Dual attention AI improves cervical cancer screening with new dataset
A Dual Cross-Attention Framework for Colposcopic CIN Grading and Swede Score Prediction Using a New Multi-Center Dataset
Computer Vision and Pattern Recognition
Summary
Cervical cancer screening can be hard because it needs expert doctors and the tests are often subjective. The authors created a smart computer method that looks at special images from multiple places to better grade the risk of cervical disease and give clinical scores. They also gathered a new collection of images and expert labels to help train and test this method. Their system works better than older ones, especially at telling disease severity and scoring important clinical factors. This could help doctors in places with fewer experts screen and treat patients more accurately.
What this means in practice
- •For healthcare providers in low-resource settings: Integrate the dual attention AI model to assist in automated cervical disease grading and risk scoring using colposcopy images.
- •For medical imaging software developers: Use the BUET multi-center colposcopy dataset and cross-attention framework to build improved colposcopy screening tools with enhanced disease grading accuracy.$Commercial implications: Enables creation of AI-powered colposcopy software products for clinical use addressing cervical cancer screening.
Authors
Dania Khan, Nuzhat Aisha Shaikh, Asfina Hassan Juicy, Raiyun Kabir, S M Shahida, Taufiq Hasan
Abstract
Cervical cancer is a major global health challenge, with disease burden falling disproportionately on low- and middle-income countries (LMICs) due to a shortage of trained specialists and the subjective nature of colposcopy-based screening. To address this challenge, we propose a novel deep learning framework for the automated grading of Cervical Intraepithelial Neoplasia (CIN) and the prediction of clinical Swede scores. We also introduce the BUET Multi-Center Colposcopy Dataset, a novel, multi-center cohort designed and annotated for Swede score prediction and CIN grading. Our proposed dual-stream cross-attention architecture mimics the visual reasoning of an expert colposcopist by explicitly fusing paired multimodal cervigrams to evaluate comparative tissue responses. Furthermore, we introduce a custom composite loss function to address severe class imbalances and scoring inconsistencies across the five Swede score components. The proposed framework achieved 71.85% accuracy and an 86.23% AUC-ROC for three-class CIN grading, outperforming existing methods. For Swede score component prediction, the architecture achieved AUC-ROC values ranging from 75.7% to 88.4%, with the composite loss function yielding consistent F1-score improvements. Finally, the total predicted Swede Score, which ranges between 0 and 10, shows a Mean Absolute Error (MAE) of 1.489. The results show that the proposed method can pave the way towards developing AI-assisted colposcopy screening tools to support risk-based triage in resource-limited healthcare settings. The dataset and source code are publicly available(url: https://github.com/mHealthBuet/BUET-colposcopy)