Assessing colposcopic accuracy for high-grade squamous intraepithelial lesion detection: a retrospective, cohort study

Reproducibility of colposcopy quality indicators-A survey among members of the European Federation for Colposcopy

INTRODUCTION

Colposcopy is an important part of the diagnostic work-up of women with an abnormal cervical screening test as it is used to guide the collection of biopsies. Although quality assurance has been used in the evaluation of screening programs, not much is known about quality indicators for the diagnostics and treatment of screen-positive women. Therefore, the European Federation for Colposcopy developed quality indicators aiming to support colposcopy practice across Europe. We performed a survey of colposcopy cases to determine if the quality indicators are understandable, relevant, and reproducible.

MATERIAL AND METHODS

We conducted a survey among all members of the European Federation for Colposcopy Quality and Standards Group from November 2022 to March 2023. Members were asked to collect information on a total of 17 quality indicators for 50 women who had been newly referred for colposcopy due to an abnormal screening test between January 1, 2020 to December 31, 2021. Results were reported descriptively.

RESULTS

We included data on 609 cases from 12 members across Europe. The majority of the quality indicators were either achieved or within reach of the agreed standard, often due to few countries with outlying data. One quality indicator had very low performance, although stratified results indicated that two countries had different clinical management of the patient type thereby skewing the results. In addition, discrepancies between the number of cases included in each quality indicator raised concerns regarding potential misunderstanding of the quality indicator and its objective.

CONCLUSIONS

Quality indicators on colposcopy must be understandable to those collecting data, highlighting the importance of validating quality indicators before data collection.

Improving diagnostic of cervical dysplasia among postmenopausal women aged ≥50 years using local vaginal oestrogen treatment prior to colposcopy: study protocol for a multicentre randomised controlled trial (the IDEAL study)

INTRODUCTION

Colposcopy is the most important diagnostic tool to detect cervical precancerous lesions and thereby prevention of cervical cancer. Due to age-dependent changes of the cervix, colposcopy is challenging in postmenopausal women, as the majority will have a non-visible transformation zone, resulting in increased risk of missing disease, a diagnostic cone biopsy and prolonged follow-up with repeated colposcopies. This study will be among the first to investigate, if treatment with vaginal oestrogen prior to colposcopy will improve the colposcopy performance, to ensure accurate and timely diagnosis of precancerous cervical lesions among postmenopausal women.

METHODS AND ANALYSIS

A randomised blinded controlled multicentre study. Enrolment will be performed at gynaecology departments in Central Denmark Region and Region of Southern Denmark. A total of 150 postmenopausal women aged ≥50 years referred for colposcopy due to abnormal cervical screening results will be randomised 1:1 to either pretreatment with vaginal application of Vagifem 30 µg or placebo once a day for 14 days prior to colposcopy. The primary outcome will be to compare the percentage of women in the two groups with a visible transformation zone at colposcopy, and biopsies representative of the transformation zone. Secondary outcomes will be the proportion of detected cervical intraepithelial neoplasia grade 2 or higher in the cervical biopsies; the proportion of diagnostics cone biopsies; the patients' report on possible side effects and compliance to the pretreatment.

ETHICS AND DISSEMINATION

The study has been approved by the Central Denmark Region Committee on Biomedical Research Ethics (1-10-72-34-22), the Central Denmark Regions' Research Unit (1-16-02-72-22) and The Danish Health Authority (Danish Medicine Agency; 2022015030). The study's EudraCT number is (1-23-456; 2022-000269-42) and it is registered on www.

CLINICALTRIALS

gov. The local Good Clinical Practice (GCP) unit will supervise and monitor the study closely before, during and after the study period. Findings will be disseminated in peer-reviewed scientific journals and presented in relevant conferences.

TRIAL REGISTRATION NUMBER

NCT05283421.

Artificial Intelligence and Colposcopy: Automatic Identification of Cervical Squamous Cell Carcinoma Precursors

Background/Objectives: Proficient colposcopy is crucial for the adequate management of cervical cancer precursor lesions; nonetheless its limitations may impact its cost-effectiveness. The development of artificial intelligence models is experiencing an exponential growth, particularly in image-based specialties. The aim of this study is to develop and validate a Convolutional Neural Network (CNN) for the automatic differentiation of high-grade (HSIL) from low-grade dysplasia (LSIL) in colposcopy. Methods: A unicentric retrospective study was conducted based on 70 colposcopy exams, comprising a total of 22,693 frames. Among these, 8729 were categorized as HSIL based on histopathology. The total dataset was divided into a training (90%, n = 20,423) and a testing set (10%, n = 2270), the latter being used to evaluate the model's performance. The main outcome measures included sensitivity, specificity, accuracy, positive predictive value (PPV), negative predictive value (NPV), and the area under the receiving operating curve (AUC-ROC). Results: The sensitivity was 99.7% and the specificity was 98.6%. The PPV and NPV were 97.8% and 99.8%, respectively. The overall accuracy was 99.0%. The AUC-ROC was 0.98. The CNN processed 112 frames per second. Conclusions: We developed a CNN capable of differentiating cervical cancer precursors in colposcopy frames. The high levels of accuracy for the differentiation of HSIL from LSIL may improve the diagnostic yield of this exam.

Evaluation of the diagnostic performance of colposcopy in the detection of cervical high-grade squamous intraepithelial lesions among women with transformation zone type 3

BACKGROUND

Inaccurate colposcopy diagnosis may lead to inappropriate management and increase the incidence of cervical cancer. This study aimed to evaluate the diagnostic accuracy of colposcopy in the detection of histologic cervical intraepithelial neoplasia grade 2 or worse (CIN2+) in women with transformation zone type 3 (TZ3).

METHODS

Records from 764 patients with TZ3 who underwent colposcopy-directed biopsy and/or endocervical curettage in Putuo Hospital China between February 2020 and March 2023 were retrospectively collected. Colposcopy was carried out based on 2011 International Federation of Cervical Pathology and Colposcopy (IFCPC) and Colposcopy nomenclature. The diagnostic performance of colposcopy for identifying CIN2 + was evaluated compared with biopsies. The Kappa and McNemar tests were used to perform statistical analyses.

RESULTS

Among the study population, 11.0% had pathologic CIN2+. The relative sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of colposcopy for histologic CIN2 + were 51.2%, 96.5%, 64.2% and 94.1%, respectively. The senior colposcopists (80.6%) had a higher colposcopic accuracy to diagnose histologic CIN2 + than junior colposcopists (68.6%). In subgroup analyses, age group ≥ 60 years (70.3%) showed lowest diagnostic accuracy when compared with age groups of < 45 years (84.4%) and 45-59 years (74.9%).

CONCLUSION

Our findings suggest an increased risk of diagnostic inaccuracy of colposcopy in identifying CIN2 + in those ≥ 60 years of age with TZ3, and the accuracy of colposcopy is required to be further improved.

Role of DNA ploidy in diagnosis and prognosis of high-grade cervical intraepithelial neoplasia: A prospective cohort study

OBJECTIVE

To compare the sensitivity and specificity of DNA ploidy with cytology, human papillomavirus (HPV) testing and colposcopy in diagnosis of high-grade cervical intraepithelial neoplasia (CIN) and to assess the role of aneuploidy in cervical lesions with the worst prognosis. A prospective observational cohort study was conducted on 254 women with altered colpocytology.

METHODS

Colposcopy, biopsy, DNA-ICM and HPV examinations were applied to cervical cytological and histological samples. Participants were evaluated every 6 months and divided into two groups: 'Harm' and 'No-harm'. Logistic regression and multivariate COX model were used to identify independent risk factors for diagnosis and prognosis of high-grade CIN, and ROC curve to assess the sensitivity and specificity of methods.

RESULTS

Variables 'age greater than or equal to 30 years', 'lesion size greater than 20%', 'aneuploidy' and 'HPV 16' were associated with diagnosis of high-grade CIN and 'aneuploidy' and 'women living with HIV', with a worse prognosis. Agreement for colposcopy was good, with a sensitivity of 79.3% and specificity of 94.4%; DNA-ICM and cytology were moderate, with sensitivity of 74.6% and 72.3% and specificity of 85.3% and 76.1%, respectively. High-risk HPV and HPV 16 tests were weak, with sensitivity of 75.0% and 43.75% and specificity of 50.0% and 88.64%, respectively.

CONCLUSIONS

In relation to high-grade CIN diagnosis, DNA-ICM presented similar sensitivity and specificity to cytology and high-risk HPV test when associated with HPV 16. Regarding prognosis, this research certifies that aneuploidy is considered a predictor of more severe cervical injury.

Effectiveness of cryotherapy delivered by nurses for treatment of cervical preneoplasic lesions

Introduction. Cervical cancer is a relevant public health problem for low- and middleincome countries. Follow-up of positive-screened women and compliance with treatment of precancerous lesions are major challenges for these settings. Objective. To evaluate the efficacy of cryotherapy delivered by nurses for cervical intraepithelial neoplasia (CIN). Materials and methods. Direct visual inspection with acetic acid and lugol iodine (VIAVILI), and colposcopy/biopsy were performed on women 25 to 59 years old, residents of low-income areas in Bogotá, Colombia. Trained nurses offered immediate cryotherapy to every woman with positive visual inspection. Colposcopy/biopsy was performed before treatment and at a 12-month follow-up. The effectiveness was measured as cure (outcome: no-lesion) and regression (outcome: CIN1) rates of CIN2/3 using colposcopic and histological verification. Results. A group of 4.957 women with VIA/VILI was valuated. In total, 499 were screen positive and 472 accepted immediate treatment. A total of 365 women (11 CIN2/3) received cryotherapy by nurses. Cure rate was 72% (95%CI: 39%-94%) and 40% (95%CI: 22%-85%) by colposcopic and histological verification, respectively. Regression rates were 100% and 60%. There were two related non-serious adverse events. Conclusions. Cure and regression rates by colposcopic verification are like those reported for cryotherapy delivered by doctors. The sample size (CIN2/3) hinders comparisons by type of verification. Our findings support the implementation of screen-and-treat algorithms by nurses among populations with limited access to health services.

Enhancing colposcopy training using a widely accessible digital education tool in China

BACKGROUND

Colposcopy is a cornerstone of cervical cancer prevention; however, there is a global shortage of colposcopists. It is challenging to train a sufficient number of colposcopists through in-person methods, which hinders our ability to adequately diagnose and manage positive cases. A digital platform is needed to make colposcopy training more efficient, scalable, and sustainable; however, current online training programs are generally based on didactic curricula that do not incorporate image analysis training. In addition, long-term assessments of online training are not readily available. Therefore, innovative digital training and an assessment of its effectiveness are needed.

OBJECTIVE

This study aimed to evaluate the short- and long-term effects of DECO (an online Digital Education Tool for Colposcopy) on trainees' colposcopy competencies and confidence.

STUDY DESIGN

DECO can be used both on laptops and smartphones and comprises 4 training modules (image interpretation; terminology learning; video teaching; and collection of guidelines and typical cases) and 2 test modules. DECO was tested through a pre-post study between September and November 2022. Participants were recruited in China, and DECO training lasted 12 days. Trainees initially learned basic theory before completing training using 200 image-based cases. Pretest, posttest, and follow-up testing included 20 distinct image-based questions, and was conducted on Days 0, 13, and 60. Primary outcomes were competence and confidence scores. Secondary measures were response distributions for colposcopic diagnoses, biopsies, and DECO training satisfaction. Multilevel modeling was used to determine improvement from baseline to posttraining and follow-up for the outcomes of interest.

RESULTS

Among 402 participants recruited, 96.8% (n=389) completed pretesting, 84.1% (n=338) posttesting, and 75.1% (n=302) follow-up testing. Colposcopic competence and confidence increased across this study. Diagnostic scores improved on average from 55.3 (53.7-56.9) to 70.4 (68.9-71.9). The diagnostic accuracy for normal/benign lesions, low-grade squamous intraepithelial lesions, and high-grade squamous intraepithelial lesions or worse increased by 16.9%, 13.1%, and 16.9%, respectively. Mean confidence scores increased from 48.1 (45.6-50.6) to 56.2 (54.5-57.9). These improvements remained evident 2 months after training. Trainees were also satisfied with DECO overall. Most found DECO to be scientific (82.5%), easy to use (75.2%), and clinically useful (98.4%), and would recommend it to colleagues (93.2%).

CONCLUSION

DECO is a useful, acceptable digital education tool that improves colposcopy competencies and confidence. DECO could make colposcopy training more efficient, scalable, and sustainable because there are no geographic or time limitations. Therefore, DECO could be used to alleviate the shortage of trained colposcopists around the world.

Detection of cervical high-grade squamous intraepithelial lesions and assessing diagnostic performance of colposcopy among women with oncogenic HPV

BACKGROUND

HPV screening tests may improve cervical cancer risk stratification and better guide decisions about follow-up with colposcopy/biopsy. This study aimed to estimate the risk of cervical intraepithelial neoplasia grade 2 or worse (CIN2+) among women with oncogenic HPV types and evaluate the performance of colposcopy in the diagnosis of histologic CIN2 + at Putuo Hospital, Shanghai, China.

METHODS

This cross-sectional survey was conducted from February 2020 to December 2022 among women who were referred to colposcopy. Women with high-risk (HR) HPV-positive, cytology testing and colposcopy-directed biopsy were included.

RESULTS

Univariate and multivariate analysis indicated that high-grade colposcopic impression ((OR, 17.61%, 95%CI: 11.54-26.85%) was associated with the highest risk for detecting CIN2+, followed by HSIL + cytology (OR, 6.90%, 95%CI: 3.56-13.37%) and HPV16/18 positive (OR, 2.91%, 95%CI: 2.12-3.99%). Overall, CIN2 + was detected in 14.6% of 2007 women. HPV16/18 had higher CIN2 + risks than other HR-HPV genotypes (30.1% vs. 10.2%, P<0.001). Among women with low-grade cytology, 24.1% had CIN2+, and the risks for HPV16/18 (58.2%) were higher than for other HR-HPV(16.8%). For those with high-grade cytology, there was no significant difference between HPV groups ( 75.0% vs. 72.9%, P > 0.05). The diagnostic performance of colposcopy in diagnosis of CIN2 + by senior and junior colposcopists was comparable.

CONCLUSIONS

The results indicated that referral to colposcopy is recommended in managing women with HR-HPV positive, and colposcopic impressions provide key clues for identification CIN2+.

Developing a predictive nomogram for colposcopists: a retrospective, multicenter study of cervical precancer identification in China

BACKGROUND

Colposcopic examination with biopsy is the standard procedure for referrals with abnormal cervical cancer screening results; however, the decision to biopsy is controvertible. Having a predictive model may help to improve high-grade squamous intraepithelial lesion or worse (HSIL+) predictions which could reduce unnecessary testing and protecting women from unnecessary harm.

METHODS

This retrospective multicenter study involved 5,854 patients identified through colposcopy databases. Cases were randomly assigned to a training set for development or to an internal validation set for performance assessment and comparability testing. Least Absolute Shrinkage and Selection Operator (LASSO) regression was used to reduce the number of candidate predictors and select statistically significant factors. Multivariable logistic regression was then used to establish a predictive model which generates risk scores for developing HSIL+. The predictive model is presented as a nomogram and was assessed for discriminability, and with calibration and decision curves. The model was externally validated with 472 consecutive patients and compared to 422 other patients from two additional hospitals.

RESULTS

The final predictive model included age, cytology results, human papillomavirus status, transformation zone types, colposcopic impressions, and size of lesion area. The model had good overall discrimination when predicting HSIL + risk, which was internally validated (Area Under the Curve [AUC] of 0.92 (95%CI 0.90-0.94)). External validation found an AUC of 0.91 (95%CI 0.88-0.94) across the consecutive sample, and 0.88 (95%CI 0.84-0.93) across the comparative sample. Calibration suggested good coherence between predicted and observed probabilities. Decision curve analysis also suggested this model would be clinically useful.

CONCLUSION

We developed and validated a nomogram which incorporates multiple clinically relevant variables to better identify HSIL + cases during colposcopic examination. This model may help clinicians determining next steps and in particular, around the need to refer patients for colposcopy-guided biopsies.

Evaluating the Feasibility of Machine-Learning-Based Predictive Models for Precancerous Cervical Lesions in Patients Referred for Colposcopy

Background: Colposcopy plays an essential role in cervical cancer control, but its performance remains unsatisfactory. This study evaluates the feasibility of machine learning (ML) models for predicting high-grade squamous intraepithelial lesions or worse (HSIL+) in patients referred for colposcopy by combining colposcopic findings with demographic and screening results. Methods: In total, 7485 patients who underwent colposcopy examination in seven hospitals in mainland China were used to train, internally validate, and externally validate six commonly used ML models, including logistic regression, decision tree, naïve bayes, support vector machine, random forest, and extreme gradient boosting. Nine variables, including age, gravidity, parity, menopause status, cytological results, high-risk human papillomavirus (HR-HPV) infection type, HR-HPV multi-infection, transformation zone (TZ) type, and colposcopic impression, were used for model construction. Results: Colposcopic impression, HR-HPV results, and cytology results were the top three variables that determined model performance among all included variables. In the internal validation set, six ML models that integrated demographics, screening results, and colposcopic impression showed significant improvements in the area under the curve (AUC) (0.067 to 0.099) and sensitivity (11.55% to 14.88%) compared with colposcopists. Greater increases in AUC (0.087 to 0.119) and sensitivity (17.17% to 22.08%) were observed in the six models with the external validation set. Conclusions: By incorporating demographics, screening results, and colposcopic impressions, ML improved the AUC and sensitivity for detecting HSIL+ in patients referred for colposcopy. Such models could transform the subjective experience into objective judgments to help clinicians make decisions at the time of colposcopy examinations.

Improving colposcopic accuracy for cervical precancer detection: a retrospective multicenter study in China

Background

Colposcopy alone can result in misidentification of high-grade squamous intraepithelial or worse lesions (HSIL +), especially for women with Type 3 transformation zone (TZ) lesions, where colposcopic assessment is particularly imprecise. This study aimed to improve HSIL + case identification by supplementing referral screening results to colposcopic findings.

Methods

This is an observational multicenter study of 2,417 women, referred to colposcopy after receiving cervical cancer screening results. Logistic regression analysis was conducted under uni- and multivariate models to identify factors which could be used to improve HSIL + case identification. Histological diagnosis was established as the gold standard and is used to assess accuracy, sensitivity, and specificity, as well as to incrementally improve colposcopy.

Results

Multivariate analysis highlighted age, TZ types, referral screening, and colposcopists’ skills as independent factors. Across this sample population, diagnostic accuracies for detecting HSIL + increased from 72.9% (95%CI 71.1–74.7%) for colposcopy alone to 82.1% (95%CI 80.6–83.6%) after supplementing colposcopy with screening results. A significant increase in colposcopic accuracy was observed across all subgroups. Although, the highest increase was observed in women with a TZ3 lesion, and for those diagnosed by junior colposcopists.

Conclusion

It appears possible to supplement colposcopic examinations with screening results to improve HSIL + detection, especially for women with TZ3 lesions. It may also be possible to improve junior colposcopists’ diagnoses although, further psychological research is necessary. We need to understand how levels of uncertainty influence diagnostic decisions and what the concept of “experience” actually is and what it means for colposcopic practice.