What question did this study set out to answer?

This research aims to evaluate a deep learning ensemble model's performance in categorizing cervical smears using the Bethesda system.

February 28, 2026Open Access

Ensemble Learning to Categorize the Bethesda System of Reporting of Cervical Cytology

Key Points

This research aims to evaluate a deep learning ensemble model's performance in categorizing cervical smears using the Bethesda system.
Developed a soft-voting ensemble deep learning model using five transfer learning architectures: VGG16, ResNet50, InceptionV3, DenseNet121, and MobileNetV2.
Processed a total of 259 cervical cytology cases with 1016 images, divided into training, validation, and test sets.
Assessed diagnostic performance using metrics such as sensitivity, specificity, accuracy, and AUROC.
The model classified 27 out of 31 NILM images and 56 out of 58 carcinoma images correctly.
NILM sensitivity was 0.871, specificity 0.977, accuracy 0.956, and AUROC 0.98.
Carcinoma sensitivity was 0.966, specificity 0.960, accuracy 0.962, and AUROC 0.99.
Intermediate categories showed variable performance, with lower sensitivity for low-grade squamous intraepithelial lesions (0.20) but high specificity (1.00).
Overall AUROC values ranged from 0.93 to 0.99 for all categories, with the ensemble model performing best.

Abstract

Abstract Aims and Objectives: To evaluate the performance of a soft-voting ensemble deep learning (DL) model incorporating five transfer learning (TL) architectures for categorizing cervical smears according to the Bethesda system. Materials and Methods: A total of 259 cervical cytology cases processed by SurePath liquid-based cytology were included, yielding 1016 representative microphotographs. Images were divided into training (70%), validation (15%), and test (15%) sets. An ensemble model comprising VGG16, ResNet50, InceptionV3, DenseNet121, and MobileNetV2 was trained using TL with soft-voting integration. Diagnostic performance was assessed using standard metrics, including 95% confidence intervals (CI). Results: In the test set, the model correctly classified 27/31 negative for intraepithelial lesions and malignancy (NILM) images and 56/58 carcinoma images. For NILM, the sensitivity was 0.871 (95% CI: 0.72–0.95), specificity 0.977 (95% CI: 0.94–0.99), accuracy 0.956 (95% CI: 0.92–0.98), and area under the curve of receiver operating characteristic (AUROC) 0.98 (95% CI: 0.96–1.00). For carcinoma, the sensitivity was 0.966 (95% CI: 0.88–1.00), specificity 0.960 (95% CI: 0.87–0.99), accuracy 0.962 (95% CI: 0.93–0.99), and AUROC 0.99 (95% CI: 0.98–1.00). Intermediate categories demonstrated variable performance, with lower sensitivity for low-grade squamous intraepithelial lesion (0.20), though the specificity remained high (1.00). Overall, AUROC values were high across all categories (0.93–0.99). Among all the models, the ensemble learning (EL) model had the best performance. Conclusions: The proposed EL approach demonstrates high diagnostic accuracy and excellent discrimination for clinically critical categories such as high-grade squamous intraepithelial lesion (HSIL) and carcinoma. This model highlights the potential of DL-based ensemble frameworks in improving the reliability of cervical cytology interpretation. Larger datasets and whole-slide image analysis may further refine the performance for intermediate-grade lesions.

Ensemble Learning to Categorize the Bethesda System of Reporting of Cervical Cytology

Key Points

Abstract

Cite This Study