Artificial Intelligence Colposcopy Models for Cervical Cancer Screening and Diagnosis: A Systematic Review and Meta-Analysis.

IF 0.5 Q4 OBSTETRICS & GYNECOLOGY

Journal of Obstetrics and Gynecology of India Pub Date : 2026-06-01 Epub Date: 2026-05-21 DOI:10.1007/s13224-026-02352-4

Swati Priya, Sunil Kumar Panigrahi, Vikash Bansal, Amiy Arnav, Manish Raj, Priyanka Rai

{"title":"Artificial Intelligence Colposcopy Models for Cervical Cancer Screening and Diagnosis: A Systematic Review and Meta-Analysis.","authors":"Swati Priya, Sunil Kumar Panigrahi, Vikash Bansal, Amiy Arnav, Manish Raj, Priyanka Rai","doi":"10.1007/s13224-026-02352-4","DOIUrl":null,"url":null,"abstract":"Background: Cervical cancer is a significant public health problem globally, with limited access to screening and diagnostic resources in developing countries contributing to its high burden. Artificial Intelligence (AI) has emerged as a promising tool to enhance the diagnostic performance of colposcopy, a procedure used to visually examine the cervix for abnormal cell growth. The findings of this study will contribute to the growing body of evidence on the role of Artificial Intelligence (AI) in cervical cancer diagnosis.Materials and methods: This Meta-analysis aims to synthesize data and provide a comprehensive evaluation of the diagnostic performance of AI-assisted colposcopy systems (sensitivity, specificity, and accuracy) for Cervical malignancy and premalignant lesions.A comprehensive literature search was conducted across multiple databases, and 12 studies were included in the Meta-analysis. Pooled estimates of sensitivity, specificity, and accuracy were calculated using a random-effects model, and heterogeneity statistics were computed to evaluate consistency across studies.Results: The pooled estimates of diagnostic performance were sensitivity, 78.1% (95% CI 71.4-84.0%); specificity, 80.6% (95% CI 74.2-85.8%); and accuracy, 83.1% (95% CI 77.8-88.2%). Subgroup analysis and heterogeneity testing were performed to explore sources of variability.Conclusion: The results suggest that AI-assisted colposcopy systems demonstrate promising diagnostic performance in Cervical malignancy and premalignant lesions diagnosis. However, substantial heterogeneity underscores the need for methodological standardization and reporting transparency in AI-based diagnostic research. Future research should focus on developing and validating AI models in diverse populations, evaluating clinical utility and cost-effectiveness, and investigating the impact of AI-assisted colposcopy on patient outcomes and healthcare systems.Supplementary information: The online version contains supplementary material available at 10.1007/s13224-026-02352-4.","PeriodicalId":51563,"journal":{"name":"Journal of Obstetrics and Gynecology of India","volume":"76 3","pages":"192-199"},"PeriodicalIF":0.5000,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13294394/pdf/","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Journal of Obstetrics and Gynecology of India","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1007/s13224-026-02352-4","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"2026/5/21 0:00:00","PubModel":"Epub","JCR":"Q4","JCRName":"OBSTETRICS & GYNECOLOGY","Score":null,"Total":0}

引用次数: 0

Abstract

Background: Cervical cancer is a significant public health problem globally, with limited access to screening and diagnostic resources in developing countries contributing to its high burden. Artificial Intelligence (AI) has emerged as a promising tool to enhance the diagnostic performance of colposcopy, a procedure used to visually examine the cervix for abnormal cell growth. The findings of this study will contribute to the growing body of evidence on the role of Artificial Intelligence (AI) in cervical cancer diagnosis.

Materials and methods: This Meta-analysis aims to synthesize data and provide a comprehensive evaluation of the diagnostic performance of AI-assisted colposcopy systems (sensitivity, specificity, and accuracy) for Cervical malignancy and premalignant lesions.A comprehensive literature search was conducted across multiple databases, and 12 studies were included in the Meta-analysis. Pooled estimates of sensitivity, specificity, and accuracy were calculated using a random-effects model, and heterogeneity statistics were computed to evaluate consistency across studies.

Results: The pooled estimates of diagnostic performance were sensitivity, 78.1% (95% CI 71.4-84.0%); specificity, 80.6% (95% CI 74.2-85.8%); and accuracy, 83.1% (95% CI 77.8-88.2%). Subgroup analysis and heterogeneity testing were performed to explore sources of variability.

Conclusion: The results suggest that AI-assisted colposcopy systems demonstrate promising diagnostic performance in Cervical malignancy and premalignant lesions diagnosis. However, substantial heterogeneity underscores the need for methodological standardization and reporting transparency in AI-based diagnostic research. Future research should focus on developing and validating AI models in diverse populations, evaluating clinical utility and cost-effectiveness, and investigating the impact of AI-assisted colposcopy on patient outcomes and healthcare systems.

Supplementary information: The online version contains supplementary material available at 10.1007/s13224-026-02352-4.

查看原文本刊更多论文

人工智能阴道镜模型用于宫颈癌筛查和诊断：系统综述和荟萃分析。

背景：宫颈癌是一个全球性的重大公共卫生问题，发展中国家获得筛查和诊断资源的机会有限，导致其负担沉重。人工智能（AI）已经成为提高阴道镜诊断性能的一种有前途的工具，阴道镜是一种用于视觉检查宫颈异常细胞生长的程序。这项研究的发现将为人工智能（AI）在宫颈癌诊断中的作用提供越来越多的证据。材料与方法：本meta分析旨在综合数据，全面评价人工智能辅助阴道镜系统对宫颈恶性及癌前病变的诊断性能（敏感性、特异性和准确性）。在多个数据库中进行了全面的文献检索，并将12项研究纳入meta分析。使用随机效应模型计算敏感性、特异性和准确性的汇总估计，并计算异质性统计以评估研究之间的一致性。结果：诊断效能的汇总估计值为敏感性78.1% (95% CI 71.4-84.0%)；特异性为80.6% (95% CI 74.2-85.8%)；准确率为83.1% （95% CI 77.8-88.2%）。进行亚组分析和异质性检验以探索变异的来源。结论：人工智能辅助阴道镜系统在宫颈恶性及癌前病变诊断中具有良好的应用前景。然而，在基于人工智能的诊断研究中，大量的异质性强调了方法标准化和报告透明度的必要性。未来的研究应侧重于在不同人群中开发和验证人工智能模型，评估临床效用和成本效益，并调查人工智能辅助阴道镜检查对患者预后和医疗系统的影响。补充信息：在线版本包含补充资料，下载地址：10.1007/s13224-026-02352-4。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

Journal of Obstetrics and Gynecology of India OBSTETRICS & GYNECOLOGY-

CiteScore

1.30

自引率

0.00%

发文量

124

期刊介绍： Journal of Obstetrics and Gynecology of India (JOGI) is the official journal of the Federation of Obstetrics and Gynecology Societies of India (FOGSI). This is a peer- reviewed journal and features articles pertaining to the field of obstetrics and gynecology. The Journal is published six times a year on a bimonthly basis. Articles contributed by clinicians involved in patient care and research, and basic science researchers are considered. It publishes clinical and basic research of all aspects of obstetrics and gynecology, community obstetrics and family welfare and subspecialty subjects including gynecological endoscopy, infertility, oncology and ultrasonography, provided they have scientific merit and represent an important advance in knowledge. The journal believes in diversity and welcomes and encourages relevant contributions from world over. The types of articles published are: · Original Article· Case Report · Instrumentation and Techniques · Short Commentary · Correspondence (Letter to the Editor) · Pictorial Essay