EUVIMEDEUVIMED entdecken
Open Access

Automated acquisition of explainable knowledge from unannotated colposcopic images for predicting the natural course of CIN2

Jun Shen, Jinhua Liao, Jiayang Wang, Hang Xing, Min Lin, Di Wu, Bo Zhang, Yuehui Zhou, Jiancui Chen

International Journal of Gynecology & Obstetrics · 2026-09 · 174 (3) · 1547-1558

DOI: 10.1002/ijgo.70976 ↗

Kurz gefasst

Fachartikel in der Zeitschrift International Journal of Gynecology & Obstetrics, online erschienen am 26.03.2026. Verfasst von Jun Shen und 8 weiteren Autor:innen. Beteiligte Einrichtung: Fujian Provincial Cervical Disease Diagnosis and Treatment Health Center, Fujian Maternity and Child Health Hospital College of Clinical Medicine for…. Hauptthemen laut MEDLINE: Uterine Cervical Neoplasms, Uterine Cervical Dysplasia, Colposcopy, Deep Learning. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 23 Literaturangaben. Der Volltext ist frei in PubMed Central verfügbar. Gefördert von: Fujian Provincial Natural Science Foundation, Fujian Provincial Federation of Social Sciences, Natural Science Foundation of Fujian Province.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Objective

Cervical intraepithelial neoplasia grade 2 (CIN2) represents a critical turning point in cervical cancer progression. While its natural regression rate reaches 50%–57%, CIN2 can also progress to higher‐grade lesions or cervical cancer. Accurately predicting CIN2 outcomes is essential for individualized management and avoiding overtreatment. The present study aimed to develop a deep learning–based prognostic model using colposcopic images to predict CIN2 outcomes: regression, persistence, and progression, and to provide clinically meaningful risk stratification.

Methods

Colposcopic images from 212 patients diagnosed with CIN2 were retrospectively collected, including unstained, acetic acid‐stained, and Lugol's iodine‐stained images. Region of interest extraction and cluster analysis were employed for feature quantification. Four machine learning models (logistic regression, XGBoost, Random Forest and Extra Trees) were constructed to predict outcomes. Performance was evaluated using confusion matrices, receiver operating characteristic (ROC) curves, precision‐recall curves, and decision curve analysis.

Results

Among 212 patients, 52.4% regressed, 31.1% persisted, and 16.5% progressed. Automated pipeline generated 42–45 informative image clusters per staining type. For unstained images, logistic regression performed best (macro‐AUC 0.884); for acetic acid‐stained images, XGBoost achieved the highest accuracy (macro‐AUC 0.933) with 70.0% sensitivity for progression; for iodine‐stained images, Extra Trees showed the highest regression sensitivity (93.9%). Decision curve analysis confirmed clinical utility, and SHAP analysis highlighted prognostic features.

Conclusion

Machine learning models based on colposcopic images can effectively predict CIN2 prognosis, offering an objective tool for risk assessment and individualized clinical management. Different staining modalities provide complementary information, aiding in balancing progression risk reduction with overtreatment avoidance.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Uterine Cervical NeoplasmsUterine Cervical DysplasiaColposcopyDeep LearningFemaleHumansRetrospective StudiesDisease ProgressionPrognosisPredictive Learning ModelsBoosting Machine Learning AlgorithmsROC CurveAdultClassification AlgorithmsLogistic ModelsMachine LearningPrediction AlgorithmsRandom Forest

Stichwörter der Autor:innen

artificial intelligencecervical intraepithelial neoplasia grade 2colposcopymachine learningprognosis prediction

Autor:innen und Institutionen

  • Fujian Provincial Cervical Disease Diagnosis and Treatment Health Center, Fujian Maternity and Child Health Hospital College of Clinical Medicine for Obstetrics & Gynecology and Pediatrics, Fujian Medical University, Fu…
  • School of Nursing, Fujian Medical University, Fuzhou, Fujian, China.
  • Department of Pathology, Affiliated Hospital of Nantong University, Nantong University, NanTong, Jiangsu, China.
  • The Alpert Medical School of Brown University, Department of Pediatrics, Women & Infants Hospital of Rhode Island, Providence, Rhode Island, USA.
  • Women's Health Care Department, Fu'an Maternal and Child Health Hospital, Fu'an, Fujian, China.
  • Fujian Maternity and Child Health Hospital College of Clinical Medicine for Obstetrics & Gynecology and Pediatrics, Fujian Medical University, Fuzhou, Fujian, China.

Förderung

  • Fujian Provincial Natural Science Foundation
  • Fujian Provincial Federation of Social Sciences
  • Natural Science Foundation of Fujian Province

Zitieren

APA

Shen, J., Liao, J., Wang, J., Xing, H., Lin, M., Wu, D., Zhang, B., Zhou, Y., & Chen, J. (2026). Automated acquisition of explainable knowledge from unannotated colposcopic images for predicting the natural course of CIN2. International Journal of Gynecology & Obstetrics, 174(3), 1547-1558. https://doi.org/10.1002/ijgo.70976

Vancouver

Shen J, Liao J, Wang J, Xing H, Lin M, Wu D, et al. Automated acquisition of explainable knowledge from unannotated colposcopic images for predicting the natural course of CIN2. International Journal of Gynecology & Obstetrics. 2026;174(3):1547-1558. doi:10.1002/ijgo.70976

BibTeX
@article{Shen2026,
  author = {Shen, Jun and Liao, Jinhua and Wang, Jiayang and Xing, Hang and Lin, Min and Wu, Di and Zhang, Bo and Zhou, Yuehui and Chen, Jiancui},
  title = {Automated acquisition of explainable knowledge from unannotated colposcopic images for predicting the natural course of CIN2},
  journal = {International Journal of Gynecology & Obstetrics},
  year = {2026},
  volume = {174},
  number = {3},
  pages = {1547--1558},
  doi = {10.1002/ijgo.70976}
}
Kostenlos starten →