Assessment of a colposcopic scoring system (Swedescore) to reduce the number of cervical biopsies: a protocol for a clinical multicentre non-randomised intervention study in Denmark Authors: Ane-Kersti Skaarup Knudsen, Tina Hovgaard Randrup, Ulrik Schiøler Kesmodel, Berit Booth, Line Gustafson, Ameli Tropé, Björn Strander, Pekka Nieminen, Anne Hammer Published: 2024 ABSTRACT AbstractIntroductionColposcopy is a standard procedure for evaluating cervical abnormalities and collecting cervical biopsies. The procedure is associated with intra- and inter-observer variation. A colposcopic scoring system, Swedescore, has been designed to standardise and facilitate colposcopy training. Swedescore has performed well in a routine clinical setting when used by expert colposcopists to find or exclude high-grade lesions. Danish clinical guidelines for colposcopy differ from other countries, as they recommend collecting four cervical biopsies in all women regardless of risk factors or colposcopy findings. Swedescore has never been examined to assess the reduction of cervical biopsies in a real-world clinical setting. This study aims to investigate whether the implementation of Swedescore can optimise the diagnostic work-up for whom the collection of biopsies can be safely omitted or reduced.Methods and analysisThe design is a clinical multicentre non-randomised intervention study in Denmark. According to a power calculation, we will need to include 586 women referred for colposcopy. Colposcopy with Swedescore will be compared with conventional colposcopy with no Swedescore. Cervical biopsies will be divided into two separate vials (target and random biopsies). The primary outcome will be normal or cervical intraepithelial neoplasia grade 1 detected in cervical biopsies. χ2 and logistic regression will be used to compare estimates between arms.Ethics and disseminationThe study protocol has been submitted to the Ethical Committee in Central Denmark region and is not notifiable to the Committee (j.no.: 1-10-72-124-22). Results will be published in a peer-reviewed journal and presented at scientific meetings.Trial registration numberNCT05870787.Protocol versionVersion 3 (date 12. November 2024).Keywords: Colposcopy, Gynaecological oncology, Biopsy STRENGTHS AND LIMITATIONS OF THIS STUDY This clinical multicentre non-randomised intervention study will examine whether the Swedescore can be used in a routine clinical setting to safely omit or reduce the number of cervical biopsies. This clinical multicentre non-randomised intervention study will examine whether the Swedescore can be used in a routine clinical setting to safely omit or reduce the number of cervical biopsies. The study provides evidence-based knowledge of a standardised and systematic colposcopic scoring tool before possible nationwide implementation. The study provides evidence-based knowledge of a standardised and systematic colposcopic scoring tool before possible nationwide implementation. Grading of cervical biopsies will be performed by pathology experts who are blinded to the colposcopic results and conclusions. Grading of cervical biopsies will be performed by pathology experts who are blinded to the colposcopic results and conclusions. Stepped wedged randomisation has not been chosen due to expectedly relatively homogenous inclusion sites. Stepped wedged randomisation has not been chosen due to expectedly relatively homogenous inclusion sites. Information bias can be a limitation due to doctors’ cross-over at inclusion sites, non-blinded clinicians and information or research sharing at regional/national meetings or conferences. Information bias can be a limitation due to doctors’ cross-over at inclusion sites, non-blinded clinicians and information or research sharing at regional/national meetings or conferences. Introduction In recent years, increasing attention has been given to reduce the number of unnecessary tests, procedures and treatments within the healthcare system. Unnecessary procedures should be avoided or minimised to reduce patient harm and lower the healthcare system’s burden, including costs. In many developed countries, new diagnostic procedures are often implemented without knowing the potential impact on patients, clinicians and the load for the overall healthcare system. Thus, it is essential to evaluate the performance of new diagnostic procedures within a real clinical setting before implementation. Colposcopy is the gold standard for evaluating cervical abnormalities, and it is used to guide the collection of cervical biopsies. Cervical precursor lesions are categorised based on the severity of cervical intraepithelial neoplasia (CIN), grade 1, mild (CIN1); grade 2, moderate (CIN2); and grade 3, severe (CIN3). Swedescore is a validated colposcopic scoring tool, which previously has shown to perform well in a routine clinical setting when used by expert colposcopists to find high-grade lesions classified as CIN2+.1 2 It has never been assessed whether Swedescore can be used to determine when biopsies can safely be omitted or reduced in a routine clinical setting corresponding to normal or CIN 1 in cervical biopsies ( 15 mm or three to four quadrants or endocervical undefinedIodine stainingBrownFaintly or patchy yellowDistinct yellowTotal score:Swedescore is a colposcopic scoring tool. All colposcopies with the ‘Swedescore methods’ will be scored for five variables in three ordered categories from 0 to 2 points. Swedescore Swedescore is a colposcopic scoring tool. All colposcopies with the ‘Swedescore methods’ will be scored for five variables in three ordered categories from 0 to 2 points. Objectives This study will investigate if Swedescore can be used in a routine clinical setting to determine when collecting of biopsies is not needed because the risk of CIN2+ is low. The potential patient benefit and desirable reduction in histopathological overload are of great interest from a national and international perspective. Hypotheses We hypothesise that cervical biopsies can be reduced or omitted when using Swedescore, compared with conventional colposcopy method without using Swedescore. Thus, we hypothesise that, compared with women in the reference arm, women in the Swedescore arm will be less likely to have disease detected in the random biopsies that are more severe than that observed in targeted biopsies. The primary outcomes will be histopathological diagnosis with normal or CIN 1 in cervical biopsies (< CIN1). Methods and analysis Study design and population We will conduct a clinical multicentre non-randomised intervention study in Denmark. We will include women referred for colposcopy due to an abnormal cervical screening test (cytology and/or high-risk HPV (hrHPV)). Women will be eligible for inclusion if they have a TZ type 1 or type 2 according to the 2011 International Federation of Cervical Pathology and Colposcopy nomenclature.41 Women will be excluded if they have a TZ type 3, atrophic cervical mucosa (impossible to use Lugol’s iodine), allergy towards Lugol’s iodine, are pregnant, had a previous hysterectomy or due to methodological errors (ie, no colposcopy performed, biopsies not collected in two separate vials, cervical biopsies are insufficient or not available in the Pathology Databank) (figure 1). Women will be recruited by colposcopic personnel (doctors and nurses) from five public colposcopy clinics and two private gynaecology clinics in the Central Denmark Region and Southern Denmark Region (table 3). All healthcare costs for the referred patients to public and private colposcopy clinics are free of charge in Denmark, and women can choose if they want to be referred to a public or private clinic. All sites follow the same Danish clinical guideline, and there is no patient fee or difference in patient rights at the included sites. Prior to recruitment, all clinicians taking part in enrolment at inclusion sites will receive standard education about HPV, the Danish cervical cancer screening programme, Danish clinical cervical cancer and precancer guidelines, and basic interpretation of colposcopy findings. Colposcopic evaluation will be performed using one of two methods: No Swedescore method (reference clinics) and Swedescore method (intervention clinics). In addition, the intervention clinics will receive extended education about the Swedescore and use of Lugol’s iodine for pattern recognition. We aim at including both private and public colposcopy clinics in each arm to ensure representativeness in the study population. As the number of referrals differ significantly between intended colposcopy clinics, sites will be allocated to Swedescore or No Swedescore taking this into account. All intended sites have agreed to participate. Reference clinics will perform conventional colposcopy as follows: colposcopy begins with cleaning of the cervix followed by application of freshly mixed acetic acid, while the cervix is inspected for a total of at least 2 min. Findings will be documented, including the visualisation of the TZ, presence/absence of acetowhitening, capillary patterns and mosaic/punctuation. Also, a clinical colposcopic impression (negative/normal, low grade, high grade, cancer) will be recorded. Intervention clinics will use the Swedescore method, which besides the conventional colposcopy includes the application of Lugol’s iodine and the collection of the score (0–2) for five variables and the total score (0–10) for Swedescore (table 2). In all clinics, enrolled women will have four cervical biopsies collected complying with Danish clinical guidelines. Cervical biopsies will be taken with a biopsy forceps and then placed in two separate vials containing formalin and marked with the number 1 or 2. The clinicians will be responsible for which vials contain which biopsies, and that the correct vial number follows the pathology request in the Pathology Databank. The clinicians decide how many biopsies (1, 2 or 3) are sent in each vial (the total number is 4 according to current Danish clinical guidelines). Each inclusion site will have a non-blinded research vial guideline, which is known by the principal investigators, and blinded for the collaborating pathology department. Biopsies will be collected from all abnormal or areas suspected of cervical precancer (targeted biopsies). Subsequently, the other biopsies will be collected from areas that are classified as normal (random biopsies). Targeted and random biopsies will be sent for histopathological examination in separate vials. In the case only random biopsies (normal colposcopy) or targeted biopsies (cervix with a large lesion or many lesions) are collected, biopsies will be sent in two separate vials as well to determine if the number of biopsies can be reduced without compromising diagnostic accuracy (figure 2). If the histologic diagnosis of the random biopsies is less severe than that in targeted biopsies, but only in the Swedescore arm and not the reference arm, the number of random biopsies can be reduced or safely omitted. This may result in a need to update the current Danish clinical guideline. Flowchart for the total study population. TZ type 3 indicates incomplete or unsatisfactory colposcopy since part or all the squamocolumnar junction is within the endocervical canal and cannot be visualised at 360°. HPV, human papillomavirus. TZ, transformation zone. Table 3Inclusion sitesNo Swedescore (reference clinics)Swedescore (intervention clinics)Department of Obstetrics and Gynecology, Regional Hospital Randers, Central Denmark Region (public hospital)Department of Obstetrics and Gynecology, Regional Hospital Gødstrup, Central Denmark Region (public hospital)Department of Obstetrics and Gynecology, Regional Hospital Horsens,Central Denmark Region (public hospital)Department of Obstetrics and Gynecology, Regional Hospital Viborg, Central Denmark Region (public hospital)Department of Obstetrics and Gynecology, Odense University Hospital,Southern Denmark Region (Public hospital)‘Guldberg Women’s Clinic’, Fredericia, Southern Denmark Region (private gynaecology practice)‘Horsens Women’s Clinic’, Horsens, Central Denmark Region. (private gynaecology practice)Inclusion sites in Denmark with colposcopy method ‘No Swedescore’ categorizedcategorised as reference clinics and ‘Swedescore’ categorizedcategorised as intervention clinics. All sites have a local investigator assigned. Both public and private colposcopy clinics with university and non-university hospitals are represented. Inclusion sites Inclusion sites in Denmark with colposcopy method ‘No Swedescore’ categorizedcategorised as reference clinics and ‘Swedescore’ categorizedcategorised as intervention clinics. All sites have a local investigator assigned. Both public and private colposcopy clinics with university and non-university hospitals are represented. Colposcopy methods with No Swedescore and Swedescore. Colposcopy with ‘No Swedescore method’ applying acetic acid to the cervix, while acetic acid and Lugol’s iodine are applied as well as the colposcopic scoring system Swedescore is used for the ‘Swedescore method’. In both methods, four biopsies are sent in two separate vials according to random and target biopsies. The clinicians decide how many biopsies (1, 2 or 3) are sent in each vial (the total number is 4 according to current Danish clinical guidelines). If only random biopsies (normal colposcopy) or targeted biopsies (cervix with a large lesion or many lesions) are collected, biopsies will still be sent in two separate vials. The pathologist is blinded regarding which vials contain which biopsies. Figure created in Biorender.com. Cytology and human papillomavirus (HPV) testing Cervical cytology will be reported according to the Bethesda 2014 grading system: ‘low-grade’ will include ASCUS and low-grade squamous intraepithelial lesion (LSIL). ‘High-grade’ will include HSIL, atypical squamous cells cannot exclude high-grade lesion (ASC-H), atypical glandular cells (AGC) and adenocarcinoma in situ (AIS). HPV type will be categorised as negative, HPV 16, HPV 18 or hrHPV non-16/18. HPV testing will be performed using Cobas 4800 DNA test (Roche Diagnostics, Switzerland). Pathology Cervical biopsies will be sent for histopathological examination at three Danish Pathology Departments (Viborg, Randers, Vejle) with extensive experience in gynaecological pathology. Samples will be processed in the pathology laboratory along with other routine samples. Each vial with biopsies will get a separate histopathological diagnosis and be reported using the CIN grading as follows: normal (including inflammation and reactive changes), CIN1, CIN2, CIN3, AIS or cancer, which also will be according to the current WHO classifications.42 When there is not enough tissue or other deficiencies present, the biopsy will be deemed not representative or inadequate. Grading of cervical biopsies will be performed by the local pathologists who are blinded from all clinical characteristics except the age of the woman, result of the current screening test, screening history and previous histopathological diagnosis. The pathologists have no knowledge of which vials contain target or random biopsies. Data collection On all women enrolled, we will state the given colposcopic method (No Swedescore vs Swedescore), and from medical records, we will collect baseline characteristics such as age, parity, screening history, HPV vaccination status, previous cervical disease including conisation, menopause status, gynaecological symptoms, body mass index and smoking status. These baseline characteristics will be self-reported information from the visit when the women are referred to colposcopy, except results on previous and current screening history, cervical biopsies and/or conisation, which will be retrieved from the Danish Pathology Databank. From medical records, we will collect information on type of TZ, colposcopic findings (acetowhitening, vessel patterns, and mosaic and/or punctuation), the colposcopic impression (normal, low grade, high grade or cancer) and number of cervical biopsies (target and random) collected. For intervention clinics, the total score of Swedescore will be obtained. The histological diagnosis will be collected for the two separate vials in the Pathology Databank. Missing data from baseline characteristics will be reported, and if relevant, a statistical imputation method will be used. All data will be entered and stored in REDCap, a secure web application for building and managing online surveys and databases.43 44 Data management and statistical analyses will be performed in STATA (V.18; StataCorp, College Station, TX). Statistical analysis Primary outcome will be