cases with HPV 16/18+ ASC-H and HPV 18+ HSIL results should be referred to colposcopy or expedited treatment and HPV 16+ HSIL patients directly to expedited treatment [14,15]. The second examined triage test approved (e.g. by the U.S. Food and Drug Administration or the Polish Society of Colposcopy and Cervical Pathology) in selected HPV-positive cases and in patients with abnormal cotesting results, is p16/Ki67 dual-staining (DS), which is characterized by high sensitivity and specificity for HSIL/CIN2+ detection [17-23].
The landscape of cervical cancer secondary prevention is evolving rapidly, driven by advances in HRHPV testing and genotyping technologies. While the use of colposcopy for all major screening abnormalities detected through primary HPV-based screening with limited genotyping is the most common management recommended in available guidelines, it is imperative to critically assess the evidence supporting this approach for all HPV 16/18+ cases. By evaluating the potential benefits and limitations demonstrated in different strategy ramifications, we aim to provide a comprehensive perspective on the need of colposcopy referral in these patients. The positive predictive value (PPV) is a valuable determinant of the accuracy of the diagnostic tests. It depends on the prevalence of the condition in the screened population and both the sensitivity and the specificity of the test [24]. PPV determines the probability that a positive test result correctly predicts the presence of the disease, so it is more important for clinicians than sensitivity and specificity [25].
This retrospective study is to evaluate performance characteristics of three different triage models, incorporating cytology alone or DS alone or in combination, in HPV 16/18 positive patients who had undergone primary HPV-based cervical cancer screening, for detecting HSIL/CIN2+ cases.
Materials and methods
Study population
The study is a post-hoc analysis including a total of 30,066 cervical cancer liquid-based screening (LBS) tests results. This dataset comprises 20,605 results from liquid-based cytology (LBC), 8331 results from HRHPV with limited HPV 16/18 genotyping, and 1130 results from DS. These LBS tests were derived from a diverse cohort of women ranging from 15 to 92 years old, with a mean age of 40.9. Data for analysis were extracted from LBS test results obtained between August 2015 and July 2020. These tests were sourced from opportunistic cervical cancer screening programs conducted at the Corfamed Woman’s Health Center, one of the largest private outpatient gynecological clinics located in an urban area of Lower Silesia, Poland. The cervical cancer screening at the Center employed three distinct models during the study period: primary cytology with reflex HRHPV, primary cotesting, and primary cotesting plus (simultaneous cotesting and DS). The reflex HRHPV test was recommended for cases with minor cytological abnormalities (ASC-US or LSIL) in the primary cytology screening. DS was conducted for ASC-US or LSIL cases in both cytology- and cotesting-based models, as well as for all positive HRHPV test results. The final study cohort consisted of 332 patients aged over 25 years, all of whom had undergone LBS testing and subsequent colposcopy with biopsy at the Center (Figure 1). This selected group of patients had no prior HPV status available. A retrospective analysis of cytologic-virologic-immunocytochemical-histologic test results was conducted, and the PPV of cytology and DS as triage tests for specific major cervical cancer screening abnormalities was evaluated. The study data were retrieved from the Center’s registry, with a detailed methodology previously described in a series of related studies [26-29]. The study received ethics committee approval (ID: 118.6120.36.2023).